INFECTION AND RESISTANCE THE MACMILLAN COMPANY 
NEW YORK • BOSTON • CHICAGO • DALLAS 
ATLANTA • SAN FRANCISCO 
MACMILLAN & CO., Limited 
LONDON • BOMBAY • CALCUTTA 
MELBOURNE 
THE MACMILLAN CO. OF CANADA, Ltd. 
TORONTO INFECTION AND 
RESISTANCE 
AN EXPOSITION OF THE BIOLOGICAL PHENOMENA 
UNDERLYING THE OCCURRENCE OF INFECTION 
AND THE RECOVERY OF THE ANIMAL BODY FROM 
INFECTIOUS DISEASE, WITH A CONSIDERATION OF 
THE PRINCIPLES UNDERLYING SPECIFIC DIAG- 
NOSIS AND THERAPEUTIC MEASURES. 
B Y 
HANS ZINSSER, M. D. 
Professor of Bacteriology and Immunity, Medical School, 
Harvard University; formerly Professor of Bacteriology at 
the College of Physicians and Surgeons, Columbia University, 
and Bacteriologist to the Presbyterian Hospital, New York; 
formerly Professor of Bacteriology and Immunity, Stanford 
University, California; Colonel, Medical Officers’ Reserve Corps, 
U. S. A. 
THIRD EDITION 
gorfe 
THE MACMILLAN COMPANY 
1923 
All rights reserved PRINTED IN THE UNITED STATES OF AMERICA 
Copyright 1914, 1918 and 1923, 
BY THE MACMILLAN COMPANY/ 
Set up and eleetrotyped. Published October, 1914. 
Reprinted January, 1916 ; January, 1917; 
New Edition Revised and Reset, May, 1918 ; 
New Edition Revised and Reset, June, 1923. 
Press ol 
J. J. Little & Ives Company 
New York. U. 8. A. TO THE MEMORY OF 
a. h. 
THIS BOOK IS AFFECTIONATELY 
DEDICATED BY HIS 
SON PREFACE TO THE THIRD EDITION 
The subject of immunity shares with other branches of biology 
the difficulties, as well as the attractiveness, of being constantly sub- 
ject to alteration of point of view and conception. The fact that it 
deals with problems closely related to medicine, while at the same 
time its observed phenomena represent, in many cases, fundamental 
biological laws, attracts workers of many varieties of training, and 
the evidence by which it grows, emanates from many different 
sources. The clinician, the statistician, the epidemiological field 
worker contribute equally with the laboratory investigator, and the 
development of the past decade has been particularly characterized 
by the increasing application of the methods of chemistry and physics 
to its problems. While our subject is, as yet, far from what we may 
speak of as an exact science, it has already made considerable head- 
way toward this possible goal by enlisting the interest of biologists, 
chemists and physicists; and no one who is following the subject can 
fail to appreciate the far-reaching effects upon it that may be looked 
for from work like that of Landsteiner on the chemical alteration of 
antigens, and of Loeb’s important studies on proteins and colloidal 
behavior. 
The worker in immunity who would seek for new methods of 
approach cannot afford any longer to neglect the guidance of the 
fundamental sciences. At no time have the experimental possibilities 
of this subject been greater or more attractive, but at no time, also, 
has there been such a rapid readjustment of point of view in many 
of its phases. Failure to appreciate this would make a treatise like 
this book thoroughly antiquated and worthless in a few years. 
In rewriting the book we have considerably altered the arrange- 
ment of material into what appears to us a more logical sequence. 
We have omitted the final chapter on colloids by Prof. Young, in- 
cluded in former editions, largely because excellent books on colloidal 
chemistry, such as those of Williams and of Bancroft, are now avail- 
able, and because the principles of this subject are, at the present 
time, more generally familiar to students of infectious disease than 
they were formerly. A considerable part of the material bearing on 
this subject has been incorporated in the text wherever applicable. 
All new material accessible to us which we have considered well 
founded on experiment and observation has been included, and new 
opinions and observations, though often still in the balance, have PREFACE TO THE THIRD EDITION 
been critically discussed whenever their bearing has seemed to us 
of sufficient importance. 
The chapters on anaphylaxis have been completely rewritten. 
When the first edition was published, this phase of the subject con- 
sisted very largely of numerous exact, but uncoordinated observa- 
tions. The work which has been done since that time has revealed 
relationships and a basis for an orderly classification of phenomena; 
and although many important points are still matters of controversy, 
it has been possible to treat the subject of hypersusceptibility in a 
more orderly and logical manner than was formerly possible. In 
subjects in which so many uncertainties are involved as in that of 
hypersusceptibility, it has often been necessary to express personal 
opinions, but whenever this has been done, we have endeavored to 
thoroughly discuss opposing views and to furnish the reader with a 
sufficient amount of material to formulate an opinion of his own; or 
at least to realize that legitimate differences of conception existed. 
The final chapters on practical therapeutic methods and the 
theories upon which they are based, have been enlarged and rewritten 
with a purpose of making them more definitely useful to those en- 
gaged in the clinical and laboratory study of infectious disease. 
While we hope that the book, as rewritten, will prove more useful 
than formerly to physicians, public health workers and laboratory 
investigators, we have adhered particularly to the original purpose, 
namely, the preparation of a critical treatise for the use of students 
of medicine and public health. We are more than ever convinced, 
from our experience in teaching such students, that immunology can 
be presented with sufficient clearness and simplicity to make it easily 
accessible to students at this stage of their careers, and that a thor- 
ough survey of the subject is almost indispensable to a proper subse- 
quent approach of the problems of infectious disease. 
No important statement has been made without as thorough a 
review of the literature as we have been able to give it within the 
allotted space. 
Purely practical laboratory methods have been detailed in only a 
few instances where they seemed essential to the understanding of 
the phenomena described, and; in the therapeutic sections at the end, 
sufficient detail of procedure has been given to guide the experienced 
practitioner; but no attempt has been made to supply with this book, 
in any sense, a manual of experimentation. 
Hans Zinssee. 
February, 1923. PREFACE 
Infectious disease, biologically considered, is the reaction which 
takes place between invading microorganisms and their products 
on the one hand, and the cells and fluids of the animal’s body on the 
other. The disease is the product of two variable factors, each of 
them to a certain extent amenable to analysis, and it is self-evident 
that no true understanding of this branch of medicine is possible 
without a knowledge of the biological principles which laboratory 
study has revealed. 
For the purpose of helping to render such knowledge easily 
accessible this book was written. While it is hoped that it may 
prove useful to the practitioner and laboratory worker, it is intended 
primarily for the undergraduate medical student. To many it will 
seem that the subject in general and our method of treatment espe- 
cially are too technical and difficult for this purpose. Our own 
experience contradicts this. During the past three years the writer 
has had the opportunity to deliver lectures and to give laboratory 
courses on this subject to medical students of 2d, 3d, and 4th-year 
classes at the Stanford and Columbia Universities. It has been a 
pleasant experience to find the medical student eager for the oppor- 
tunity to obtain this knowledge and, under the present increased 
requirements for preliminary training at our best schools, fully 
capable of assimilating it. It is not a good plan to attempt too 
extensively to simplify material that, in its close analysis, presents 
complex phenomena and intricate reasoning. For this reason no 
attempt has been made to write an A B C of immunity as a quick 
road to comprehension. hTo true insight into any branch of medi- 
cine or, for that matter, into any other science, can be attained 
without a certain amount of labor; however the concepts of this 
subject are, indeed, relatively simple after the first principles have 
been mastered, and the writer has attempted, therefore, at the risk 
of seeming pedantic in places, to treat the subject critically, sep- 
arating strictly those data which may be accepted as fact from those 
in which legitimate differences of opinion prevail. 
As far as was feasible every chapter has been written as a sep- 
arate unit. This has necessitated occasional repetition, but, it is 
hoped, will add considerable to clearness of presentation in each 
individual subject. Theories have been discussed with as little 
prejudice as the possession of a personal opinion in many cases has 
permitted. PREFACE 
The chapter on Colloids was written especially for the hook by 
Prof. Stewart W. Young, of Stanford University. Since so many 
analogies between serum reactions and those taking place between 
colloidal substances generally have been observed, it has seemed best 
to devote this chapter entirely to the elucidation of the principles 
governing colloidal reactions, so that its contents may be utilized as 
explanatory of the many allusions made to colloids in the rest of 
the text. 
All available sources of information have been freely used. In 
the large majority of cases we have had access to the original papers 
and monographs. However, we acknowledge much aid from careful 
reading of the admirable summaries, written by acknowledged 
authorities, in the works edited by Kolle and Wassermann, and by 
Kraus and Levaditi. Similar acknowledgment is made to equally 
important sources in Weichhardt’s Jahresbericht, the Bulletins of 
the Pasteur Institute, and in such text-books as those of Paul Theo. 
Muller, Emery, Adami, Gideon Wells, Marx, Dieudonne, and others. 
It is needless to acknowledge the use of such classics as that of 
Metchnikoff or of the many critical writings of Bordet and of 
Ehrlich—masters who have helped to shape the thoughts of all men 
working in this field. 
The writer takes pleasure in acknowledging many helpful sug- 
gestions from his associates, Drs. Hopkins and Ottenberg, and much 
aid, in the verification of references, from Mr. Walter Bliss, Fellow 
in the Department of Bacteriology. CONTENTS 
PAGE 
Chapter I.—Infection and the Problem of Virulence .... 1 
Scope of subject. Conception of infection. Attributes of pathogenic 
microorganisms. Forms of infection. Influences of biological adapta- 
tion. Classification of parasites on the basis of invasive properties. 
Factors which determine the power to invade. Fluctuations in virulence. 
How microorganisms defend themselves against destruction. Serum 
fastness, arsenic fastness, capsule formation, inagglutinability, etc. 
Resistance to phagocytosis. Development of offensive properties on the 
part of bacteria. Accidental factors favoring infection. Types of in- 
fection. Specificity of infections. Factors determining localization. 
Incubation time. Effects of symbiosis. Presence of bacteria in tissues 
in a latent condition. 
Chapter II.—Bacterial Poisons 33 
Part played by bacterial poisons in clinical manifestations. Ptomaines. 
Importance of ptomaines in disease. True toxins or exotoxins. Funda- 
mental differentiation of two types of antigens. Endotoxins. Chemo- 
tactic bacterial extracts. Basie properties of true toxins. Other sub- 
stances biologically similar to them. Analogy to enzymes. Snake * 
venoms. Incubation time of toxins. Conception of antitoxins. Work 
of Vaughan. Researches of Friedberger. Uncertainties regarding the 
endotoxin problem. “X” substances. Absorption of toxins. Selec- 
tive action of toxins. Distribution of tetanus poison. Causes under- 
lying selective action in general. Injury done during the excretion of 
toxins. Union of toxins with susceptible cells. Importance of cell 
lipoids. Summary. 
Chapter III.—Our Knowledge Concerning Natural Immunity, Acquired 
Immunity and Artificial Immunity 58 
Natural resistance against infection. External defenses. Skin secre- 
tions. Natural species immunity. Factors on which species immunity 
rests. Examples of relative susceptibility of species to various diseases. 
Racial differences discussed. Inheritance of immunity. Individual re- 
sistance in the same race. Acquired immunity. Lasting immunity con- 
ferred by some diseases and not by others. Early observations on im- 
munization. Jen8?" and smallpox. Pasteur and chicken cholera. An- 
thrax. Methods of artificial immunization. Immunization with living 
but attenuated cultures. Immunization with virulent cultures in small 
doses; with dead bacteria and extracts. Methods of extracting bacteria. 
Immunization with toxins. Passive immunization. The conception of 
specificity. 
Chapter IV.—The Mechanism of Natural Immunity and the Phenom- 
ena Following upon Active Immunization .... 88 
Investigations on problems of inflammation. Metchnikoff’s earlier 
studies. Concentration of attention upon properties of the blood. The 
development of the humoral school. Early opposition of cellular and 
humoral points of view. Behring’s early summary of the problem. 
Phenomena following upon active immunization. Early theories. Ex- 
haustion theory. Retention theory. Alkalinity theory. Osmotic theory. 
Discovery of specific antibodies by Behring and collaborators. Ehr- CONTENTS 
PAGE 
lich’s study on ricin. Antitoxins. Pfeiffer’s discovery of lysins. Ag- 
glutinins, Precipitins, Opsonins. Tropins. Conception of antibodies 
ras a whole. Generalization of the facts discovered in the case of bac- 
teria. Haemolysins. Cytotoxins. Haemagglutinins. Precipitins to 
unformed proteins. Conception of an antigen. Nature of antigens. 
Approach to a chemical understanding of the antigen. Influence of 
racemization. Work of Wells and Osborne. Landsteiner’s work on 
changes in specificity by chemical substitution. Conception of the 
‘ ‘ Haptene. ’ ’ The problem of the non-protein antigens. Species 
specificity and organ specificity. Drug tolerance. The origin of anti- 
bodies. 
Chapter Y.—Toxin and Antitoxin 120 
Nature of the reaction. Earlier views. Calmette’s work on snake 
poison. Piltration experiments. Morgenroth’s work on HCl-toxin 
modifications. Ehrlich’s ricin neutralization. Development of the 
neutralization ideas by Ehrlich and Behring. Conception of antitoxin 
unit. Instability of toxin. Ehrlich’s experiments. The conceptions 
of M.L.D., L0 and L+ doses. Discrepancy between L0 and L + . Toxoids 
and toxons. Method of partial absorption. Toxin spectrum. Opinions 
of Arrhenius and Madsen. Bordet’s opinion. The Danysz effect. Anal- 
ogy of antigen-antibody reaction with reactions of trypsin and pepsin 
with their substrates. The Side Chain Theory. Early work of Knorr. 
Ehrlich’s analogy of cell with chemical substances. Analogy with fer- 
ments. Weigert’s law of overcompensation. Antibodies and cell re- 
ceptors. Theoretical conclusions drawn from work with tetanus poison. 
Importance of lipoidal substances in brain tissue. 
Chapter VI.—Bactericidal Properties of Blood Serum, Cytolysis 
and Sensitization 153 
Early analysis of the protective functions of the blood. Active and pas- 
sive theories of immunity. The Pfeiffer phenomenon and analysis of 
the bactericidal effect. Lytic substances. Hemolysis. Bordet’s work 
and the analysis of Ehrlich and Morgenroth. The amboceptor con- 
ception. Reasons why the amboceptor conception is not tenable. Rami- 
fications of Ehrlich’s reasoning. Bordet’s objections. Zone phenomena. 
Quantitative relationship of sensitizer and alexin. Bovine colloid or 
conglutinin. 
Chapter VII.-—Development of Our Knowledge Concerning Comple- 
ment or Alexin. Complement Fixation 184 
Origin of alexin. Microcytase and Macrocytase. Anti-lysins. Alexin 
in oedema fluids, etc. The question of alexin in the circulating blood. 
Alexin and the thyroid. Alexin and the liver. Chemical nature of 
complement and alexin. Cobra-lecithid. Enzyme-like nature of alexin. 
Filtration of alexin. Inactivation by heat. Effect of salt concentra- 
tions. Action of acid and alkali. Complement or alexin splitting. Re- 
turn to activity of inactivated alexin on standing. Inactivation by 
shaking. Photoinactivation. Alexin Fixation. Bordet-Gengou experi- 
ments. Theoretical explanation of these facts. Albuminolysins of 
Gengou. Alexin fixation by precipitates. Views of earlier writers. 
Nicoll’s view. Writer’s opinion. Conception of ‘‘Bordet-antibody.” 
Nonspecific alexin fixation. Importance of lipoids. Fixation by un- 
sensitized cells, substances in suspension. Anticomplementary prop- 
erties of serum. 
Chapter VIII.—Practical Applications of Complement-Fixation Meth- 
od. The Wassermann Reaction 216 
Historical. Early work on monkeys. First use of human beings. 
Theories of the Wassermann reaction. Methods of preparing antigen. CONTENTS 
PAGE 
Titration of antigen. Titration of haemolytic sensitizer. Alexin titra- 
tion. Performance of the test. Noguchi’s modification. Modifications 
of Bauer, Stern and others. Results of test. Reliability. Spinal 
fluid, etc. Direct precipitation methods in syphilis. Sachs-Georgi re- 
action. Kahn reaction. Complement or Alexin Fixation for the De- 
termination of Unknown Protein. Neisser-Sachs method. Principles 
of the method. Performance of test. Complement-Fixation Tests in 
Diagnosis of Malignant Neoplasms. Historical. Yon Dungern’s 
method. Results obtained. Complement-fixation in glanders. Comple- 
ment-fixation in gonococcus infections. Complement-fixation in tuber- 
culosis. The preparation of bacterial antigens for complement fixation 
as advised by writer. 
Chapter IX.—The Phenomenon op Agglutination 240 
Discovery. Applications of clinical methods. Clinical usefulness. Rela- 
tion of motility. Passive role of bacteria Bordet’s discovery of the 
importance of electrolytes. Nature of agglutinogen. Alterations by 
heat. Alterations in agglutinability. Reasons for agglutinability. 
Specificity. Biological relations between bacteria parallel to agglutinins. 
Castellani’s method of absorption. Normal agglutinins. Agglutinoids. 
Inhibition zones. Bordet’s views. “Two-phase” theory. Physical 
interpretation. The work of Neisser and Friedemann. Recent work of 
Northrop and deKruif. Acid agglutination. Iso-agglutinins. Func- 
tional importance of agglutination in infectious diseases. 
Chapter X.—The Phenomenon of Precipitation 272 
(Precipitins) 
Discovery by Kraus. Extension to bacterial antigens. Conception of 
the precipitinogen. Chemical definition of precipitinogens. Specificity. 
Methods of producing precipitins in animals. Species specificity and 
zoological significance. Nuttall’s work. Forensic precipitin tests. 
Alteration of precipitinogen by heat and other physical factors. Organ 
specificity. Ehrlich’s conception of the precipitins. Physical concep- 
tions and colloidal analogies. Rapid precipitin production. Simul- 
taneous presence of precipitinogen and precipitin in the same serum. 
Chapter XI.—Iso-antibodies 297 
Ehrlich and Morgenroth’s early work on iso-hemolysins. Similar work 
on other animal species. Iso-antibodies in human blood. Landsteiner’s 
early work. Classification of Jansky and of Moss. Hereditary charac- 
ter of blood grouping. Medical and biological importance. 
Chapter XII.—Further Consideration of the Nature of Antibodies . 306 
General consideration of biology of antibody formations. Nature of 
substances which alter the control of cell reactions on successive ad- 
ministration. Basic function of tissue cell in this regard. The cell con- 
sidered as a chemical and physical unit. Suggestion of relationship be- 
tween antigenic function and non-diffusibility. The union of antigen 
with antibody. Purely chemical and physical ideas. Influence of Loeb’s 
recent work on these conceptions. Relationship of antigen-antibody 
union and hydrogen ion concentration. Influence of cells. Dissociation 
of antigen and antibodies. Landsteiner’s work. Recent development of 
this by Huntoon. On the essential identity of the antibody. General 
development of the writer’s idea concerning the identity of antibodies. 
Attempts at physical measurement of the antibody reaction. Bordet- 
Danysz phenomenon. The epiphanin reaction. The meiostagmin reac- 
tion. Colloidal gold reaction. CONTENTS 
PAGE 
Chapter XIII.—Phagocytosis. Chemotaxis ... , 333 
Early investigations. Metchnikoff’s first studies. Phagocytosis in 
lower animals. Its significance. Importance in the development from 
the larva to the adult. Its importance in resorption of degenerated 
cells. Varieties of phagocytosis. Giant cells. Leucocytosis in response 
to the presence of bacteria. In the peritoneum. Phagocytosis in 
tuberculosis. Chemotaxis. Botanical studies. Early studies of Leber. 
Early studies of Buchner. Methods. Theories of chemotaxis. Im- 
portance of surface tension. 
Chapter XIV.—The Relationship of the Leucocytes and of Phagocy- 
tosis to Immunity 357 
Metchnikoff’s attempt to establish parallelism between phagocytic ac- 
tivity and immunity. The development of the phagocytic school of 
immunity. Earlier theories of leucocytosis. Views on the leucocytic 
origin of alexin and antibodies. Is there alexin in the circulating blood? 
Leucocytic bacterial substances. Leucoprotease. Effects of leucocytic 
substances upon infection. Leucocytic extracts. The problem of 
leucocytosis. 
Chapter XV.—Factors Determining Phagocytosis 373 
Opsonins, tropins and the opsonic index. Metchnikoff’s early idea of 
stimulins. The work of Denys and Leclef. The introduction of 
Leischmann’s technique. The work of Wright and his collaborators on 
the opsonins. Neuf eld’s work on the immune opsonins or tropins. Re- 
lationship of opsonins to alexin and to other antibodies. Difference 
between normal and immune opsonins. The opsonic index. Methods 
of determining it. Wright’s technique described. Wright’s attempt 
to establish relationship between clinical condition and opsonic index. 
His work on staphylococcus infections. Applications of these ideas to 
other conditions. General appraisal. 
Chapter XVI.—Anaphylaxis 404 
Historical survey. Description of the development of our knowledge of 
hypersusceptibility. Varieties and classification of phenomena of hyper- 
susceptibility. True anaphylaxis, or hypersusceptibility of proteins. 
The criteria. The antigen, its chemical and physical differentiation. 
Specificity. Active sensitization. Methods, quantitative relations, in- 
cubation time. Differences in methods of sensitizing various animals. 
Reinjection for the production of anaphylactic shock. Methods and 
quantitative considerations. Duration of the hypersensitive state. Pas- 
sive sensitization. Various methods of accomplishing this. Duration of 
passive sensitization. Relationship of passive sensitization to the anti- 
body contents of serum. The necessity of an interval between injection 
and sensitiveness in passive sensitization. Nature of the antibody in- 
volved. Symptoms of anaphylactic shock; in guinea pigs. Analysis 
of symptoms. Difference between physiological reactions of guinea 
pigs and other animals; rabbits, dogs, monkeys, mice. Relationship of 
these differences to distribution of smooth muscle. Desensitization or 
anti-anaphylaxis. Non-specific desensitization. 
Chapter XVII. Anaphylaxis (continued) 446 
Early theories considered. The site of the reaction. The cellular site. 
Experiments to prove the cellular site of the anaphylactic reaction. 
Transfusion experiments. Experiments with isolated tissues. Dale’s 
method. The Arthus phenomenon, or local anaphylaxis. The factor of 
injury in anaphylaxis. Humoral theories and the development of the 
anaphylatoxin Idea. The anaphylatoxin conception. Development of 
these ideas by Friedemann, Friedberger and others. CONTENTS 
PAGE 
Chapter XVIII.—Serum Sickness, Food Idiosyncrasy, Hay Fever, 
Drug Idiosyncrasy 469 
Serum Sickness. Early observation. Description of symptoms. The 
work of Von Pirquet and Schick. Analysis of the condition by various 
investigators. Is serum sickness anaphylaxis? Opposing views. De- 
sensitization in serum sickness. Differentiation between true anaphy- 
laxis and idiosyncrasies. Food Idiosyncrasy. Description of the con- 
dition. Conditions under which it occurs. Relationship of general 
hypersensitiveness and cutaneous hypersusceptibility. Cases of multiple 
sensitiveness. Theories to explain the condition. Is it true anaphylaxis? 
Congenital factor. Problem of passive transfer. Desensitization in food 
idiosyncrasies. Schloss’s experiments. Hay Fever, and Asthma. First 
observations and occurrence. Plants involved. Varieties. Search for 
antibodies. Problem of passive transfer. Desensitizing treatment. 
Drug Idiosyncrasy. Statement of condition and of problems. Different 
drugs in which it has been noticed. 
Chapter XIX.—Bacterial Anaphylaxis, Tuberculin and Allied Reac- 
tions, Bacterial Anaphylatoxin Theories, Toxin Hyper- 
susceptibility, Toxicity of Normal Serum, General Correla- 
tion 493 
The anaphylaxis problem in connection with bacteria and bacterial pro- 
teins. Early work. Analogies with protein anaphylaxis. Tuberculin 
and similar reactions, and their relationship to anaphylaxis. Bacterial 
extracts and the nature of the substances with which skin reactions can 
be produced. Anaphylatoxin theories as applied by Friedberger to bac- 
terial infection. Toxin hypersusceptibility. Description and discussion. 
Toxic action of normal sera. Its relationship to anaphylaxis. General 
attempt at coordination of the various phenomena of hypersusceptibility. 
Chapter XX.—Therapeutic Immunization 519 
Therapeutic use of diphtheria antitoxin. Antitoxin in normal human 
blood. Kellogg’s modification of Roemer’s method of measuring anti- 
toxin. Speed of absorption of injected antitoxin. Practical problems 
of dosage as based on experimental evidence. Practical considerations 
connected with diphtheria antitoxin production and standardization. 
Active immunization in diphtheria with mixtures of toxin and anti- 
toxin. The mixture actually used. The intracutaneous method of de- 
termining toxin and antitoxin values. Roemer’s method. Tetanus 
antitoxin and its standardization. Therapeutic use of tetanus antitoxin. 
Antitoxin against snake poison. Serum treatment in botulisinus. Anti- 
toxin treatment in anaphylactic wound infection. Welch bacillus, Vi- 
brion Septique and Oedematiens. Passive immunization in bacillary 
dysentery. Passive immunization in diseases caused by bacteria which 
do not form soluble toxins. General considerations. Serum treatment 
of meningitis. Serum production and method of administration. Serum 
treatment in streptococcus infections. Serum treatment in pneumonia. 
Serum production and methods of administration. Serum treatment in 
typhoid fever. Serum treatment in plague. Passive immunization in 
anthrax. Poliomyelitis. 
Chapter XXI. Active Immunization in Man, Prophylactic and Thera- 
peutic 570 
General considerations. Production and standardization of vaccines. 
Technique of vaccine production. Prophylactic immunization in typhoid 
and the paratyphoid fevers. Therapeutic vaccine treatment in typhoid 
fever. Prophylactic immunization against cholera. Prophylactic im- 
munization against plague. Prophylactic vaccination in pneumonia. 
Vaccines in streptococcus infections. Vaccination in staphylococcus in- 
fections. Vaccine treatment in gonorrheal infections. Small-pox vacci- 
nation. Prophylactic immunization in rabies. Immunity in syphilis. CONTENTS 
PAGfc 
Chapter XXII.—Non-specific Effects of Protein and Protein Deriva- 
tives 625 
Serum enzymes. Leucocytic enzymes. The influence of injections of 
non-specific substances upon infectious disease. General considerations. 
Effects. Substances used. Serum and leucocytic enzymes. Nature of, 
and fluctuation. Consideration of the Abderhalden reaction. INFECTION AND RESISTANCE INFECTION AND RESISTANCE 
CHAPTER I 
INFECTION AND THE PROBLEM OF VIRULENCE 
The early history of our knowledge of infectious disease is that 
of fermentation. It was a philosopher, Robert Boyle, writing in the 
17th century, who prophesied that the problem of infectious dis- 
ease would be solved by him wrho elucidated the nature of fermenta- 
tion. His prediction wras fulfilled 200 years later by the train of 
investigations begun by Cagniard-Latour and by Schwann, and car- 
ried to a brilliant culmination by Pasteur. It was the discovery of 
the living nature of ferments and the specific nature of the various 
micro-organisms which caused the several forms of fermentation, 
and especially of putrefaction, which made possible rational investi- 
gations in the field of infectious disease and led by analogy, first to 
logical speculation—then to actual experimental proof of the etiolog- 
ical relationship between the minute forms of life and the com- 
municable diseases. 
It is not much more than 50 years since Pollender described the 
anthrax bacillus in the blood and spleens of animals dead of this 
disease. In this short period the large number of maladies of ani- 
mals and human beings caused by micro-organisms belonging both 
to the varieties spoken of as bacteria and to those classified as 
protozoa has necessitated the segregation of this branch of knowl- 
edge into a separate chapter. 
The period of etiological investigation is now approaching its 
maturity. The causative agents of most of the more common infec- 
tious diseases have been discovered, and the biology of many of the 
pathogenic micro-organisms has been thoroughly studied both in 
their artificial cultures and in the infected animal body. In spite 
of a considerable accumulation of facts, however, the science of 
immunity, that is, the study of the defensive powers of the living 
animal body against infection, is still in its infancy, and the practi- 
cal therapeutic successes based on this science are disappointingly 
out of proportion to the really large amount of detailed knowledge 
of cellular and serum reactions at our disposal. 
The study of putrefaction and of fermentation—though furnish- 
ing the basic analogy from which the first impulse was obtained-^ 
1 2 
INFECTION AND RESISTANCE 
presented after all a problem infinitely more simple than that of the 
infection of living tissues with bacteria. For, given any organic 
material containing suitable nutritive constituents, with favorable 
environmental conditions of moisture and temperature, and spon- 
taneously or experimentally inoculated with germs of a proper 
species, and the phenomena which ensued were merely those of bac- 
terial growth, in which an active part was played by the bacteria 
only, the dead organic materials serving simply as a passive men- 
struum for these activities. 
During the earlier days of the development of bacteriology, 
therefore, when the attention of investigators was concentrated pri- 
marily upop the discovery of the specific causal agents of various in- 
fectious diseases, it seemed that the simple bringing together of 
pathogenic germ and susceptible subject should suffice for the ac- 
complishment of an infection. We have learned, however, that the 
process is much more involved, and that, fortunately for the sur- 
vival of the higher animals and man, the conditions which determine 
infection are intimately dependent upon a variety of secondary 
modifying factors. 
Throughout nature bacteria are abundant, and the environment 
of man and animals, the outer integuments of skin and hair, and the 
mucous membranes of the conjunctive, the intestinal and respira- 
tory tracts, are constantly inhabited by a thriving bacterial flora. 
The distribution of certain species in definite localities is often suffi- 
ciently constant to be regarded as a normal condition. Thus the 
Bacillus xerosis is a characteristic inhabitant of the conjunctiva, 
certain cocci and spirilla are always present in the mouth and 
pharynx, as is Doderlein’s bacillus in the vagina. The fact that 
bacilli of the colon group are invariably present in the bowels of ani- 
mals and man from the first few days or hours after birth has even 
been interpreted by some investigators as a physiologically beneficial 
condition. In the course of ordinary existence, therefore, and much 
more so during the course of accidental exposure to individuals in 
whom infection is present, the bodies of the higher animals are in 
intimate contact, not only with ordinarily harmless bacteria (sapro- 
phytes), but also with many varieties of the micro-organisms spoken 
of as “pathogenic” or disease-producing. Perfectly normal indi- 
viduals have, then, on occasion, been found to harbor diphtheria 
bacilli in nose and pharynx, meningococci have been found in simi- 
lar localities, and tetanus bacilli, the bacillus of malignant edema, 
the Welch bacillus, and other distinctly pathogenic germs have been 
isolated from the intestinal contents of individuals who showed no 
evidence of disease. In fact, the problem of the so-called bacillus car- 
riers—persons who, though themselves apparently well for the time 
being, harbor within their bodies and distribute to their environ- 
ment bacteria capable of causing disease in others—is, as we shall THE PROBLEM OF VIRULENCE 
3 
see, now recognized as one of the most important difficulties of sani- 
tary prophylaxis. In the case of typhoid fever this is particularly 
true, for it is now well known that a perfectly healthy individual may 
harbor typhoid bacilli in the gall-bladder for years and constitute, 
through all this time, a constant focus of danger to the public health. 
The accomplishment of an infection, then, is not determined 
merely by the fact that a micro-organism of a pathogenic species 
finds lodgment in or upon the body of a susceptible individual, but 
it is further necessary that the invading germ shall be capable of 
maintaining itself, multiplying and functionating within the new 
environment. An infection, then, or an infectious disease, is the 
product of the two factors, invading germ and invaded subject, each 
factor itself influenced by a number of secondary modifying circum- 
stances, and both influenced materially by such fortuitous conditions 
as the number or dose of the infecting bacteria, their path of en- 
trance into the body, and the environmental conditions under which 
the struggle is maintained. 
We have in truth, then, a battle of two opposed forces, the result 
of which is infectious disease. And it is the systematic analysis of 
these forces in their variable conditions, and the laws which govern 
them, which constitutes the science of immunity. It is the initial 
skirmish between the two which determines whether or not a foot- 
hold shall be gained upon the body of the subject and an infection 
thus established, and it is the balance between them which decides 
the eventual outcome of recovery or death. And though it is un- 
fortunately true that much of the knowledge gained by such studies 
has yielded few direct therapeutic results, the facts that have been 
revealed are fundamental to the pathology of infectious disease and 
as essential to the clinical understanding of these maladies as is the 
knowledge of the mechanism of the circulation, the chemistry of 
metabolism, or the structural changes of the tissues to the compre- 
hension of other pathological conditions. 
And from this point of view the study of infectious diseases can 
be made an eminently logical one, in that, knowing the criteria 
which govern the infection of a human being with a given germ, 
knowing the probable path of entrance, manner of distribution, and 
biological activities of the micro-organism, and the peculiarities of 
the mechanism of resistance set in motion in the body by this par- 
ticular infection, definite clinical deductions can often be made. 
Pathogenicity and Saprophytism.—One of the most fundamen- 
tal facts, immediately apparent on considering the problems of 
infection, is the phenomenon that among the innumerable varie- 
ties of bacteria and protozoa present in nature there is a very 
limited group which is capable of becoming parasitic upon' the 
body of higher animals, and among these a still smaller propor- 
tion which is capable of being “pathogenic” or causing disease. 4 
INFECTION AND RESISTANCE 
We have used the terms pathogenic and non-pathogenic as practi- 
cally synonymous respectively with “parasitic” and “saprophytic.” 
But, as we shall see, although as a rule a micro-organism must 
be parasitic to possess pathogenic powers, some of the true sapro- 
phytes or so-called half-saprophytes may be pathogenic under cer- 
tain conditions, and the terms do not cover each other absolutely. 
It is reasonable to suppose that all micro-organisms were origi- 
nally in the condition which we designate by the term “saprophytic.” 
By this term we imply that these germs maintain themselves only 
upon dead organic matter and do not thrive in or upon the living 
animal tissues. The class of saprophytes is widely distributed and 
constitutes, of course, the most important group of bacteria in na- 
ture, since upon the activities of these germs depends the unlocking 
of nitrogen and carbon from the organic complexes in the dead bod- 
ies and waste products of animals and plants. Such bacteria if 
strictly saprophytic, that is, entirely unable to maintain themselves 
upon living tissues, have little importance as producers of disease, 
or, expressed in technical terms, have little “pathogenicity.” Never- 
theless, there are cases in which strict saprophytes may cause dis- 
ease by lodging upon and growing in animal tissues which have 
been killed by other causes, so-called necrotic areas; and these, still 
being in relation with the body as a whole through the blood and 
lymph channels, furnish an area of saprophytic growth from 
which products of putrefaction or even bacterial poisons may be 
absorbed. While, as a rule, the disease following the invasion 
of necrotic tissue—such as gangrenous amputation stumps, old 
unhealed sinuses, diabetically gangrenous areas, etc., may be caused 
by a large variety of saprophytic bacteria, there are a few very 
important and specifically pathogenic bacteria which are, strictly 
speaking, saprophytes. Thus the form of meat poisoning caused 
by the Bacillus hotulinus is due entirely to the poison formed by 
this bacillus outside of the body within the substance of the dead 
foodstuff, and disease ensues as the result of subsequent ingestion of 
this poison with the food. In the same way the tetanus bacillus and, 
less strictly speaking, the diphtheria bacillus, at least in its ordinary 
mode of attack, are rather closer to the class of saprophytes than to 
that of the parasites, since neither of these bacteria, under usual 
circumstances, invades the substance of the tissues beyond the point 
of initial lodgment, causing disease only by the production of 
specific poisons, a condition known as “toxemia” or intoxication. 
The tetanus bacillus, moreover, is not usually capable of maintain- 
ing itself and multiplying even at the point of initial lodgment 
unless the tissues have been injured by trauma or irritated by the 
presence of foreign bodies. Bacteria of such characteristics, there- 
fore, though pathogenic—that is, incitant of disease—remain never- 
theless essentially saprophytes living upon the dead animal tissues, THE PROBLEM OF VIRULENCE 
5 
not invading the living cells or body fluids. It is true that investi- 
gations of Frosch 1 have shown that diphtheria bacilli may often be 
found in blood and organs of diphtheritic patients, and tetanus 
bacilli have occasionally been found in the spleen. However, such 
distribution is not necessary for the production of disease by these 
bacteria, and the essential point remains that they may cause violent, 
often fatal, disease without truly departing from their saprophytic 
mode of life upon dead tissues. Between the saprophytes and the 
true parasites or invaders of living tissue many transitions occur, 
and the condition of parasitism is probably a form of specific adap- 
tation. 
How such transition may be biologically developed is probably 
well illustrated by the investigations of Italian bacteriologists upon 
tetanus bacilli.2 Tarozzi 3 inoculated guinea pigs and rabbits with 
tetanus spores subcutaneously and found that these spores were 
rapidly transported to the liver, spleen, and kidneys, where they 
could maintain a latent existence for as long as 51 days. If during 
this period trauma or any injury of the organs was practiced which 
led to the formation of necrotic tissue the spores would develop upon 
this basis and cause acute or chronic tetanus. Canfora,4 continuing 
these studies, likewise found that tetanus spores inoculated under the 
skin are rapidly distributed throughout the circulation. If no 
trauma has taken place at the point of inoculation the locally lodged 
spores may be rapidly destroyed, probably by phagocytosis. In the 
circulation they appear to be less rapidly eliminated and may be 
present for from ten to thirteen days. If, during this period, there 
is produced a small wound, blood clot, or necrotic area in the body— 
this may serve as a focus for development and tetanus may ensue. 
After ten or more days the spores disappear from the blood, but may 
then take up a latent existence in some of the organs—as stated by 
Tarozzi. Apart from their importance as constituting a sort of 
transitional condition between pure saprophytism and parasitism, 
these investigations would seem to have much bearing upon the so- 
called cases of “cryptogenic tetanus.” 
Resistance of Tissue Cells to Invasion.—True infection, that is, 
the invasion of one species by individuals of another, and the 
ability of the latter to multiply and functionate within the cell 
complexes of the former, is a process quite out of keeping with the 
ordinary plans of nature, throughout which there seems to be a 
distinct opposition to the colonization and functionation of one 
living being within the living substance of another. Thus, as 
1 Frosch. Zeitschr. f. Hyg., Yol. 13, 1893. 
2 Belfanti, quoted from Canfora, Centralblt. f. Bad., I. Ori<>\ Yol 45 
1908. 
3 Tarozzi. Centralblt. f. Bact.} Orig. Yol. 38, 1905. 
4 Canfora. Centralblt. f. Bad., Orig. Yol. 45, 1908. 6 
INFECTION AND RESISTANCE 
Bail 5 has pointed out, a mass of frogs’ eggs will remain entirely 
uninvaded while alive, though the water surrounding it may swarm 
with bacteria of many varieties, but when by some accident such 
a mass of eggs ceases to live, it immediately falls prey to bacterial 
infection. The same point is illustrated by the rapidity with which 
intestinal bacteria will spread throughout the body after death, 
when during life they have remained confined to the lumen of 
the intestine, or, at most, get into the portal circulation, to be de- 
stroyed in the liver. 
We must also bear in mind that invasiveness on the part of 
micro-organisms may take the form merely of an ability to lodge 
in the intercellular spaces. Secondary cell destruction, then, may 
be first from the outside as foci are established, and invasion of 
the cellular protoplasm in such cases follows only after toxic and 
pressure effects may have brought about cell death. 
By the living cell, therefore, opposition is usually offered to 
invasion by bacteria, a vital function which Bail has attempted to 
make clearer by formulating it as a law, referring to it as “Das 
Gesetz der Lebensundurchdringlichkeit.” Upon which cell func- 
tion this vital resistance to invasion depends is to a large extent a 
mystery. It would seem to rest in principle upon the fact that the 
invading cell meets the invaded one under conditions peculiarly 
adapted to the activities of the latter, and is overcome before condi- 
tions suitable for its own activities have been established. The con- 
ditions here are not unlike those observed in the case of digestive 
enzymes, a comparison which becomes more than an illustrative anal- 
ogy when we consider that apart from the mere mechanical disturb- 
ance created by the presence of bacteria as foreign bodies the struggle 
between invader and tissue is largely one of enzyme against enzyme. 
Thus, for instance, the gastric juice does not act upon the mucous 
membrane of the stomach during life—but after death, at autopsy, 
partial digestion of this membrane by the pepsin is often seen. 
The function upon which the resistance of the living cell de- 
pends will probably not be understood until we have a clearer con- 
ception of what constitutes cell death. Morphologically, until de- 
composition has set in, a recently killed cell looks exactly like a 
living one, and the difference which permits one cell to continue 
metabolism and cell division, and inhibits this in another, is, so far 
at least, too intricate to yield to analysis. Of some interest in 
this connection, though vague, is the observation made by us some 
years, ago, namely, that the degree of hydrogen ion concentration 
which changes the normally negative charge carried by bacteria in 
neutral solutions, to a positive one, corresponds roughly to the 
degree of acidity at which multiplication ceases. 
6 Bail. “Das Problem der Bakt. Infection.” Klinkhardt, Leipzig, 1911. THE PROBLEM OF VIRULENCE 
7 
Much more definite, and of greater possible significance are the 
more recent observations of Osterhout,6 who has found that during 
the process of cell death electrical conductivity undergoes a definite 
change by means of which the process may be followed with pre- 
cision. By means of careful experiments with, among other things, 
Laminaria, placed in an JSTaCl solution of a conductivity equal to 
that of sea water, he found within less than six minutes, a fall of 
resistance to 94.6 per cent, of the resistance it had in sea water. 
Replacement of the tissue in sea water brought about a return to 
practically normal resistance. His researches indicate that in the 
course of injury and death, electrical resistance diminishes progres- 
sively and if the exposure is sufficiently prolonged, death together 
with extensive fall of resistance occurs. If exposure to the injury 
is arrested sufficiently early, complete recovery of normal cell life 
and normal resistance are reestablished. And of remarkable in- 
terest in this connection is that between the two, a degree of ex- 
posure to injury may be practiced which permits a recovery which 
is incomplete. It is by researches of this kind that we may hope 
eventually, perhaps, to gain some knowledge of the vital cell re- 
sistance to injury and invasion to which Bail has merely given a 
name. And it would be of the greatest interest to utilize the 
methods of Osterhout for similar studies on the effect of bacterial 
poisons of various kinds upon the resistance of body cells.* 
Whenever this vital resistance or opposition is overcome, and 
micro-organisms enter the tissues or cells, an abnormal process is 
taking place, and this process is, strictly defined, infection. Never- 
theless, it is by no means necessary that such infection should al- 
ways be accompanied by manifestations of disease. It is true that, 
in most cases, the natural resistance is such that a struggle ensues 
by which the invader is destroyed or thrown off, or in which the 
invaded subject is functionally injured or even killed, and the ac- 
companying evidences of such a struggle constitute what we know 
as infectious disease. But there are special cases, cases of adapta- 
tion, biologically speaking, in which neither invader nor host is seri- 
ously harmed.7 In the field of protozoology, especially, there are 
many examples of true parasites, that is, invaders truly maintaining 
their metabolism at the expense of the tissues and body substances 
of the host, which do not arouse reactions sufficiently vigorous to be 
termed “disease.” Thus the Trypanosoma Lewisi may be found in 
the blood of rats 8 without noticeably affecting the health of the ani- 
6 Osterhout. Jour. Biol. Chem., 23, 1915, 67; Proceed. Amer. Philosophical 
Soc., 55, 1916, 533; Jour. Gen. Physiol., 3, 1920, 15, and 3, 1920, 145. 
7 See also Bail, loc. cit. 
8 Doflein. “Die Protozoen als Krankheitserreger.” 
* In this discussion, however, we must not forget that there are a number 
of diseases caused by micro-organisms which characteristically invade the cells 8 
INFECTION AND RESISTANCE 
mals, and other protozoa have similarly been found in organs and 
blood stream of a number of other apparently healthy animals. Al- 
though such conditions have been frequently spoken of as “infection 
without infectious disease/’ the distinction is probably one of degree 
only—there being some reaction on the part of the host even in the 
mildest cases, if only in the weakening by withdrawal of body sub- 
stance, which distinguishes the infected from the uninfected animal. 
In other cases there may even be advantage to the host, following the 
infection, to the detriment of the invading micro-organism, a phe- 
nomenon most clearly illustrated by the invasion of the root hairs of 
leguminous plants by the Nitrogen-fixing “root-tubercle” bacilli, a 
condition in which, as Fischer says, the plant may be regarded as 
parasitic upon the bacteria. 
In cases of so-called chronic septicemia in which bacteria may 
be again and again isolated by blood culture from the circulation it 
is more than likely that the organisms are constantly present, not 
because they multiply or maintain themselves within the circulation, 
but rather because they are being continuously discharged into the 
blood from an established focus in the tissues—as, for instance, on a 
heart valve. We have examined the serum of patients with subacute 
and chronic septicemia (endocarditis), and often found powerful 
opsonic action against the invading germs even when the patient’s 
own serum and leukocytes were used in the tests, evidence that the 
bacteria were probably being successfully disposed of after they had 
gained entrance into the blood stream. In rabbits, too, in our ex- 
perience and in that of Miss Gilbert in this laboratory, it would 
seem that protracted septicemia is present only when secondary foci 
have been established from which the bacteria are constantly being 
discharged into the blood. This we believe is rather the rule and 
the establishment of a balance within the blood stream an exception. 
When bacteria do succeed in withstanding successfully the opposing 
forces active within the circulating blood their rapid accumulation, 
the collapse of the defensive mechanism, and death of the patient are 
probably the most common course. 
The point of view which we have expressed in the preceding 
paragraph has been impressed upon us with particular insistence by 
the observation of certain cases of bacteriemia following infections of 
the middle ear, mastoid processes, and thromboses of adjacent veins. 
In such cases it appears that the blood may be flooded with bacteria 
which, nevertheless, disappear after the focus of infection has been 
of the body directly. Such is the case of the malarial plasmodia, with some of 
the Leischmannia, which cause Kala-Azar, which enter the endothelial cells 
of the spleen, liver and bone marrow, and is also characteristic of the Rickettsia 
bodies associated with typhus and trench fever. This also probably applies to 
some of the filtrable viruses, notably those of smallpox, encephalitis and 
herpes. THE PROBLEM OF VIRULENCE 
9 
removed. We have recently had the opportunity to observe this in 
a ease of septicemia caused by Streptococcus mucosus, in which blood 
culture plates showed very numerous colonies, and in which recov- 
ery followed promptly upon complete excision of the thrombosed 
veins. It would seem to us, therefore, that bacteriemia offers a 
rather better prognosis than was formerly supposed, at least in 
cases in which the focus is surgically accessible. 
The same principle is illustrated in the ordinary clinical course 
of typhoid fever in the human being. Here the disease begins as a 
bacteriemia. Very rapidly, usually within two weeks, the bacteria 
disappear from the blood stream and a high serum immunity is 
established in the patient. Nevertheless, the bacteria remain actively 
growing within definite foci in the tissues, where they are to a cer- 
tain extent protected or inaccessible to the defensive powers so suc- 
cessfully active in the blood stream. At any rate the patient remains 
diseased and the bacteria can be isolated from the spleen, gall-blad- 
der, and intestines at a stage when they are no longer present in the 
blood stream, and during which a measurement of the bactericidal 
and opsonic powers of the patient will reveal a serum immunity 
much higher than normal. Just why the organisms are protected 
from these influences in certain tissues and not in others we do 
not know. 
The Problem of Adaptation in Infection.—We have seen, then, 
that a micro-organism may be pathogenic and still be purely sapro- 
phytic in its mode of life. In order that this can occur, however, 
it is necessary that it should possess the power of producing at 
the place of lodgment a poison or toxin which can be absorbed and 
cause disease. The condition which ensues is not, properly speak- 
ing, an infection, but rather a “toxemia,” differing from the tox- 
emias resulting from the ingestion of drugs or other poisons only 
in so far as the toxins are manufactured at some point of bacterial 
lodgment within the body of the victim. Typical tetanus and diph- 
theria, for instance, can be produced as readily by injection of the 
bacteria-free culture filtrates as by inoculation with the bacteria 
themselves. And although these bacteria may, on occasion, become 
invasive and thereby satisfy the criteria of true infection, this is not 
necessary for their pathogenicity. 
In the large majority of bacterial diseases, however, it is neces- 
sary that the germs shall be capable of producing a true infection 
before they can become pathogenic, and it is our task therefore to 
attempt to analyze those bacterial attributes upon which the invasive 
power or virulence may be said to depend. 
In the realm of infectious micro-organisms a wide range of cul- 
tural variations is encountered which indicates that some of these 
germs have adapted themselves very closely to the specific environ- 
mental conditions found in the living animal body, while others can 10 
INFECTION AND RESISTANCE 
take up with ease and under the simplest cultural conditions a purely 
saprophytic existence. 
Many pathogenic micro-organisms have so far defied all attempts 
at cultivation in artificial media. These we cannot use for examples 
since it may well be that the failure of attempts in many of them 
may hinge upon such simple alterations of method as the exclusion 
of oxygen, the addition of fresh tissue, or the supplying of amino- 
acids, which have made possible the cultivation of the spirochseta 
pallida and the leprosy bacillus. But among those which we can 
cultivate there are many which require for successful cultivation 
the production of artificial conditions simulating closely those ob- 
taining in the living body. Thus malarial plasmodia can be made to 
multiply only if furnished with uninjured human red blood cells, 
within which they can develop. The gonococcus requires, in its 
first cultures outside the body, a medium containing human protein; 
and the hemophile bacteria, among them the influenza bacillus, re- 
quire hemoglobin. Other organisms like pneumococci, many strep- 
tococci, diphtheria bacilli, and many others, though easily grown on 
artificial media, are still fastidious in their requirements and de- 
velop sparsely or not at all unless definite conditions of nutrient 
materials, temperature, reaction, and osmotic pressure are observed. 
On the other hand, typhoid, anthrax, and dysentery bacilli, staphylo- 
cocci and numerous other pathogenic germs grow easily and luxuri- 
antly on the simplest laboratory media and within a wide range 
of environmental variations. 
Biologically considered, we could arrange the scale of adaptation 
to parasitic conditions on this basis and it would seem, a priori, that 
those bacteria which had thus adapted themselves most closely to the 
living body should be the most infectious. There is not, however, 
such parallelism, since many of the most powerfully invasive or 
virulent germs, for instance, the anthrax bacillus, have retained 
their capacity for saprophytic life to the fullest extent. It is more 
logical, therefore, to classify parasites, not according to their ability 
to revert to saprophytic conditions, but rather, as Bail 10 has done it, 
on the basis of their relative powers of invading the living body. 
His classification, of course, implies that the position of each micro- 
organism in this scale must be determined with reference to a given 
animal species, since a germ which is highly infectious (“parasitic” 
in Bail’s sense) for one species may be a “half-parasite” or even a 
pure saprophyte for another. 
Briefly reviewed, his classification is as follows: 
I. Pure Saprophytes.—(1STecroparasites, superficial parasites, 
or external parasites.) 
Micro-organisms which under no circumstances can be made to 
develop within the living tissues of a given animal. This does not 
10 Bail. Loc. cit. THE PROBLEM OF VIRULENCE 
11 
exclude their pathogenicity for this animal, since, like the diphtheria 
or tetanus bacillus, they may develop and produce toxins on the 
basis of a localized area of dead tissues. 
II. Pure Parasites.—Organisms like the anthrax bacillus or 
the bacilli of the hemorrhagic septicemia group which, implanted in 
small quantity in an animal, will rapidly gain a foothold, thrive, 
and spread throughout the body. 
III. Half parasites, organisms which may be infectious if in- 
troduced into the animal body, but, not possessing this invasive 
power to the same degree as the preceding class, require the inocula- 
tion of considerable quantities, often a special mode or path of in- 
oculation, or even possibly a preliminary reduction of the local and 
general resistance of the infected individual in order that they may 
multiply and become generalized. This class includes the large 
majority of the bacteria pathogenic for man. 
This property of invasive power is spoken of as virulence in 
contradistinction to toxicity—the latter implying merely the abil- 
ity to produce poisons, and not necessarily being associated with the 
power to invade. 
However we may classify these conditions (and it must be re- 
membered that a classification of complex scientific facts should 
never be regarded as a rigid definition of limitations, but merely 
as a convenient working basis), the facts of the matter are that 
many infectious micro-organisms, like, let us say, the treponema 
pallidum and the gonococcus, have adapted themselves so rigidly 
to the human body in cultural and other requirements, that they 
cannot be cultivated outside the body with any degree of success, 
and can be so cultivated only when, in nutritive and temperature 
conditions, the circumstances prevailing in the human body are to 
some extent simulated in the cultural procedure. This is par- 
ticularly true of many of the filtrable virus varieties and the pro- 
tozoan infecting agent. When such organisms can be cultivated, 
as, for instance, the gonococcus and the influenza bacillus, pro- 
longed cultivation upon artificial media gradually increases their 
ability to grow upon them. Conversely, with this increased ability 
to grow outside the body more easily and in simpler media, as a 
result of prolonged cultivation, the pathogenicity may often decrease. 
While both of these facts are true of individual cases, still they 
do not constitute a general rule. Many organisms, like the anthrax 
bacillus, typhoid bacilli, plague bacilli, cholera spirilla, and a host 
of others, can invade the bodies of one or more species of animals 
with great violence and still retain the capacity for a saprophytic 
existence to the fullest extent, are easily cultivated on artificial 
media and unless specific efforts are made to reduce their virulence, 
they retain their infectious power almost indefinitely if preserved 
on artificial media. Thus, virulent cultures of anthrax, of plague, 12 
INFECTION AND RESISTANCE 
of typhoid bacilli, etc., can he kept in sealed tubes, in the cold 
and in the dark, almost indefinitely, without losing either their 
ability to grow or their ability to cause specifically characteristic 
infection when again introduced into the susceptible animal. The 
evidences of special adaptation of different micro-organisms to par- 
ticular species of animals is dealt with extensively in the chapter 
on Natural Resistance. 
Of the greatest importance in illustrating the adaptation factor 
in infection is the ease with which fluctuations of virulence can be 
artificially produced in certain species of bacteria. The ease with 
which such variations can be produced would seem to furnish an- 
other argument in favor of the conception that the property of in- 
vasiveness is a biological attribute of relatively recent acquisition. 
Such variations are particularly noticeable in the cases of strep- 
tococcus and pneumococcus, organisms in which no two strains need 
be alike in infectiousness, and in which the injection of some strains 
into susceptible animals may produce no result whatever, while other 
strains will kill if administered in the smallest measurable quanti- 
ties. As a rule, the virulence of such strains can be enhanced for 
a given species of animal by successive passages through animals 
of this species. In the case of streptococci, strains which may not 
kill mice in quantities of a cubic centimeter or more, of an 18-hour 
broth culture, can, by many passages from mouse to mouse, be so 
enhanced that one-one hundred thousandth of a cubic centimeter of 
a similar culture will kill promptly. This cannot be done with all 
strains of streptococci, and of a dozen strains or so, one may succeed 
with two or three only. 
Among the earliest observations on this point are those of Pas- 
teur 11 in his work on rabies. He found that the virus of hydro- 
phobia when successively passed through rabbits gained in viru- 
lence until a degree of maximum infectiousness was attained beyond 
which it could no longer be enhanced. After only three passages 
through monkeys, however, the virulence of this “virus fixe” for 
rabbits was reduced almost to extinction. His experience with 
swine plague was similar. Swine plague bacilli successively passed 
through rabbits and pigeons gained enormously in virulence for 
these animals respectively, but lost in virulence for hogs. 
There are numerous methods, moreover, by which the virulence 
of micro-organisms can be attenuated by laboratory manipulations, 
and since many of them are of great importance in the active 
immunization of animals we will reserve their detailed discussion 
until we come to consider the methods of immunization themselves. 
Suffice it to say in this place that most methods of attenuation 
consist in subjecting the bacteria, in artificial culture, to deleterious 
influences, either of unfavorably high temperature, exposure to light 
i1 Pasteur and Thuilljer, Compt. rend, de Vacad, des. scv Yol. XII, 1883. THE PROBLEM OF YIRULENCE 
13 
or harmful chemical agents, or allowing them to remain in pro- 
longed contact with the products of their own metabolism by in- 
frequent transplantation. As a rule the attenuation which in- 
evitably follows any form of artificial cultivation in the case of 
bacteria like streptococci or pneumococci can be delayed by pre- 
serving them in media containing sera or tissues. In the case of 
the pneumococcus, for instance, one of the best methods of con- 
serving virulence in storage is to keep them either in a soft rabbit- 
serum-agar mixture, as practiced by Wadsworth, or, better still, to 
store them within the spleen of a mouse dead of pneumococcus 
infection, as recommended by Heufeld. The mouse is autopsied 
and the spleen kept in the dark and cold in a desiccator, under 
sterile precautions. This, again, as well as the enhancement of 
virulence on passage through the same species of animal—or the 
reduction of virulence for one species by passage through another 
•—shows that such fluctuations are dependent upon a very delicate 
biological adaptation. 
It is interesting, moreover, to look upon this process of adapta- 
tion as a sort of immunization of the bacteria against the defensive 
powers of the host, a conception early suggested by Welch. For just 
as the animal body may become more resistant to the offensive 
weapons of the invaders, so it is reasonable to suppose that the bac- 
terial body may gradually develop increased resistance to the de- 
fensive mechanism of the host. And this, if it occurs, would of 
course lead to an increase of its invasive power or virulence. The 
increase of virulence by passage through animals would alone lead 
us to suspect that such acquired resistance to destructive agents on 
the part of the bacteria might be responsible for the enhancement, 
but additional evidence pointing in this direction has been brought 
by experiments in which it was shown that bacteria cultivated in the 
serum of immune animals not only gained in resistance to destruc- 
tion by the serum constituents, but at the same time were rendered 
more highly pathogenic. Experiments of this kind were carried out 
by Sawtchenko,12 by Danysz,13 and by Walker.14 The results of 
Walker are especially instructive. He worked with a typhoid ba- 
cillus which he cultivated for a number of generations upon the 
serum of a typhoid-immune animal, and found that after such treat- 
ment the organism had gained in virulence and lost in agglutin- 
ability by immune serum, and that a larger amount of specific 
immune serum was necessary to protect animals against it than 
sufficed for protection against normal typhoid strains not thus 
cultivated. We will refer to these results in a later chapter, since 
the conception will be easier to grasp when we have considered more 
12 Sawtchenko. Ann. Past., Yol. 11, 1897. 
13 Danysz. Ann. Past., 14, 1900. 
** W&lker, Jour, of Path, and Bact,} Yol, 8, 1903, 14 
INFECTION AND RESISTANCE 
fully the mechanism of defence by which the animal body is pro- 
tected against invasion. 
That this power of gaining resistance against deleterious influ- 
ences on the part of bacteria is not confined to their resistance to the 
animal defences alone is well shown by the experiments of Danysz 13 
upon the immunization of anthrax bacilli against arsenic. In in- 
oculating series of 50 tubes containing arsenic dilutions (ranging 
from 1 to 10,000 to 1 to 200) with anthrax bacilli Danysz found 
that up to 1 to 5,000 the arsenic increased the growth of the bacilli; 
in concentrations higher than this growth was inhibited. By grad- 
ually progressive cultivation of the organisms in increasing concen- 
trations of arsenic he finally succeeded in obtaining growth in solu- 
tions five times more concentrated than those in which they would 
develop at first. 
It is intensely interesting also that Danysz found, both in the 
case of his serum-resistant and arsenic-resistant strains, that, as 
they became less sensitive to the deleterious effects of these agencies, 
they were altered morphologically in that they developed capsules. 
Similar in significance to this is the very important observation that 
certain strains of spirocliseta pallida may acquire resistance against 
salvarsan or “606.” 16 These so-called arsenic-fast strains are ap- 
parently unaffected by the injection of this preparation into the 
patient. 
Relationship of Infectiousness to Capsule Formation.-^-The ex- 
periments of Danysz were probably the first to call attention to the 
possible relationship of bacterial capsule formation to virulence, 
and this particular phase of the subject has since then been ex- 
tensively studied. It is a matter of common observation that micro- 
organisms like the pneumococcus, the anthrax bacillus, some 
streptococci, and a number of other germs which are capable of pro- 
ducing capsules under suitable conditions are most virulent in the 
capsulated stage. As the strains are passed through animals and 
their virulence increases their ability to form capsules becomes more 
and more apparent—whereas the diminution of virulence which 
takes place on artificial media is accompanied by a gradual loss of 
capsule formation. Organisms like the Friedlander bacillus which 
retain their ability to form capsules almost indefinitely in artificial 
culture moreover do not lose their virulence to any great extent as 
long as this property is preserved. It is also well known that cap- 
sulated bacteria are peculiarly insusceptible to the ordinary aggluti- 
nating powers of specific immune sera. This has been noticed, not 
only in the case of heavily capsulated bacteria like those of the Fried- 
lander group or the streptococcus mucosus, but, in the case of plague 
bacilli—where capsulation is usually present only in cultures taken 
15 Danysz. Loc. cit. 
16 Oppenheim. Wien. hi. Woch., 23,1910, No. 37. THE PROBLEM OF VIRULENCE 
15 
directly from the animal body and cultivated at 37° C., Shiba- 
yama 17 has found a direct relation between non-agglutinability and 
a slimy condition of the cultures. Cultures kept at 5° to 8° C. in 
the ice-chest were easily agglutinable and lacked the slimy property. 
Cultures kept at 37.5° C. were slimy and thready in consistency and 
were not as easily agglutinated by the same immune serum. 
Porges 18 later showed that inagglutinable, capsulated bacteria can 
be made amenable to the agglutinating action of the serum—which 
we may assume to indicate vulnerability by the serum if the capsule 
is previously destroyed by heating at 80° C. for about 15 minutes in 
*4 normal acid. 
Against the cellular defences, the leukocytes, capsulated bacteria 
seem also to be more resistant than are the non-capsulated. This 
has been especially studied by Gruber and Futaki,19 who find that a 
capsulated bacillus is rarely taken up by a phagocyte even when 
these cells are apparently normal and able to take up the uncap- 
sulated organisms. They go so far as to claim that, in the case of 
anthrax in rabbits, the development or absence of a capsule deter- 
mines whether or not infection can take place. The same conclusion 
is reached in similar studies by Preisz,20 who does not believe that 
anthrax bacilli can ever cause infection unless they possess the 
power of forming capsules. All this experimental evidence points 
strongly toward a probable direct relationship between capsule for- 
mation and virulence, in the sense that a thickening of the ectoplasm 
may in some way protect the bacteria from the destructive forces 
aimed at them by the cells and fluids of the invaded body. 
As a matter of fact, even when no distinct capsule is visible, it 
is nevertheless possible that ectoplasmic changes may take place. 
This phase of the subject has been thoroughly discussed by a number 
of writers, more especially by Eisenberg.21 It appears that many 
bacteria, in which true capsule formation has not been observed, may 
show swelling or enlargement under conditions in which their of- 
fensive activities in the infected animal body are called into play.22 
Radziewsky 23 has noticed such swelling of B. coli in fatal guinea- 
pig infections, and spoken of it as “one of the characteristic signs 
of infectiousness.” Kisskalt 24 has described the same thing in the 
case of streptococci, and Eisenberg interprets this as signifying an 
ectoplasmic hypertrophy comparable in principle to capsule forma- 
tion. He looks upon the ectoplasmic zone as a protective layer, and 
17 Shibayama. Centralbl. f. Bad., Orig. Vols. 38, 1905, and 42, 1906. 
18 Porges. Wien. klin. Woch., p. 691, 1905. 
19 Gruber and Futaki. Munch. med. Woch., 6, 1906. 
20 Preisz. Centralbl. f. Bakt., Yol. 49, 1909. 
21 Eisenberg. Centralbl. f. Bakt., I, 45, 1908, p. 638. 
22 These forms Bail has spoken of as “thierische Bazillen.” 
23 Radziewsky. Zeitschr. f. Hyg., Yol. 34. 
24 Kisskalt. Cited after Eisenberg, loc. cit. 16 
INFECTION AND RESISTANCE 
calls attention to the observation of Liesenberg and Zopf,25 who 
showed that capsulated strains of leukonostoc mesenteroides will 
withstand 85° C., a temperature at which uncapsulated forms are 
rapidly killed. 
Upon this point our own recent work has seemed to us of con- 
siderable interest, although we cannot yet prove our point. In 
another part of this book (see page 110), we have described sub- 
stances that are probably not whole proteins, which are rapidly 
exuded from all bacteria that we have examined, into the surround- 
ing fluid. These so-called residue antigens, in spite of their ap- 
parent low molecular structure, react powerfully with antibodies 
and represent, we believe, metabolic products which ooze out from 
the bacteria as the result of their own chemical exchange. From 
bacteria that are capsulated, like the pneumococcus, from menin- 
gococci which are difficult to subject to agglutination, these sub- 
stances can be obtained by simply washing the bacteria. In the 
case of other bacteria, like typhoid bacilli, tubercle bacilli, etc., it 
is necessary to grind and extract the bacteria in order to obtain 
them. It is our opinion that in capsulated bacteria these substances 
form a protective layer about the bacterial cell which, because of its 
ability to unite with antibodies, can divert these sensitizing sub- 
stances from the bacterial cell itself. 
Acquired Resistance to Antibody Effects.—Pertinent, also, to 
this discussion of adaptation 'is work in which the actual re- 
sistance of micro-organisms to reaction with antibodies has been 
shown to result from prolonged contact of the bacteria with blood 
or exudate from animals or human beings or with immune sera. 
This seems to be shown especially by the experiments of Walker 
and Morishima, 25a which are referred to in other places, in which 
it was found that bacteria grown on immune sera gain a certain 
amount of resistance against the injurious properties of these sub- 
stances, and evidence more directly bearing upon the question is 
furnished by the studies on the typhoid carrier state in rabbits 
made by Chirolanza,26 Blackstein,27 Johnston,28 and recently by 
Gay and Claypole.29 The last-named writers found that they could 
regularly produce the typhoid carrier state in these animals if they 
first cultivated the typhoid bacilli upon a medium containing de- 
fibrinated rabbits’ blood. Even in these cases, however, it is not 
at all improbable that the typhoid bacilli establish a permanent 
focus from which they are discharged into the blood stream. 
25 Liesenberg and Zopf. Centralbl. f. Bakt., Yol. XII, 1892. 
25" Morishima. Jour, of Bacter., 6, 1921, 275. 
28 Chirolanza. Ztschr. f. Hyg., Vol. 62,1909. 
27 Blackstein. Bull. Johns Hop. Hosp., 1891. 
28 Johnston. Journ. Med. Bes., 27, 1912. 
29 Gay and Claypole. Arch, of Int. Med., 1.2, 1913. THE PROBLEM OF VIRULENCE 
17 
It is not likely, however, that this merely passive increase of the 
resistance to injury on the part of the bacteria accounts for the 
entire train of phenomena included in an enhancement of virulence. 
It has been suggested by a number of observers that definite active 
offensive characteristics distinguish the virulent from the avirulent 
bacteria, in that the former may secrete, within the living body, 
substances by which the destructive powers of serum and leukocytes 
are neutralized or held at bay. A very definite suggestion of such a 
possibility wTe find expressed in the now classical paper of Salmon 
and Smith 30 on hog cholera immunity, published in 1886. They 
say: “. . . the germs of such maladies are only able to multiply in 
the body of the individual attacked, because of a poisonous principle 
or substance which is produced during the multiplication of these 
germs.” 31 Bouchard formulated such a theory in 1893 by speaking 
of the “produits secretes par les microbes pathogeniques,” substances 
which he found in cultures of virulent bacteria, and which seemed 
to reenforce the invasive powers of the germs. Kruse 32 also within 
the same year developed a similar idea. He assumed that bacteria 
may secrete enzyme-like substances which paralyze the destructive 
properties of animal serum, and in this way gain the power to 
invade. As a matter of fact we have learned, since that time, that 
staphylococci may secrete soluble substances, “leukocidins,” which 
injure white blood cells, and that many bacteria produce similar 
poisons, “hsemotoxins,” which specifically injure red blood cells—- 
thereby causing amemia and reducing the resistance of the host. 
However, the correlation and further elaboration of these thoughts 
of Salmon and Smith, of Bouchard and of Kruse was left to Bail,33 
in what is known as his “aggressin theory.” Bail maintains on the 
basis of careful experimentation that virulent bacteria can produce 
within the animal body substances which he calls “aggressins,” upon 
which depend their invasive powers or virulence. These substances 
are secreted only under stress of the struggle against the unusual 
defences, are not demonstrable in test-tube cultures, and are in 
themselves, according to Bail, entirely non-toxic. 
He obtains these aggressins by injecting virulent bacteria into 
the peritoneal cavity of a guinea pig and immediately after death 
removing the exudate. This he centrifugalizes, removes the bac- 
teria and cells, and sterilizes the supernatant liquid by the addition 
of small quantities of chloroform. The action of the exudates in 
which aggressins have been produced by the bacteria is the fol- 
30 Salmon and Smith. Proc. Biol. Soc., Washington, D. C., Ill, 1884, 
6, p. 29. 
31 A typewritten copy of this paper was kindly put at my disposal by 
Prof. Theobald Smith. 
32 Kruse. Ziegler’s Beitrage, Yol. XII, 1893. 
33 Bail. Archiv f. Hyg., Vols. 52 and 53, 1905; Folio serologica, Yol. 7, 
1911. 18 
INFECTION AND RESISTANCE 
lowing: (We take this tabulation from Bail’s own paper on typhoid 
and cholera aggressins in the Archiv filr Hygiene, Vol. 52, p. 342.) 
1. Sublethal doses of typhoid bacilli or cholera spirilla become 
lethal when the aggressin is injected with them. 
2. Lethal doses of bacilli which ordinarily would cause a slow 
infection only cause a rapid and severe infection when aggressins 
are added. 
3. The addition of aggressin neutralizes the bacteria-destroy- 
ing power of immune serum in the peritoneal cavity of a guinea pig. 
4. The injection of aggressin alone produces subsequent im- 
munity. 
It is impossible to discuss with completeness the arguments ad- 
vanced for and against the correctness of Bail’s views until we have 
described in detail the mechanism of protection at the disposal of 
animals. But the main objection brought against this theory is that 
of Wassermann and Citron,34 who claim that all these properties of 
the aggressive exudates can be explained by the fact that they con- 
tain extracts of the bacteria (endotoxins), which, injected with 
a sublethal dose of bacteria, merely enhance their action in the same 
way that this would have been accomplished by the injection of 
additional dead bacterial bodies. It will require much further work 
before this point is settled, and the problem is peculiarly involved 
and difficult. However, the recent work of Rosenow 35 on pneumo- 
nococci seems to bring some reenforcement to the ranks of those who 
maintain the existence of a special offensive substance at the com- 
mand of virulent bacteria. Rosenow extracted pneumococci grown 
on serum broth and found that such extracts when made from 
virulent strains would protect avirulent strains from engulfment 
by phagocytes. The non-virulent strains left in these extracts for 
24 hours became virulent. He believes, therefore, that the viru- 
lence of pneumococci depends largely upon the possession of these 
substances which he calls “virulins,” and which in function at least 
are conceived as very similar to the “aggressins.” 
Recent results obtained by the writer36 with Dwyer seem to 
indicate that anaphylatoxins produced from the typhoid bacillus 
possess some of the properties claimed for his aggressin by Bail. 
It is not impossible that the “aggressins” obtained by him were of 
this nature. 
Virulence, then, may be analyzed into two main attributes: one a 
purely passive property of resistance or self-preservation on the 
part of the bacteria, perhaps morphologically expressed in ectoplas- 
mic hypertrophy and capsule formation; the other an actively of- 
fensive weapon in the form of substances of the nature of the “ag- 
34 Wassermann and Citron. Deutsche med. Woch., Vol. 31, 28, 1905. 
35 Rosenow. Jour, of Inf. Dis., Vol. 4, 1907. 
36 Zinsser and Dwyer. Proc. Soc. Exp. Biol, and Med., Feb., 1914. THE PROBLEM OF VIRULENCE 
19 
gressins” of Bail or the “virulins” of Rosenow. The extent of our 
present knowledge of details does not warrant a statement of the 
case in more definite terms. 
From what has been said above, then, it appears that mutual 
adaptation in the biological sense between the invading germs and 
the invaded body must play an extremely important role, not only 
in determining whether or not an infection is to take place, but 
also in influencing the degree of infection, its acuteness or chro- 
nicity, and through these facts, the eventual outcome. That many 
different types of adaptation may occur and that the balance struck 
must depend to some extent upon the relative speed with which 
either the invaders or the invaded can modify their biological prop- 
erties to the new conditions, is obvious. Theobald Smith 37 has 
given this problem the most thorough analysis. A micro-organism 
that has been closely adapted to the human body may have de- 
veloped particularly its offensive and defensive powers to such a 
degree that when it enters a new human body, it proceeds on its 
invasive course with speed and violence, and the infected tissues 
react with a powerful effort, to rid themselves of the foreign in- 
vader. The result is violent and acute inflammation and disease. 
This violent reaction on the part of the body would be evidence of 
a mobilization of protective functions to a living foreign substance' 
to which no adaptation has taken place, and which the body attempts 
to get rid of. As long as this reaction on the part of the bacteria 
carries the upper hand, the increase in virulence will continue. 
Conversely, however, as Theobald Smith points out, the chronicity 
of a disease is largely dependent upon an adaptation between host 
and invader in which the reaction to the parasite is less violent, 
and in which a condition approaching more and more closely to a 
sort of symbiosis, is established. As Smith points out in his re- 
cent publication, “there is a struggle on the part of the parasites 
to adapt themselves and to establish some equilibrium between them- 
selves and their host”; and, again, “the final outcome is a harmless 
parasitism or some disease of little or no fatality, unless other 
parasites complicate the invasion.” 
Such an adaptation probably takes place. On the other hand, it 
is not likely to occur unless from the beginning the balance is 
such that host and invader may be in contact for a period longer 
than duration of the ordinary acute infectious disease. In cases 
where the organism is either rapidly gotten rid of, or in which acute 
death ensues within a short time, such developments can hardly be 
expected. However, through gradual transitional stages and many 
successive infections, the condition might be produced from an 
originally acute form of disease to a more subacute or chronic one. 
37 Smith, Theobald. Trans. Soc. Amer. Phys., 1921, see also Jour. A.M.A., 
May, 1913, 60. 20 
INFECTION AND RESISTANCE 
Such a conception would assign the slow and gradual but pro- 
gressively invasive powers of such diseases as tuberculosis, leprosy, 
and syphilis in which systemic symptoms are manifest only after 
the disease has gained an extensive foothold, to the lack of acute 
physiological reaction resulting from the presence of the invading 
micro-organism. In the case of such infections as those caused by 
some of the yeasts or blastomyces we have seen foci of blastomycotic 
lodgment in the kidney and other organs surrounding which there 
was neither an accumulation of mobile cells—(leukocytes or lympho- 
cytes)—nor any evidence of cloudy swelling or other injury, by 
poisons, of adjacent parenchyma cells. Here, as in tuberculosis or 
leprosy, the reaction induced by the presence of the micro-organisms 
is slow and gradual—expressed in an eventual fixed tissue-cell reac- 
tion and giant-cell formation—similar to that induced by insoluble 
foreign bodies. And it may well be that the progressive ability to 
multiply without arousing the invaded body to rapid and powerful 
reaction may account for the prolonged period of apparent well- 
being in the early stages of such infections and permit the invaders 
to pervade the body so extensively. 
Criteria which determine Actual Infection.—In order that a 
micro-organism may be a true parasite in Bail’s sense—or invasive 
—for any given species of animal it must of course possess certain 
basic cultural attributes which enable it to grow in the environ- 
ment furnished by the host. For instance, a micro-organism which 
does not grow at temperatures below 37.5° C. cannot very well be- 
come parasitic upon cold-blooded animals. An excellent illustration 
of this influence of body temperature upon the invasive powers of 
bacteria is furnished by the different races of acid-fast bacilli which 
invade the bodies of man and of birds. The avian tubercle bacillus, 
for instance, is non-pathogenic for man and in cultures will not 
develop at temperatures below 40° C., which is about the body 
temperature of most birds. The human tubercle bacillus, on the 
other hand, is non-pathogenic for birds and ceases to grow in arti- 
ficial cultures when the temperature is raised above 40° to 41° C. 
This is merely one of a number of examples which might be cited 
to demonstrate the necessity of simple cultural adaptation, as it 
influences the property of virulence. Again, it is probable that 
in order to develop in the animal body it is necessary that a micro- 
organism shall be capable of developing with little or no free oxy- 
gen. While this point is not definitely certain, it is not probable 
that any of the virulent bacteria can be strict aerobes. As a matter 
of experience none of the pathogenic bacteria at present known are 
absolute aerobes—though many of them grow better in artificial cul- 
ture when oxygen is freely present than when it is absent. 
The invading bacteria, then, must be culturally and in other re- 
spects adapted to development in or upon the tissues of the invaded THE PROBLEM OF VIRULENCE 
21 
subject—a general capacity which is prerequisite to the property of 
“pathogenicity 
Furthermore, the conditions encountered by bacteria as they 
enter the animal body will vary considerably according to the path 
by which they gain entrance. Organisms entering by the intestinal 
canal are subjected to conditions of acidity or alkalinity, the action 
of digestive juices, of bile, and to competition with other intestinal 
bacteria, forces to which many pathogenic germs will succumb, while 
others may survive there and thrive. Those entering into the tissues 
by way of the skin and mucous membrane, on the other hand, en- 
counter an immediately mobilized protective mechanism which, suc- 
cessfully resisted by some of them, might easily and quickly dispose 
of small quantities of other bacteria more resistant to conditions in 
the bowel. It is but natural for this reason that the accomplishment 
of an infection by any. given germ must depend to a great extent 
upon its gaining entrance to the body by the path best adapted to its 
peculiar requirements. 
The mechanical protection afforded by the coverings of skin and 
mucous membranes is as a rule sufficient to prevent the penetration 
of any bacteria which by chance may have found lodgment upon 
them. In the case of the most usual pyogenic cocci and many bacilli 
such protection is probably absolute, and a distinct break of con- 
tinuity, such as a bruise or a wound, even though this may be too 
small to attract attention, is necessary for successful infection. In 
the case of a very limited number of diseases infection seems to take 
place even through the unbroken skin, and the method, often spoken 
of as the vaccination method of Kolle, employed in many instances 
when it is desired to produce experimental plague infection in rats 
or guinea pigs, consists in merely rubbing a small amount of cul- 
tural material into a shaven area of the skin. However, in this 
case, as well as in other instances where mere massage of bacteria 
into unbroken skin has led to successful inoculation, it is more 
than likely that success has depended upon either microscopic 
lesions or possibly the violent introduction of the organisms into the 
sebaceous glands, the sweat glands, or hair follicles. The defence 
of intact mucous membranes, however, is by no means impervious. 
While many organisms can be implanted upon mucous membranes 
with impunity, there are a number of others that can cause local 
inflammations upon these and can further pass through them into the 
deeper tissues and thence into the general system. Thus gonorrhea 
is ordinarily a disease of implantation upon a mucous membrane, 
and diphtheria bacilli and streptococci give rise to localized disease 
on the pharyngeal and nasal mucosse, the latter not infrequently 
penetrating from the initial point of lodgment upon the mucosa into 
the deeper tissues and the circulation, causing a condition of “septi- 
cemia” or “bacteriemia.” For the experimental determination of 22 
INFECTION AND RESISTANCE 
the penetrative power of organisms through mucous membranes the 
conjunctiva has been a favorite test object, and it has been shown 
that plague 38 and glanders,39 as well as hydrophobia, may be trans- 
mitted by simple instillation of infectious material into the unin- 
jured conjunctival sac. In the case of hydrophobia 40 it is related 
that in Paris a young man contracted hydrophobia by rubbing his 
eyes with a finger contaminated with the saliva of a rabid dog. In 
the case of syphilis, though often claimed, there is no positive proof 
to show that infection may take place through the uninjured sur- 
faces. It has been definitely shown, however, that tubercle bacilli 41 
may pass into the lymphatics through the intestinal mucosa without 
there being any traceable injuries on this membrane. 
It may well be, however, that even without the existence of 
demonstrable morphological lesions penetrability by micro-organisms 
may presuppose local physiological or functional injury, such as 
congestion or catarrhal inflammation. 
Thus it is seen that the mechanical obstacle to the entrance of 
micro-organisms offered by skin and mucous membranes, though 
important and not to be underestimated, is by no means a perfect 
safeguard. 
However, it is only very definite species of micro-organisms 
which can cause disease at all when introduced into the body by 
these paths. For, although the rubbing of plague bacilli into the 
skin, or the inoculation of a cut surface with streptococcal or glan- 
ders bacilli, will rapidly lead to progressive infection, similar inocu- 
lation with the typhoid bacillus or the cholera spirillum would lead 
to no such result. And, though the swallowing of pus cocci, pneu- 
mococci, and a number of other micro-organisms would be entirely 
without effect, similar ingestion of the typhoid and cholera organism 
would usually result in typical infection. 
The path of introduction, therefore, is an important considera- 
tion in determining whether or not a given micro-organism may give 
rise to disease. It is necessary that the manner of gaining entrance 
be suited to the cultural and other peculiarities of the germ in ques- 
tion. In the case of cholera, for instance, the spirillum which causes 
this disease is peculiarly susceptible to the deeper defences residing 
in the body fluids and cells, and cutaneous infection by the small 
numbers of bacteria likely to be introduced in this way would 
promptly be checked by these agencies. In the intestinal mucosa, 
however, the cholera spirillum finds conditions most favorable for 
rapid multiplication and the disease is caused by the inflammation 
and destruction of the mucous and submucous tissues by the poison- 
38 Germ. Plague Com. Arb. a. d. kais. Gesundheitsamte, Yol. 16, 1899. 
39 Conte. Rev. veterin., Yol. 18, 1893. 
40 Galtier. Compt. rend, de la soc. biol., 1890. 
41 Bartel. Wien. Klinikhandt, 1906-1907. THE PROBLEM OF VIRULENCE 
23 
ous substances emanating from the large numbers of cholera spirilla 
which die and are disintegrated, as well as by the absorption of these 
poisons into the circulation. The bacteria themselves, however, never 
gain a permanent foothold within the blood or other organs. In the 
case of typhoid fever the conditions are somewhat similar, although 
here, during the earlier weeks of the disease, we have an actual, 
penetration of the bacilli into the circulation. This, however, prob- 
ably takes place only after intraintestinal proliferation has taken 
place, which then, on the injured mucosa, represents a dose out of 
all proportion great when compared with the quantities that would 
spontaneously come into contact with the external surface of the 
body. 
This leads us to another important factor concerning the invad- 
ing forces, in the determination of successful infection, namely, that 
of the quantity introduced or the dosage. 
In order to cause infection, even when the bacteria are of the 
variety known to produce disease or “pathogenic,” and are brought 
into contact with the body by a path suitable to their peculiar re- 
quirements, the initial quantity introduced must be sufficiently large 
to preclude complete annihilation by the first onslaught of the de- 
fensive powers of the body. It is plain, therefore, that in the case 
of bacteria weak in power to cause disease, given the subject of in- 
fection and his defences as a constant, the quantities to be introduced 
must be larger than in the case of micro-organisms of violent disease- 
producing properties. The dosage necessary to cause infection, 
therefore, is in inverse proportion to that property of bacteria spoken 
of as their “virulence.” Thus we measure the degree of the so-called 
virulence of bacteria by determining the smallest quantity, measured 
by dilution of platinum loops or by fractions of agar slant cultures 
(both very inexact methods), which will still cause infection and 
death in susceptible animals of a standard weight. In the case of 
micro-organisms of extreme virulence, such as the anthrax bacillus 
or bacilli of the hemorrhagic septicemia group, the inoculation of a 
very small number of bacteria may suffice to initiate infection. In7 
deed, it has been claimed for the anthrax bacillus that the injection 
of a single bacterium will produce fatal disease in a susceptible 
animal. The inverse relation existing between the degree of viru- 
lence and the number of bacteria inoculated is well illustrated by the 
experiments of Webb, Williams, and Barber,42 carried out upon 
white mice with anthrax, by the method of inoculation devised by 
Barber.43 This technique consists in picking up single organisms 
with a capillary pipette under microscopic control, from a very thin 
emulsion of bacteria and injecting directly from the pipette through 
a needle puncture in the skin. While requiring a considerable de- 
42 Webb, Williams, and Barber. Jour. Med. Res., 1909, Yol. XV. 
43 Barber. Kansas Univ. Science Bulletin, March, 1907. 24 
INFECTION AND RESISTANCE 
gree of skill, the method, when successful, permits an actual accurate 
count of injected bacteria instead of the merely approximate esti- 
mate which can be made by consecutive dilutions of thicker emul- 
sions. In their experiments with anthrax in white mice Webb, Wil- 
liams, and Barber found that the inoculation of a single thread of 
anthrax bacilli (3 to 6 individuals) taken directly from the blood 
of a dead animal (that is, in the most virulent condition) would 
regularly cause death, and it was impossible for this reason to 
immunize with such bacilli. On the other hand, if taken from 
12-hour agar cultures of the same strain such small quantities 
would often fail to kill. The brief period of growth under arti- 
ficial conditions had sufficiently lessened the virulence of the bacilli 
so that 2, 3, and more threads could be injected without harm. 
And after several generations of such cultivation as many as 27 
and more threads could be inoculated with impunity. 
Another example of the measurement of relative degrees of 
virulence, by a method more commonly employed, may be illustrated 
as follows: The problem in which this particular measurement was 
used consisted in the comparison of the virulence of two strains of 
pneumococcus, one (N2) successively passed through white mice, the 
other (Nt) kept alive for several weeks on serum-agar. To accom- 
plish this graded quantities of 18-hour broth cultures of the two 
strains were injected into mice of approximately the same weight, 
as follows: 
Result 
0.1 e. c. “ dead 24 hrs. 
0.05 c. c. — lives 
0.02 c. c. — lives 
0.01 e. e. — lives 
N2 Result 
0.1 c. c. = dead 24 hrs. 
0.05 c. c. — dead 24 hrs. 
0.02 e. e. rr dead 24 hrs. 
0.01 c. e. = lives 
Accidental Factors Favoring Infection.—In discussing the prob- 
lems of virulence we must not forget that there are many acci- 
dental factors which may make an infection likely where, other- 
wise, it might not have occurred. Trauma is perhaps the most 
important of these. It is well known that injury in which tissue 
is killed will favor the development of streptococci and staphylococci, 
which in the same dose and with the same degree of virulence en- 
tering into wounds incised with a minimum tissue destruction, might 
have been easily overcome by the rapidly mobilized defences of 
the body. In tetanus and other anaerobic infections, this is, of 
course, a well-known state of affairs, and has been again and again 
proved in connection with war wounds. In these infections, more- 
over, the additional danger of concomitant infection with other 
organisms is of importance. If, for instance, staphylococci are in- THE PROBLEM OF VIRULENCE 
25 
jected at the same time with tetanus spores, the chances of the 
development of tetanus are vastly increased, as has been shown by 
workers in the United States Hygienic Laboratory. There are 
certain epidemic infections, too, which are more dangerous in con- 
nection with their ability to prepare the way for secondary invasion, 
than they are in themselves. Most important among these are 
measles and influenza. Pure influenza is a mild disease, but an 
influenzal infection of the respiratory tract seems to render the 
subject susceptible to secondary invasion with streptococci and 
pneumococci to an extreme degree, a fact which is responsible for 
by far the greater part of the mortality in influenza epidemics. 
This is almost equally true of measles, where the disease itself is 
rarely fatal, but prepares the way for fatal respiratory infection 
with streptococci and pneumococci, particularly, and perhaps for 
subsequent tuberculosis. To a less marked degree, the same thing 
is true of whooping cough where secondary influenza bacillus, pneu- 
mococcus and streptococcus invasion may follow very promptly upon 
the initial Bordet-Gengou infection. A considerable number of 
analogous examples could be cited. 
Types of Infection.—From the facts we have discussed in the 
preceding paragraphs it now becomes manifest that the elements 
which determine the nature of an infectious disease are twofold. 
On the one hand each variety of infectious germs possesses certain 
biological and chemical attributes which are specific and peculiar 
to itself; by these its predilection for path of entrance and mode 
of attack is determined, and upon these depends the nature of 
the reaction called forth in the animal body. On the other hand 
the degree of infection in each case, the severity of the reaction and 
the ultimate outcome are determined by the balance which is struck 
between the virulence of the entering germ and the protective mechan- 
ism opposed to it. 
The specific properties of each micro-organism are the factors 
which account for the clinical uniformity (within definite limits) 
which is observed in the maladies produced in different individuals 
by the same species of bacteria. Thus a severe typhoid fever is, in 
essential characteristics, entirely similar to a mild case—since in 
both instances the path of entrance, through the intestine, is the 
same, the distribution of the germs after entrance differs only in 
degree, and the reactions, local and systemic, which are called forth 
are alike. And cases of this disease in general differ as a class from 
the maladies caused by, let us say, the group of clinical conditions 
resulting from anthrax infection, where entrance is through the 
skin, and generalized infection of the blood ensues without definite 
or regular localization in any given organ. Again, a localized 
staphylococcus abscess will differ materially from an equally local- 
ized focus of tuberculosis, because the chemical constituents of these 26 
INFECTION AND RESISTANCE 
bacteria respectively call forth each a characteristic response on the 
part of the defensive mechanism. 
Such specificity of the various micro-organisms may of course 
be due partly to their mode of attack and distribution, and partly, 
as we shall see, to the pharmacological action of the poisonous prod- 
ucts given out by them. 
That both factors contribute seems beyond doubt; but recent 
work, especially that of Friedberger, which is fully discussed in 
another place indicates the possibility that clinical differences 
depend much less than was formerly supposed upon specificity of the 
intracellular poisons, and much more upon distribution and localized 
accumulation of the germs, conditions which are determined rather 
by the mode and extent of invasion than by chemical differences of 
poison production. This problem, rather difficult to discuss on the 
limited basis of the facts so far outlined, will become clearer as we 
proceed, but we need only refer at present to the essential clinical 
uniformity of the various forms of septicemia, where organisms 
freely circulate in the blood—with often a focus of distribution on a 
heart valve—conditions in which it is rarely possible to determine 
the species of the responsible germ except by blood culture. Or, 
again, as Friedberger 44 points out, there is great similarity between 
the ordinary pneumococcus pneumonia and that caused by the Fried- 
lander bacillus. In both cases the distribution and mode of attack of 
the bacteria are essentially the same, though the micro-organisms 
themselves are biologically very dissimilar. 
One and the same micro-organism, on the other hand, may cause 
entirely different clinical conditions, and here the type of infection 
depends purely on the degree of invasion possible in the given case—• 
that is, the balance between virulence and resistance. A germ may 
enter the body and cause an inflammatory reaction at the point of 
entrance, the process remaining purely localized. In such cases the 
defensive forces have been so efficient, the invasive properties of the 
germ so relatively weak, that progression beyond the point of en- 
trance is prevented and the resultant disease takes the form merely 
of a localized abscess. This is the case when a healthy individual 
is infected with an attenuated organism or by one whose species’ 
characteristics do not include a powerful invasive property. Thus 
streptococci, if entering the tissues of a normal subject in small 
numbers or in attenuated form, may produce a purely localized in- 
fection, and ordinarily non-pathogenic germs like proteus, subtilis, 
or colon bacilli may produce localized abscesses in weak and debili- 
tated individuals, though implanted upon a healthy subject they 
would be rapidly disposed of without gaining even a preliminary 
foothold. Such tendency to localization is the common form of in- 
fection in the case of a number of germs. It is the most usual type 
44 Friedberger. Deutsche med. Woch., No. 11, 1911. THE PROBLEM OF VIRULENCE 
27 
of staphylococcus infection, for instance, in which the degree of 
virulence of the strains ordinarily met is such that the balance struck 
by them with the average defensive powers of man results in localiza- 
tion. However, the same micro-organism, enhanced in virulence, or 
gaining entrance in unusual numbers in a weakened individual, may 
rapidly spread from the point of inoculation, at first by contiguity, 
then by invasion of the blood and lymph channels, and become 
generalized. 
When organisms become generalized and circulate in the blood 
the resulting condition is spoken of as septicemia or hacteriemia. 
This is the form of infection commonly caused by streptococci, 
bacilli of the hemorrhagic septicemia group, anthrax bacilli, and 
many others. It implies a powerful invasive property and always 
constitutes a condition of great gravity when persistent. We are 
learning of recent years, however, that in many infectious diseases 
formerly regarded as purely localized a temporary entrance of the 
bacteria into the circulation is a usual occurrence. Thus Fraenkel 45 
has shown that lobar pneumonia is almost always accompanied dur- 
ing the acute stages of the disease by pneumococcus septicemia, and 
in typhoid fever we now know that the organisms circulate freely in 
the blood during the first two weeks of the disease, and often longer 
than this. 
In these and other conditions the bacteria may be gradually de- 
stroyed and disappear from the blood stream as the immunity of the 
subject increases. In other cases the bacterial activities may be 
partially checked, the process becoming slower and more chronic. 
This is especially often the case when micro-organisms after entrance 
to the circulation have found a secondary lodgment upon a heart 
valve, from which a continuously renewed supply of bacteria can 
be given off to the blood. A special form of such “malignant endo- 
carditis” caused by the Streptococcus viridans is particularly apt to 
take this chronic course. 
The presence of bacteria in the blood is not, therefore, as for- 
merly supposed, an invariably fatal condition. 
Adami’s recent work would indicate, moreover, that bacteria may 
normally enter the portal or even the general circulation from the 
intestine during health. This condition of “sub-infection,” as he 
calls it, is more fully discussed on p. 256. That colon and other in- 
testinal bacteria may often penetrate into the portal circulation is 
indicated by the occasional occurrence of colon bacillus abscesses 
after trauma of the liver. In most septicemias, however, caused by 
virulent bacteria the invasion of the blood stream persists, rapid 
multiplication occurs and leads to death. 
From the circulation the bacteria may gain lodgment in various 
organs and cause the formation of secondary abscesses. This condi- 
45 Fraenkel. F. Leyden Festschr., 1902. 28 
INFECTION AND RESISTANCE 
tion is known as “pyemia,” and may be caused by almost any bac- 
teria which are capable of producing septicemia. Thus staphylo- 
cocci, streptococci, or pneumococci may lodge in bones, joints, brain, 
or kidneys, in fact in any organ in which they can gain a foothold. 
However, there are evidences of distinct tissue predilections on the 
part of certain germs. Thus the virus of rabies and that of polio- 
myelitis, though to some extent universally distributed, seem espe- 
cially to concentrate in the nervous system; cholera spirilla and 
dysentery bacilli appear to find conditions most favorable for de- 
velopment in the intestinal mucosa; amebic abscesses are most 
common in the liver; gonococcus infections when generalized find 
secondary localization with particular frequency on heart valves and 
joints; leprosy bacilli have a predilection for the nerve sheaths; and 
glanders bacilli injected into the peritoneum of a male guinea pig 
localize with such regularity in the testicles that the experiment has 
diagnostic value (Strauss test). Conversely it is only explicable on 
the assumption of such selective lodgment that tubercle bacilli, even 
though otherwise universally distributed through the body, will be 
absent from striped muscle tissue, and rare in the walls of the stom- 
ach. Such selection, as far as we can account for it all, seems to 
depend upon the varying cultural conditions encountered by the 
germs in different organs. 
Among the most specific and curious of the selections of special 
tissues is that recently worked out with the probably closely related 
infectious filtrable agents of herpes, encephalitis and the salivary 
viruses first described by Levaditi. Into the same general class prob- 
ably belongs the virus of small-pox and of chicken-pox. These, as 
yet uncultivated and probably, also, unseen micro-organisms, can 
be inoculated upon animals, but appear to be limited in their in- 
fectious powers to cornea, skin and nervous system, rigidly so in 
some cases, less so in others, all of these tissues being, in their embry- 
ological origin, ectodermal. 
On the other hand, localization may also be dependent upon 
accidental conditions such as trauma. Infections in which the en- 
trance of bacteria is coincident with injury—as in the case, for in- 
stance, of compound fractures—will be able to spread throughout 
the injured region much more easily than they could enter the 
healthy tissue. In fact, it is well known that local tissue injury at 
the point of inoculation favors infection since it furnishes a rich 
substratum for growth in the form of dead cells or blood clot and 
interferes with the accomplishment of a normal protective reaction. 
In cases in which bacteria are circulating in the blood mechanical 
injury may create a focus of reduced resistance on which the in- 
vaders can gain a foothold. It is in this way perhaps that, among 
other things, we can explain tuberculosis of joints or bones which 
present a history of injury preceding the development of the infec- THE PROBLEM OF VIRULENCE 
29 
tion—or the pleurisy and lobular pneumonias which have been known 
to ensue upon the fracture of a rib. 
It is also possible that bacteria may be distributed in various 
organs directly from the initial focus by embolism or by the massive 
invasion of a blood vessel. It is by such breaking into a vein that 
Weigert explains the generalization of miliary tuberculosis. 
The inflammatory reaction which usually ensues at the point of 
entrance of bacteria is merely a result of the local struggle between 
invader and tissues, and the violence of this reaction is in a large 
measure an indication of the resistance of the infected subject. 
When, for instance, a streptococcus of moderate virulence gains 
lodgment in the skin of a healthy individual the rapid mobilization 
of leukocytic and other defences may prevent further invasion by 
the bacteria and lead to a struggle which is clinically evidenced by 
severe local symptoms. Did the virulence of the streptococci far 
overbalance the powers of resistance the local struggle might be 
reduced to a minimum, the infection progressing without any, or 
with but a slight local, reaction. The fact that pneumococci lodging 
in the human lung ordinarily cause lobar pneumonia is merely an 
evidence of a considerable degree of resistance to these germs on the 
part of the average human being. Pneumococci introduced into the 
pulmonary alveoli of very susceptible animals (rabbits) may pass 
directly through into the circulation, causing fatal septicemia with- 
out leading to a more than mild and temporary reaction in the lungs 
themselves. If, as in Wadsworth’s 46 experiments, the rabbits are 
partially immunized—that is, their resistance increased before the 
pulmonary inoculation is carried out—a violent local reaction, anal- 
ogous to lobar pneumonia, may follow, the severity of the reaction 
at the portal of entry being manifestly an evidence of more energetic 
opposition to further penetration of the bacteria. 
The entrance of bacteria into the deeper tissues, and even the 
circulation, without any, or with but slight, local evidences of infec- 
tion at the point of entrance is by no means rare. The innocent 
appearance of the site of the entrance of the bacteria in generalized 
streptococcus infection is a common surgical observation, and a strep- 
tococcus-infected wound of the hand or leg in a patient dying of septi- 
cemia may appear but slightly inflamed and edematous and incom- 
parably milder in appearance than a staphylococcus boil with which 
the patient is walking about and suffering hardly any systemic dis- 
turbance. 
Incubation Time.—Between the time of entrance of the bacteria 
into the body and the first appearance of symptoms of disease there 
is always a definite interval which is spoken of as “incubation time.” 
This period is made up of two definite divisions—one the time 
necessary for growth, distribution, and accumulation of the bacteria, 
48 Wadsworth. Am. Jour, of the Med. Sc., Vol. 27, 1904. 30 
INFECTION AND RESISTANCE 
the other the time necessary for the action of the toxin or poison 
which may be secreted. The latter, the incubation time of the toxin, 
is a subject which is still unclear in many of its phases, and will 
be discussed in the following chapter (see p. 43). The former, 
however, is easily comprehended, in fact, is to be expected. For 
the small number of bacteria* which gain entrance to the tissues in 
spontaneous infection is entirely inadequate in itself to produce 
symptoms. It is necessary that multiplication shall take place until 
the bacteria have accumulated in number sufficient to cause notice- 
able physiological disturbance. That the interval necessary for 
this must vary according to the number of bacteria originally in- 
troduced, the virulence of these, and the specific resistance of the 
patient goes without saying. Von Pirquet and Schick have sug- 
gested also that the incubation time may correspond roughly to the 
interval during which the subject is becoming “allergic” or hyper- 
susceptible to the bacteria or virus. This will be discussed at greater 
length in the chapter on anaphylaxis.47 
But within the limits of the variations introduced by these fac- 
tors the incubation time of each infectious disease—if spontaneously 
acquired—is sufficiently uniform to be characteristic. Thus the pri- 
mary lesion in syphilis follows the inoculation after an interval of 
two or three weeks, rabies follows inoculation with street virus after 
about four to six weeks, the period being somewhat dependent on the 
location of the bite; typhoid fever takes about two weeks to develop; 
gonorrhea about five to seven days; small-pox about two weeks; 
yellow fever three to five days; and scarlet fever and diphtheria 
about two to six days. In general, it may be stated that within the 
limits observed for each particular infection the shorter the incuba- 
tion time the more severe is the infection. Thus if tetanus follows 
inoculation with the tetanus bacillus within seven days the prognosis 
is far more grave than when the incubation time has occupied two 
or three weeks. And if localized and general symptoms follow rap- 
idly (within twenty-four to forty-eight hours) after a streptococcus 
infection it is likely that the process is a very severe and virulent 
one. 
Effects of Symbiosis upon Infection.—That the influence of 
various species of bacteria growing in the same culture is of great 
mutual importance to the growth and life of the individual species 
is, of course, well known to bacteriologists. Such interdependence 
may be noticeable as antagonism, usually due to inhibition of one 
species by the metabolic products of the other. On the other hand, 
even in culture, two different species may enhance each other’s 
activity. This has been particularly noticed with diphtheria bacilli 
and streptococci in culture by Hilbert.48 While the condition of 
47 Yon Pirquet u. Schick. Wien. kl. Woch., 16, 1903, pp. 758 and 1244. 
48 Hilbert. Zeit. f. Hyg., 29, 1895. THE PROBLEM OF VIRULENCE 
31 
mutual enhancement in culture is rare and perhaps questionable, 
even in the cases described, in infections of the body it is common 
and logically so, since one micro-organism may be able to ward 
off the defensive mechanism of the body from the other, or by its 
own destructive processes, create conditions more favorable to the 
development of the offences of its mates. Examples of such symbiotic 
enhancement of virulence are the association of diphtheria bacilli 
and streptococci in the throat, and a similar association of the 
organisms of Vincent’s Angina with diphtheria; in both of these 
cases the clinical diphtheria is apt to be considerably more 
severe than when the diphtheria bacillus is found alone with the 
ordinary inhabitants of the nose and throat. An extremely im- 
portant example is that which has been much investigated, in con- 
nection with tetanus infection. Here it has been shown that the 
destructive effects upon tissues exerted by simultaneous staphylo- 
coccus or streptococcus infections in the wound, may be instrumental 
in creating the conditions that favor the development of tetanus. 
The association of streptococci with scarlet fever is not so clear 
an example since there is still a possibility that the streptococci may 
be more intimately related to the disease than merely as symbiotic 
accompaniments. Upon this point no opinion is, as yet, entirely 
justified. Of the greatest epidemiological importance in this connec- 
tion is the incontrovertible fact that diseases such as influenza and 
measles pave the way for severe and fatal invasion of the respiratory 
tract and lungs with streptococci and pneumococci; in fact, so much 
is this the case that the original diseases, themselves, rarely kill, 
but have a high death rate owing to the secondary invaders for which 
they prepare the field. Of probably similar significance, also, are 
the conditions prevailing in hog cholera and in typhus fever where, 
in the former case, the so-called hog cholera bacillus is almost 
always associated with the disease, although we know that the specific 
etiological agent is quite a different organism, a filtrable virus. Tn 
typhus fever the frequent presence of the Plotz bacillus and other 
organisms may represent a similar state of affairs. Thus, while 
we know very little about the mechanism in most cases, the con- 
comitant presence of a number of infectious agents may very severely 
increase the ability of either or of both, to invade. 
Presence of Bacteria in Tissues in a Latent Condition.—A very 
interesting fact connected with the reactions between bacteria and 
tissues is the occasional local balance struck between infectious 
agents and the tissues in which they lie. In such diseases as 
tuberculosis and syphilis this phenomenon is quite common. The 
organisms may remain in a definite place for a month and even 
longer, without giving rise to any signs of disease, and yet, at a 
given moment, often without apparent cause, a characteristic in- 
flammatory process may be initiated. In the case of experimental 32 
INFECTION AND RESISTANCE 
syphilis, we have observed a rabbit inoculated with virulent trepo- 
nema in the testis which showed absolutely no signs of reaction for 
three and one-half months, at which time a typical process began 
to appear. The phenomenon has been observed with many different 
bacteria, even with those ordinarily causing acute disease. Not 
long ago we saw a patient that was extensively incised in the course 
of a hemolytic streptococcus infection of the hand. A second opera- 
tion made for purposes of improving function at a time when no 
inflammation whatever existed in the part, revealed the presence 
of the same hemolytic cocci which had apparently remained latent 
in the tissues for the entire interval. Many instances of the same 
phenomenon may be cited, and on the basis of these, it is quite easy 
to understand how trauma or other accidents may occasionally give 
rise to what we speak of as cryptogenic infections. CHAPTER II 
BACTERIAL POISONS 
When bacteria have gained a foothold anywhere within the 
animal body the local and general disturbances which follow, in all 
but the mildest and most trifling cases, are such that we cannot 
account for them solely on the basis of mechanical injury. 
It may well be that the obstruction of capillaries and lymphatics 
and the pressure upon parenchyma cells, always incident to inflam- 
matory reactions, contribute materially to local destruction, and 
thereby indirectly to systemic effects. However, even in diseases 
like anthrax, in which the body of the victim after death is found 
flooded throughout with masses of bacteria, these factors cannot fully 
explain the clinical manifestations. And such cases, indeed, are 
extreme examples, since, in the large majority of bacterial diseases, 
the illness resulting in the patient is severe out of all proportion to 
the extent of the tissue area invaded. 
Moreover, all infections, if at all severe, whatever their nature 
or localization, give rise to fever, and this symptom alone, if care- 
fully observed from hour to hour, may be sufficiently characteristic 
to indicate the specific micro-organism which is causing the illness. 
With this there occur alterations of the blood picture, either a 
numerical increase of white blood cells (leukocytosis) or a change in 
the relative proportions of the different kinds of leukocytes—or 
again an anemia caused by the destruction of red cells. There may 
also be degenerative changes in parenchyma cells of organs far re- 
moved from the actual site of bacterial lodgment. All these facts 
indicate very definitely that, apart from localized tissue destruction 
or purely mechanical interference with function by capillary ob- 
struction or pressure, there is at the same time an absorption of 
poisonous substances emanating from the bacteria. 
Ptomains.—From the earliest days of logical investigation into 
the nature of infectious disease, as soon, in fact, as cultural methods 
had been introduced, bacteria were studied with the purpose of 
throwing light upon this phase of their activity. As a result of 
such investigations Selmi,1 in 1885, described certain basic toxic tub- 
stances which he obtained from putrefying human cadavers and for 
which he suggested the designation “ptomain” (from tttwhcl — dead 
body). These poisons were later more extensively studied by 
1 Selmi. Cited from Hammarsten, ‘‘Textbook of Physiol, Chem.,” p. 16, 
33 34 
INFECTION AND RESISTANCE 
Brieger,2 Gautier,3 Griffiths,4 and others, and it was at first sur- 
mised that the formation of such substances in the infected animal 
might he held responsible for the toxemic manifestations which 
accompany bacterial disease.5 
This, as we shall see, is not the case. Ptomains are probably not 
formed in traceable quantity in the living tissues and are not in any 
way identical with the specific bacterial poisons which are respon- 
sible for the toxemia of infectious diseases. Nevertheless, they have 
some pathogenic significance, since they are invariably products of 
the proteolysis caused by bacteria and can give rise to illness when 
ingested with putrefying foodstuffs. It is important, therefore, that 
we discuss them briefly and consider their fundamental distinction 
from the true bacterial poisons. 
Whenever dead organic material, meat, fish, vegetable refuse, 
etc., is left to itself under suitable conditions of moisture and tem- 
perature, putrefaction sets in. As a result of bacterial growth the 
protein is broken up and among the intermediate products of such 
proteolysis ptomains appear. Chemically6 these substances are 
basic nitrogenous compounds which may or may not contain oxygen. 
Because of their basic and often highly toxic properties they have 
been spoken of as “animal alkaloids.” Many of them contain only 
C, H, and N, and are ammonia substitution products. (See 
Vaughan and Novy, loc. cit., p. 248.) Thus some of the simpler ones 
are: 
Methylamin=( CH3 ) NH2 
Dimethylamin= ( CH3 ) 2 NH 
Trimethylamin=:(CHl3)3 N 
Among those somewhat more complex are: 
Putrescin==NH2—CH2—CH2—CH2— CH2-NH2 
and Cadaverin=NH2—CH2 -CH2 -CH2 —CH2—CH2—NH2 
Samuely classifies the ptomains according to their nitrogen con- 
tents as follows: 
1. Those with one nitrogen atom (C8HUN) (C8H13N) 
(C10H15N) 
2. Those with two nitrogen atoms such as putrescin (C4Hj2N2) 
and cadaverin (C5H14N2) and 
2 Brieger. “Die Ptomaine,” Berlin, 1885; Virchow’s Archiv., Yols. 112 
and 115; Berl. klin. Woch., 1887, 1888. 
3 Gautier. Cited after Pick, Bull, d-e Vacad. de med., 1886. 
4 Griffiths. Compt. Rend, de Vacad. des sc., Yol. 113. 
5 For a historical outline of our knowledge of these poisons, as well as 
for a thorough treatment of their nature, see Yaughan and Novy, “Cellular 
Toxins.” 
6 For a discussion of the chemistry of the ptomains see Vaughan and 
Novy, “Cellular Toxins,” Lea Bros., Philadelphia, 1902. See also Samuely in 
Oppenheimer’s “Handbuch der Biochemie,” Vol. I, pp. 794 et seq.; and 
“Chemical Pathology,” Saunders, Philadelphia, 1907, BACTERIAL POISONS 
35 
3. Those with three nitrogen atoms such as methyl guanidin 
(C2H7H3). 
4. Finally there is an important group which contains oxygen, 
such as the substance sepsin obtained by Faust from 
putrefying yeast cells. 
They are not in all cases protein cleavage products, since bodies 
of the cholin group, cholin, neurin, and muscarin, the two last named 
highly toxic, are lecithin derivatives, and Samuely points out that 
other lipoid cleavage products, always present in decomposing tis- 
sues, may well contribute to ptomain production in the presence of 
a source of nitrogen. It is interesting to note also that the vegetable 
poison muscarin, isolated by Schmiedeberg from mushrooms, is 
chemically identical with a toxic base found by Brieger in decom- 
posing fish. 
The ptomains are not poisonous in every case. The chemically 
simpler ones like methylamin, di- and trimethylamin possess little 
or no toxicity. Others chemically more complex—like cadaverin 
and putrescin—may be capable merely of causing local necrosis, 
while sepsin, closely related to cadaverin in chemical constitution, 
but containing oxygen, is a powerful poison which acts violently 
upon the intestinal blood vessels, causing capillary dilatation, con- 
gestion, and diapedesis.7 The presence of oxygen seems indeed to be 
necessary for the development of strong toxicity (Brieger, Vaughan, 
and Novy). Again, the lecithin derivative, cholin, is but weakly 
toxic, while neurin is exceedingly poisonous. In putrefying mix- 
tures these toxic bodies appear on or about the fifth or seventh day 
after putrefaction sets in, and disappear, by further cleavage, more 
or less rapidly, yielding less complex nitrogenous substances that are 
non-toxic. 
With the limited knowledge regarding bacteria and infectious 
diseases at the disposal of the earlier investigators it was but natural 
that the discovery of ptomains in cultures of putrefactive bacteria 
aroused the suspicion that these bodies were responsible for the 
toxemia of infectious disease. 
The search for poisonous substances in pure cultures of patho- 
genic bacteria was, therefore, assiduously taken up by Brieger and 
his pupils, and, in truth, ptomains were actually found as products 
of some of the disease-producing micro-organisms, just as they had 
been found in the mixed cultures involved in the putrefaction of 
meat. Thus cadaverin was found in cultures of the cholera spiril- 
lum, another nitrogenous poison, typhotoxin, in those of typhoid 
bacilli, and still another in tetanus cultures, all of them producing 
more or less severe illness when injected into animals. 
In spite of this evidence, however, we have been forced to con- 
clude that the ptomains cannot properly be held responsible for bac- 
7 Meyer and Gottlieb. “Experim. Pharmakologie,” 2d ed., p. 262. 36 
INFECTION AND RESISTANCE 
terial toxemia as manifested in disease. In the first place it is 
doubtful whether ptomains, in noticeable quantity, are ever produced 
within the living infected body. Then, again, potent ptomains are 
produced in culture by many bacteria having absolutely no patho- 
genic power, while highly pathogenic bacteria may produce little or 
no ptomains. Ptomain production, moreover, is not specific, since 
the same ptomains may be produced by many different bacteria or 
mixtures of bacteria, provided the conditions of nutrient materials 
and temperature are favorable for growth. We cannot therefore 
account for bacterial toxemia, in which the poison produced by an 
individual species is characteristic and invariably the same, under 
varying cultural and environmental conditions, by the production of 
ptomains. And even when ptomains are produced in culture fluids 
by pathogenic bacteria their physiological action is usually quite 
different from that of the poisons produced by the same micro-organ- 
isms in the infected subject. 
Briefly summarized, therefore, the ptomains are poisons elab- 
orated by all bacteria that are capable of producing protein cleavage, 
if planted on suitable nutrient materials under conditions favoring 
growth. The matrix of these poisons is the protein nutriment; they 
are not products of intracellular metabolism specifically characteris- 
tic of the bacteria which produce them. 
Their importance in the production of disease, therefore, is really 
an indirect one. They may cause disease if putrid meat or other 
material is ingested, and with it preformed ptomains, which may be 
taken in and further elaborated by continued putrefaction in the 
intestines. This form of meat poisoning, without bacteriological in- 
vestigation, may be difficult to distinguish from such bacterial forms 
of meat poisoning as those caused by the Gartner bacillus or the 
bacillus botulinus. Hovy 8 believes that true ptomain poisoning of 
this kind is rather less frequent than formerly supposed. However, 
in such cases as those of Vaughan, who isolated a poisonous ptomain 
“tyrotoxicon” from cheese and milk, their importance seems rea- 
sonably certain. It is also probable that certain forms of auto- 
intoxication may be caused by the production in the intestinal ca- 
nal of ptomains resulting from bacterial putrefaction incident to 
faulty digestive conditions. It is the antagonism to such intes- 
tinal putrefaction by the acid production of the bacillus Bul- 
garicus which is probably the basic cause of any favorable thera- 
peutic effects which have attended the soured milk therapy of 
Metchnikoff. Again the growth of saprophytes in necrotic tissues 
such as gangrenous extremities in diabetes or amputation stumps, 
may lead to the formation of ptomains which, after absorption, can 
cause disease. In all such cases the process is one determined by the 
bacterial putrefaction of dead organic materials, and the absorbed 
8 Novy in Osier’s “Modern Medicine,” Vol. 1, p. 223. BACTERIAL POISONS 
37 
poisons are not true bacterial toxins, since they do not emanate 
specifically from the cell substance of the micro-organisms but rather 
represent incidental cleavage products of the nutrient materials. 
Therefore, also, the ptomains are unspecific—their formation a com- 
mon attribute of a large variety of saprophytic organisms, their 
production, as to quantity and kind, primarily dependent upon 
the nature of the nutrient materials on which the bacteria are 
grown. 
Other Non-Specific Toxic Effects.—The fermentative and putre- 
factive activities of many bacteria not ordinarily classified as patho- 
genic, may give rise on occasion to disease if they take place in 
the intestinal canal by virtue of a temporary or chronic alteration 
of the flora normal to this location. It is well known, of course, 
since the studies of Escherich,9 Herter,10 Kendall,11 Rettger,12 and 
others that the infant bowel, sterile at birth, rapidly begins to har- 
bor bacteria. For the first few days there is a period of irregular 
accidental flora which is soon transformed into a characteristic 
infantile intestinal flora, changing only as diet changes in the course 
of years. In breast-fed children, the upper part of the small in- 
testine, for instance, will usually contain various Gram positive 
cocci which predominate over the Gram negative bacilli. Further 
down, organisms of the Colon, Lactis aerogenes type appear, and in 
the lower part of the cecum, Gram positive bacilli, like the Bifidus 
of Tissier and other anaerobes predominate. As breast feeding is 
discontinued, the flora changes by the superimposition of Colon 
bacilli and the acidophilic group. For the purpose of making our 
point it is not necessary to go into details concerning the actual 
flora, for which we refer the reader to the original Monograph and 
to a brief summary in our own “Textbook of Bacteriology.” It 
is sufficient to indicate that the ingestion at any time of milk or 
other food, very highly contaminated with organisms capable of 
powerful fermentative action upon carbohydrates with the forma- 
tion of various acids and their by-products, or with masses of putre- 
factive organisms capable of splitting proteins to various degrees, 
may give rise, on the one hand, to the absorption of many abnormal 
cleavage products of food materials, considerable amounts of mod- 
erately toxic bacterial products and, in addition to this, interfere 
with digestion and the normal movement of the intestinal wall by 
distention with large volumes of gas. In this sense, not only 
ptomains but perhaps acid products of carbohydrates, digestion, etc., 
9 Escherich. “Darmbakterien des Saiiglings,” Stuttgart, 1886. 
10 Herter. “The Common Bacterial Infections of the Digestive Tract,” 
Harvey Lecture, 1906-1907. 
11 Kendall. “Bacteriology, General, Pathological and Intestinal,” Lea & 
Febiger, Philadelphia, 1916, p. 580. 
12 Rettger. Jour. Biochem., 2, 1906. 38 
INFECTION AND RESISTANCE 
may readily give rise to intoxication, constipation or diarrheal dis- 
ease, or, as is often the case, constipation followed by violent diarrhea 
as the stagnated material is subjected to further bacterial action. 
It is this form of non-specific disease production on the part of 
many bacteria ordinarily not considered pathogens, which is per- 
haps most neglected and overlooked in clinical work. Judging from 
the work of Park, Holt and their co-workers, upon infantile diarrhea 
in relation to heavily contaminated milk, this mechanism may play 
an important role in the summer mortality of children. 
Specific Bacterial Poisons.—In contradistinction to the ptomains, 
the specific bacterial poisons, in the technical meaning of the term, 
are substances which are characteristic for each individual species 
of bacteria and truly the products of bacterial metabolism in that 
they emanate from the cell itself, either as a secretion or excretion 
during cell life, or as an inherent element of the cytoplasm lib- 
erated after death (or possibly as a cleavage product of the dis- 
integrating bacterial protein).13 They are dependent upon the 
nature of the culture medium only in so far as this favors or retards 
the normal development of the micro-organisms. While, therefore, 
a diphtheria bacillus undoubtedly produces the largest quantities 
of its specific poison on bouillon suitably prepared for this particular 
purpose, it will also, in smaller amount, produce qualitatively the 
same poison on all media on which its growth is free and uninhibited,. 
even on a medium such as that of Uschinsky, which is entirely 
devoid of proteins. The toxins are, therefore, elements of intra- 
cellular metabolism, permanently or transiently constituent parts 
of the cell body. 
A specific bacterial toxin was first obtained from the diphtheria 
bacillus by Roux and Yersin 14 in 1889. They discovered that if, 
diphtheria bacilli were grown on veal broth and the cultures filtered 
through porcelain candles, after seven days at 37.5° C. the filtrates 
were highly toxic, producing the same symptoms and autopsy find- 
ings in rabbits, guinea pigs and birds which followed the injection 
of the living bacilli themselves. The poison was therefore a soluble 
product of the bacteria during the period of their vigorous growth, 
apparently given up by them to the culture fluid. Very soon after 
this, in 1891, Kitasato 15 discovered a similar specific toxin in cul- 
ture filtrates of the tetanus bacillus, and it was the hope of bacteri- 
ologists that analogous poisons could be determined for all patho- 
genic bacteria. 
This hope, however, has been disappointed. It was soon found 
that cultures of cholera spirilla, typhoid bacilli, and many other 
13 In connection with this read the discussion on anaphylaxis in chapter 
XVII. 
14 Roux and Yersin. Ann. de VInst. Pasteur, Vol. 2, 1889. 
15 Kitasato. Zeitschr. f. Hyg., 1891, Vol. 10. BACTERIAL POISONS 
39 
germs did not yield toxic filtrates of tliis kind but that the poisons 
in these cases seemed to be firmly bound to the bacterial bodies dur- 
ing life, and given up to the surrounding media only after death 
and disintegration of the cells. 
Pfeiffer 16 was the first one to formulate this conception in his 
studies upon cholera poisons. He found that when cholera spirilla 
were grown upon broth and filtered after 6 or 7 days, the filtrate was 
but slightly toxic, but that, in this case, unlike the conditions pre- 
vailing in diphtheria and tetanus cultures, the residue of bacterial 
cell bodies, even after they had been killed by chloroform, thymol, 
or drying, were powerfully poisonous. 
We have then two main classes of specific bacterial poisons. One 
—typified by diphtheria and tetanus poisons—is produced during 
the period of energetic growth by the living bacteria, is given off to 
the surrounding culture fluid as a secretion or excretion, and can be 
obtained in bacteria-free filtrates at a time when few, if any, of the 
micro-organisms have died or disintegrated. These are spoken of as 
“true toxins” or “exotoxins.” 
The other group—typified by the cholera poisons as described by 
Pfeiffer—is apparently an intracellular, constituent part of the bac- 
terial body—not given off during life and not, therefore, obtained in 
filtrates of young living cultures. If the cultures are preserved until 
cell death has taken place and the dead bodies have been extracted 
by the culture fluid, the filtrate becomes gradually more toxic. The 
bodies of such bacteria are in themselves powerfully toxic when 
injected, dead or alive. These poisons for obvious reasons Pfeiffer 
has named the “endotoxins,'’ since he regarded them as specific and 
definite substances, present as such in the living bacterial cell. 
In addition to the endotoxins the bacterial protein contains sub- 
stances which attract and lead to the accumulation of leukocytes. In 
other words, they exert a positive chemotactic influence. This was 
first observed in 1884 by Leber,17 who induced the formation of pus 
by injecting dead staphylococcus cultures, and, later, found that the 
same effect resulted from the injection of alcoholic extracts of 
staphylococci. These chemotaxis-inducing substances were later 
particularly studied by Buchner. Buchner 18 extracted them from 
many varieties of bacteria, independent of pathogenicity. Although 
there are quantitative differences, all bacteria seem to contain such 
substances, and Buchner believed the chemotactic property to be a 
general attribute of the bacterial protoplasm. He speaks of his ex- 
tracts as bacterial proteins. 
Exotoxins or True Toxins.—The true toxins or exotoxins, then, 
appear to be products of living bacteria given off from these very 
16 Pfeiffer. Zeitschr. f. Hyg., Yol. II, 1892. 
17 Leber. “Uber die Entziindung,” Leipzig, 1884. 
18 Buchner. Berl. klin. Woch., 1890. 40 
INFECTION AND RESISTANCE 
much as are the ferments and enzymes by which micro-organisms 
cause cleavage of carbohydrates or proteins—and indeed the French 
school, from the first, compared these toxins to enzymes, with which, 
as we shall see, they have much in common. The endotoxins—on 
the other hand—at least as conceived by Pfeiffer, are structural in- 
gredients of the bacterial protoplasm which are toxic when brought 
into solution as the cells break up. 
Concerning the accuracy of this conception, however, much doubt 
has recently arisen, as a result of researches which will be discussed 
below. 
These two types of poison, moreover, differ from each other not 
only in mode of origin but in biological characteristics far more 
fundamental than this. 
The discovery of diphtheria toxin by Roux and Yersin was fol- 
lowed by diligent investigations into the toxic properties of all 
known pathogenic bacteria, and it was soon found that a few only 
of these germs could produce poisons biologically similar to that 
found in diphtheria cultures. It was in the course of investigations 
of this kind, indeed, that Pfeiffer, failing to discover an exotoxin in 
cultures of cholera and other germs, formulated his endotoxin theory. 
The list of true toxin or exotoxin producers, then, is short. 
Among the more important are, in addition to the diphtheria and 
tetanus bacilli—which have been mentioned above—the Bacillus 
botulinus,19 the Bacillus pyocyaneus,20 and that of symptomatic an- 
thrax.21 It has also been claimed that similar toxins are formed by 
the cholera spirillum (Brau and Denier),22 by the dysentery bacillus 
of the• Shiga-Kruse type (Kraus and Doerr) 23 and the Bacillus ty- 
phosus (Arima).24 In the cases of the three last-named organisms, 
however, the secretion of a true exotoxin has not been accepted as a 
fact by all observers. Indeed, even though such substances may pos- 
sibly be produced by these bacteria in small amounts it is not likely, 
in the light of our present knowledge, that they play more than a sec- 
ondary role in the toxemic manifestations of cholera, dysentery, and 
typhoid, the important poisons in these cases being those derived 
from the bacterial cell bodies. 
Similar in essential properties to the true exotoxins also are the 
erythrocyte poisons (hemotoxins) produced by many bacteria which 
cause hemolysis of red cells, and the leukocyte-destroying poison 
(leukocydin) which is a product of the Staphylococcus aureus. 
All of these “true bacterial toxins” or exotoxins, apart from sim- 
19 Kempner. Zeitschr. f. Hyg., Vol. 26, 1897. 
20 Wassermann. Zeitschr. f. Hyg., Vol. 22, 1896. 
21 Grassberger and Schattenfroh. Wien Deuticke, 1904. 
22 Brau and Denier. Ann. de VInst. Past., Vol. 20, 1906. 
23 Kraus and Doerr. Wien kl. Woch., 42, 1905. 
24 Arima. Centralbl. f. Bakt., I, Vol. 63, 1912. BACTERIAL POISONS 
41 
ilarity of origin, as soluble secretions of the living bacteria, possess 
certain common biological characteristics which sharply differentiate 
them from the “endotoxins.” These characteristics they share with 
a number of non-bacterial substances such as the vegetable poisons 
ricin, crotin, and abrin, with animal poisons like snake venom and 
spider poison (arachnolysin), and, in certain important respects, 
with the substances spoken of as enzymes. 
Thus the bacterial true toxins are not biologically unique sub- 
stances. Both in themselves and in regard to the reactions they 
elicit when injected into the animal body, they share certain cardinal 
properties with analogous substances derived from the higher plants 
and from animals. And it is important to recognize at once that we 
are dealing here, as in other phases of the study of bacterial immu- 
nity, with broad biological laws, which find application not only in 
bacteriology, but in general pathology and in the phenomena of pro- 
tein metabolism in general. It so happens that these phenomena 
have been studied and are most easily elucidated in connection with 
bacteria. But their general significance must not be lost sight of. 
The cardinal characteristic which unites all of these substances 
into a single well-defined biological group is their property of in- 
ducing the formation of antitoxins when injected into animals. This 
property is so important and its thorough comprehension so essential 
that we may be permitted to digress briefly in order to make it 
clear.23 
As we shall see, in subsequent chapters, all substances which lead 
to the formation of specifically reacting antibodies in the treated 
animal are spoken of as “antigens” or “antibody-inducing sub- 
stances.” The class of “antigens” is a large one, including all 
known proteins, and possibly some of the higher protein split prod- 
ucts, and protein-lipoid combinations, though the “antigenic” prop- 
erties of the last two are still in controversy. But among this 
large group of substances it is only the bacterial true toxins (exo- 
toxins), obtained in broth filtrates of living cultures, together 
with the vegetable poisons and other substances we have classified 
with them above, which induce in the blood of the treated animal a 
neutralizing antibody—(antitoxin)—which inhibits quantity for 
quantity the activity of the injected toxin or vegetable or animal 
poison. This property of eliciting the production of antitoxin in the 
animal body alone separates these substances sharply from all other 
23 We suggested in a Harvey Lecture some years ago that the differ- 
entiation between the exotoxins, enzymes, snake venoms, ricin, abrin, etc., 
by reason of their production of a neutralizing antitoxin antibody in the 
treated animal, was such a sharp one from the general protein antigens 
derived from bacteria and other sources, that it might add considerably to 
clearness if we separated them in a class under the name of antitoxinogens. 
This would seem one of the few instances in which it might be desirable to 
add a new term to the nomenclature of immunology. 42 
INFECTION AND RESISTANCE 
antigens, toxic or otherwise, and, in this respect, they differ sharply 
from the so-called “endotoxins’’ against which no antitoxins can he 
produced. 
As an important secondary characteristic of this group of sub- 
stances we may regard their chemically indefinable nature. In the 
case of none of them have we any definite knowledge of chemical 
constitution except in so far as it has been hitherto impossible to 
separate them from the protein molecule. The intensive chemical 
study of the toxins has universally resulted in failure to obtain a 
protein-free product which has the characteristic toxic properties of 
the original filtrate, or its antitoxin-inducing power. Concerning 
the methods which have been employed in the study of the chemistry 
of these substances we will have more to say in another place.26 It 
is safe to summarize all this work for our present purposes, by stat- 
ing that, whatever the method employed, until now all of the prep- 
arations obtained have given one or another of the protein 
type-reactions, and that none of them can be positively accepted as 
protein-free. The results here obtained have been entirely analogous 
to those obtained in similar investigations upon enzymes. (See also 
discussion of antigens, chapter IV.) 
The analogy with enzymes is indeed a striking one and noted by 
the first investigators of a true toxin, Roux and Yersin. Biolog- 
ically, of course, we have the cardinal similarity in that the injec- 
tion of toxins into animals induces the production of antitoxin, and 
treatment with enzymes induces specific and neutralizing anti-en- 
zymes. In addition to this, they are alike in their susceptibility to 
heat (both being destroyed when in solution by temperatures over 
80° C.), in their gradual deterioration on standing, and their mys- 
terious activity in small quantities upon disproportionately larger 
masses of the substances they attack. There is, however, one impor- 
tant difference between the two in their mode of action. For, while 
the toxins are apparently bound or neutralized by the tissues they 
attack, the action of an enzyme seems rather to be a process in which 
the enzyme unites with the substance it acts upon, is released as the 
result is attained, and freed for further action, without noticeable 
loss of quantity. Such catalytic properties have not yet been satis- 
factorily demonstrated for the bacterial toxins. However, there are 
other modifying factors which may account for lack of similarity in 
this respect, and in all other important points the two classes of sub- 
stances are closely analogous. 
The property of heat sensitiveness, which is a characteristic of 
bacterial exotoxins and enzymes, is shared with them by all of the 
substances mentioned above except snake venoms. Snake venoms 
are not destroyed completely until the temperature is raised to 75° to 
26 An extensive and authoritative summary of this phase of the subject is 
that of E. Pick in “Kolle u. Wassermann Handbuch,” etc., 2d ed., Yol. 1, BACTERIAL POISONS 
43 
80° C. The earlier contention of Leclainche and Vallee, that the 
toxin of symptomatic anthrax possessed similar heat stability has 
been satisfactorily refuted by Grassberger and Schattenfroh,27 who 
find that heating it to 50° C. for an hour completely destroys it. 
There is another important attribute of the true toxin which 
deserves discussion, though we are by no means in a position to offer 
any satisfactory explanation for it. We refer to the incubation time 
which elapses between the administration of a toxin and the occur- 
rence of symptoms. Here again snake poisons form an exception— 
since local manifestations may appear within an extremely short 
period after the injection of the venom or as the result of a snake 
bite. Such absence of incubation time, also, seems to be true of the 
toxin of the Vibrion Septique by which guinea pigs and rabbits 
are killed by moderate doses almost without any latent period, what- 
ever. However, in the case of all other toxins there is a definite 
lapse of time between the entrance of the poison and the first symp- 
toms, local or general. This interval is longer when small doses are 
given—shorter when the doses are large—but is never entirely elim- 
inated—even when many times the fatal dose is given. 
In the case of tetanus poison, for instance, injections into a horse 
may not cause symptoms for as long as four or five days. In mice, 
animals that are extremely susceptible, the incubation time may be 
shortened from 36 to 12 hours if we inject 3,600 lethal doses, but, 
in any case, whatever the dose, this interval cannot be shortened 
below 8 or 9 hours.28 Many attempts have been made to explain 
this. Ehrlich, as we shall see, assumes that the action of a poison 
depends upon two occurrences: one, the union of the poison with 
the vulnerable cell, the other the gradual injury of the cell by the 
toxic atom groups in the poison molecule. The time necessary for 
the institution of this process, he believes, explains the interval. 
Richet has suggested that the toxin itself may not be potent until 
acted upon by the body of the recipient and transformed into a 
potent form. His views are more directly related to the phenom- 
enon of anaphylaxis and are discussed in another section. De Waele 
has recently advanced a theory which implies that the incubation 
time represents the period necessary for the gradual concentration 
of the poisons in the vulnerable tissues, a process which depends 
either upon chemical affinities or solubility of the toxins in the cell* 
lipoids. A little at a time would then be absorbed by the vulnerable 
cells as they come in contact with the poison, through the circulation, 
and the symptoms would not appear until a definite intracellular 
concentration had been attained. His views are so closely bound up 
with the theories on the selective action of the toxins upon individual 
27 Grassberger and Schattenfroh. “Uber das Rauschbrandgift, etc./’ 
Wien. Deutieke, 1904. 
28 De Waele. Zeitschr. f. 1mm., Vol. 4, 1910, 44 
INFECTION AND RESISTANCE 
tissues and organs that they will be rendered clear as we proceed 
with a discussion of the latter. 
An important point to remember is that there are only a limited 
number of bacteria which form true toxins, that is, poisons like the 
ones described which cause antitoxin formation. The most im- 
portant of these are: Diphtheria bacillus (Behring and Wernicke), 
Tetanus bacillus (Behring and Kitasato), B. Chauvii, or Bacillus 
of Symptomatic Anthrax (Grassberger and Schattenfroh),29 Bacillus 
botulinus (Kempner),30 Bacillus pyoeyaneous (Wassermann),31 
Welch bacillus (Bull and Pritchett),32 Vibrion septique (Robert- 
son),33 Bacillus oedematiens (Weinberg and Seguin).34 Toxin 
formation has been claimed, and is likely for the Shiga type of 
dysentery bacillus.35 Toxin production has been claimed, but is 
doubtful in the case of the cholera spirillum, the typhoid bacillus 
and a number of other organisms. 
Within the limits of our definition of true toxins of bacterial 
origin, also are the leucocidins of the Staphylococcus aureus and the 
hemolysins of the streptococci, staphylococci and some other bacteria. 
Toxic Effects in the Cases of Bacteria which do not form True 
Exotoxins.—The majority of pathogenic bacteria do not, as we have 
seen, produce true toxins or exotoxins. Cultures of cholera spirilla, 
plague bacilli, and of many other bacteria do not yield toxic filtrates 
until the cultures have been allowed to stand for prolonged periods 
during which extraction and possibly autolysis have occurred. In 
these cases, moreover, definite toxic properties can be demonstrated 
in the dead cell bodies or in extracts prepared by various methods. 
In no case, however, is the injection of these “endotoxins” followed 
by the production of antitoxins. It was very natural to suppose 
that in micro-organisms of this class the toxic principle might be 
present in the form of a preformed intracellular poison which could 
be extracted or which became free as cell-death occurred and dis- 
integration ensued. 
It was assumed that, when bacteria entered the animal body and 
were destroyed by the action of the serum or cells, these endotoxins 
were liberated and poisoning resulted. The very protective action 
of the serum, which prevented the extension of the infectious in- 
vasion, by limiting bacterial growth, was thus looked upon as the 
agency by which the endotoxins, toxalbumins, were set free. Ex- 
periments by Radziewsky and others, in which it was shown that large 
29 Grassberger and Schattenfroh. Munch, med. Woch., 1900 and 1901, 
30 Kempner. Zeit. f. Hyg., 26, 1897. 
31 Wassermann. Zeit. f. Hyg., 22, 1896. 
32 Bull and Pritchett. Jour. Exper. Med., 26, 1917, 876. 
33 Robertson. Jour. Pathol, and, Bacter., 1920. 
34 Weinberg and Seguin. La Gangrene Gazeuse, Masson, Paris, 1920. 
35 Todd. Brit. Med. Jour., 2, 1903, 1456; Olitsky and Kligler, Jour. 
Exper. Med., 31, 1920, 19. BACTERIAL POISONS 
45 
doses of bacteria injected into immunized animals were violently 
toxic and more rapidly fatal than corresponding amounts injected into 
normal animals, were taken to mean that in the immune animals a 
more powerfully bacteriolytic property of the serum led to a more 
rapid liberation of the endotoxins. 
This was the conception of Pfeiffer and, in more recent theoret- 
ical discussions, that of Wolff-Eisner. Its essential features con- 
sisted in the assumption that the poisons were preformed and were 
contained within the cell body as such, and that they were specific for 
each micro-organism, determining to a certain extent its pathogenic 
properties. Thus typhoid endotoxin, cholera endotoxin, or dysen- 
tery endotoxin was supposed each to possess its own particular 
pharmacological properties by which the clinical manifestations of 
the respective diseases were partially determined. 
It is chiefly the work of Vaughan36 which has begun to throw 
doubt upon Pfeiffer’s original views, in that Vaughan has shown 
that all proteins, bacterial or otherwise, would yield, upon cleavage 
with alkalinized alcohol, toxic split products which possessed many 
of the pharmacological properties of the so-called endotoxins. In 
fact, Vaughan succeeded in producing, in animals, fever and other 
symptoms which are generally associated with infection, merely by 
injecting into them graded quantities of toxic split products obtained 
from bacterial protein. 
Following Vaughan, Friedberger succeeded in showing that 
toxic substances similar to Vaughan’s split products are formed 
when baqteria of various species are subjected to the action of nor- 
mal or immune sera, and that such poisons were pharmacologically 
alike and produced with equal ease from pathogenic and non-patho- 
genic micro-organisms. These phenomena are discussed in greater 
detail in our section on bacterial anaphylaxis. It is necessary, how- 
ever, to point out in this place the uncertainty in which these re- 
searches have left the conception of endotoxins. They suggest that 
the toxic effects following upon the introduction of pathogenic bac- 
teria into the animal body are not due to endotoxins, but are rather 
the result of the action of toxic cleavage products formed in the re- 
action between blood plasma and bacterial cell. These split products 
are not conceived as specific for individual bacteria but may be 
formed from all bacterial proteins, both the pathogenic and the non- 
pathogenic. The differences in pathogenicity between bacteria of 
this class would then depend entirely upon their powers to invade— 
not at all upon their possession of individually peculiar cell poisons. 
The differences in clinical course and toxemic manifestations would 
be taken to depend entirely upon the accumulation and the distribu- 
tion of the invading germs, and the consequently variable energy in 
36 For a complete discussion of Vaughan’s work see Vaughan, “Protein 
Split Products,” Lea & Febjger, Phila, and N. Yv 1913, 46 
INFECTION AND RESISTANCE 
the production of the toxic split products from them. Considerable 
experimental evidence has accumulated in favor of this point of 
view. But the subject is an involved one and will be considered 
more extensively in a later chapter. 
The entire question of “endotoxins,” or rather the problem of 
the mechanism by which such bacteria as typhoid bacilli, plague 
bacilli (and other organisms which do not produce exotoxins) poison 
the animal body, must be subjected to experimental revision. In 
addition to the idea of toxic split products in the sense of Vaughan 
and Friedberger, there are other alternatives. Jobling and Petersen 
have shown that bacteria injected into the circulation may absorb 
lipoidal substances which ordinarily act as anti-enzymes. In con- 
sequence of this, serum protease may be liberated to act upon the 
plasma itself, and produce toxic substances. 
Again, it is well known that the bacterial cells are relatively 
poor in coagulable protein, and we have shown with one of our stu- 
dents (Aronovitch) that primary and secondary proteoses may be 
obtained in considerable quantities in bacterial extracts. It is not 
impossible that these in themselves may have toxic functions when 
liberated, without further splitting. This particular subject finds a 
more extensive discussion in the chapter on bacterial anaphylaxis. 
Considerable difficulty in the explanation of bacterial intoxica- 
tion has been encountered in infections caused by organisms like 
the streptococci, staphylococci, anthrax bacilli, etc., in which no true 
exotoxins and no toxicity of the cell bodies themselves have been 
demonstrable. The streptococcus question is perhaps the most inter- 
esting and illustrative of these problems. Individuals infected with 
streptococci, if the infection is at all severe, often show powerful 
toxic symptoms, and then lymphangitis which extends upward from 
the infected part in permanent red lines, is very likely a sign of 
toxic irritation rather than of ascending bacteria growth, inasmuch 
as a ligature applied across the part very rapidly causes a fading 
of the distal red line. Many authorities, Marmorek, Braun, Aranson 
and, more recently, Clark and Fenton and Havens have described 
true exotoxins for hemolytic streptococci to explain these phenomena. 
There is no doubt about the fact, as these studies as well as more 
recent studies by the writer and Miss Kuttner, and studies made 
in our laboratory by John Rice have shown that certain strains of 
hemolytic streptococci do produce a toxic substance which, in our 
own work, is designated as the “X” substance. This substance, how- 
ever, is never quantitatively as powerful as many of the exotoxins 
and cannot, for the present at least, be classified with the true exo- 
toxins. Incidentally, it can be washed off the surfaces of strepto- 
cocci freshly grown on agar by rapid shaking with salt solution and 
filtration. Similar “X” substances can be washed away from fresh 
agar cultures of many other bacteria. Just how much toxicological BACTERIAL POISONS 
47 
significance these substances have in the animal body, remains to 
be shown.37 
A summary of bacterial toxemia would not be complete, we 
believe, unless we also mention a type of toxic reaction during 
bacterial infection wdiich is dealt with more extensively in the chapter 
on bacterial anaphylaxis and the tuberculin reaction.38 It appears 
from our own studies that in the course of many infections be- 
ginning within two weeks, the animal body becomes extremely hyper- 
sensitive to substances diffused out from the bacterial growth into 
the circulation. This is the basis of tuberculin, mallein, typhoidin, 
etc., reactions, and since these materials are actively produced by 
the growing bacteria, it is not at all unlikely that in infections 
sufficiently prolonged or repeated, much of the toxemia may be 
due to the action of these materials which are hardly toxic for the 
normal animal, but highly so for the animal sensitized by injection. 
J. T. Parker,39 working with the influenza bacilli, found that 
certain strains recently isolated after growing on chocolate broth 
for as little as 6 to 8 hours, will show in the broth filtrate a poison 
probably of this type, which may be so powerful that 1 to 2 c.c. 
will kill a rabbit acutely after 45 to 90 minutes. These poisons 
are thermolabile and while we do not believe that they belong to 
the exotoxin class, are, nevertheless, sufficiently powerful to perhaps 
explain many of the symptoms of the toxemia accompanying in- 
fluenza bacillus infections. 
In order to do injury to the infected individual the bacterial 
poisons must be produced in such locations that they can easily enter 
the physiological interior of the body. Few of the poisons that 
have been so far investigated can produce injury when introduced 
into the alimentary canal. In this location they are, as a rule, de- 
stroyed, or they pass through without doing harm. Neither diph- 
theria toxin nor tetanus toxin will produce symptoms when in- 
troduced intra-intestinally.40 Even cholera poison does not pass 
through the uninjured intestinal wall. Kruse 41 assumes, and Kolle 
and Schiirmann 42 seem to agree with him, that the absorption of 
cholera poison does not occur until the intestinal wall has been in- 
jured by the actual growth of the living bacteria. Kruse calls atten- 
tion to experiments by Burgers in which enormous quantities of 
37 For references on these poisons and a discussion of their significance, 
see Zinsser. Jour, of Immunol., Yol. 5, p. 265, 1920. 
38 Zinsser. Jour. Exper. Med., 34, 1921, 495. 
39 Parker, J. T. Jour. Immunol., 4, 1919, 331. 
40 Meyer and Gottlieb. “Exp. Pharmakol.,” Urban & Schwartzenberg, 
Berlin, 1911. Ransom. Deutsche med. Woch., No. 8, 1898. Nencki. Cen- 
tralbl. f. Dakt., Vol. 23, 1898. Carriere. Ann. de VInst. Past., Yol. 13, 1899. 
41 Kruse. “Allgemeine Mikrobiologie,” Yogel, Leipzig, 1910, p. 934. 
42 Kolle and Schurmann in “Kolle u. Wassermann Handbuch,” 2d Ed., 
Vol. 4. , . 48 
INFECTION AND RESISTANCE 
cholera poison, i.e., 200 cultures of dead or living cholera bacilli, 
could be administered to healthy guinea pigs and rabbits by mouth 
without harm in spite of the fact that these animals are definitely 
susceptible to the poisons and although the poisons are not injured 
by the intestinal ferments. It is likely therefore that the absorption 
of poison begins only after the bacteria have extensively invaded the 
intestinal mucosa and, by injuring tissue, have opened paths for 
absorption. In the case of diphtheria probably a similar condition 
exists in that the localized injury to the mucous membrane at the 
point of lodgment of the primary infection prepares a portal of 
entry. The poison of the Bacillus botulinus seems to form an ex- 
ception to this rule,43 since this substance, though apparently a 
true bacterial toxin, is absorbed directly from the intestinal canal. 
With most bacteria this problem does not arise, since the poisons are 
elaborated within the tissues, where resorption is a necessary result. 
Certain snake poisons, especially the viper poisons, may cause a 
violent gastro-intestinal irritation if taken by mouth, and the veg- 
etable toxins, such as ricin, are characteristically of powerful toxicity 
if ingested. The toxic injury of the intestinal tract produced in 
the course of such infections as bacillary dysentery, typhoid fever, 
and some others is probably not due entirely to the action of the 
toxins directly in the intestine, but also to injury occurring in the 
course of excretion through the intestinal mucous membrane. There 
has been some controversy as to the probable reason for the in- 
nocuousness of most of the bacterial poisons after ingestion. Prob- 
ably the action of intestinal bacteria is of relatively little importance 
as compared with the destructive action of the intestinal enzymes, 
particularly trypsin. There may also be an antagonistic action on 
the part of some of the constituents of the bile. 
Like alkaloids and other organic as well as inorganic drugs, the 
action of many bacterial poisons is largely selective. Most of these 
poisons may excite inflammatory reactions if concentrated in any 
part of the body, but, in addition to this, there is a specific distribu- 
tion after introduction which indicates that the poison goes into 
selective relationship with certain tissues and cells. This fact is most 
clearly illustrated by the bacterial hemotoxins which specifically in- 
jure the red blood cells of the infected individual and by such sub- 
stances as the leukocidin produced by the Staphylococcus aureus, a 
poison which directly and visibly injures the white blood cells. Here 
the action is specifically aimed at a well-defined variety of body cell. 
In considering this problem in connection with infectious disease, 
it is of great importance to distinguish between selective injury by the 
poisons transported through the body by the lymph, blood, and other 
channels, on the one hand, and the selective lodgment of the micro- 
organisms themselves on the other. The latter may occasionally de- 
43 Madsen in “Kraus u. Levaditi, etc.,” Yol. 1. BACTERIAL ROISONS 
49 
pend on local cultural advantages for the particular bacteria in one 
organ or another, bnt may just as often be determined by the peculiar 
manner of entrance to the body which is most suitable for lodgment 
of the germs in question, and the degree of local resistance at the 
point of entrance, which determines whether or not the infection shall 
be locally limited or permitted to invade beyond this point. In the 
case of a disease like acute anterior poliomyelitis, where our.knowl- 
edge of the micro-organisms which cause the disease is yet in its in- 
fancy, it is impossible to decide whether the injuries noted in the 
motor areas of the cord and medulla are due to toxins or the lodg- 
ment of the germs themselves. In the case of rabies it seems reason- 
ably sure that the micro-organisms themselves select the nervous sys- 
tem. In such instances as the injury of the motor areas by tetanus 
poison, that of certain peripheral nerves by diphtheria toxin, or even 
the characteristic lesions of post-syphilitic maladies like tabes, we 
can be reasonably sure that we are dealing with the specific action 
of the poisons, independent of actual localized growth of the infec- 
tious agents. 
Diphtheria toxin, after distribution through the body, may act 
upon many different tissues, as is evident by degenerations in the 
heart muscle, liver, and kidney, and the petechial hemorrhages in 
serous surfaces. In addition to this general action, however, there is 
a very marked selection of certain nerve centers. By Meyer and 
Gottlieb 44 diphtheria toxin is classed as a specific vascular poison. 
Its action results in a rapid sinking of the blood pressure with final 
cardiac death in spite of artificial respiration. These manifestations 
seem to have a central origin, with particular action upon the vagi 
and the phrenic nerves. Apparently also the localization of the 
diphtheritic lesion may influence the selection of individual nerves, 
the most concentrated action taking place upon the nerves whose 
endings are distributed in this particular region, for, as Meyer and 
Ransom 45 have shown, this poison, like tetanus toxin, may be ab- 
sorbed into the nerves directly through the nerve endings. An in- 
teresting selective action also of diphtheria poison is the apparently 
specific alteration of the suprarenal glands which is regularly no- 
ticed, as enlargement and congestion, in diphtheria-infected guinea 
pigs, and which has been associated by many workers with the char- 
acteristic drop in blood pressure which accompanies all severe cases 
of the disease. Abramow 46 has studied this lesion particularly, and 
believes that it consists in a degeneration and final disappearance of 
44 Meyer and Gottlieb. “Pharmacology Trans. Halsey/’ Lippincott, 1914, 
p. 556. 
45 Meyer and Ransom. Arch, de pharmacodyn., Yol. 15, 1905, also Meyer, 
Berl. klin. Woch., 25 and 26, 1909, also Arch. f. exp. Path. u. Ther., Yol. 
60, 1909. 
48 Abramow. Zeitschr. f. 1mm., Yol. 15, 1912. 50 
INFECTION AND RESISTANCE 
the chromaffin substance and of the medullary cells. He believes 
that this, together with degeneration of the heart muscle itself, is of 
great importance in causing the characteristic vascular failure. 
In botulinus poisoning there is, as Marinesco 47 and Kempner 
and Pollack 48 have shown, a direct effect upon the cells of the an- 
terior horns with degenerative changes in the Nissl granules. 
Tetanus poison, which has been studied extensively by pharma- 
cologists, shows a very marked affinity for the nervous system, as, in 
fact, the symptoms of tetanus indicate. Indeed, while many of the 
bacterial poisons are distributed by the blood stream to the point of 
final attack, in tetanus the absorption of the toxin from the lesion or 
the point of injection takes place entirely by the path of the nerves, 
entering by way of the motor nerve endings. 
That this method of poison distribution might be, among others, 
an important one was suggested as early as 1892 by Bruschettini,49 
who found tetanus toxin in the nerves but not in the adjacent muscle 
and other tissues surrounding the point of subcutaneous injection. 
Similar results were obtained subsequently by Hans Meyer, whose 
experiments were confirmed and extended by Marie and Morax.50 
Finally Meyer and Ransom 51 furnished complete proof that the 
poison was absorbed from the blood and tissues by the peripheral 
nerve endings alone and was transported centripetally only by the 
paths of the neurons. The experimental facts elicited may be sum- 
marized as follows: 
1. When tetanus toxin is injected into the thigh muscles of a 
guinea pig the poison is found at first only in the sciatic nerve of the 
same side and in the blood. (The determination of poison was made 
by injecting macerations of the respective tissues into mice.) If ex- 
amination was delayed until the symptoms had become generalized, 
the poison was found in the opposite sciatic, but the muscle bundles, 
fat, etc., from the vicinity of the injection area were poison-free.52 
2. When a nerve is cut poison absorption ceases as soon as axis 
cylinder degeneration has set in. 
3. If the nerve is cut before the poison is injected the distal 
end contains poison, the proximal end does not. This again shows 
that the nerve absorbs the toxin not from its capillaries but solely 
through the end organs. 
4. If a nerve which already contains poison is severed, toxin 
47 Marinesco. Compt. rend, de la soc. de biol., Yol. 3, 1896. 
48 Kempner and Pollack. Deutsche med. Woch., 32, 1897. 
49 Bruschettini. Riforma medica, 1892. 
60 Marie and Morax. Ann. de VInst. Past., 1902. 
61 Meyer and Ransom. Archiv f. exp. Path. u. Pharm., 49, 1903. 
62 In view of our discussion of the importance of fats in the absorption 
of tetanus toxin, it seems inconsistent that the toxin does not concentrate 
in fatty as well as in nervous tissues. This Meyer explains by the inactive 
and poorly vascularized condition of the fat tissues. BACTERIAL POISONS 
51 
will disappear rapidly from the proximal end, since it no longer 
obtains a renewed supply from the periphery. 
5. If antitoxin is injected into the nerve, above the point of 
injection, it will successfully bar the way for the ascending toxin. 
6. Severing of the spinal cord prevents the passage of the 
poison from below upward. 
These facts ascertained in the case of tetanus find their parallel in 
the phenomena of the distribution of rabic virus 53 as well as in that 
of poliomyelitis, in both of which there seems to be a progressive cen- 
tripetal transportation through the nerves. 
However, in these conditions we are probably dealing not with a 
poison but with a living virus and, though analogous, the conditions 
are not entirely comparable. 
From the practical point of view these facts regarding tetanus 
may explain the frequent failure of therapeutic success attending 
the injection of tetanus antitoxin after the symptoms of the disease 
have set in, since in such cases the poison is already distributed to 
the nerves and is largely inaccessible to the antitoxin. They also 
have pointed a way toward a more hopeful therapy, namely, the 
method of injecting the antiserum directly into the nerves about the 
point of injury. It is not surprising, however, in view of the stated 
facts, that even this is unsuccessful when done at too late a time, 
after a considerable amount of poison has already passed above the 
point of injection to the spinal centers. 
Such selective action on the part of the bacterial poisons is en- 
tirely analogous to the similar specific action of alkaloids, narcotics, 
and other drugs. In order that the poison may act upon a cell we 
must, of course, assume that it has either chemical or physical affin- 
ity for this cell. The problem, as many writers have pointed out, is 
strongly analogous to that of tissue staining. A dye must be able to 
form a chemical union with the cell or it must be soluble in the cell 
substance in order to stain it. The chemical difference between cells 
is a delicate one and not often definable by our present methods. We 
can obtain an insight into the principles probably underlying selec- 
tive action only by inference from the relation between the chemical 
constitution of drugs or their physical properties, solubility, etc., and 
their respective tissue affinities. These problems are difficult and, to a 
large extent, obscure. They cannot be directly investigated upon 
bacterial poisons since these are themselves of chemically unknown 
nature. But the study of drugs of known constitution has revealed 
certain definite relations of this kind which have furnished analogies 
from which the general principles of selection in bacterial poisons 
can be surmised. 
It is a well-known fact to pharmacologists that there is a definite 
63 Di Yestea and Zagari. Fortschr. d. Med., Yol. 6, 1888. 52 
INFECTION AND RESISTANCE 
relation between chemical structure and toxicity. Fraenkel 54 ex- 
presses it as follows: “By the addition of identical atom groups in 
an identical manner, similarly acting substances are obtained.” lie 
cites the well-known example of curare; whichever the path by which 
this poison is injected it leaves intact the tissues with which it comes 
in contact, but after general distribution acts specifically upon the 
nerve endings. It had been discovered by Brown and Fraser 55 that 
by introducing methyl radicles (CII3) into molecules of various alka- 
loids, strychnin, morphin, atropin, and others, substances were ob- 
tained which paralyzed nerve endings, and this irrespective of their 
previous physiological action. It appears that the combination of 
four methyl radicles attached to the nitrogen atom (quaternary bases) 
universally possesses this paralyzing action. Tertiary bases on the 
other hand lack this property. 
Ammonium base” 
“Tertiary base” 
Quaternary 
Subsequently Bohm 56 discovered that curare contains two bases 
--the one, “curin,” is slightly toxic and is a tertiary base; the other, 
which possesses the typical curare action, “curarin,” is an “ammo- 
nium base.” By “metliylizing” curin, curarin could be obtained. 
From these and other examples it is clear that in a certain num- 
ber of cases actual chemical affinity must play a part in toxic action; 
on the other hand, there are mapy cases in which toxic action seems 
to depend merely upon physical conditions such as solubilities. 
Meyer and Overton’s well-known theory of narcosis maintains that 
certain narcotics exert their action by passing out of blood and 
lymph solution into solution by the fat-like, lipoidal substances 
(lecithin, cholestrin, etc.) contained in the nerve cells, because the 
latter are better solvents for them than is the blood plasma. This 
theory of Meyer and Overton has stimulated much investigation and 
speculation, and it is not unlikely that it is valid in the case of many 
narcotics, although it does not explain the action of narcotics in gen- 
eral; for Dickson notes that chloral hydrate, for instance, is more 
54 Sigmund Fraenkel. “Arzneimittel Synthese,” 2d Ed., Springer, Ber- 
lin, 1906. 
55 Brown and Fraser. Trans. Royal Soc. of Edinburgh, 25, 1868, cited 
from Fraenkel. 
56 Bohm. Arch, de Pharm., cited from Fraenkel. See also Dickson, A 
Manual of Pharmacology/’ E. Arnold, London, 1912. BACTERIAL POISONS 
53 
soluble in water than in oils, and some narcotic drugs like alcohol 
exert definite action on proteins and are oxidized in the body. These 
are pharmacological questions of which we cannot speak with author- 
ity. We wish merely to point out that the action of poisons upon the 
body may depend in some cases upon mere physical or mechanical 
relationship between the two.57 
As regards bacterial poisons the union between poison and sus- 
ceptible cell is extremely firm and difficult to dissociate in many in- 
stances, and this points to the possibility that, in these cases at least, 
true chemical union takes place rather than merely a loose combina- 
tion like that of the solution of one substance in another. Further- 
more, the complete inactivation of some poisons by mixture with the 
cells of tissues capable of binding them would likewise point to more 
than mere physical union.. Nevertheless, it does not by any means 
exclude the thought that the poisons may, in fact, go into selective 
relationship with special cells because of physical properties, such as 
solubility in the lipoidal cell membranes,58 and may subsequently 
be bound chemically or destroyed by oxidation or enzymotic hydroly- 
sis after such entrance. In such a case the actual specificity would 
yet depend on purely physical properties. 
In addition to the specific physical and chemical affinities be- 
tween the poisons by certain cells there are probably also certain 
fortuitous factors connected with the distribution and local accumu- 
lation of the poisons which have some weight in determining the 
location of injury. For the specific selection is not absolutely strict 
and there are probably few parenchyma cells in the body that are 
entirely insusceptible to injury if the poisons are sufficiently con- 
centrated upon them. Thus, to cite an analogy from the toxicology 
of non-bacterial poisons, in lead poisoning, as Meyer and Gottlieb 
point out, the paralysis of the extensors of the arm occurs chiefly in 
adults who use these muscles in the exercise of their professions 
(painters, type-setters), while in children and in animals, in which 
no such selective use of particular muscle groups is habitual, lead 
paralyses are atypical, attacking legs as well as arms. It is not un- 
likely that the frequent injury of the heart muscle by bacterial poi- 
sons or the irregular parenchymatous changes in various organs is 
determined by analogous fortuitous factors, in that functional activ- 
ity and increased metabolism may predispose to injury. 
Bacterial poisons also may produce their lesions in the course of 
excretion. This seems likely in the case of typhoid poisons in which 
57 Ivar Bang. “Biochemie der Lipoide,” Bergmann, Wiesbaden, 1911. 
Meyer and Gottlieb. “Experimentelle Phannakologie,” 2d Ed., Urban & 
Sehvvartzenberg, Berlin, 1911. 
58 For Overton’s theory of osmosis see R. Hober, “Physikalische Chemie 
der Zelle u. Gewebe,” Leipzig, Engelmann, 1911. Compare also, regarding 
this entire question, the discussion in P. Th. Miiller, “Vorlesungen iiber 
Immunitat, etc.,” Fischer, Jena, 1910. 54 
INFECTION AND RESISTANCE 
we have often seen bloody diarrhea in rabbits within a few hours 
after intravenous injection of powerfully toxic culture filtrates. In 
connection with the dysentery bacillus Flexner and Sweet59 have 
studied the conditions carefully. They succeeded in showing first 
that the introduction of the dysentery poison into the lumen of the 
intestine does no harm and that the toxin is slowly destroyed by peptic 
and tryptic digestion. They concluded that probably no absorption 
of the poison through the uninjured intestinal mucosa takes place. 
They then showed that the toxin after intravenous administration 
is excreted by the intestine and that the inflammatory reactions and 
injury of the mucosa are incident to this act of elimination. 
Whether or not the kidneys are injured in the same way it is 
difficult to decide. In many infectious diseases, of course, the bac- 
teria themselves pass through the kidney into the urine, and renal 
injury may result from the actual presence of the bacteria in the 
kidney; however, renal injury may also occur without this, and it 
is not at all impossible that the conditions here are similar to those 
just described for the intestine. 
All the facts which we have considered indicate that, although 
most bacterial poisons can injure many different tissues, yet in some 
cases there is a particular susceptibility on the part of an individual 
tissue which is independent of accidental factors and seems to be 
due to specific chemical or physical affinity. It seems even that in 
tetanus, botulismus, and a few other conditions there is a differential 
selection of particular areas within a tissue like the nervous system, 
just as this occurs in the case of certain drugs. As stated above we 
have no satisfactory scientific explanation for this, but a great deal 
of work has been done to show that the bacterial poisons actually 
unite with and are taken up by the susceptible tissues. 
Indirectly, proof of this has been brought by the demonstration 
of the rapid disappearance of various toxins from the blood streams 
of susceptible animals and their persistence in the circulation of 
animals insusceptible to them. Thus Donitz60 has shown that 
tetanus toxin injected into the blood stream of a susceptible animal 
rapidly diminishes in quantity, and Knorr,61 in similar experiments, 
showed that the demonstrable disappearance of such toxins out of 
the blood stream is synchronous with the appearance of symptoms, a 
fact which excludes disappearance by excretion. Conversely Asa- 
kawa 62 showed that in pigeons, which are but slightly susceptible, 
tetanus poison could be demonstrated in blood, liver, spleen, kidneys, 
and muscles six days after injection, but not in the brain, showing 
59 Flexner and Sweet. Jour, of Exp. Med., Yol. 8, 1906. 
60 Donitz. Deutsche med. Woch., No. 27, 1897. 
61 Knorr. Fortschr. der Medizin, 1897, No. 17, and Munch, med. Woch., 
1898, Nos. 11 and 12. 
62 Asakawa. Centralbl. f. BaTct., Yol. 24, pp. 166 and 234. BACTERIAL POISONS 
55 
that in this organ, at least, there must have been either a union or a 
destruction of the poison. Similar to these results are those of 
Metchnikoff,63 who found the poison unchanged after two months in 
the circulation of insusceptible animals (lizards). 
Direct evidence of union between susceptible tissues and poison 
has been furnished by the experiments of Wassermann and Takaki,64 
who showed that the brain and cord tissues of rabbits and guinea 
pigs, mixed with tetanus toxin before injection, served to neutralize 
its harmful effects. And it appears that the toxin-neutralizing prop- 
erty of the brain substances of various animals is proportionate to 
their individual susceptibility to the poison. Thus Metchnikoff 65 
not only confirmed the results of Wassermann and Takaki for rab- 
bits and guinea pigs, but showed further that the brains of chickens, 
animals that are but moderately susceptible, possess a correspond- 
ingly slighter neutralizing power, and, further, that brain tissues of 
entirely insusceptible cold-blooded animals, turtles and frogs, pos- 
sess absolutely no neutralizing properties. 
The original interpretation by Wassermann of these facts was 
based on the assumption that the poison was bound to the brain tissue 
just as it is bound to antitoxin. Experiments by Besredka 66 have 
cast some doubt upon this. This worker’s experiments seem to indi- 
cate that a brain emulsion which has been saturated with the toxin 
can be rendered capable of absorbing more toxin if tetanus antitoxin 
is mixed with it. In other words, the affinity of the antitoxin for the 
toxin is stronger than that of the brain substance for the poison, and 
that the union toxin-brain tissue is very easily dissociated; as indeed 
it should if the union were purely a physical one depending on solu- 
bility. 
After it had been shown that the poisons which acted specifically 
upon certain cells were actually taken up by these cells, a number of 
attempts were made to determine chemically the tissue element which 
united with the poisons. Eoguchi 67 showed that cholesterin and 
alcoholic extracts of blood serum neutralized tetanolysin. The same 
thing was later shown by Muller,68 and Landsteiner 69 showed that 
ether extracts of red blood cells likewise neutralized this poison. 
In a later study by Landsteiner and von Eisler 70 the relation of the 
tissue lipoids to various toxic substances was still more definitely 
established. They studied first the various hemolysins and found 
that extraction of blood cells with ether rendered the stromata less 
83 Metehnikoff. “L’lmmunite clans les maladies Infect./’ Paris. 
64 Wassermann and Takaki. Berl. klin. Woch., 1898, No. 1. 
65 Metehnikoff. Ann. de Vlnst. Past., 1898, p. 81. 
68 Besredka. Ann. Past., 1903, p. 138. 
67 Noguchi. TJniv. Pa. Med. Bull., Nov., 1902. 
68 Muller. Centralbl. f. Bakt.,. Vol. 34, 1903. 
89 Landsteiner. Wien. kl. Rundschau, 13, 1905. 
70 Landsteiner and von Eisler. Centralbl. f. Bakt., 39, p. 318, 1905. 56 
INFECTION AND RESISTANCE 
capable of binding the hemolytic substances. The same thing they 
showed for bacteriolysins, in the latter case demonstrating at the 
same time that the ether extracts of bacterial bodies possessed slight 
binding properties for the bactericidal substances of the serum. These 
experiments have, of course, a merely indirect significance in the 
present connection, since they do not deal with the type of poisons 
we have discussed. However, Landsteiner and Botteri 70“ also 
worked with tetanus toxin, and found that protagon obtained by 
alcoholic extraction of human brain possessed a powerfully neutral- 
izing effect upon the toxin, markedly greater than that possessed by 
other lipoidal substances. 0.15 gram neutralized 120 minimal lethal 
doses of the toxin. 
Takaki,71 who investigated these relations in great detail, iso- 
lated an alcohol-soluble element, cerebron, from nerve tissues, a sub- 
stance to which he ascribes the toxin-binding properties. Overton 
and Bang 72 found, furthermore, that cholesterin and lecithin inhibit 
the action of cobra venom, a poison which is in so many ways similar 
to those produced by bacteria. Taking into consideration all avail- 
able evidence, we are forced to admit that the lipoids seem to play ail 
important role in determining the selective action of the nervous sys- 
tem by the bacterial poisons. It may not, of course, be an influence 
depending merely upon the solubility of the harmful substances in 
the lipoids themselves. Tor, as Bang expresses it, “the lipoids pos- 
sess to a high degree the property of altering by their presence the 
solubilities of other bodies,” and it is quite possible that in the tis- 
sues they are present as lipoid-protein combinations. Their action 
in determining the solubility of toxins in a given cell may therefore 
be a purely indirect one. 
It is of some interest in this connection to recall the experiments 
of De Waele,73 which bring out another clear analogy between alka- 
loids and bacterial poisons in their relation to lecithin. He found 
that the addition of small quantities of lecithin increases the activity 
of both toxins and alkaloids in the animal body, whereas larger 
amounts inhibit both. 
From the foregoing, then, it is plain that in order to gain en- 
trance to the body and establish themselves there, certain criteria 
as to the pathogenic nature of the species of bacteria involved, and 
the virulence of the infecting strain as balanced against the sus- 
ceptibility of the host, must coincide. According to the biological 
attributes of the bacteria, the path of entrance must be by a certain 
route adapted to their pathogenic habits. The infection thus having 
come about, further progress takes place, varying again according to 
70a Landsteiner and Botteri. Centralbl. f. Baht., 42, 562. 1906. 
71 Takaki. Beitr. zur chem. Phys. u. Path., 11, No. 19, 1908. 
72 See Ivar Bang, “Biochemie der Lipoide,” Bergmann, Wiesbaden, 1911. 
73 De Waele. Zeitschr. f. Immunity Yol. 3, 1909, p. 504. BACTERIAL POISONS 
57 
the biological properties of the bacteria; in some cases consisting 
very largely of local growth, tissue infiltration, extension and 
eventually, perhaps, entrance into the blood stream, distribution 
through the body, and in certain varieties of infections, secondary 
localization in the organs, with or without abscess formation. In 
other cases the organisms may remain localized, injuring the body 
by the production of a powerful poison. Between the two extremes, 
both factors may be active. Injury to the body, on the one hand, 
is by virtue of the local struggle which takes place between invaders 
and the tissue cells and fluids of the host, with injury to function 
and afterwards tissue destruction; and, on the other hand, by the 
absorption of toxic products of various kinds incident to the growth 
of the bacteria. These toxic products may be either the powerful 
and specific exotoxins, they may be products absorbed after libera- 
tion from the destroyed bacterial cells, or they may even be toxic 
substances produced by proteolytic cleavage of bacterial and tissue 
elements. The cause of death varies considerably. In the case of 
the true exotoxin formers, like diphtheria and tetanus, death will be 
due to the severe intoxication and the injury of specific cellular ele- 
ments necessary to life, the local, mechanical causes being negligible. 
However, even in diphtheria, when the localization of the process 
is such that it obstructs the respiratory passages, especially the 
larynx, a purely mechanical death may be possible. This, however, 
is a special case. 
The cause of death in the cases of bacteria in which the toxin 
formation is of a less severe degree is not absolutely clear. There 
may be extensive tissue degenerations in organs like the liver, the 
kidney, the heart muscle, etc., which eventually lead to death. There 
may be, as in some cases, as very severe malaria, perhaps anthrax 
and other infections in which the parasitic organisms become enor- 
mously increased in number, an actual interference with capillary 
circulation. In many cases, such as meningococcus infection, there 
may be actual mechanical and local toxic injury to parts of the 
central nervous system which lead to death. 
The causes of death by bacterial infection, therefore, may be 
manifold, but it is not likely that even in cases like streptococcus, 
anthrax and staphylococcus infections where we have but very rudi- 
mentary knowledge of poison formation, that the toxic element is 
ever entirely absent. CHAPTER III 
OUR KNOWLEDGE CONCERNING NATURAL IMMU- 
NITY, ACQUIRED IMMUNITY, AND ARTIFICIAL 
IMMUNIZATION 
NATURAL RESISTANCE AGAINST INFECTION 
In the preceding chapters we have confined ourselves largely to 
the consideration of those properties of the bacteria which determine 
their ability to infect. In this discussion, however, we have repeat- 
edly emphasized the fact that every infectious disease is the result of 
a struggle between two variable factors—the pathogenic powers of 
the bacteria on the one hand, and the resistance of the subject on the 
other, each of these again modified by variations in the conditions 
under which the struggle takes place. Thus a given micro-organism 
may be capable of causing fatal infection in one individual but may 
be only moderately virulent or even entirely innocuous for another. 
Conversely the same individual may be highly susceptible to one 
variety of bacteria, but highly resistant to others. Even in reactions 
with one and the same micro-organisms, the susceptibility or resist- 
ance of the individual may be determined by variations in the physi- 
ological state or by the environmental conditions under which the 
two factors—invader and invaded—are brought together. There- 
fore, the conceptions “resistance,” “immunity,” and its opposite 
“susceptibility,” are relative terms which can never be properly dis- 
cussed without careful consideration of all modifying conditions 
which influence them. 
The science of immunity deals with a detailed analysis of these 
variables. Its ultimate practical aim is the determination of meth- 
ods by which an original susceptibility can be transformed into 
resistance or even immunity. And the rational method of approach- 
ing this subject consists in a careful study of the conditions of 
susceptibility and immunity as they exist naturally in the animal 
kingdom. 
The mere fact that both animals and man are in constant con- 
tact with infectious micro-organisms, many of them in a high state 
of virulence, indicates in itself that the animal disposes normally 
over a defensive mechanism of considerable efficiency. 
To a certain extent, of course, this escape from harm is due to 
the external defences of skin and mucous membrane which, in the 
58 NATURAL IMMUNITY 
59 
healthy state, mechanically prevent the entrance of the micro-organ- 
isms into the body. For we have seen, in another place, that few of 
the bacteria can pass through the uninjured surfaces. Moreover, 
added to this, there is some protection in the bactericidal properties 
of the secretions. An example of this is the inhibitory power exer- 
cised by the acidity of the normal gastric juice upon the cholera 
spirillum. In order to infect the intestinal canal of guinea pigs 
with these organisms Koch found it necessary to neutralize the gas- 
tric juice with sodium carbonate solutions, and other observers have 
found it necessary to inject directly into the duodenum. But even 
after entrance into the animal tissues a second line of defence is 
normally encountered by all invading germs which tend to inhibit 
their further progress more or less perfectly. This active opposition 
to the bacteria after their entrance is expressed chiefly in the anti- 
bacterial (bactericidal) activity of the blood serum, and the pha- 
gocytic powers of leukocytes and other cells. To a certain extent 
these forces are active against all bacteria in all animals, but they 
may vary in different species, races, or even individuals in potency 
against any given infectious agent, and, to a certain extent, varia- 
tions in resistance may be referable to this. The analysis of these 
forces, both in the normal and in the artificially immunized animal, 
forms the substance of the systematic discussions which are to fol- 
low, and, for the present, we will confine ourselves to an examination 
of the facts that have been gathered regarding the actual differences 
in normal resistance or “Natural Immunity” between various 
species of animals. 
And if we glance over the list of diseases to which different spe- 
cies and races of animals are victim, it is immediately evident that 
some animals are never spontaneously infected with many of the 
micro-organisms that cause extensive and fatal ravages in others. 
Also, within the same race or species, an epidemic sweeping through 
a community will kill many individuals and leave others unscathed. 
Such differences point to variations in the defensive mechanism, 
since the invader in these cases is the same. We speak, therefore, of 
Natural Immunity which is an attribute of species, that which, 
within the same species, is racial, and that which, within the same 
race, is individual. And the attempts to discover the causes under- 
lying such differences in natural resistance have elucidated many of 
the fundamental principles of immunity in general. 
Instances of natural immunity which appear to depend on spe- 
cies are common. We have pointed out, above, that in order to make 
infection at all possible, it is necessary that the invading germ shall 
find suitable cultural conditions in the body of the host. It is this 
simple principle which probably explains the fact that bacteria which 
cause disease in warm-blooded animals cannot, as a rule, cause dis- 
ease in those that are cold-blooded, and vice versa. Thus frequent 60 
INFECTION AND RESISTANCE 
attempts to produce anthrax in turtles, frogs, and other cold-blooded 
species have failed. Also among wann-hlooded animals differences 
in body temperature have been shown to influence susceptibility. 
Thus avian tuberculosis does not develop in mammals, nor do the 
human and bovine types of tubercle bacilli infect birds. And this is 
probably due to the fact that the avian bacillus has become adapted 
to growth at from 40° to 45° C., about the normal temperature of 
birds, while the mammalian bacilli cease to grow when the tempera- 
ture is raised above 40° C. Another observation which clearly illus- 
trates the influence of body temperature upon susceptibility is that 
made by Gibier 1 upon anthrax. Frogs are ordinarily resistant to 
this disease. When they are kept in water at 35° C. a fatal infec- 
tion can be produced. Nutt all’s 2 experiments with plague infection 
in lizards illustrate the same point. Kept at 16° 0., no infection 
could take place. Warmed to 26° C., they could be readily infected. 
It is ordinarily assumed that these results are explicable upon the 
basis of purely cultural and temperature considerations. And this, 
indeed, is most likely. It is possible, however, that an additional 
factor involved in this may be the lowering of the general resistance 
of cold-blooded animals when warmed, just as warm-blooded animals 
can be rendered susceptible by chilling. 
It is for similar simple cultural reasons, possibly, that diseases 
which occur spontaneously in carnivora do not occur in purely 
herbivorous animals. The relative resistance of dogs to anthrax 
and to tuberculosis may possibly be accounted for in this way. 
However, there are many micro-organisms which infect easily 
both carnivorous and herbivorous animals, and it may well be that 
the frequently cited cases we have mentioned above depend on fac- 
tors more complicated than mere cultural conditions incident to 
metabolic differences. In most cases of species resistance, indeed, 
simple nutritional conditions alone do not serve as valid explanations. 
Species resistance may be so perfect that it amounts to an ab- 
solute immunity. This is apparently so in the cases cited above, 
namely the immunity of the cold-blooded species to certain diseases 
of warm-blooded animals. However, such examples are exceptional. 
When we are dealing with diseases of warm-blooded animals only, 
natural resistance, in all but a limited number of cases, is sufficient 
only to prevent the spontaneous occurrence of the particular disease, 
or to prevent infection when experimental inoculation with moderate 
doses is practiced upon normal animals. In most of these cases, 
however, when the dose experimentally administered is excessive, or 
the resistance is lowered artificially, by chilling or any other 
form of local or general injury, infection can be accomplished. In 
the case of protozoan diseases species adaptation is much more rigid 
1 Gibier. Compt. rend, de I’acad. des sc., Yol. 94, 1882. 
2 Nuttall. Centralbl. f. Bakt., Yol. 22, 1897 NATURAL IMMUNITY 
61 
and parasites that infect one species are very often restricted en- 
tirely to that class, being unable to infect any other animal, even 
though no striking difference in temperature or metabolism exists. 
We may convey the clearest conception of all such species differ- 
ences by a tabulation of some of the more important infectious dis- 
eases of man with a statement in each case concerning its transmissi- 
bility to animals, as follows: 
Tuberculosis, human type, spontaneously infects man. It is 
very often observed in monkeys kept in captivity. Cattle, swine, and 
sheep are probably never spontaneously infected; guinea pigs are 
highly susceptible to experimental inoculation. Cattle, swine, sheep, 
and rabbits are relatively very resistant to experimental infection. 
Dogs and goats are still more so. Birds seem to be entirely 
refractory. 
Tuberculosis, bovine type.—Spontaneous infection occurs in 
domestic animals, chiefly cattle; it is less frequent in sheep, hogs, 
and horses; it has been reported in dogs and goats. In man infec- 
tion does occur, but only a small percentage of human tuberculosis 
is of the bovine type, and these cases are almost exclusively in chil- 
dren. In tabulating 1,042 cases which have been carefully studied, 
Park and Krumwiede 3 report the following figures: 
Cases of Tuberculosis in Man (1042) 
Over 16 years 
Human type 677, bovine type 9. 
5 years to 16 years 
Human type 99, bovine type 33. 
Under 5 years 
Human type 161, bovine type 59. 
The large majority of bovine infections were abdominal or in- 
volved cervical lymph nodes. 
Experimental infection is successful in rabbits and guinea pigs, 
both of these animals succumbing more rapidly to this than to the 
human bacillus. In fact, the relative resistance of rabbits to the 
human bacillus is such that rabbit inoculation is one of the most 
important methods of differentiating between the two types. Birds 
are refractory. 
Tuberculosis of the avian type occurs spontaneously in birds. 
It may be experimentally produced in rabbits (Strauss and Gama- 
leia). Injected into cattle it causes a local reaction only. 
Tuberculosis of cold-blooded animals is not transferable to 
warm-blooded animals. 
Syphilis spontaneously occurs in man only. It can be inoculated 
into chimpanzees, in which primary and secondary lesions develop, 
corresponding mildly to human syphilis. Primary lesions can be 
3 Park and Krumwiede. Jour, of Med. Res., Yol. 23, 1910. 62 
INFECTION AND RESISTANCE 
produced in lower monkeys. It can be transferred by intratesticular 
inoculations to rabbits. 
Gonococcus infection occurs spontaneously in man only. No 
typical lesions can be produced in experimentally inoculated ani- 
mals, though death can be caused by large doses, probably by toxic 
action. 
Influenza bacillus spontaneously infects man only. Experi- 
mental infection is partly successful in monkeys only. (Pfeiffer 
and Beck, Deut. med. Woch., 1893.) 
Glanders.—Spontaneous infection occurs in horses and mules; 
less frequently in sheep, goats, and camels. This disease, like plague, 
may be regarded as primarily a disease of animals, but man may be 
infected by direct or indirect contact with the diseased animal. All 
domestic animals may be infected experimentally with ease, except 
cattle and rats, in which cases large doses are necessary. Birds show 
local reactions only. (Wladimiroff—in “Kolle und Wassermann 
Handbuch,” Vol. 5, 2d Ed.) 
Plague occurs spontaneously chiefly in man and in rats. It has 
also been found in California ground squirrels and in hogs during 
plague epidemics in Hong Kong. It is highly infectious for guinea 
pigs and white rats—slightly less so for mice; rabbits are much less 
susceptible than guinea pigs. Dogs, cats, and cattle are relatively 
resistant. Birds appear to be immune. Cold-blooded animals are 
immune unless artificially warmed. (See above.) 
Malta fever occurs spontaneously in man and in goats. It is 
pathogenic for all mammals, but it is not fatal for lower animals 
when the organisms are directly cultivated out of the human body. 
Diphtheria occurs spontaneously in man only. Experimental in- 
oculation is fatal in guinea pigs, rabbits, dogs, cats, and birds. Rats 
and mice are highly resistant. The typical pseudomembranous in- 
flammation can be produced in susceptible animals only after pre- 
vious injury of the mucous membrane, and then it rarely shows any 
tendency to spread. 
Tetanus is spontaneous in man, horses, cattle, and sheep. It is 
found rarely in dogs and goats. Birds are highly resistant to ex- 
perimental inoculation. 
Anthrax is primarily a spontaneous infection of cattle, sheep, 
and horses; it occurs in man largely through direct or indirect con- 
tact with these animals. Guinea pigs, rabbits, and white mice are 
very susceptible to experimental inoculation. Rats and hogs are less 
susceptible, and dogs are relatively resistant, though they can be 
regularly killed by moderate doses intravenously injected. Birds 
and cold-blooded animals are highly resistant. 
Asiatic cholera develops spontaneously in man only. Rabbits 
and guinea pigs can be killed by injections of cultures, but die prob- 
ably of toxemia. In rabbits a cholera-like condition has been pro- NATURAL IMMUNITY 
63 
duced by injection of the spirilla into the duodenum after ligation 
of the common bile duct. (Nikati and Rietsch, Ref. in Deut. med. 
Woch., Yol. II, 1884, p. 613.) Ordinarily no multiplication takes 
place in the animal body. Pigeons are insusceptible, a fact which 
helps to distinguish this organism from Spirillum metchnikovi and 
other similar bird-pathogenic spirilla. 
Typhoid fever occurs spontaneously in man only. It has recently 
been produced in a mild form in chimpanzees. Animals are suscep- 
tible to the endotoxins and can therefore be killed by injections of 
bacilli and extracts, but the organism is not invasive as in the case of 
the lower animals. Typhoid septicemia can be produced in rabbits 
by inoculating them with especially virulent cultures of the bacilli, 
or cultures previously grown on rabbit-blood agar (Gay). The 
typhoid-carrier state may ensue for considerable periods in such 
animals. 
Pneumococcus infection in various forms occurs spontaneously 
in man. Rabbits, mice, and guinea pigs are highly susceptible. 
Rats, dogs, cats, cattle, and sheep are relatively resistant. 
Straphylococcus and streptococcus infections may occur in al- 
most all of the warm-blooded animals, chiefly as abscess producers. 
In horses a severe form of pleuropneumonia is caused by them. 
Leprosy occurs spontaneously in man only. Lesions simulating 
human leprosy have been produced in monkeys by inoculation, and 
partially successful experiments have been made upon the Japanese 
dancing mouse. Other animals are immune. 
Scarlet fever occurs spontaneously in man only. Monkeys may 
possibly be susceptible, though not all observers have been successful 
in such experiments. (Draper and Handford, Journ. of Exp. Med., 
Yol. 17, 1913.) Landsteiner and Levaditi (Ann. Past., Vol. 25, 
1911) have succeeded in producing the disease in the chimpanzee, 
though they failed with lower monkeys. 
Small-pox occurs spontaneously in man only. It is probably 
identical with cow-pox. (See reasons for this assumption given by 
Hacius as cited by Paul in “Kraus and Levaditi Handbuch,” etc., 
Vol. 1.) It can be experimentally produced in monkeys. 
Measles develops spontaneously only in man. Macacus rhesus 
has been successfully inoculated by Anderson and Goldberger (U. S. 
Pub. Health Reports, 26, 1911). Other animals are immune. 
Typhus fever occurs in man only. Experimentally it has been 
produced in chimpanzees, Macacus, Cercopithecus, Ateles, and My- 
cetes monkeys. Anderson has succeeded in producing temperature 
reactions in piinea pigs by injecting blood from typhus patients or 
from other similarly infected guinea pigs. More exact information 
concerning this disease will probably be available soon, if the 
reported cultivation of the organism of the disease by Plotz, is, 
&ut;b<?hti<?ated. 64 
INFECTION AND RESISTANCE 
Yellow fever up to the present has been observed in man only. 
Poliomyelitis is spontaneous in man only. Can be transmitted 
to monkeys and—in a doubtful form—to rabbits. No other animals 
are known to be susceptible. 
The above represents an incomplete tabulation of che variations 
in susceptibility in the animal kingdom for infections which occur 
spontaneously in man. They will illustrate sufficiently, however, 
the facts of variable species susceptibility as we have stated them. 
We might, with equal profit, tabulate the infections occurring spon- 
taneously in any single species of animal and show how variable 
would be their pathogenic powers for other animals and for man. 
Thus man is immune to the organism which causes cattle plague, 
and to that of chicken cholera, and probably to many other diseases 
peculiar to animals, though, of course, in the case of infections of 
the human being we are entirely dependent for such information 
upon observed immunity to spontaneous infection, and upon a few 
instances of accidental inoculation. 
In regard, also, to differences of susceptibility between various 
races, within the same species, many interesting facts have been ob- 
served. Thus gray mice are, as a rule, more resistant to strepto- 
coccus and pneumococcus infection than are white mice. Algerian 
sheep are said to be more resistant to anthrax than are European 
sheep. Of black rats inoculated by Muller 4 with anthrax over 79 
per cent, survived, while of white rats similarly inoculated only 14 
per cent, survived. 
In man, too, racial differences are marked. The extraordinary 
susceptibility of the negro to tuberculosis is familiar to all American 
physicians, and it is well known that Eskimos transported to tem- 
perate climates and civilized conditions are particularly prone to 
contract this disease. Small-pox is considered a relatively mild 
disease in Mexico. Dr. James Carroll 5 stated that whites are more 
susceptible to yellow fever than are negroes, and that among the 
latter those living nearest the equator are less susceptible than are 
the more northern races. There seems to be no doubt about the 
actual occurrence of such racial differences, although, as Hahn 6 
very justly points out, many instances formerly regarded as racial 
differences of susceptibility may have been simulated by racial, 
or often religious, differences of custom that influence sanitary con- 
ditions, and consequently the incidence of epidemic disease. 
Apart from the explanations furnished in a few instances by 
gross physiological differences such as body temperature, the factors 
determining species resistance are largely a mystery, and in the 
4 Muller. Fortschr. der Med., 1893. Cited from Sobernheim, in “Kolle 
u. Wassermann Handbuch,” 2d Ed., Vol. 3. 
5 Carroll in “Mense, Tropenkrankheiten,” Vol. 2, p. 124. 
6 Hahn in “Kolle und Wassermann’s Handbucb,,, Vol, 1, NATURAL IMMUNITY 
65 
matter of racial variations, of course, we have no instances in which 
such very obvious physiological factors play a part. In attempting 
to find causes for differences of resistance or susceptibility in gen- 
eral, the nature of the problem makes it necessary for us to examine 
it from a number of different points of view. A micro-organism 
may be infectious for a given species of animal more than for 
another, because of special adaptation to the conditions, nutritive 
and otherwise, encountered in the tissues of these animals. Such 
adaptation is illustrated in the experience of Pasteur with “rouget” 
and with rabies, where passage through one variety of animal en- 
hanced the virulence for this species but reduced it for others; and 
the same thing is easily demonstrated in the laboratory with so many 
bacteria that it may be accepted as a principle underlying enhance- 
ments of virulence in general. This adaptation implies that, to a 
certain extent, the part played by the animal body in determining 
its own susceptibility is passive. Gonococcus, for instance, infec- 
tious for man only, requires human protein for growth, at least in 
its first generations outside the body. Its ability to cause disease 
in man may be largely dependent upon its cultural need of human 
protein. The resistance of other animals to this disease, then, is, in 
part, due to their failure to supply proper nutriment. This, as Kolle 
points out, is analogous to Atrepsie, a term used by Ehrlich, in 
speaking of the insusceptibility of one species to cancerous growths 
originating in another. 
Again, “adaptation” on the part of the bacteria may imply, not 
only an increased ability to meet altered cultural conditions, but an 
actual acquisition of greater offensive or invasive powers with which 
to meet the particular defences opposed to it by the given animal. 
Thus the increased virulence of typhoid bacilli after cultivation in 
immune sera would point toward an increased ability to survive 
under the adverse conditions encountered in the animal body. An 
organism may possibly acquire particular infectiousness for one 
species to the exclusion of others, by a succession of spontaneous 
inoculations—comparable to the experimental passage of the micro- 
organism through animals of the same species. This is especially 
probable in diseases such as gonorrhea, syphilis, and some others 
where infection is usually direct from one person to another. And 
it is these diseases particularly in which infectiousness is rather 
strictly limited to the human species. 
Inheritance of Immunity.—Regarding the matter purely from 
the point of view of the animal body and the factors which determine 
its powers to ward off a given infection, we may justly assume that 
natural resistance may be largely a matter of inheritance. Whether 
this is to be interpreted as purely an instance of survival of the 
fittest or whether immunity acquired by an individual can be wholly 
or in part transmitted to the offspring is an open question—at pres- 66 
INFECTION AND RESISTANCE 
ent in the same state of unclearness as are other questions relating 
to the transmissibility of acquired characteristics. However this 
may be, there are a number of facts available which indicate that 
inheritance plays an important part. It is apparent in the case 
of many diseases afflicting human beings that infection takes a 
milder course in those races among which it has long been endemic 
whereas the same disease, suddenly introduced among a new peo- 
ple, is relatively more severe and spreads more rapidly. This seems 
to be the case with yellow fever and tuberculosis, and in measles 
and small-pox, too, the principle seems to hold good. Syphilis when 
hist described authentically—as epidemically sweeping through 
Europe toward the close of the 15th century—appears to have been 
a far more acute and violent disease than it is among us to-day. 
It may well be that this depends upon a gradual elimination (elimi- 
nation in this case, especially as far as reproduction is concerned) 
o those individuals that are fortuitously more susceptible and, by 
natura! selection, a higher racial resistance is graduallv developed. 
Whether or not direct inheritance of the individually acquired 
immunity can be considered at all as a contributing factor is difficult 
to decide. That immunity can be transmitted from mother to off- 
spring was observed by Chauveau 7 as early as 1888. Lambs thrown 
by anthrax-immune ewes possessed a higher resistance against this 
infection than did the lambs of normal ewes. The extensive experi- 
ments of Ehrlich carried out chiefly upon mice with the vegetable 
poisons ncm and abrm, showed that in these cases immunity may 
be transmitted from mother to offspring, but depends upon a passive 
transfer of the specific antitoxins both by the blood and the milk 
of the mother. The sperm of the father did not seem to have any- 
thmg to do with inherited resistance, since no immunity followed 
m the offspring when immunized males were paired with normal 
cmales; I rom the complete absence of immunity in the second 
generation (grandchildren) of the immunized female, and from 
he short duration (2 to 3 months) of its persistence, he concluded 
that the ovum itself had no influence, but that the entire phenomenon 
was attributable to a passive transference of antitoxins from mother 
““ ””8 gestation and lactation. He interpreted, in the same 
sense, Chauveau s anthrax experiments, and similar experiments of 
omas and Kitasato with symptomatic anthrax, suggesting 
t lat, here also, a transfer of antibodies from mother to offspring had 
taken place. The experiments of Ehrlich permit of no doubt as to 
le validity of Ins conclusions. However, we must remember that 
they wore carried out with antitoxic immunity only, in which the 
7 Chauveau. Ann. Pasteur, 1888. 
8 Ehrlich. Zeitschr. f. Hyg., 1892, Vol. 12. 
9 Thomas. Compt. rend, de Vacad. des sc., Yol. 94, cited by Ehrlich loc cit 
l0Kitasate. Cited by Ehrlich, loc. cit, NATURAL IMMUNITY 
67 
resistance is purely dependent upon the circulating antibody and 
is never, even in actively immunized individuals, a permanent state. 
The ricin work of Ehrlich, then, constitutes nothing more than 
an example of the transfer of antibodies from mother to offspring, 
a passive immunization which in no sense can be regarded as a true 
somatic inheritance of immunity. This question of the transfer- 
ence of antibodies from mother to offspring has given rise to a con- 
siderable amount of investigation, a most extensive piece of work 
on the transmission of hemolytic antibodies emanating from the 
laboratory of Famulener,11 who came to the conclusion that in goats, 
at least, such antibodies are chiefly transmitted with the milk, and 
not through the placenta. The question is of considerable impor- 
tance in connection with the transfer of diphtheria antitoxin from 
mother to child in human beings, where we shall see in a later chap- 
ter that an amount of antitoxin sufficient to protect is often found 
in the new-born child, and lasts in adequate amount up to, or nearly 
up to the first year of life. Since Famulener’s work, most pedia- 
tricians have assumed that such transfer in human beings took place 
through the milk, more especially through the colostrum during the 
first days of feeding, and much weight was given to this opinion 
by the observations of Theobald Smith 12 concerning the importance 
of colostrum in preventing a premature colon bacillus infection of 
the bowel in calves. Apparently calves that never have had 
colostrum, but were brought up from the beginning without nursing 
from the mother, in many cases succumb to a disease called scours, 
consisting of a generalized colon bacillus infection. A recent investi- 
gation in our own laboratory by Drs. Kuttner and Ratner 13 has 
shown that in human beings the conditions are unlike those existing 
in the calf. Careful measurement of antitoxin values in colostrum 
and milk of mothers and in the cord blood have shown that what- 
ever antitoxin is transmitted, passes through the placenta, hardly 
any being found in the colostrum. To our minds, in spite of much 
that has been written to the contrary of recent years, these experi- 
ments, carefully carried out, should settle the question of maternal 
transmission to the child through the placenta, at least in human 
beings. It is quite likely that the conditions in human beings may 
differ from those prevailing in many animal species, because of the 
anatomical differences between the human placenta and that of 
many lower animals. In the human being there is only one layer 
of connective tissue between the maternal of the fetal blood in the 
placental layers; whereas, in animals, two, three and even four cell 
layers may intervene.14 
11 Famulener. Jour. Inf. Bis., X, 332, 1912. 
12 Theobald Smith. Jour. Exp. Med., August, 1922. 
13 Kuttner and Ratner, in press. 
14 Grosser. “Eihaiite u. Plazenta,” Braumiiller, Leipzig, 1909. 68 
INFECTION AND RESISTANCE 
In immunity such as that acquired against typhoid fever, plague, 
cholera, and other diseases after recovery from an attack, the in- 
dividual remains relatively resistant long after the demonstrable 
antibodies have disappeared from the circulation, and we must 
assume that this permanent resistance depends upon a physiological 
alteration—inexplicable for the present, but surely residing in the 
body cells. In such cases it is by no means certain that there may 
not be a very slight, but through generations gradually acumulating, 
inheritance of immunity. At any rate the experiments of Ehrlich 
do not disprove such a possibility. Moreover, in this connection 
it must not be forgotten that natural immunity, unlike acquired 
immunity, cannot be passively transferred from one animal to 
another, and implies therefore a fundamental cellular difference 
rather than a condition depending merely upon antibodies circulat- 
ing in the blood. 
For this last reason also it has been unsatisfactory to attempt 
explanations of natural immunity purely upon grounds of bacteri- 
cidal and other properties of the blood serum. These points we will 
take up at greater length when we discuss the mechanism of resist- 
ance in general. 
An important observation upon the inheritance of serum prop- 
erties is that which has been made by Ottenberg and Epstein 15 in 
connection with the iso-agglutinins. We shall see in another section 
that the blood serum of one human being will often possess the 
property of agglutinating the human blood cells of another indi- 
vidual. These iso-agglutinating constituents of the serum are ap- 
parently transmitted from parents to offspring. Von Dungern and 
Hirschfeld,16 in studying these iso-agglutinins in 72 families, upon 
348 people, not only confirmed the observations of the preceding 
workers, but showed that such inheritance follows Mendelian laws. 
Hot only is this of great biological interest, but it is of great im- 
portance in connection with our present discussion in showing that 
such properties as agglutinating powers of serum can be influenced 
by inheritance from the father as well as from the mother. 
Individual Differences of Resistance in Same Race.—The indi- 
vidual differences in resistance which unquestionably exist among 
members of the same species and races are very difficult to 
but, as far as we can tell anything about them at all, they seem to 
depend upon variation in what is popularly spoken of as “general 
condition.” The laboratory animals with which most experimenta- 
tion is done, rabbits and guinea pigs, if healthy, show very slight 
individual variations. In fact, the astonishing uniformity of re- 
action on the part of guinea pigs of similar age and weight against 
measured quantities of bacterial toxins has alone made it possible 
15 Ottenberg and Epstein. Proceedings of the N. Y. Path. Soc., 1908. 
16 Yon Dungern and Hirschfeld. Zeitschr. f. ImmunitatsYol. 4, 1910. NATURAL IMMUNITY 
69 
to utilize these animals in the standardization of antitoxins. Pneu- 
mococcus and streptococcus cultures can be measured with reason- 
able accuracy upon white mice of approximately uniform weight, 
and the same animals are relatively uniform in their reactions to 
identical amounts of tetanus poison. Many other examples might 
be cited which make it clear that healthy animals of the same species, 
kept under the same conditions, fed upon the same food, and of ap- 
proximately the same age and weight, differ but slightly from each 
other in reaction to the same infectious agent. This would indicate 
that the individual differences in resistance displayed so plainly by 
human beings are due, not to any fundamental individual variations, 
but rather to such fortuitous factors as nutrition, metabolic fluctua- 
tions, temporary physical depression, fatigue, or chilling. A person 
suffering from functional impairment of any kind is more likely 
to permit the invasion of a pathogenic micro-organism than is a per- 
fectly healthy well-nourished individual of the same species. 
Most of these facts we know from the accumulated experience of 
clinicians who also have given us much valuable information con- 
cerning the susceptibility to infection on the part of chronically dis- 
eased persons, especially diabetics and nephritics. In the case of a 
few of these influences, chilling and fatigue, experimental data on 
animals are available. It is, however, extremely difficult to analyze 
the causes underlying such depression of resistance. For instance, 
with fatigue or chilling there may he temporary congestion of mucous 
surfaces, due to vasomotor influences, which alter the secretions on 
mucous surfaces, or interfere with the normal mobilization of 
leukocytes, permitting penetration of bacteria where ordinarily 
they would have been held back. Our ignorance is nowhere more 
clearly illustrated than in the fact that we know practically nothing 
concerning the relation between a thorough chilling and the acquisi- 
tion of wdiat is spoken of as a common “cold.” We can only assume 
that there is interference in some way with the normal bactericidal 
and phagocytic mechanisms, making possible the penetration and 
lodgment of small quantities of bacteria, ordinarily destroyed imme- 
diately after entrance or prevented from entering at all. 
Of course we must except those individual differences of sus- 
ceptibility which may be dependent upon inheritance. We know, 
for instance, that in such diseases as diphtheria, where resistance 
depends upon antitoxins circulating in the blood, there may be a 
passive immunity, conferred from mother to offspring, which lasts 
for several weeks or months after birth. It is important to remem- 
ber such a possibility in the selection of guinea pigs for diphtheria 
antitoxin standardization, as Anderson has pointed out. Whether 
or not a tendency to tuberculosis can be inherited is still an open 
question. In most cases it is more than probable that the supposedly 
inherited tendency to tuberculosis is not really an inherited sus- 70 
INFECTION AND RESISTANCE 
ceptibility, but rather an actual infection acquired during childhood 
from the parents. Cornet and Kossel,17 who have recently sum- 
marized the statistics dealing with this problem, have come to the 
conclusion that this factor, namely, infection from the parents, 
probably is the cause of the greater frequency of tuberculosis among 
children of tuberculous parents, and that there is no definite proof 
of inherited susceptibility. 
ACQUIRED IMMUNITY AND IMMUNIZATION 
We have outlined in the preceding pages the differences in sus- 
ceptibility to various diseases apparent among different species of 
animals, and have noted that the degree of resistance of some ani- 
mals to infection with germs rapidly fatal to others is often suffi- 
ciently well-marked to be termed “immunity.” Such immunity, 
because it is a natural biological attribute of the species, as much a 
characteristic property as are its anatomical or physiological prop- 
erties, has been spoken of as “Natural Immunity 
It is a matter of common knowledge, however, that among 
species of animals readily susceptible to certain infections resist- 
ance, or even extreme resistance, i. e., immunity, may he acquired 
by an attack of the disease. Thus human beings who have recovered 
from plague, small-pox, typhoid fever, cholera, the exanthemata, 
mumps, typhus, yellow fever, and a number of other conditions do 
not ordinarily contract the disease a second time. In some of these 
conditions, notably cholera, plague, typhoid fever, and small-pox, 
the rule is almost invariable. In others, such as measles, scarlet 
fever, and mumps, a second attack may occur, though it is rare. A 
few infections like pneumonia and influenza may recur at relatively 
short intervals. 
The following table briefly indicates infectious diseases in which 
permanent immunity follows an attack: 
Infectious Diseases in Which One Attack Conveys Lasting Immunity 
Plague. 
Typhoid—second attack rare—about 2.4 per cent. (Curschmann). 
Cholera. 
Small-pox—second attack very rare. 
Chicken-pox—second attack very rare. 
Scarlet fever—second attack about 0.7 per cent. 
Measles—second attack uncommon, hut less rare than scarlatina. 
Yellow fever. 
Typhus fever. 
Syphilis—reinfection rare, though more common than formerly supposed. 
Mumps—second attack rare (Kraus). 
Poliomyelitis. 
17 Cornet and Kossel in “Kolle u. Wassermann,” Yol. 5, 2d Ed. ACQUIRED IMMUNITY 
71 
No lasting immunity is conferred by one attack in: 
Infection with the Pyogenic cocci 
Gonorrhea 
Pneumonia 
Influenza 
Glanders 
Dengue fever 
Diphtheria in general protection, second attack in 0.9 
per cent, cases—0.01 antitoxin unit per c. c. of circu- 
lating blood protects 
Recurrent fever 
Tetanus 
Erysipelas 
Beri beri 
Tuberculosis 
There is another group of diseases in which the immunological 
conditions after infection are as yet not clear—namely, protozoan 
infections like malaria and trypanosomiasis, and treponema diseases 
like syphilis. In these conditions reinfection seems to be impossible 
only so long as the individual still harbors the micro-organism, but no 
lasting immunity is conferred. We have discussed these conditions 
in extenso in the section on syphilis. In tuberculosis, also, resist- 
ance to superinfection is dependent on the presence of a focus. 
These observation's actually form the point of departure of that 
entire branch of medical science which devotes itself to the study of 
resistance to infection, serum diagnosis, and specific therapy, and 
it will be seen that all the facts that have been gathered upon these 
subjects are the fruits of detailed analysis of this phenomenon of 
acquired immunity. 
Its occurrence in many instances has been so striking that 
ancient observers, long before the birth of rational medicine, re- 
ferred to it, and often drew from it conclusions of great hygienic 
importance. Thucydides, in the second book of his account of the 
Peloponnesian Wars, in describing the plague at Athens, notes the 
apparent safety from reinfection of those who had recovered, sug- 
gesting the possibility of their being therefrom immune against 
disease in general. The literature of the Middle Ages and of earlier 
modern times contains numerous further references which indicate 
that acquired resistance was clinically recognized as a result of re- 
covery from many diseases. The phenomenon was not only observed, 
but put to practical utilization by the ancients of China and India. 
Thus the practice of inoculating children with small-pox material 
from the active pustules of patients, or making them sleep in beds or 
wear the shirts of sufferers was a dangerous practice but logical, ou 
the reasoning that the disease conveyed to a person in full health 
and good condition would probably take a mild course and confer 72 
INFECTION AND RESISTANCE 
immunity, while the naturally acquired disease, contracted often 
because of the weak and debilitated condition of the individual, 
would be more apt to end fatally. 
Such methods, though barbaric and eventually unjustified by the 
naturally high mortality incident upon them, were actually brought 
to Europe from the East, and for a time practiced in European 
countries. 
The first great advance which bridged the gap between the obser- 
vations regarding naturally acquired immunity and rational experi- 
mental imunization was made by Edward Jenner. It had been 
noticed before Jenner began his work that milkmaids and others who 
had contracted cow-pox in the course of their occupations were usually 
spared when a small-pox epidemic occurred in their community. Spo- 
radic attempts had been made to put this observation to practical use, 
but no one with sufficient intelligence, persistence, and training had 
taken up the matter seriously. Jenner, interested by the reports of 
this nature and by his own observations, was especially impressed by 
the similarity between the local manifestations of small-pox, cow-pox, 
and a disease of horses spoken of as “grease.” Though at first disin- 
clined to identify small-pox with cow-pox (at present the prevailing 
opinion is that the second is an attenuated form of the former), Jen- 
ner thoroughly investigated cases of alleged protection by cow-pox, a 
claim which before this had been hardly more than a rumor, and 
finally, with the encouragement of John Hunter, proceeded to the 
vaccination of human beings with cow-pox, testing the result by 
subsequent inoculation of the same individual with small-pox. His 
report to the Royal Society in 1796 and his subsequent publications 
incorporate the results of these experiments by means of which the 
practice of vaccination against small-pox was introduced and the 
virtual eradication of the disease from civilized communities was 
attained. 
The principles underlying small-pox vaccination are extremely 
simple. The attenuated virus after inoculation incites a mild and 
localized form of the disease, from which the subject recovers rapidly 
and completely. The recovery implies the mobilization of certain 
protective forces and a specific physiological alteration of the body in 
such a way that a permanently, or at least prolongedly, increased re- 
sistance against the disease remains. In consequence, if the indi- 
vidual is subsequently exposed to spontaneous infection with this dis- 
ease, his acquired specific resistance suffices to prevent invasion by 
the virus. This is merely an artificial imitation of the conditions 
which obtain when an individual recovers from an attack of a disease 
and is rendered immune thereby. In this case, however, the attenua- 
tion of the virus lias eliminated the dangers attendant upon an actual 
attack. The immunity thus conferred is probably never as perfect nor 
as lasting as that following a seizure of the disease in its unattenuated ACQUIRED IMMUNITY 
73 
form; however, it suffices, as a rule, to prevent spontaneous infec- 
tion which is never as severe a test as experimental inoculation. 
In contrast to the “Natural Immunity’’ which is an inherited at- 
tribute of race or species, we speak of such increased resistance in a 
member of an originally susceptible race as “Acquired Immunity.” 
When the immunity has been attained as the result of an attack of 
the disease itself it is spoken of as “Naturally or Spontaneously 
Acquired ImmunityWhen produced by some form of treatment 
with the virus of the disease, altered in such a way that an actual 
attack is averted, we speak of it as “Artificially Acquired Immunity 
The premises of Jenner’s reasoning were valid as his experiments 
were convincing. But knowledge regarding infectious disease and 
its causation by living germs was not developed until almost one 
hundred years later, by the work chiefly of Pasteur. For this reason 
no direct continuation of Jenner’s work appeared until Pasteur 
made his communication upon Chicken Cholera to the Parisian 
Academy of Medicine in 1880. Though his investigations differed 
entirely from those of Jenner both in method and the nature of the 
disease with which they dealt, Pasteur recognized the similarity of 
the fundamental principles underlying both discoveries. 
His observations took origin in a purely accidental occurrence. 
Cultures of chicken cholera which had been allowed to stand without 
transplantation and under aerobic conditions for periods of several 
months were found to have diminished in virulence. Inoculated into 
chickens, they failed to kill, giving rise in many cases to localized 
lesions only. It occurred to Pasteur that inoculation with such an 
attenuated culture might protect against subsequent infection with 
fully virulent strains and, indeed, experimental investigation of this 
idea proved to be correct. He developed a method of “vaccination” 
against chicken cholera which consisted in injecting first a very 
much attenuated culture of the organism (premier vaccin), and, 
after 12 or 14 days, another less perfectly attenuated (deuxieme 
vaccin), since he observed that a single inoculation was often in- 
sufficient to confer protection. After two inoculations a degree of 
immunity could be attained which sufficed to protect against spon- 
taneous infection as well as against experimental inoculation with 
doses of the virulent germs, fatal for untreated animals. 
These experiments, simple as they are, constitute the beginnings 
of the science of Immunity, since, for the first time, an investigator 
working with a pure culture of a pathogenic micro-organism had 
succeeded, in planned and purposeful experiments, in conferring 
artificial immunity. The path was now clearly indicated and the 
years immediately following were fruitful in the development of 
many methods by which pathogenic bacteria may be attenuated and 
changed in such a way that they can be used to confer immunity 
without causing more than a transient and harmless reaction in the 74 
INFECTION AND RESISTANCE 
subject. Most of the earlier discoveries of this kind came from 
Pasteur himself and from members of his school. 
Since in all these methods the inoculated animal attains its in- 
creased resistance by reason of the activities of its own tissues, these 
processes are spoken of as “Active ImtnunizationNo protective 
factor is conferred directly. The disease itself is inoculated, though 
in an altered form, and the subsequent immunity is purely the result 
of the physiological reaction occurring as the subject struggles against 
and overcomes the injected virus, bacteria, or their products. Such 
“Active Immunization,” we shall see, is in contrast to “Passive 
Immunization” a procedure in which the serum of an actively im- 
munized animal is injected into another, carrying with it certain 
substances by which protection is conferred. The recipient here is 
passively protected by products of the active reaction which has 
taken place in the body of the donor. 
After his success in 'active immunization against chicken cholera 
Pasteur applied the principles here learned to experiments upon the 
protection of animals against anthrax. This problem was fraught 
with considerable difficulty because of the great virulence of the 
anthrax bacillus. However, successful attenuation was attained by 
a method which depended upon the cultivation of anthrax cultures at 
temperatures above the optimum for its growth. Toussaint18 had 
shown that the resistance of sheep could be increased if they were 
inoculated with blood from animals dead of anthrax after this had 
been heated to 55° C. for ten minutes. Toussaint’s idea had been 
that by heating the blood in this way the bacteria themselves were 
killed. Pasteur 19 showed, however, that this was not the case, buti 
that what actually occurred was a reduction of the virulence of the 
strain by the exposure to heat. As a matter of fact, morever, the 
method of Toussaint did not furnish a reliable means of attenuating 
anthrax, and Pasteur succeeded in developing a far more satis- 
factory procedure on which he based a practical method for the pro- 
tective vaccination of sheep and cattle. 
His method was as follows: 20 Virulent anthrax bacilli were cul- 
tivated at 42° to 43° C. on neutral chicken bouillon (Sobernheim 
states that horse or beef broth—or even agar—answers the same pur- 
pose). Cultivated under these conditions a gradual and progressive 
reduction of virulence occurs. After about 12 days of such cultiva- 
tion the culture as a rule no longer kills rabbits, but is still virulent 
for guinea pigs and mice. After twenty-four or more days the 
virulence for rabbits and guinea pigs is lost and mice only can be 
18 Toussaint. Compt. rend, de Vacad. des sc., 1880. 
19 Pasteur, Chamberland and Koux. Compt. rend, de Vacad. des sc., Vol. 
91, 1881. 
20 Cited from Sobernlieim. “Kraus und Levaditi Handbuch der Technik, 
etc.,” Yol. 1, 1909. ACQUIRED IMMUNITY 
75 
killed with it. The latter—the most fully attenuated strain—was 
called premier vaccin by Pasteur, and, in the immunization of cattle 
or sheep, is first injected. After 10 or 12 days the stronger deuxieme 
vaccin is administered. This is the method which Pasteur used in 
his now classical experiments at Pouilly-le-Fort, in which he con- 
vinced a hostile audience of the efficacy of his immunization. Sheep 
were protected in the manner indicated, and 14 days after the last 
injection a fully virulent culture was inoculated and the animals 
found capable of successfully resisting it. 
In the train of this work many other methods of producing active 
immunity have been devised—all of them of considerable theoretical 
interest and many of them practically adapted to some special case. 
We may conveniently classify these methods as follows: 
I. Immunization with Living but Attenuated Cultures 
(1) Methods in which the attenuation is obtained by heating. 
This is the method of Toussaint as outlined above, in which anthrax 
blood was heated to 55° C. for 10 minutes, and is probably the least 
efficient or reliable method for the attenuation of the anthrax bacillus. 
It has been applied to rabies by Babes (cited from Kraus in “Kraus 
u. Levaditi Handbuch, etc.,” Yol. 1, p. 708), who attenuated 
the virus by heating to 58° C. for periods varying from 2 to 40 
minutes. 
(2) Attenuation by prolonged cultivation of the bacteria at 
temperatures above the optimum for their growth. This is illustrated 
by Pasteur’s anthrax immunization as described in the preceding 
paragraphs. 
(3) Attenuation by passage through animals. Examples of this 
are Pasteur’s experiments with the “rouget” organism, in which pas- 
sage through rabbits diminished the virulence for hogs. The attenu- 
ation of rabic virus by passage through monkeys is another instance, 
and Jennerian vaccination is also an example of this, although here 
the attenuation by passage through cattle is attained naturally and 
not by experimental procedures. Based on the same principle is 
Behring’s 21 method 22 of immunizing cattle against tuberculosis by 
inoculating them with tubercle bacilli of the human type. 
(4) Attenuation by prolonged growth of bacteria on artificial 
media in the presence of their own metabolic products. This is the 
method first employed by Pasteur in chicken cholera, as described 
above, and is applicable to many organisms, such as pneumococci, 
streptococci, and others. In fact, it is difficult to maintain the viru- 
lence of many of these bacteria unless special methods of cultivation 
21 Behring. “Therapie der Gegenwart,” April, 1907. 
22 See also Romer, “Kraus u. Levaditi Handbueh,” lgt Suppl., p. 310. 76 
INFECTION AND RESISTANCE 
or passage through animals are practiced. Pasteur believed that 
free access of oxygen to the cultures increases the rapidity of the 
attenuation. 
(5) Attenuation by drying. The classical example for this 
method is the Pasteur method of prophylactic immunization against 
rabies. Rabbits are inoculated with virus fixe, and their spinal 
cords dried for varying periods in bottles containing KOI! at a tem- 
perature of about 25° C. The virus grows progressively weaker 
with each day of drying. Greater details concerning this method are 
given in another place. 
(6) Attenuation by the use of chemicals.—Chamberland and 
Roux23 succeeded in attenuating anthrax by growing it in the 
presence of various antiseptics. They used carbolic acid 1 to 600, 
bichromate of potassium 1 to 1,500 and sulphuric acid 1 to 200, 
and found that, after a short time of cultivation under such condi- 
tions, the bacilli lost their ability to form spores and became aviru- 
lent for sheep. Behring 24 and others have applied this method to 
the attenuation of diphtheria toxin ; Behring adds terchlorid of iodin, 
Roux potassium iodid—iodin solutions. The principle, of course, 
is not exactly the same in the last cases, since here the attenuation 
is not of the bacteria themselves, but rather of the toxin. 
(7) Attenuation by cultivation under pressure. This method 
is difficult to apply, and has no striking advantages over other pro- 
cedures. It was employed by Chauveau 25 for the attenuation of 
anthrax. He succeeded in accomplishing this by cultivation of 
anthrax bacilli at 28° to 39° C. at a pressure of 8 atmospheres. 
II. Active Immunization with Fully Virulent Cultures in 
SuBLETIIAL AMOUNTS 
The original methods of Pasteur carried out with chicken cholera 
and anthrax were aimed particularly at diminution of virulence, 
since these organisms, as isolated from the diseased animal, are so 
extremely infectious that it would he very difficult—(in the case 
of many animals, impossible)—to inoculate with the unattenuated 
germs without producing fatal disease. However, in the case of 
many other infections it has been found feasible to inoculate normal 
animals with the fully virulent germs in such small quantities that 
the body can successfully overcome them, and, in doing so, acquire 
specific resistance. It is obvious that this method is more easily 
carried out with the organisms which Bail terms “half parasites” 
than with organisms as highly infectious as anthrax. Ferran 20 
23 Chamberland and Roux. Compt. rend de Vacad. des sc., 96, 1882, 
24 Behring and Wernicke. Zeitschr. f. llyg., 12, 1892. 
25 Chauveau. Compt. rend, de Vacad. des sc., Yol. 98, 1884, 
28 Ferran. Compt. rend, de Vacad. des sc,, 1885, ACQUIRED IMMUNITY 
77 
applied this method both to animals and to human beings with broth 
cultures of cholera spirilla. Ilogyes27 has introduced a similar 
procedure for immunization against rabies by injecting dilutions of 
fully virulent rabic virus, beginning with a dilution of 1 to 10,000 
and rapidly working up to a dilution of 1 to 10. In tuberculosis 
immunization with fully virulent cultures in small amounts has been 
attempted by Webb, Williams, and Barber,28 using the Barber 
method of isolation, and giving a single micro-organism at the first 
injection. That such a method is feasible, if carried out with suffi- 
cient care, even with the most virulent germs, was demonstrated by 
the same workers. They succeeded in immunizing animals against 
anthrax (with cultures kept 12 hours on agar) 29 by injecting a 
single thread (3 to 6 bacilli) as the first dose, and then gradually 
increasing the amount. 
In the general laboratory immunization of animals treatment 
with virulent bacteria in sublethal doses is of considerable value and 
frequently employed. 
It would seem that possibly this method or some modification of 
it will be found to have very definite advantages over methods in 
which either attenuated or dead bacteria are employed. Bail’s work 
upon the aggressins and upon anti-aggressin immunity (see chapter 
I, page 17) has opened the possibility that virulent bacteria pro- 
vide, within the living body, specific aggressive substances which are 
not produced in the test tube. If this proves to be true, and the 
question is by no means settled, it may be necessary in such cases 
to immunize with organisms which are in a condition capable of 
producing these aggressins. Sublethal doses of fully virulent or- 
ganisms would furnish these conditions more perfectly than at- 
tenuated avirulent strains, in which the invasive (aggressive) power 
is considerably diminished. 
The methods of active immunization so far described differ from 
those which are to follow in that the preceding were all based upon 
the use of living bacteria or virus, whereas the methods to be de- 
scribed below depend upon the treatment of animals with dead bac- 
teria or bacterial products. It is well to call attention in this place 
to the fact that a number of recent investigations seem to point to 
the greater efficiency of immunization with living germs. This 
method has recently given hopeful results in the case of plague in 
the hands of Strong,30 and Metchnikoff and Besredka,31 in their 
attempt to vaccinate chimpanzees against typhoid fever, make the 
27 Hogyes. “Lyssa Nothnagels Handbuch, etc.,” Vienna, 1897. 
28 Webb, Williams, and Barber. Jour, of Med. Res., Vol. 15, 1909. 
29 This was not possible where the organisms were taken directly from the 
blood of a dead mouse. In such cases even a single thread caused fatal disease. 
30 Strong. Jour, of Med. Res., May, 1908. 
31 Metchnikofif and Besredka, Ann, Past., Vol. 25, 1911. 78 
INFECTION AND RESISTANCE 
statement that vaccination with dead typhoid bacilli or autolysates 
does not confer adequate protection, hut that this can be attained by 
treatment with small doses of the living bacilli. 
In speaking of this subject it is well to mention recent ob- 
servations upon immunization with “sensitized” bacteria,32 although 
this necessitates anticipatory reference to subjects not so far dis- 
cussed. It is a matter of common experience in laboratories that 
rabbits and other animals will withstand relatively large amounts of 
pathogenic bacteria if these are first treated with heated specific 
immune serum (sensitized). This is probably due to the fact that- 
such “sensitized” micro-organisms are very rapidly taken up by 
phagocytes. In spite of the phagocytosis, immunity is developed. 
Metchnikoff and Besredka, in the communication alluded to above, 
state that typhoid vaccination with unaltered living bacilli is efficient, 
but is attended by severe local and general reactions. If the living 
bacilli are first “sensitized” no such severe reaction occurs and im- 
munization is nevertheless successful. The recent work of Gay points 
in the same direction, and it is at least possible that by the practice of 
sensitization we may be able to employ living unattenuated or- 
ganisms for purposes of immunization more extensively than we have 
in the past. 
III. Active Immunization with Dead Bacteria, and Bacterial 
Extracts 
This method is the one most extensively practiced in the labora- 
tory immunization of animals. It is usual in most experiments of 
this kind to inject dead organisms once or twice before living bac- 
teria are administered. High degrees of resistance can in some 
instances be attained by progressively increasing doses of dead cul- 
tures only. This method is not only useful in experimental work, 
but is clinically employed in the active immunization of human 
beings as introduced by Wright and as applied, before Wright, to 
tuberculosis (tuberculin treatment). But it is very probable that 
the immunity so attained is not entirely comparable to the immunity 
following an attack of a disease, nor even that produced by the in- 
jection of living bacteria. 
The method employed for killing the bacteria is of considerable 
importance since, both by excessive heating as well as by too vigorous 
chemical treatment, the immunizing properties of the bacterial 
protein may be destroyed. In employing heat it is a safe rule never 
■to expose the bacteria for prolonged periods to temperatures which 
considerably exceed the thermal death point. As a rule, heating non- 
?2 See discussion on “sensitization” of antigens in chapter VR ACQUIRED IMMUNITY 
79 
spore-forming bacteria to a temperature of from 65° to 70° C. for 
thirty minutes will suffice to kill them without too radically altering 
the immunizing properties of the protein constituents.33 
If the temperature is not raised above 60° C., and this is ad- 
vised by many workers, the suspensions must be carefully controlled 
by cultural tests before they are used, at least for the treatment of 
human beings. As wTe shall see in a later section, the best results 
have been obtained when heating was not carried beyond 53° to 
55° C. 
When bacterial death is to be accomplished by chemicals the 
antiseptics most commonly used are carbolic acid (0.5 per cent.), 
toluol (removed before use of vaccine by filtration or evaporation), 
chloroform, and formaldehyd (1 per cent.). 
Pfeiffer, who was one of the first to practice the immunization 
of animals with dead bacteria on an extensive scale, believed that, 
in the case of bacteria which were toxic by reason of their intra- 
cellular constituents (endotoxins), the injection of the cell protein 
itself, whether dead or alive, was the sole essential for successful 
immunization. The method developed by Kolle 34 and by Pfeiffer 
and Marx 35 for the prophylactic immunization of human beings 
against cholera depends upon the injection of cholera cultures 
emulsified in salt solution, killed by exposure to 58° C. for one hour, 
and further insured against contamination by the addition of 0.5 
per cent, phenol. The application of this method to other diseases, 
both prophylactically and therapeutically, is more fully discussed 
in another place. 
Since the essential point in such immunization is the introduc- 
tion of the bacterial protein, it is often customary to inject bacterial 
extracts instead of the whole cells. This has been especially desir- 
able in the case of such insoluble micro-organisms as the tubercle 
bacillus, where the injection of the whole dead organism produces 
localized reactions similar to those caused by the living bacteria.36 
Thus “Old Tuberculin,” as commonly used, is a glycerin-broth ex- 
tract of tubercle bacilli. The method has been extensively used and 
a variety of procedures have been devised for bacterial extraction. 
These have included simple autolysis of the bacterial bodies in alka- 
line broth, shaking in salt solution in mechanical shakers, trituration 
with salt or sand, trituration after freezing, digestion with proteo- 
lytic enzymes, and extraction by pressure in a Buchner press. 
33 In a subsequent chapter we shall see that the physical changes pro- 
duced in an antigen by heat result in differences in the antibodies formed 
after animal inoculation. This point has practical significance in the present 
connection. See also the chapter on agglutinins, the work of Joos there dis- 
cussed, and Friedberger and Moreschi, Centralbl. f. Bakt., 1905, Yol. 39. 
34 Kolle. Deutsche med. Woch., 1897, p. 4. 
35 Pfeiffer and Marx. Deutsche med. Woch., 1898. 
36 Prudden and Hodenpyl. N. T. Med. Journal, 1891. 80 
INFECTION AND RESISTANCE 
We may mention some of the more important methods for pre- 
paring bacterial extracts for purposes of immunization and antigen 
production in general as follows: 
A. Extraction of Bacteria by Permitting Them to Remain for 
Prolonged Periods in Fluid Media 
The bacteria may be grown upon slightly alkaline bouillon and 
kept at incubator temperature for one to two months. They are 
then filtered through Bcrkefeldt or other suitable filters. This is the 
common method of producing antigen for precipitin reactions, in fact 
the method employed by Kraus in the discovery of the bacterial pre- 
cipitins. It is by no means certain whether the antigens prepared in 
this way represent simple extractions or autolytic products of the 
bacteria; probably both processes take place. The antigenic value 
of the fluids obtained in this way is never very great. From such 
filtrates Brieger and Mayer, Pick, and others have attempted to 
obtain the antigen in a purified form by chemical precipitation. 
Pick37 precipitates the bouillon filtrate by saturation with am- 
monium sulphate; the precipitate is redissolved in water and again 
precipitated with ammonium sulphate and the resultant precipitate 
dried on a filter. It is then dissolved in water and precipitated with 
alcohol. The sticky substance which comes down represents the 
antigen. 
Suitable extracts can occasionally be obtained also by emulsify- 
ing agar cultures in physiological salt solution and allowing them 
to stand for twenty-four hours or more at incubator temperature. 
In our own experience we have found this method rather inefficient 
for yielding strong extracts. More efficient extraction is usually ob- 
tained when the bacteria are suspended in alkaline fluids such as 
N . 
— sodium hydrate. Lustig and Galleotti digest the bacterial mass 
for 24 hours with 1 per cent. KaOH, then precipitate with am- 
monium sulphate, dry in vacuo and pulverize.38 
Recently, also, Uhlenhuth 39 has employed the proprietary prep- 
aration “antiformin”40 for the production of antigens. This 
37 Pick. “Hoffmeister’s Beitrage, etc.,” Vol. 1, 1902. For an extensive 
discussion of the various methods employed for the production of bacterial 
antigens by chemical methods see Pick in Kraus und Levaditi, etc., Yol. 1, and 
the same author in Kolle u. Wassermann, etc., 2nd Ed., Yol. 1. 
38 See Pick. Loc. cit. 
39 Uhlenhuth. Centralbl. f. Baht., I, Ref. Vol. 42, Beilage, p. 62. 
40 “Antiformin” is a substance largely employed for the cleansing of 
pipes and vats of organic matter because of its powerfully solvent action. 
Its value in concentrating tubercle bacilli out of sputum and other mixtures 
depends upon its power to dissolve the tissue elements and all bacteria except ACQUIRED IMMUNITY 
81 
substance thoroughly dissolves all but the acid-fast bacteria when 
used in concentrations of 2.5 per cent. Since it is alkaline it is 
necessary to neutralize it with hydrochloric or sulphuric acid before 
use. 
For the preparation of antigen from pneumococci Neufeld 41 has 
utilized the solvent action upon these organisms of bile. He adds 
the bile and broth cultures just as this is done in the diagnostic 
“bile test” (0.1-0.2 c. c. of fresh bile to a broth culture; sodium 
taurocHolate solution can also be used). Many bacteria can also be 
broken up by emulsifying them in 17 per cent, salt solution and 
allowing them to stand for some time in this medium and then 
diluting this to 0.85 per cent. 
B. Extraction by Mechanical Methods 
One of the most useful methods for obtaining extracts of bacteria 
within a relatively short time is that which Besredka 42 has applied 
mainly for the preparation of typhoid (endotoxin), 24-hour agar 
cultures washed up in very small quantities of physiological salt 
solution, killed by heat at 60° to 65° C. and dried in vacuo. The 
dried mass is mixed with a measured quantity of dry salt and the 
mixture thoroughly triturated in a mortar for a considerable time. 
While triturating distilled water is added in small quantities until 
the fluid represents a 0.85 per cent, salt solution. This is allowed 
to stand for anywhere from a few hours to a week, and the bacteria 
are then removed by centrifugalization. This method has been modi- 
fied by many observers and gives good results whenever thorough 
trituration is practiced. It is also probable that the exposure to the 
hypertonic salt solution in the earlier stages of the trituration may 
aid considerably in breaking up the bacteria. 
Trituration after freezing is a method which has yielded excel- 
lent results in the hands of Macfadyen and others. This requires a 
rather complicated piece of machinery originally described by Mac- 
fadyen and Rowland. The principle of this is one of mechanical 
trituration in a steel cylinder which is surrounded by an ice-brine 
mixture so that the bacteria and sand may be kept frozen during the 
process. 
those that are acid-fast. Rosenau (“Preventive Medicine and Hygiene,” 
Appleton, 1913, p. 1020) gives its composition as follows: “Antiformin 
consists of equal parts of liquor sod® chlorinate of the British Pharmacopoeia 
and a 15 per cent, solution of caustic soda. The formula for liquor sod® 
chlorinate is as follows: 
Sodium carbonate 600 
Chlorinated lime 400 
Distilled water 4,000” 
41 Neufeld. Zeitschr. f. Hyg., Yol. 34, 1900. 
42 Besredka. Ann. de I’Inst. Past., 19-20, 1905, 1906. 82 
INFECTION AND RESISTANCE 
Mechanical trituration is also the principle of the production of 
the new tuberculins as advised by Koch. 
One of the earliest methods of obtaining bacterial substances 
by mechanical means was that used by Buchner and Hahn 43 in the 
production of their “plasmines.” The bacteria were grown in quan- 
tity on large agar surfaces, the moist bacterial masses triturated 
together with quartz and were then subjected to high pressure in an 
especially constructed press spoken of as the “Buchner press.” 
Mechanical breaking up and extraction of the bacteria also under- 
lies in principle the use of the variously constructed shaking ma- 
chines. There are many models of such machines on the market, all 
of them designed to accomplish prolonged agitation of bacterial emul- 
sions. In many cases the apparatus can be placed inside of an incu- 
bator and shaking carried on at 37.5° C. The bacteria are suspended 
for this purpose in distilled water salt solution, weak alkali, or in 
serum, and glass beads or sand may be added to aid in their mechan- 
ical injury. Shaking must be continued for 24 hours or more in 
order to give good results. 
IV. Active Immunization with Bacterial Products (Toxins) 
As soon as the investigations of Roux and Yersin had shown that 
in some diseases, at least, the injury sustained by the infected animal 
was largely due to the soluble toxins produced by the bacteria, it 
was logical to attempt to immunize animals with such products. 
Probably the first attempts in this direction were those made by 
Salmon and Smith in hog cholera. The experiments of these writers 
have attained much historical importance since they represent the 
first purposeful attempt to immunize animals with the products of 
bacterial metabolism. In the actual experiment, however, the im- 
munization practiced by Salmon and Smith was probably a combina- 
tion of immunization by bacterial products and by dead bacteria. 
Nevertheless, the thought of immunization with bacterial products 
was the underlying one in their experiments. Working with the 
hog cholera bacillus which they had recently discovered they im- 
munized pigeons in the following way: The bacilli were grown in 
broth for two weeks, and the cultures were killed by exposure to 58° 
to 60° C. for several hours. One and one-fifth cubic centimeter of 
this culture liquid was then injected into pigeons, and after three 
such injections the inoculated pigeons withstood, without harm, doses 
of the bacilli which rapidly killed untreated animals. Salmon and 
Smith 44 stated distinctly in their conclusions that: “Immunity may 
43 Buchner and Hahn. Munch, med. Woch., 1897. 
M Salmon and Smith on “A New Method of Producing Immunity from 
Contagious Disease,” Proc. Biol. Soc., Wash., D. C., Ill, 1884, 6, p. 29, 
printed Feb. 22, 1886. ACQUIRED IMMUNITY 
83 
be produced by introducing into the animal body such chemical 
products (results of bacterial growth in culture fluids) that have been 
produced in the laboratory.45 
Similar attempts to immunize rabbits against certain forms of 
septicemia by the injection of culture filtrates were made by Cham1 
berland and Roux 46 in 1888,47 and the same investigators applied 
this method to anthrax immunization just prior to the discovery of 
diphtheria toxin by Roux. However, neither in hog48 cholera49 
nor in the other infections upon which this method was first tried do 
the bacteria produce a true soluble toxin, and the immunization 
which was accomplished depended probably upon the injection of 
bacterial extracts. Nevertheless, these attempts had shown the way 
in a new direction, and bore immediate fruit in the investigations of 
Brieger and Fraenkel,50 and more especially in those of Beh- 
ring 51 and his collaborators. Fraenkel, though following the method 
of injecting filtered diphtheria culture fluids, came to the erroneous 
conclusion that the toxin and the immunizing substances in the 
cultures were not identical (loc. cit., p. 1135).52 The degree of 
immunity obtained in his experiments, moreover, was a slight one 
only. 
Behring, in his first work, in collaboration with Kitasato, suc- 
ceeded in immunizing animals with culture filtrates and with pleural 
exudates of diphtheritic animals. Similar results were accomplished 
with tetanus. Since the publication of these results—especially in 
consequence of the epoch-making discovery of passive immunization, 
to which they were the immediate guides, toxin immunization has 
been investigated and accomplished in all cases in which a true solu- 
ble toxin can be demonstrated. It has accordingly been carried 
out with the exotoxins of pyocyaneus53 bacilli, the bacilli of 
symptomatic anthrax 54 and botulinus,55 the specific leukocidins 56 
45 For a copy of the original paper by Salmon and Smith I am indebted 
to Professor Theobald Smith. 
46 Chamberland and Roux. Ann. Past., Yol. 1, 1888. 
47 Op. cit., Vol. 2, 1889. 
48 Joest in “Kolle u. Wassermann Handbuch, etc./’ Yol. 3, p. 632. 
49 Karlinski. Zeitschr. f. Hyg., Vol. 28, 1898. 
50 Brieger and Fraenkel. Berl. Id. Woch., 1890, Nos. 11, 12 and 49. 
51 Behring and Kitasato. Deutsche med. Woch., No. 49, 1890. Behring. 
Deutsche med. Woch., No. 50, 1890. Behring and Wernicke. Zeitschr. f. 
Hyg., 1892. ... 
52 De Sehweinitz, indeed, who further studied hog cholera immunization 
(Medic. News, 1892; Centralbl. f. Bakt., Yol. 20, 1896), claimed that in 
sterilized milk the bacillus produced “enzymes” with which immunization 
could be accomplished. 
53 Wassermann. Zeitschr. f. Hyg., Yol. 22. 
54 Kempner. Zeitschr. f. Hyg., Vol. 23, 1897. 
55 Grassberger and Shattenfroh. Deuticke, Wien, 1904. 
66 Denys and Van der Velde. “La Cellule,” Vol. 2, 1895. 84 
INFECTION AND RESISTANCE 
produced by staphylococci, and with various bacterial hemolytic poi- 
sons (tetanolysin and other bacterial hemotoxins). The result of all 
this work has been the very important determination that susceptible 
animals may be actively immunized both against the effects of the 
toxin alone, as well as against the virulent bacteria themselves, by 
systematic treatment with culture filtrates containing the toxins. 
Since in many cases the effects of the toxins were so powerful that 
their attenuation was desirable, Behring and others have advised 
the addition of iodin terchlorid and other chemicals to the first 
injections. 
PASSIVE IMMUNIZATION 
In the logical development of the fundamental facts regarding 
immunization, with attention focused early on the blood and body 
fluids as the probable carriers of immunity, it was hut a rational 
step from active immunization to the conception that such acquired 
immunity might he transferred from a treated to a normal animal 
by injecting blood from the former into the latter. This was prob- 
ably the underlying thought of Toussaint’s57 early work with 
anthrax, in which he heated anthrax blood to 55° C. and injected it 
into other animals, wrongly believing that the bacteria had been 
killed by the heating. The method of Toussaint, however, was vague' 
in its conception, and in no way constitutes an example of true 
passive immunization. The beginning was made in a purposeful 
and clearly conceived way by Richet and Hericourt.58 
These investigators actively immunized dogs against staphylo- 
cocci, and then attempted to transfer the immunity to normal rabbits 
by injecting defibrinated blood from the immune dogs. Their suc- 
cess was a partial one only, for reasons that we will discuss directly. 
Reasoning similar to that of Richet and Hericourt was applied 
by Babes and Lepp 59 to rabies immunization. When the blood of 
rabies-immune dogs was injected into normal dogs and rabbits, and 
these inoculated with rabies several days later, the treated animals 
regularly survived the controls, but in one dog only was the occur- 
rence of rabies absolutely prevented. Since their animals were not 
experimentally inoculated, but subjected to the more uncertain 
method of allowing them to be bitten by a mad dog, and since the 
series included 4 animals only (2 treated and 2 controls), Babes and 
Lepp were unable to draw definite conclusions. The establishment 
of passive immunization as a proved scientific fact was finally 
57 Toussaint. Compt. rend, de Vacad. des sc., 1880. 
58 Richet and Hericourt. Compt. rend, de Vacad. des sc., 1888, Yol. 107, 
p. 750. 
69 Babes and Lepp. Ann. Past., Yol. 3, 1889. ACQUIRED IMMUNITY 
85 
accomplished in 1890-1892 by Behring and Kitasato,60 and by 
Behring and Wernicke. The results of this work—the direct out- 
come of their success in actively immunizing with soluble toxins, is 
summarized in their first paper as follows: “The blood of tetanus- 
immune rabbits possesses tetanus-poison-destroying properties; these 
properties are demonstrable in the extravascular blood and in the 
serum obtained from this; these properties are of so lasting a nature 
that they remain active in the bodies of other animals, so that one 
is enabled to obtain positive therapeutic results by transfusing the 
blood or injecting the serum. These tetanus-poison-destroying 
properties are absent from the blood of non-immune animals, and 
when the tetanus poison is inoculated into normal animals it can be 
demonstrated as such in the blood and other fluids of these animals 
after death.” 
With these researches begins the therapeutically practicable 
method of passive immunization which is now in such widespread 
and successful use in the treatment of diphtheria, in the prophylactic 
treatment of tetanus, and, to a less common and less successful 
degree, in the treatment of dysentery, typhoid fever (Besredka), 
plague, and a number of other bacterial diseases. The same method 
has been successful in the treatment of various diseases of domestic 
animals. The principle was also applied by Ehrlich 61 to ricin and 
crotin immunity, in the formulation of which he succeeded in work- 
ing out passive immunization on a quantitative basis, showing that 
the degree of immunity in such cases could be directly referred to 
the amounts of the specific antitoxin present in the blood of the 
immunized animal. Calmette,62 and Physalix and Bertrand,63 then 
succeeded in producing passive immunization against snake venoms. 
To summarize the success of passive immunization in general we 
may say that it has achieved its greatest usefulness in the case of 
those diseases in which the pathogenesis depends upon a true exo- 
toxin—which, as we have mentioned before, leads to the formation 
of an antitoxin in the immunized animal. In these cases the passive 
immunization is accomplished by the transfer of the antitoxins from 
the treated to the normal animal. 
In the case of bacterial infections in which no true toxin is 
formed—where no antitoxin results and the immunity depends, as 
we shall see, upon an enhancement of the bactericidal and phagocytic* 
properties of the blood and the cells, passive immunization has not 
been a practical therapeutic success. The probable reasons for this 
cannot be properly discussed until we have examined more closely 
into the mechanism by which the immune animal is protected after 
60 Behring and Kitasato. Deutsche med. Woch., No. 49, 1890. 
61 Ehrlich. Deutsche med. Woch., 1891; Fortschr. d. Med., p. 41, 1897. 
62 Calmette. Compt. rend, de la soc. de biol., 1894. 
63 Physalix and Bertrand. Compt. rend, de la soc. de biol., 1894. 86 
INFECTION AND RESISTANCE 
specific treatment with bacteria or their products. The moderately 
beneficial effects of the various antiplague sera and the limited 
success attending the use of antistaphylococcus, antistreptococcus, 
and antipneumococcus sera probably depend, as recent work tends to 
show, not upon the direct action of antitoxin bodies, but rather upon 
the indirect opsonic action 64 which renders the bacteria more easily 
amenable to phagocytic action. These points we shall discuss at 
greater length in a succeeding chapter. 
The Conception of Specificity. In speaking of methods of 
immunization in the preceding sections we have frequently em- 
ployed the terms “specific” and “specificity,” without sufficiently 
defining them. It will be necessary to explain them since the 
principle of specificity is at the same time one of the most important 
and one of the most mysterious of the phenomena of immunity. 
When an individual has recovered from an attack of typhoid he is 
thereafter immune to typhoid—but to no other disease—similarly 
with plague, cholera, small-pox, etc. The same principle governs 
artificial immunization. Vaccination against anthrax protects 
against anthrax only—and active or passive immunization in any of 
the infectious diseases produces a resistance which is “specifically” 
aimed only at the particular infectious agent with which the original 
active immunity was produced. The principle points to an exquisite 
chemical difference between the protein substances which constitute 
the bacterial cell bodies or their metabolic products. For although 
by chemical methods we can detect no differences between them—- 
yet the reactions of immunity are sharply differentiating. When we 
come to consider the antibodies which specifically precipitate the 
substances by which they are incited we shall see that the delicacy 
and consequent differential value of these reactions far outstrip any 
known chemical methods, and it is upon this principle indeed—- 
inexplicable as it is—that the great diagnostic value which these 
reactions have attained depends. The conception of the specificity 
of the causes of infectious disease, as well as that of the specificity of 
toxins, has become so common and self-evident to us that we are too 
apt to forget how fundamental to progress the establishment of this 
fact was in the early days of bacteriological research. When, in 
1878, Koch published his treatise on the “Etiology of Wound Infec- 
tions” specificity was not generally accepted, and the supposed 
metamorphosis of bacterial species, as asserted by Hallier and 
others,65 had first to be scientifically refuted by Cohn, Koch, and 
their pupils, before it could be assumed that a given infectious 
64 Neufeld. Deutsche med. Woch., No. 11, 1897. Neufeld and Rimpau. 
Deutsche med. Woch., No. 40, 1904. Bail and Kleinhans. Zeitschr. f. 1mm., 
Yol. 12, 1912. Weil. Zeitschr. f. Hyg., 75, 1913. 
65 Hallier. Cited from Behring, “Bekampfung der Infektionskrank- 
heiten,” Leipzig, 1894. ACQUIRED IMMUNITY 
87 
disease was always the result of infection with a definite and constant 
species of bacteria. The same applied to the specificity of toxins— 
and rational investigations into the reaction of the animal body 
against bacterial poisons was not possible until the works of Roux 
and Yersin on diphtheria and that of Kitasato on tetanus had 
differentiated between the true, specific bacterial poisons and the 
unspecific ptomains and “sepsins” of Selmi, Nencki, and Brieger. 
The problem of the reasons for specificity has been and still 
remains one of the unsolved mysteries of immunology. We have 
seen in a preceding section that Abderhalden 66 in attempting to 
form an analogy by which the infinite variety of specific relations 
can he visualized, used the simple arithmetical consideration of the 
enormous variety of combinations that can be made by rearrange- 
ments of the twenty-one amino acids of which all proteins seem to 
be built up. The question, however, does not seem to he as simple 
as this. 
In order to avoid needless repetition, we refer the reader to the 
section on Antigens on page 105, where we discuss in some detail the 
investigations of Ohermeyer and Pick, and of Landsteiner and his 
collaborators, on the alterations produced in the specificity of pro- 
teins by the simple substitution of various inorganic chemical radi- 
cles, in various places in the protein molecule. Apparently a rela- 
tively slight chemical change by the introduction of diazo groups, 
iodin or bromin, or methyl radicles, in the protein molecule com- 
pletely changes specific relations; and not only is the mere intro- 
duction of such groups important, but the manner and place in which 
they are introduced may cause further alterations. It is through 
researches like those of Landsteiner that a closer approach to the 
problem of the reason for specificity must be sought, and the infinite 
varieties of possible specific differences need not surprise us if we 
consider that some of the substances which Landsteiner introduced 
into the protein molecule and which completely changed its specific 
relations, entered only into the tyrosin and histidin nuclei of the 
proteins which constitute only 3 per cent, of the particular protein 
molecule with which the experiments were done. 
Studies like those of Wells upon the chemistry of different 
antigenic proteins are also of considerable importance in this con- 
nection. They are dealt with in the section on Anaphylactic 
Antigens, particularly. 
66 Abderhalden, Munch, med. Woch., 14, 1913. CHAPTER IV 
THE MECHANISM OF NATURAL IMMUNITY AND THE 
PHENOMENA FOLLOWING UPON ACTIVE 
IMMUNIZATION 
ANTIGENS AND ANTIBODIES. THE ORIGIN OF ANTIBODIES 
Pasteur’s work on active immunization was carried out in the 
later seventies and the early eighties. During and immediately after 
this time it was very natural that the attention of investigators 
should have concentrated upon the elucidation of the causes under- 
lying both the natural resistance against bacteria observed in animals 
and man, and the changes which during active immunization were 
fundamentally responsible for the acquisition of resistance. 
It was easily determined that there were no anatomically and 
physiologically determinable differences between the various mam- 
malia which could account for the observed striking variations of 
susceptibility, nor could gross anatomical or histological changes be 
noted in an animal which had been artificially immunized. Morpho- 
logically such an animal was indistinguishable both in the size and 
appearance of its organs, and in the arrangement and structure of its 
cells from any other individual of the same species not subjected to 
treatment. 
It was a natural development of the investigations brought to 
bear upon this problem that attention should, for a time, be concen- 
trated upon the phenomena of inflammation, processes which were 
regularly associated with infections of all kinds and seemed indeed 
to represent a sort of local expression of tissue resistance to the 
invading micro-organisms. 
It was in the course of investigations upon the nature of inflam- 
mation that Metclmikoff first became interested in problems of resist- 
ance. In 1883 he presented a paper at the Naturalists’ Congress 
at Odessa, in which he referred the absorption of dead or foreign 
corpuscular elements in the bodies of invertebrates to a process of 
intracellular digestion carried out by phagocytic cells. As early as 
1874 Panum had suggested that possibly resistance against invading 
micro-organisms might be due to a similar intracellular destruction, 
and Metclmikoff, soon after his first communication, extended his 
phagocytic studies to phenomena of infection. His first investiga- 
88 PHENOMENA FOLLOWING IMMUNIZATION 
89 
tions concerned themselves with an infections disease caused by a 
form of yeast in a small crustacean—the daphnia or water flea. He 
showed that recovery or death from the disease depended upon the 
completeness with which the invading micro-organisms were taken 
up by the white blood cells found in the body cavity of the daphnia. 
Immediately subsequent studies, carried out with the aid of numer- 
ous pupils, embraced an extensive material throughout the animal 
kingdom in which he attempted to show parallelism between natural 
immunity and the phagocytic activities mobilized by the body against 
the invading germs. 
Meanwhile studies along another path were in progress. It had 
been observed many years before this by the physician, Hunter, that 
the shed blood of animals was not as easily subject to putrefactive 
change as were many other organic substances. Similar observations 
by Traube and, in 1881, by Lord Lister 1 (the latter reported at a 
time when Pasteur’s experiments were reaping their first practical 
results) further stimulated investigation of the blood as the possible 
seat of the increased antibacterial property. For, indeed, these 
observations seemed to imply that by resisting decomposition, even 
when inoculated with putrefying material, the blood must possess 
definite means of inhibiting or even destroying the putrefactive 
bacteria. 
In 1884, in a dissertation submitted at Dorpat, Grohman 2 stated 
that cell-free blood plasma inhibited the growth of micro-organisms. 
But Grohman was unable to determine actual bacterial destruction. 
Similar, but inconclusive, observations were published by Yon 
Fodor 3 in 1887. In 1888, however, Nuttall,4 who was investigating 
the validity of the phagocytic theory of Metchnikoff, experimentally 
determined that normal blood possessed the property of killing 
bacteria—a property now spoken of as “bactericidal” power. The 
attitude taken by Nuttall, and others of the Fliigge school, toward 
Metchnikoff’s opinions was one of doubt as to the fundamental 
significance of phagocytosis in determining resistance. They argued 
that Metchnikoff had not yet proved that living bacteria were taken 
up by the phagocytic cell, and that the action of these cells might 
therefore be interpreted as merely a process of removal of the dead 
bacteria, after these had been killed by other influences. Huttall, 
accordingly, repeated some of Metchnikoff’s experiments on anthrax 
in frogs and rabbits, essentially confirmed the basic observations, but 
showed also that the cell-free defibrinated blood of these and other 
animals possessed definite bacteria-destroying properties (bacteri- 
1 Lister. Trans. Intern. Med. Congress, London, 1881. 
2 Grohman. Cited from Lubarsch, Centralbl. f. Bald., 6, 1889. 
3 Fodor. Deutsche med. Woch., No. 34, 1887. 
4 Nuttall. Zeitschr. f. Hyg., Yol. 4, 1888. 90 
INFECTION AND RESISTANCE 
eidal power) for many different micro-organisms. He detected 
similar properties in pleural exudates, pericardial fluids, and aqueous 
humor, and determined that this property was “inactivated” or 
destroyed when the fluids were heated to 55° C. for 10 minutes or 
longer. Buchner5 then confirmed Nuttall’s results and showed 
further that the bactericidal property resided, not only in defibri- 
nated blood, peptone blood, and plasma, but was present also in the 
serum obtained after clotting. He applied the term “Alexin” to 
this active constituent of the blood—likening its action to that of a 
ferment. 
The immediate theoretical result of these discoveries was an 
attempt, begun by Fliigge’s school, to base natural as well as acquired 
resistance upon the bactericidal properties of the blood and body 
fluids in general. For tke observations of Nuttall and Buchner were 
soon extended to peritoneal and other exudates by Stern,6 and to 
ascitic fluids by Prudden.7 By these two groups, that of Fliigge- 
Nuttall-Buchner on the one hand, and that of Metchnikoff on the 
other, there were founded the two schools of immunity—the humoral 
and the cellular, both originating in attempts to explain natural 
immunity, and later extending to problems of acquired resistance. 
And it is to the diligent and ingenious intellectual and experimental 
conflict between these schools that we owe much of the knowledge we 
now possess concerning the phenomena of immunity. A bridge 
between them was early established when Buchner himself—(even 
before Metchnikoff)—suggested the possible leukocytic origin of the 
bactericidal serum constituent (alexin). The later work of Denys, 
of Gruber and Futaki, of Wright, of ISTeufeld, of Bail, and of others 
has demonstrated, as was to be expected, the inadequacy of either 
point of view by itself, and the intimate interdependence of the 
humoral and the cellular processes. 
As concerns the relation of bactericidal serum effects and natural 
immunity, it could be unquestionably shown by Nutt all, Buchner, 
Nissen,8 and their immediate followers that the blood of most ani- 
mals possessed bactericidal properties against many micro-organisms, 
their experiments being so planned that the participation of leuko- 
cytes could be absolutely excluded. However, a parallelism between 
bactericidal power and the degree of natural resistance could not 
be established. Lubarsch,9 writing during the early periods of the 
controversy, stated that “he would regard the (purely humoral) 10 
experiments of Nuttall as decisively contradicting the phagocytic 
5 Buchner. Centralbl. f. Baht., 1889. 
6 Stern. Zeitschr. f. klin. Med., Vol. 18 (cited from Hahn) 
7 Prudden. Med. Bee., Jan., 1890. 
8 Nissen. Zeitschr. f. Hyg., Yol. 6. 
9 Lubarsch. Centralbl. f. Bakt., Vol. 6, 1889. 
10 Bracketed phrase our own. PHENOMENA FOLLOWING IMMUNIZATION 
91 
theory if the bactericidal action of the blood (for anthrax bacilli) 
could be shown to be more potent in immune than in susceptible 
animals.” Metchnikolf 11 himself, taking this point of view, called 
attention to the fact that the blood serum of rabbits, animals that 
are highly susceptible to anthrax, is more powerfully bactericidal 
for these micro-organisms than is the blood of dogs or even that of 
immunized calves, both of which are much more resistant than are 
rabbits. Nuttall answered this by reporting that the blood of 
anthrax-immunized calves is actually more powerfully bactericidal 
than is that of normal calves. Although this argument of Nuttall 
was perfectly valid in principle, it exerted little influence on opinions 
at this time because anthrax happens as a matter of fact to 
belong to that group of infections in which bactericidal protection 
is actually secondary to phagocytic, and Lubarsch could show that 
the differences observed by Nuttall were often less than those 
obtaining between specimens of blood taken from individual normal 
rabbits. 
Lubarsch himself, then, in carefully planned experiments, 
showed that rabbits and cats could be killed with quantities of 
anthrax bacilli far less than the number which the extravascular 
blood of these animals can destroy. He concluded that the resistance, 
in these cases at least, is certainly not parallel with the bactericidal 
properties of the blood, and suggested the possibility that the intra- 
vascular blood does not possess bactericidal power to the same degree 
in which it is possessed by the extravascular plasma or serum. This 
point, first raised by Lubarsch—namely, the possibility of a differ- 
ence between the intravascular blood and the extravascular blood 
serum or plasma in bactericidal functions—soon became one of the 
focal points of the controversy, since Metchnikoff, admitting the 
bactericidal power of the shed blood, assumed that this was purely the 
result of substances given off by the leukocytic cell-bodies after 
extravascular injury. 
The Metchnikoff school defended its premise by the dual method 
of attempting on the one hand to establish a parallelism between 
phagocytic activity and natural resistance, and, on the other hand, 
by showing that the cell-free blood serum of naturally resistant ani- 
mals often furnished an excellent culture medium for the bacteria 
in question. Thus Wagner showed that anthrax bacilli grow well 
in the blood of fowls at 42° C., and Metchnikoff himself called 
attention to the fact that pigeons’ blood is an excellent medium for 
the cultivation of the Pfeiffer bacillus, whereas the living pigeon is 
entirely insusceptible to influenzal infection. Arguments based on 
such observations, however, have lost much of their original weight, 
for we have since then learned more about the delicate quantitative 
conditions and the difficulties of accurate measurements obtaining in 
11 Metchnikoff. Virch. Archiv, Vol. 97, 1884. 92 
INFECTION AND RESISTANCE 
experiments upon in vitro bactericidal phenomena. For although a 
specimen of the blood of a naturally immune animal may be capable 
of destroying a considerable number of bacteria of a given species, the 
implantation of such a specimen with a slight excess of the bacteria 
would soon exhaust the active serum constituents and profuse growth 
could then take place. Furthermore, the conditions of temperature 
established in cultural experiments lead rapidly to a deterioration 
of the alexin necessary for bactericidal action, and any bacteria 
remaining alive at the end of a number of hours would then have 
unopposed opportunity to multiply. 
The attempts to establish parallelism between phagocytic activity 
and natural immunity, though somewhat more successful than the 
analogous efforts of the humoral school, nevertheless also failed to 
furnish complete explanation for existing conditions, and, as we shall 
see, no adequate generalizations could be made until later years 
revealed the close cooperation between cells and fluids. We must 
postpone any attempts to do justice to this phase of the problem, 
therefore, until we are in a position to discuss the question of 
phagocytosis on the basis of a fuller knowledge of the phenomena 
which influence it. 
In the chapter on phagocytosis we will see that experiments by 
Kyes and others have shown that the natural immunity of some 
animals (birds) to pneumococci can be largely explained by phago- 
cytic activity on the part of such fixed tissue cells as the Kupfer 
cells of the liver. It is not unlikely that in all processes of natural 
immunity the phagocytosis by tissue cells is an important protective 
factor. 
The clear thinking and unprejudiced logic brought to bear upon 
this controversy by some of the great bacteriologists of this time are 
nowhere more instructively illustrated than in a short introduction 
published by v. Behring 12 to his second article on diphtheria. He 
says: “Neither deduction nor theorizing can at present decide 
whether a compromise will be found in the future between the two 
hypotheses (humoral and cellular), or whether the one or the other 
alone will be found correct. As yet the opinions of many experiment- 
ing bacteriologists are in direct opposition in this respect. Mean- 
while, for the purposes of medical advancement and therapeutic 
success it is not necessary to await a decision of this question. . . . 
It is indeed of advantage to the cause if the struggle against infec- 
tion is undertaken from the most varied points of view; attempts to 
make proselytes for a dogma have never led to progress. In this, 
sense I will try to summarize those experimental results which 
support the humoral point of view without attempting particularly 
to detract from the importance of opinions which I do not share.” 
12 v. Behring. Zeitschr. /. Hyg., Yol. 12, 1892. PHENOMENA FOLLOWING IMMUNIZx\TION 
93 
THE PHENOMENA FOLLOWING UPON ACTIVE IMMUNI- 
ZATION 
The cellular and humoral points of view, formulated largely upon 
the facts of natural immunity, were equally applied, almost from 
the beginning, to the explanation of active immunization. The light 
thrown upon these phenomena by the efforts of both schools rapidly 
led to a complete abandonment of those earlier theories of immunity 
which had conceived the acquired resistance of animals against 
bacteria as a purely passive development in the body of conditions 
unfavorable for bacterial growth. 
Among these earlier theories, now of historical interest only, are 
the “Exhaustion Theory” of Pasteur and the “Retention Theory” of 
Nencki,13 Chauveau, and others. 
Pasteur’s views, defended for a time also by Garre, held that the 
growth of any given variety of bacteria in the animal body exhausted 
certain specific nutritive substances necessary for this growth. Sub- 
sequent lodgment in the same body was impossible owing to the 
absence of proper nutrient material. It is interesting to note, as 
Kolle 14 points out, that this theory is in principle very similar to 
the “Atrepsie” idea of Ehrlich advanced in explanation of species 
immunity to cancer. 
The hypotheses of Chauveau, of Nencki, and others were the 
converse of those of Pasteur. They were based purely on inference, 
assuming that conditions occurring in the test tube could be applied 
also to those existing in the animal body. Baumann 15 had shown 
that, among other things, phenol was produced as a result of bacterial 
putrefaction. Nencki had noticed the inhibition of bacteria id 
culture by the products of their own metabolism. Wernicke,16 too, 
had demonstrated the presence of phenol, phenylacetate, skatol, and 
other aromatic compounds harmful to bacteria in putrefying 
mixtures. The reasoning which formulated the so-called “Retention 
Theory,” therefore, was the following: Bacteria growing in the 
animal body produce certain substances peculiar to their own metab- 
olism, which eventually lead to inhibition of their growth. By the 
retention of these products the animal is rendered immune. Chau- 
veau’s adherence to this theory was largely based on the fact that he 
had observed immunity in the lambs born of Algerian ewes which 
had recovered from anthrax shortly before or during parturition. He 
explained this by a transference of the retention products from 
13 Nencki. Jour f Prakt. Chem., May, 1879, cited from Sirotinin, 
Zeitschr. f. Hyg., Yol. 4, 1888. 
14 Kolle in “Kolle n. Wassermann Handbuch,” 2d Ed., Yol. 1. 
15 Baumann. Zeitschr. f. physiol. Chem., Vol. 1. 
18 Wernicke. Virch. Archiv, Vol. 78. 94 
INFECTION AND RESISTANCE 
mother to offspring. As a matter of fact the observation could just 
as well have been utilized as support for the Exhaustion Theory. 
Both the theory of “Exhaustion” as well as that of “Retention” 
could not long withstand experimental criticism. Theories which 
were not so easily disproved and which have given rise to much 
investigation are the “Alkalinity Theory,” tirst formulated by 
v. Behring,17 and the “Osmotic Theory” of Baumgarten.18 In the 
former an attempt was made to demonstrate a parallelism between 
blood alkalinity and bactericidal action—the latter was based on the 
supposition that the destruction of bacteria in the body was largely 
due to harmful osmotic conditions. Neither of these theories was 
long seriously maintained. Behring himself took an active part in 
the subsequent development of our present views. Baumgarten 1!' 
still clings to his own opinion in a modified way, in that he main- 
tains that the only effect produced by specific antibodies upon cells— 
bacterial or otherwise—is that they change the permeability of the 
cell membranes and render them more vulnerable to osmotic injury. 
However crude or vague these theories may seem to us now, it 
must not be forgotten that they were conceived at a time when no 
knowledge had been gained regarding specific “antibodies.” The 
phagocytic powers to which Metchnikoff attributed natural immunity 
and the bactericidal powers of the blood, regarded in the same light 
by the Fliigge school, were general properties possessed by many 
animals toward many different micro-organisms. That immuniza- 
tion could specifically increase these functions toward the particular 
micro-organisms used for treatment seemed indicated by the experi- 
ments of Nuttall in which higher bactericidal power was found in 
the blood of anthrax-immune calves than in that or normal animals. 
However, no definite and conclusive work on the specific increase of 
measurable serum or cell properties was available. 
This great advance, giving new energy and pointing out new 
paths of investigation, came in 1890-1892 with the publication of the 
work of Behring and his collaborators, Kitasato and Wernicke, on 
immunity to diphtheria and tetanus. As we have indicated in a 
preceding paragraph, the fundamentally important points of this 
work were as follows: 
1. The establishment of the fact that animals may be actively 
immunized with products of bacterial metabolism—true toxins or 
exotoxins. 
2. The discovery that such active immunity was dependent upon 
specific antibodies formed in the treated animal and circulating 
freely in the blood; and, 
17 v. Behring. Centralbl. f. klin. Med., 1888, No. 38. 
18 Baumgarten. Berl. klin. Woch., 1899, 1900. 
19 Baumgarten. Lehrbuch, etc., 1912. PHENOMENA FOLLOWING IMMUNIZATION 
95 
3. That, by the transfer of the blood or the blood serum contain- 
ing these specific antibodies, other normal animals could be passively 
protected—not prophylactically only, but even after active disease 
had set in. 
These observations were rapidly confirmed for tetanus by Tiz- 
zoni and Cattani, and by Vaillard, and, similar but less successful 
attempts at passive immunization were made in other diseases by 
Foa, Emmerich, Bouchard, and many others. The discovery of 
passive immunization established the fact of specific alteration of 
the blood by active immunization, and represented, for the time, a 
distinct triumph for the humoral hypothesis. 
Summarizing the knowledge of immunity as it stood at the close 
of this period, Behring says: “In the case of natural immunity no 
generally applicable explanation has as yet been found. (By this he 
referred to the lack of complete parallelism between natural im- 
munity and either the bactericidal or the phagocytic activities.) For 
artificial immunization, however, it has now been shown, in a number 
of carefully studied infections, that we can surely attribute it to 
properties of the cell-free blood.” 
Within a very short time after Behring and Ivitasato’s first paper 
Ehrlich 20 demonstrated that the principle discovered by them was 
not limited to bacterial poisons. He was investigating immuniza- 
tion against ricin in mice, and showed that here, too, the blood of 
the immune animals contained a body which would antagonize the 
toxic action of ricin, and which, injected into normal mice, would 
passively protect them. He spoke of this blood constituent as 
“antiricin.” 
It is natural that extensive generalization followed these discov- 
eries. However, while it was found that the blood of all actively 
immunized animals possessed a certain degree of protective power 
for normal individuals, it was soon shown that this was not due in 
all cases to antagonism to the bacterial poisons on the part of the 
immune blood serum. In immunity to the Vibrio Metchnikovi—in 
pneumococcus and cholera immunity—Sanarelli,21 Isaeff,22 Pfeiffer 
and Wassermann,23 and a number of others showed that here, unlike 
diphtheria and tetanus, the protective power of the immune serum 
did not rest on “antitoxic” properties, but rather on antagonism to 
the bacteria themselves. It soon became definitely established that 
antitoxic immunity resulted only in the cases of those bacteria in 
which a true soluble exotoxin was produced, and where the disease 
following infection was primarily due to the absorption of these 
poisons. The antibodies incited in the blood of toxin-immune 
20 Ehrlich. Deutsche med. Woch., No. 32, 1891. 
21 Sanarelli. Ann. Past., Yol. 7, 1893. 
22 Isaeff. Ibid, and Zeitschr. f. Hyg., Yol. 16, 1894. 
23 Pfeiffer and Wassermann, Zeitschr. f. HygYol. 14, 1893, 96 
INFECTION AND RESISTANCE 
animals were therefore spoken of by Behring and Ehrlich as 
“antitoxins” and their action—after a number of false hypotheses 
—was finally recognized as a direct neutralization of the bacterial 
poisons. 
The strict specificity of these antibodies was, from the first, clear 
to v. Behring, who observed that diphtheria-immune serum and 
tetanus-immune serum acted each upon its respective toxin only. It 
was recognized at the same time that the passive immunity produced 
by injecting antitoxic sera is almost immediately established; that, by 
proportionately increasing the amount of antitoxin, immunity can 
be produced against any amount of toxin; and that this passive or 
transferred immunity is of relatively short duration. 
The antitoxins, then, as we shall see in the more detailed analysis 
of their action (in Chapter V), are specific poison-neutralizing anti- 
bodies formed in the blood of animals immunized with a true 
bacterial toxin or exotoxin—conferring resistance or immunity, not 
by influencing the bacteria, but by rendering innocuous the specific 
bacterial poisons. 
The therapeutic successes of passive immunization achieved with 
tetanus and diphtheria very naturally led to a careful inquiry into 
the antitoxic properties of the blood of animals immunized with all 
known pathogenic bacteria and bacterial products, and with many 
poisons of animal and vegetable origin. 
Contrary to earlier expectations, however, the list of bacteria 
against which antitoxic immunity can be achieved has remained 
relatively small, limited in fact, as we have previously stated, to those 
species which produce a soluble exotoxin. The inciting of a specific 
neutralizing antibody (antitoxin), however, is also a property of 
many other substances of proteid nature which are for this reason 
classified biologically with the true toxins or exotoxins. In fact, 
the one absolutely constant attribute which defines our conception 
of the “true toxins” and the substances classified with them is their 
antitoxin-inciting power. We classify a bacterial product as a 
“toxin” or “exotoxin” only if it incites a neutralizing “antitoxin” in 
the serum of an immunized animal. 
The first discovery of a non-bacterial antitoxin-stimulating sub- 
stance was, as we have stated, that of ricin by Ehrlich,24 1891, and 
this was soon followed by similar determinations for abrin and robin 
—other vegetable poisons. In 1894 Calmette,25 and Physalix and 
Bertrand 26 extended the principle to poisons of animal origin by 
demonstrating antitoxin formation against snake poison. And that 
similar specific neutralizing bodies were formed in response to 
24 Ehrlich. Deutsche med. Woch., 1891; Fortschr. d. Med., 1891, 1897. 
25 Calmette. Ann. Past., Yol. 8, 1894. 
26 Physalix and Bertrand. Compt. rend, de la soc. de biol.s 1894. PHENOMENA FOLLOWING IMMUNIZATION 
97 
immunization with ferments was shown in 1900 by Mor- 
genroth.27 
The more important individual substances which may be bio- 
logically grouped together because of their property of inciting a 
specific antitoxin (or toxin-neutralizing body) in the blood of im- 
munized animals may be tabulated as follows: 
Diphtheria toxin—(loc. cit. Behring & AYernicke). 
Tetanus toxin—(loc. cit. Behring & Kitasato). 
The Toxin of the Bacillus of Symptomatic Anthrax—(Grassberger & 
Shattenfroh, Miinch. Med. Woch., 1900, 1901 and 10 c. cit.). 
The Toxin of the Bacillus Botulinus—Kempner, Zeitschr. f. Hyg., Yol. 
26, 1897). 
The Toxin of the AVelch Bacillus (Bull & Pritchett loc. cit.). 
The Toxin of the Vibrion Septigue (AVeinberg & Seguin loc. cit.). 
The Toxin of the Bacillus Pyocyaneus—(Wassermann, Zeitschr. f. Hyq., 
Vol. 22, 1896). 
The Toxin of the Dysentery Bacillus (?) Shiga-Kruse type—(Kraus u. 
Doerr, Wien. klin. Woch., 1905). 
The leukocyte poison of the Staphylococcus pyogenes aureus, Leucocidin 
— (Denys & Van de Velde, La cellule, 1895). 
The Hemolytic Poisons of Various Bacteria (see Pribram in “Kraus und 
Levaditi Handbuch,” Vol. II, p. 223). 
Proteolytic Ferments of the Hog Cholera Bacillus (De Schweintz, Medical 
News, 1892). 
The Toxin of the Cholera Spirillum (?) Brau & Denier, Compt. rend, de 
I’acad. des sc., 1906, Kraus, Centralbl. f. Bakt., 1906, and Wien. klin. Woch., 
1906). 
Kicin—(Ehrlich, loc. cit.). 
Abrin—(Ehrlich, loc. cit.). 
Krotin—(Ehrlich, loc. cit.). 
Snake venom—(Calmette, loc. cit.). 
Spider poison—(Sachs, “Hoffmeister’s Beitrage,” 1902, and Ehrlich, 
“Gesammelte Arbeiten,” etc.). 
Lab. enzyme—(Morgenroth, loc. cit.). 
Pepsin—(Sachs, Fortschr. d. Med., 1902). 
Trypsin—(Achalme, Ann. Past., 1901). 
Leukocytic ferments Leukoprotease—(Jochmann & Muller, Miinch. med. 
Woch., 1906).28 
The period of investigation which was initiated by the discovery 
of the specific antitoxins was replete with efforts to determine true 
toxins and, consequently, antitoxic immunity for all pathogenic 
bacteria. We have already mentioned that in many cases these 
efforts were futile—the bacteria in question being found to secrete 
no exotoxin and the immunity established against them developing 
without the formation of demonstrable antitoxin. Metchnikoff 29 
27 Morgenroth. Centralbl. f. Bakt., 26, 1899. 
28 This list includes Ml the important antitoxin-inciting substances. For 
a more complete tabulation see Wassermann in “Kolle u. Wassermann Hand- 
buch, etc.,” Yol. IY, 1st Ed., p. 498. Our own list is adapted from the one 
there given. 
29 Metchnikoff. Ann. Bast., 1892, 98 
INFECTION AND RESISTANCE 
showed this to he the case with hog cholera as early as 1892, and the 
investigations of Sanarelli, Isaeff, and Pfeiffer and Wassermann 
pointed in the same direction. 
Perhaps the clearest definition of the conditions prevailing dur- 
ing immunization of animals with non-toxin-forming bacteria was 
that formulated at this time by Pfeiffer. The importance of the 
bactericidal power of serum, as discussed before this by Fliigge, 
ISTuttall, and others, had dealt largely with variations of this general 
property in relation to natural immunity, but had failed to recognize 
clearly a specific increase in these powers during active immuniza- 
tion. Pfeiffer with Wassermann 30 had studied the pathogenicity of 
cholera spirilla for guinea pigs, and had come to the conclusion that 
the animals died of toxemia (and not of bacteriemia, as claimed by 
Gruber and Wiener), and that this toxemia was due to the liberation 
of poisons from the dead bodies of cholera vibrios, killed by the 
serum of the infected animals. Pfeiffer 31 now showed that the in- 
jection of cholera spirilla killed with chloroform brought about a tox- 
emia identical with that following inoculation with living cultures. 
He further determined that the resistance of animals against cholera 
was due to the bactericidal effects of the serum, which killed the 
injected cholera spirilla, and not to any poison-neutralizing property. 
Isaeff,32 one of Pfeiffer’s pupils, continuing this work, expresses 
his own and Pfeiffer’s conceptions as follows: “Guinea pigs vac- 
cinated against cholera, in spite of high immunity to infection with 
living spirilla, do not develop any immunity to cholera [endo] 33 
toxins. The blood of immunized guinea pigs possesses no antitoxic 
properties. The maximal dose of cholera Toxin’ which immunized 
guinea pigs can withstand is not higher than that which can be borne 
by normal animals, and but slightly higher than the maximal dose 
of living spirilla, which they can survive. The blood of cholera-vac- 
cinated guinea pigs possesses strong specific protective powers. The 
same specific immunizing properties are demonstrable in the blood 
of cholera convalescents toward the end of the third week of the 
disease.” 
The path was thus cleared for a definite conception of cholera 
immunity, and this was formulated, in their next communication, 
by Pfeiffer and Isaeff.34 35 36 In this paper they showed that the 
cholera spirilla injected into the peritoneum of a cholera-immune 
30 Pfeiffer and Wassermann. Zeitschr. f. Hyg., Yol. 14, 1893; also 
Pfeiffer, Zeitschr. f. Hyg., Yol, 16, 1894. 
31 Gruber and Wiener. Archiv f. Hyg., Vol. 15, 1893. 
32 Isaeff. Zeitschr. f. Hyg., Yol. 16, 1894. 
33 Bracketed word our own. 
34 Pfeiffer and Isaeff. Zeitschr. f. Hyg., Yol. 17, 1894. 
35 Pfeiffer. Ibid., Yol. 18, 1894. 
38 Pfeiffer and Isaeff. Deutsche med. Woch., No, 13, 1894. PHENOMENA FOLLOWING IMMUNIZATION 
99 
guinea pig were subjected to a rapid dissolution, a process which 
could be observed by taking small quantities of exudate out of the 
peritoneum, at varying intervals, with capillary pipettes. No such 
dissolution occurred in normal pigs or with normal serum. But the 
same rapid swelling, granulation, and, finally, dissolution occurred 
when the spirilla were injected into the peritoneal cavity of a nor- 
mal guinea pig, together with the serum of an immunized animal. 
The process took place apparently without the cooperation of the 
leukocytes or other cells, and was absolutely specific. For instance, 
no “lysis” occurred when the vibrios “Nordhafen,” “Massauah,” and 
other cholera-like organisms were injected into cholera-immune pigs, 
but took place regularly when true cholera strains, from various 
sources, were used in the experiment. The immunity of cholera- 
treated animals, therefore, was found to be an antibacterial and not 
an antitoxic one. Cholera spirilla introduced into a normal animal 
were permitted to multiply and accumulate until a sufficient number 
were present to furnish, upon cell death, a fatal dose of poison. In 
immunized animals the small quantities of bacteria first introduced 
succumbed rapidly to the lytic properties of the serum and accumu- 
lation was prevented. 
By these experiments, now commonly spoken of as the “Pfeiffer 
Phenomenon,” it was definitely proved that active immunization 
with bacteria incites in the serum of the treated animal a potent in- 
crease of bactericidal properties—an increase which is entirely spe- 
cific in that the bactericidal power toward bacteria other than those 
employed in the immunization does not exceed the normal. The 
immunity in these cases, then, is not antitoxic, but rather “antibac- 
terial,” and depends on the development, in the immune sera, of anti- 
bodies quite distinct from the “antitoxins ” These immune serum 
constituents were spoken of by Pfeiffer as “bacteriolysins” or “spe- 
cific bactericidal substances.” 
Not long after the discovery of the specific bacteriolysins another 
property of immune sera was described by Gruber and Durham.37 
They had been studying bacteriolytic phenomena with colon and 
cholera organisms, and noticed that these bacteria were rapidly ag- 
glomerated and gathered in small clumps when emulsified in homolo- 
gous immune serum. Similar clumping had indeed been described 
before. Metchnikoff, Isaeff, Washburn, and Charrin and Roger had 
described it on various occasions, but had not recognized it as a 
specific property of immune serum.38 Gruber and Durham studied 
it carefully, determined that it was present to a degree roughly pro- 
portionate to the degree of immunization attained, and that its 
specificity was such that it could be utilized for bacterial differen- 
37 Gruber and Durham. Miinch. med. Woch., 1896. 
38 jror references see chapter on Agglutinins. 100 
INFECTION AND RESISTANCE 
tiation. They believed that the substances in the immune serum 
responsible for this agglutination were independent of other serum 
constituents and applied to them the term “agglutinins ” 
The problems of immunization had now considerably expanded 
and the nature of the new serum reactions was assiduously studied. 
Primarily the phenomenon of agglutination was regarded as a part 
of the struggle of the body against the living bacteria and Gruber 
himself believed that it depended upon a swelling or “klebrig wer- 
den” of the microorganisms which tended to cause their sticking 
together, and rendered them more readily amenable to the action of 
the bactericidal powers of the serum. Bordet,39 however, early con- 
ceived the process as a physical phenomenon in which the bacteria 
themselves were entirely passive, and, indeed, Widal 40 soon demon- 
strated that bacteria killed by heat were equally as agglutinable as 
the living germs. 
This naturally suggested that the reaction between specific agglu- 
tinating serum and bacteria was based on individual peculiarities of 
the bacterial proteins, and it occurred to Kraus,41 accordingly, to 
investigate whether or not the immune sera would cause any sort of 
reaction when mixed with the dissolved body substances of homolo- 
gous bacteria. Working at first with cholera and plague, he pre- 
pared solutions of bacterial proteins, both by allowing broth cultures 
to stand for varying periods and by emulsifying agar cultures in 
alkaline broth. The extracts were then filtered through Pukal filters 
to remove the bacterial bodies. When the sera of immunized ani- 
mals were added to these clear filtrates—cholera serum to cholera 
filtrate, and plague serum to plague filtrate, slight turbidity devel- 
oped and was followed within twenty-four hours by the formation of 
small flakes. In other words, it was found that the mixture of a 
clear filtrate of a bacterial culture with the serum of an animal 
immunized against these bacteria resulted in the formation of a 
precipitate. The reaction was found to be as strictly specific as that 
of agglutination. 
Although, from the beginning, Paltauf 42 attempted to associate 
the phenomena of agglutination and precipitation, the property of 
precipitating homologous culture filtrates was attributed by Kraus 
and others to specific antibodies in the immune sera, distinct and 
independent of those previously described, and spoke of them as 
“precipitins 
The discovery of the various “antibodies” so far discussed 
resulted from the study of the direct action of blood serum upon 
39 Bordet. Ann. Past., 1896. 
40 Widal. La semaine medicate, No. 5, 1897. 
41 Kraus. Wien. Bin. Woch., No. 32, 1897. 
42 Paltauf. “Discussion of Kraus’ Paper,” Wien. kl. Woch., No. 18, 1897, 
p. 431. PHENOMENA FOLLOWING IMMUNIZATION 
101 
bacteria and bacterial products. This did not, however, completely 
deflect the attention of investigators from the unquestionable impor- 
tance of phagocytosis in the defence of animals against bacterial in- 
vasion. Metchnikoff and his school continued diligently to pursue 
this other phase of the study of immunity and, although the increas- 
ing knowledge of serum antibodies continued to strengthen the prem- 
ises of the purely humoral point of view, it had still to be admitted 
that in some diseases—particularly anthrax and the pyogenic coccus 
infections, phagocytosis must largely be held responsible for recov- 
ery. It was found, moreover, by the later investigations of Denys, 
Wright, Neufeld, and others that phagocytosis in immunized animals 
was far more extensive and efficient than in normal ones, and that 
this depended on specific constituents of the immune serum which 
rendered the bacteria more amenable to the phagocytic action of the 
cells. These further antibodies we will discuss in a subsequent chap- 
ter, under the terms “opsonins” and “bacieriotropins” designations 
applied to them by their discoverers. 
We have thus reviewed briefly the various specific properties 
which develop in the serum of an animal when it is systematically 
treated (actively immunized) with bacteria or bacterial products. 
These serum activities have been attributed to the development in 
the serum of substances which we speak of as “antibodies 
In our discussion of the first of these antibodies, antitoxin, we 
call attention to the fact that the principle discovered in the case of 
bacterial toxins was rapidly extended to vegetable poisons, snake 
venom, spider poison and enzymes. It was found that the power of 
inciting antitoxins when injected into animals was an attribute be- 
longing to a large group of substances in nature, and not limited to 
bacteria alone. A similar generalization of conception has been pos- 
sible with other antibodies. Specific lysins, agglutinins, and pre- 
cipitins may be produced by the treatment of animals with many 
substances not of bacterial nature. 
The first observation of this kind was made almost simultaneously 
by Bordet 43 and by Belfanti and Carbone.44 They observed that 
the serum of an animal that had been treated with the red cells of 
another species acquired the power of laking these cells. That the 
normal serum of one species is often toxic to, and causes the laking. 
of, the erythrocytes of another species is an observation that dates 
back to the earliest experiments on transfusion, and had been studied 
in considerable detail by Landois as early as 1875. The phenom- 
enon possesses much interest in its bearing on the problems of ana- 
phylaxis and will be discussed more particularly in that connection. 
We mention it in this place to show' that, like bactericidal bodies, 
“hemolytic” (erythrocyte laking) properties may be present in nor- 
43 Bordet. Ann. Past., Yol. 12, 1898 . 
44 Belfanti and Carbone. Giorn. della B. Acad, di Torino, July, 1898. 102 
INFECTION AND RESISTANCE 
mal sera, though irregularly and by no means occurring in every 
species of animal. Incidentally it may be stated that this is true 
also of agglutinins and of opsonins which may be found in consider- 
able amounts in normal sera. Of precipitins, however, this does not 
seem to be true. 
By the work of Bordet it was found that “hemolysins” could be 
specifically 45 incited in an animal by systematically treating it with 
the red blood cells of another species. Apart from the great interest 
attaching to this discovery in itself, it has had a very profound influ- 
ence upon investigations on immunity generally, since it has fur- 
nished a method of studying lysis far more simple and easily con- 
trolled than is the analogous phenomenon of bacteriolysis. And 
since, in fundamental principles, bacteriolysis and hemolysis are 
essentially alike, much of our knowledge regarding the former has 
been arrived at by experiments upon the latter. The specific hemo- 
lysins, then, are antibodies formed in response to “immunization” 
with red blood cells, analogous to the similarly produced “bacterio- 
lysins.” Because both of these antibodies exert definite injury upon 
cells, we speak of them by the group names of “cytolysins” or ‘"cyto- 
toxic” substances. 
The discovery of hemolysins naturally suggested the use of other 
cells, and the following years brought forth many reports of further 
specific cytotoxins. In 1899, Landsteiner,46 and very soon after- 
ward Metchnikoff,47 described specific “ spermotoxins” which ap- 
peared in the blood of animals treated with spermatozoa. Von 
Dungern 48 obtained analogous substances by injecting ciliated epithe- 
lium from the trachea. ISTeisser and Wechsberg 49 produced “leuko- 
toxin” by injecting leukocytes; Delezenne 50 produced “neurotoxin” 
and “hepatotoxin,” and Surmont,51 pancreas cytotoxin. Subsequent 
years have added to these “gastro-toxin” (Bolton),52 thymotoxin 
(Slatineau),53 adrenal cytotoxin (Gildersleeve),54 placentar cyto- 
toxin (Frank),55 corpus luteum cytotoxin (Miller),56 and a number 
45 By the use of the word specific in this case we imply that an animal 
immunized with any given variety of red blood cells will form hemolysins 
for this variety only. Thus an animal treated with ox blood will form ox 
blood hemolysins only, and his serum, though strongly hemolytic for ox 
blood, will not lake sheep cells, dog cells, human cells, etc. 
46 Landsteiner. Centralbl. f. Baht., Vol. 25, p. 549, 1899. 
47 Metchnikoff. Ann. Past., Yol. 13, 1899. 
48 Von Dungern. Munch, med. Woch., p. 1228, 1899. 
49 Neisser and Wechsberg. Zeitschr. f. Hyg., Yol. 36, 1901. 
50 Delezenne. Ann. Past., 1900; Compt. rend, de Vacad. des sc., 1900. 
51 Surmo nt. Compt. rend, de la soc. de biol., 1901. 
52 Bolton. Lancet, 1908. 
53 Slatineau. Cited from Roessle, loc. cit. 
54 Gildersleeve. Cited after Roessle. 
65 Frank. Jour. Exp. Med., 1907. 
50 Miller. Centralbl. f. Bakt., 47, 190S. PHENOMENA FOLLOWING IMMUNIZATION 
103 
of others. In fact, as Roessle 57 puts it, in a review of the literature, 
there is no organ in the body for which it has not been claimed that 
specific cytotoxins can be formed by the injection of homologous 
macerated tissues. 
Recent critical study of these organ-cytotoxins has revealed, how- 
ever, that the specificity of a serum produced with the tissues of one 
organ is not strictly limited to this organ alone, and that the serum 
may injure other organs as well. It is true, indeed, that there are 
certain cells and tissues in the body such as the spermatozoa, the 
tissues of the testicles, the ovary, the lens of the eye, and, possibly, 
the placenta which have chemical characteristics so well defined and 
individual that the cytotoxic sera induced by them have definite 
organ specificity. The same to a more limited extent seems true 
of kidney substance (Pearce). In most cases, however, in which 
originally a specific cytotoxin was claimed, it has been possible to 
show subsequently that the apparently selective injury was due not to 
organ specificity alone but to the fact that the injection of tissue- 
macerates, even when sufficiently freed from blood, induced the for- 
mation of considerable amounts of hemagglutinins and hemolysins. 
Pearce 58 expresses it as follows: “. . . it is evident that the 
cells or the various organs of the body, while differing in morphology 
and function, have certain (receptor) characteristics in common, 
and that one type of cell may therefore produce antibodies affecting 
several cells of differing morphology, but with like (receptor) 
groups. This is shown by the sera prepared from washed liver, 
kidney, pancreas, and adrenal, all of which may agglutinate and 
hemolyze red blood cells and may cause degenerative changes also 
in the liver and the kidneys. Some of these cytotoxic sera have no 
effect upon organs for which they are supposed to have a morpho- 
logical affinity, but exert a powerful lytic influence upon other cells. 
Aside from nephrotoxin, which has a distinct injurious action upon 
renal epithelium, the various cytotoxins studied (kidney, liver, pan- 
creas, and adrenal) have no specific action in the morphological 
sense.” 
This opinion seems to be in harmony with that of most observers 
who have studied the problem recently, at least as regards most of 
the organ cytotoxins. Much of the promised light upon pathological 
processes—looked for when cytotoxins were first studied, has faded, 
moreover, since it has been found that cytotoxins cannot be produced 
by injection into an animal of cells, tissues, or fluids from its own 
body. “Autocytotoxins” in general cannot be produced, a question 
discussed at greater length in the chapter on lysis, in connection 
with Ehrlich’s work on the isolysins. 
The work outlined in the preceding paragraphs had thus ex- 
57 Roessle. “Lubarsch und Ostertag,” Yol. 13, 1909. 
58 Pearce. Jour, of Med. Res., N. S., Yol. 7, 1914, p. 13. 104 
INFECTION AND RESISTANCE 
tended the principles of antitoxin and lysin production beyond the 
scope of pure bacteriology, and had shown them to possess the sig- 
nificance of general biological laws. Similar generalization was soon 
attained in the case of the agglutinins and in that of the precipitins. 
In the former, the nature of the reaction limited it to observations 
upon cells in suspension, and, in connection with the earlier experi- 
ments upon hemolysis it was soon discovered that the erythrocytes 
were often clumped before lysis could take place, when brought to- 
gether with a hemolytic serum of moderate or feeble potency, or 
when solution, for other reasons, was delayed. 
The first observations on the general significance of the precip- 
itin reaction we owe to Tschistovitch 59 and to Bordet.60 Tschisto- 
vitch was studying the toxic action of eel serum upon rabbits. This 
serum, as Kossel 61 had shown, is toxic for rabbits and possesses the 
property of causing hemolysis of rabbit erythrocytes. Its similarity 
to ricin, in this respect, stimulated attempts to produce an antitoxic 
substance against eel serum, even as Ehrlich had produced an an- 
tiricin. In the course of such experiments Tschistovitch observed 
that, when eel serum was mixed with the serum of a rabbit which 
had received several injections of this substance, the mixture became 
rapidly opalescent and soon a flocculent precipitate was formed. 
Coincident with this discovery Bordet made a similar observation. 
He had injected chicken blood into rabbits in the course of experi- 
ments upon hemagglutination. He found that the serum of the rab- 
bits so treated acquired the property not only of producing hemolysis 
and hemagglutination of chicken cells, but also of giving a precipi- 
tate if mixed with chicken serum.62 Soon after this precipitins were 
produced by injecting rabbits with milk (Bordet), egg albumen 
(Ehrlichj Uhlenhuth), and many other substances, and the speci- 
ficity of such reactions was demonstrated by Fish,03 Wassermann 
and Schiitze,64 Uhlenhuth, and many others. 
It is apparent from the preceding paragraphs that the discovery 
of specific antitoxins merely constituted the first step in the formu- 
lation of a fundamentally important biological law. There is, then, 
a large group of substances of animal and vegetable origin which 
call forth the formation of specific reacting bodies when injected 
into animals. In order to elicit this response it is necessary that 
these substances shall penetrate to the physiological interior of the 
50 Tschistovitch. Cited by Bordet, loc. cit., and also Ann. Past., 13, 1899. 
60 Bordet. Ann. Past., Yol. 13, 1899. 
01 Kossel. Berl. klin. Woch., No. 7, 1898. 
02 This, we know now, was due to the fact that the blood cells injected 
were not washed free of chicken serum. Thus chicken serum precipitin was 
formed as well as were hemagglutinin and hemolysin. 
63 Fish. St. Louis Med. Cour., 1900. Cited from Uhlenhuth. 
04 Wassermann and Schiitze. Deutsche med. Woch., No. 30, 1900. 
Yereinsbeilage. PHENOMENA FOLLOWING IMMUNIZATION 
105 
body in a relatively unchanged condition. For this reason any form 
of injection, subcutaneous, intravenous, or into a serous cavity, is 
followed, with regularity, by antibody formation, whereas feeding 
or other means of intraintestinal administration is negative in result, 
unless abnormal conditions prevail which permit entrance into the 
blood before the digestive enzymes have decomposed the ingested 
materials. 
The substances with which antibody-formation may be induced 
are collectively spoken of as “antigens.” 
The Conception of the Antigen.—Antigens are all substances 
which, injected into the animal body, induce specific antibody forma- 
tion. They form a large group in nature and are chemically pro- 
teins; indeed, we may say that all known proteins may act as 
antigens. Whether or not this term may also include lipoid-protein 
combinations, lipoids or the higher protein derivatives is as yet 
uncertain and need not in the present connection concern us. 
We may divide antigenic substances into two main classes.65 
One of these comprises all of those substances of bacterial, animal or 
vegetable origin which, injected into the animal body, give rise to 
specific neutralizing or antitoxic properties in the blood of the in- 
jected animal. These are the bacterial exotoxins, the snake venoms, 
some powerful vegetable poisons and proteolytic and other enzymes 
of animals and plants. They are all substances which are power- 
fully active—some of them strongly toxic to the living animal, others 
true enzymes or ferments. Indeed all of them possess properties 
which at least suggest our placing them into the class of enzymes 
in general. The number of such substances known is limited. The 
reaction they call forth in the animal body seems aimed directly at 
the specific neutralization of their respective activities, and is so 
unique and different from that induced by other antigens that it 
would be convenient had we another term like “ antitoxinogen’ to 
set them apart by themselves. 
The other class of antigens comprises all proteins which are in- 
active, showing in themselves neither toxic nor enzyme-like prop- 
erties. Introduced into the animal body parenterally, they call 
forth a response of a nature entirely unlike that of the antitoxins, 
and which as far as we can fathom its purpose seems aimed merely 
at the assimilation or the removal of the infected substance. For 
the cells of the animal cannot utilize the foreign protein as such, 
and thus it is only foreign proteins injected into an animal that act 
antigenically, and no antibodies are formed when homologous mate- 
rial is injected. 
This large group, composed of all formed and unformed sub- 
65 See also Zinsser, “The More Recent Developments in the Study of Ana- 
phylactic Phenomena.” Arch, of Int. Med., Yol. XYI, 1915, pp. 223-256. 
Harvey Lecture. 106 
INFECTION AND RESISTANCE 
stances in nature in which a protein structure is involved, does not 
induce the formation of anything like the neutralizing antitoxins 
spoken of above. The antibodies appearing in animals treated with 
such substances have been spoken of as cytolysins or cytotoxins— 
precipitins—and in the case of formed antigens like bacteria or 
blood cells—agglutinins and opsonins. As we shall see in another 
section, it is our opinion that all these various antibodies are iden- 
tical in structure and significance. 
We must not forget, however, that the observation of antibodies 
in the circulating blood is but one of the changes that have taken 
place in the treated animal. Much has been made of this phase of 
the problem because serum antibodies are readily studied in vitro; 
but their origin of course must be sought in the body cell, in which 
the original and most profound changes must necessarily have taken 
place during such treatment, changes the nature of which are to a 
large extent still a mystery, but on which ultimately depend the im- 
portant physiological difference between treated and untreated ani- 
mals. For such changes—whether we refer to those immediately 
under discussion, namely, those of allergy or anaphylaxis, or whether 
we think of the so-called immunity remaining after attacks of many 
diseases—remain present long after the circulating antibodies have 
disappeared and must therefore be regarded as associated with pro- 
found alterations in the ultimate tissue unit, the body cell. 
A good deal of our knowledge of the chemical nature of antigens 
has been developed recently, since the more delicate anaphylactic 
reactions have been available. For it is often possible to demon- 
strate antigenic properties by the anaphylactic experiment when 
the more gross determinations in the test tube are too weak for defi- 
nite interpretation. There are certain chemically true proteins 
which have no antigenic properties. The most important of these 
is gelatin. Starin 66 found that the chief difference between gelatin 
and other proteins consisted of a deficiency in the gelatin of 
tryptophan and tyrosin, and a low content of phenylalanine. Wells 67 
believes it possible that the deficiency in aromatic amino acids may 
be responsible for the lack of antigenic power. Wells and Osborne 68 
have found, on the other hand, that zein from corn lacks glycin and 
tryptophan. Gliadin of wheat and hordein of barley contain neither 
glycin or lycin, and very little arganin or histidin. These proteins 
are powerfully antigenic, and yet extremely poor in diamino acids. 
Wells suggests that since this is the case, and since, conversely, there 
are protamines which are poor antigens and consist chiefly of 
diamino acids, that the diamino acids are of little or no importance 
as antigens. 
66 Starin. Jour. Infec. Dis., 23, 1918, 139. 
67 Wells. Phys. Rev., 1, 1921. 
68 Wells and Osborne. Jour. Infec. Dis., 8, 1911, 66. PHENOMENA FOLLOWING IMMUNIZATION 
107 
The question of the antigenic properties of protein cleavage 
products has been largely investigated, and in general we may say 
that the products of protein hydrolysis have not been found anti- 
genic. Zunz 69 and, more recently, Fink 70 working in Wells’ labora- 
tory, have given this subject particular attention. Zunz produced 
heteroalbumose and protoalbumose by peptic and tryptic digestion 
of fibrin, and claims that he was able to sensitize guinea pigs and 
rabbits to the anaphylactic reaction with these substances. At- 
tempts to repeat his work by Friedberger and Joachimoglu 71 were 
unsuccessful and the amounts which it was necessary for Zunz to 
use in the second injections in his guinea pigs and rabbits made it 
impossible to exclude the possible primary toxic action of his albu- 
moses. Moreover, we know that many of the protein-split products 
injected into animals give rise to symptoms not at all unlike ana- 
phylactic reactions. Thus, Zunz’s work is not decisive. Hailer 72 
studied the digestion products of beef muscle, chiefly, with negative 
results. When he did get reactions, they were lacking in specificity 
and were always extremely feeble. Fink, working with the cleavage 
products of hydrolyzed egg white, found slightly antigenic prop- 
erties in fractions precipitated at three-quarters and complete satura- 
tion with ammonium sulphate, but not for those obtained with lower 
concentrations of the salt. Results of all these investigators which 
suggest possible antigenic properties of such cleavage products may 
well be explained by Wells’73 observation that beef serum digested 
as long as 16 months with trypsin, did not completely lose all of its 
coagulable material, in spite of the fact that the Biuret reaction 
had disappeared. It is, therefore, extremely difficult to be sure that 
there was not a certain amount of whole undigested protein left in 
the preparations used by the observers mentioned. 
Investigation of the antigenic properties of altered proteins has 
led to a considerable amount of important information which 
eventually may prove to clear up the antigen question. Of great 
importance is Ten Broeck’s 74 discovery that when protein is race- 
mized with alkali by Dakin’s method and is, in consequence, so 
changed in configuration that it is optically altered and no longer 
susceptible to enzyme action, it loses its antigenic properties. 
Perhaps the most interesting and important progress in this line 
of work is that initiated by Obermeyer and Pick.75 These investi- 
gators altered proteins by various chemical methods. They distin- 
69 Zunz. Zeit. f. Immunitats., 16, 1913, 518. 
70 Fink. Jour. Infec. Dis., 25, 1919, 97. 
71 Friedberger and Joachimoglu. Zeit. f. Immunitats., 24, 1914, 522. 
72 Hailer. Arb. a.d.k. Gesamtsht., 527, 1914. 
73 Wells. Jour. Infec. Dis., 6, 1909, 506. 
74 Ten Broeek. Jour. Biol. Chem., 17, 1914, 369. 
75 Pick. Biochemie der Antigene, Jena, Fischer, 1912; Kolle & Wasser- 
mann Handbuch, Wien. Klin. Woch., 22, 1903, 10, 1904, 12, 1906. 108 
INFECTION AND llESISTANCE 
guishcd between methods which simply changed physical constitu- 
tion and others which actually interfere with the chemical structure 
of the protein. Heating, the action of toluol, chloroform, simple 
acids and alkalis, produced modifications of antigenic properties, 
but did not alter species specificity. This is in keeping with certain 
investigations upon the cocto-precipitins made by Schmidt and by 
ourselves, which are discussed in the precipitin chapter. However, 
when they introduced N02 groups by treatment with nitric acid or 
diazotized the protein with nitrous acid, or iodized it by treatment 
with Lugol’s solution, they found that the specificity of the antigenic 
properties of the protein was actually altered. Thus, an immune 
serum produced by the treatment of animals with any one of these 
altered proteins, no longer reacted with the native horse serum from 
which it had been produced, but reacted strongly with similarly 
altered proteins from other species. The species specificity, there- 
fore, had given way to a specificity of chemical structure. Since 
in these reactions the change produced was due to some substitution 
in the aromatic radicals of the proteins, they suggest that such 
changes are of paramount importance to antigenic function. This 
would fall into agreement with Wells’ suggestion that the deficiency 
in aromatic amino-acids is responsible for the lack of antigenic prop- 
erties found in gelatin, and fortify the conclusions drawn by Wells 
and Osborne from their very careful structural studies of antigens, 
that chemical structure is the important factor since chemically 
similar gliadin and hordein, even when derived from different plants, 
may be antigenically very similar to each other. Landsteiner76 
with many collaborators, has given this subject a considerable amount 
of attention. It is quite impossible in the confines of this book to 
go into the details of all the important and laborious work which 
Landsteiner has published on this subject. We must, therefore, 
select from all his experiments, a few to illustrate their work. One 
of the processes used particularly was diazotizing the proteins. An 
example of their procedure as given by them, is as follows: 100 c. c. 
of horse serum were treated with 100 c. c. of normal soda solution, 
and a small amount of a 1 per cent, solution of diazo-benzol pro- 
duced from analin was added. The solution was cooled and a drop 
tested for alkalinity with phenolphthalein. Then it was allowed to 
stand in ice-water for about 10 minutes and precipitated with hydro- 
chloric acid. The precipitate was caught on a filter paper, taken 
up in a little water and brought into solution by the gradual addi- 
tion of normal soda solution and then precipitated with alcohol. 
This precipitate was again dissolved in water, and again precipitated 
76 Landsteiner. Zeit. f. Immunitdts., Vol. 20, 1913, 211 and 618; Vol. 21, 
193; Yol. 26, 136; Biochem. Zeit., Vol. 58, 362, 61, 67, 74, 86, 1918, 343 
and 93, 1919, 106; Proc. Boyal Academy of Sciences, Amsterdam, 25, No. 1, 
p. 34. PHENOMENA FOLLOWING IMMUNIZATION 
109 
with alcohol, and this again taken up in water by the addition of a 
small amount of soda solution. It was then neutralized to litmus, 
and brought to a volume of about 200 e. c. in 1 per cent, salt solu- 
tion to which }/2 per cent, phenol had been added. The alcohol 
precipitation was employed in order to remove toxic substances 
which were found to be present unless this was done, and in some 
cases this had to be repeated in order to permit injection into 
animals. For further particulars, we refer to Landsteiner’s paper 
in the Biochemische Zeitschrift for 1918, page 343. The general 
conclusions of his work may be stated as follows: The specificity 
of antigens is apparently determined by the chemical structure of 
relatively small parts of the large antigen molecule. This is par- 
ticularly evident from the fact that as Pauli 77 has shown, the diazo 
bodies, introduced among other things by Landsteiner, enter only 
the tyrosin and histidin groups of protein, which would constitute 
about 3 per cent, of the proteins used. In the course of this work, 
Landsteiner did a large number of reactions with 23 different cases 
or antisera produced with 33 different azo-proteins. Six of the 
immune sera reacted only with antigens which had been produced 
in a homologous way, in regard to the azo components. In some 
of the others there were group reactions. Obermeyer and Pick’s 
observation that the species specificity was entirely altered, and a 
chemical specificity now substituted, was definitely confirmed, and 
it was further shown that the particular place in the aromatic ring- 
in which the new radical, that is the azo-group was introduced, was 
also important in determining specificity. 
With such slight chemical differences determining antigenic 
properties, it is not at all impossible that the differences in the species 
specificity of animal proteins may depend purely on structural 
variations in the aromatic amino groups, which, because of the 
isomeric nature of proteins, it would be almost impossible to define 
with chemical exactitude at the present time. 
Moreover, such investigations also make it quite easy to under- 
stand that the same serum or protein of any kind is a mixture of a 
number of different and separately specific antigens. This has been 
often found, and especially emphasized in the work of Dale and 
Hartley 78 who separated the euglobulin, pseudoglobulin and albu- 
moses of horse serum, and found that each fraction could be made 
to sensitize separately. Similar wTork has been done in the separa- 
tion of protein constituents of egg albumin. 
Before leaving the subject of antigens it seems advisable to add a' 
brief discussion of a form of substance which is antigenic only in 
the sense that it reacts specifically with antibodies, but with which 
the formation of antibodies by injection into the animal has not yet 
77 Pauli. Zeit. f. Physchem., 42, 1904, 512, Vol. 94, 284 and 428, 
78 Dale and Hartley. Biochem, Jour,, 10, 1916, p, 110, 110 
INFECTION AND RESISTANCE 
been demonstrated. Some years ago, Landsteiner suggested that 
tuberculin might be such a substance, largely because, in the case of 
this material, specific reactions are produced in the injected animal 
body, but attempts to produce antibodies with it have failed. Our 
own later researches have lent strong probability to this suggestion 
of Landsteiner. From a considerable number of bacteria, tubercle 
bacilli, pneumococci, staphylococci, influenza bacilli, meningococci, 
we have been able to obtain substances which remain in the residue 
after boiling with acid, and an attempt to remove all of the ordinary 
protein materials. These substances are precipitable with alcohol, 
easily soluble in water, and in extremely minute amounts give power- 
ful reactions with anti-sera, both precipitating and fixing alexin 
specifically. Although our work is not finished, we have not been 
able to produce antibodies with these materials when sufficiently 
purified. 
In view of the chemical studies detailed above, it seems quite 
likely that the antigenic function of specific union with the anti- 
body may be dependent upon a relatively small nucleus of the pro- 
tein molecule, which perhaps we have been able to split off by the 
methods used by us. Permitting ourselves a tentative speculation, 
our work suggests to us a teleological conception of antibodies in 
the following wray: Antibodies are formed in the animal body only 
upon injection of entirely non-diffusible substances like the true 
proteins. These substances being of large molecular size and non- 
diffusible, can, therefore, react with the cellular elements of the 
body only by cell surface relations, and in order to go into reaction 
with the tissues, antibodies are necessary and, therefore, formed. 
Since the molecule becomes smaller, and relation with the cell by a 
certain amount of diffusion becomes possible, antibody formation 
becomes less and less necessary as a physiological reaction. In a 
vague way this conception associates antibody formation with 
molecular size. These considerations are still largely speculative, 
but are rapidly taking form in experimental work which we have 
not been able to finish at the present writing.79 
For substances such as those described above, namely, with the 
ability to unite with antibodies, but the apparent lack of ability, 
possibly based on small molecular size, of stimulating antibody 
formation in the animal body, Landsteiner has proposed the term 
Haptene. 
Since the phenomenon of antibody formation is not at all limited 
79 It must not be supposed in this connection that the function of anti- 
bodies in the disposal of foreign proteins or bacteria can be regarded as a 
process of direct ingestion. Proteolytic and other enzyme activities have 
not been shown to be involved in the reaction between antibodies and their 
antigens, whether or not alexin took part in these reactions. This has been 
particularly investigated by Jobling and Petersen. See Jour. A. M. A,, 65, 
1915, 515 and 66, 1916, 1753; Archiv. Inter. Med.} 15, 1915, 286. PHENOMENA FOLLOWING IMMUNIZATION 
111 
to bacteria or bacterial derivatives, it cannot be looked upon merely as 
a mechanism existing for the primary purpose of protecting the body 
against infectious disease. This latter function is important, indeed, 
but the process undoubtedly has a much more general biological 
significance. 
In the course of normal existence substances which are not di- 
rectly assimilable as such—foreign proteins, for instance—do not 
penetrate directly into the blood and tissues. Taken into the ali- 
mentary canal, they are first hydrolized into peptons, albumoses, 
polypeptids, and probably amino-acids before absorption, to be recon- 
structed from these cleavage products (“Bausteine” is Abderhalden’s 
expression for the amino-acids) into protein biologically identical 
with that of the tissues. Digestive and other accidents, however, on 
numerous occasions during life permit the direct entrance of these 
materials unchanged or insufficiently changed into the circulation. 
It is probably by the action of digestive powers of the serum—or, in 
the case of the entrance of the undissolved foreign particles, by the 
activity of the phagocytic cells—that such substances are then dis- 
posed of and assimilated. For each particular variety of substance 
(antigen) a specific mechanism is called into play, and when this 
mechanism is repeatedly called upon—as in successive injections of 
foreign proteins—this mechanism, whatever it may consist of, is en- 
hanced in efficiency—i. e., increased in quantity. How this increase 
of specific antibodies is theoretically conceived wre will discuss later 
in connection with Ehrlich’s side-chain theory. 
The phenomena of antibody formation against bacteria on this 
basis may be taken to constitute, then, a mechanism for the dis- 
posal 80 of a foreign protein which has penetrated into the tissues 
and, because of its living state, increases within the body by multi- 
plication, furnishing progressive stimulation to the antibody-pro- 
ducing function. Infectious disease, therefore, from this point of 
view may be looked upon as an invasion of the body by a living 
foreign protein which must be assimilated and disposed of; which, 
in some cases, has a primary toxicity per se; and which is variously 
distributed among the organs and tissues according to the biological 
peculiarities of the particular micro-organism in question. This 
general conception will become more clear as we analyze the phe- 
nomena associated with the individual antibodies. It is, of course, 
quite plausible as far as it refers to the phagocytic functions, or even 
bacteriolytic and cytolytic phenomena. It has been less clear in 
connection with the agglutinins and precipitins in which a direct de- 
fensive or bacteria-destroying value is not apparent. However, in 
our discussions of these phenomena we will have occasion to point out 
many reasons for assuming that, even in these phenomena, there are 
60 It will be seen in a later section that the action of antibody and alexin 
on an antigen does not imply proteolysis or any form of digestion. 112 
INFECTION AND RESISTANCE 
features which fall into direct correlation with the views we have 
just expressed. 
The substances which possess antigenic properties—that is, which 
give rise to antibody production—with the exception of a few isolated 
and contested cases, are all of them protein in nature. Well-trained 
chemists have exerted themselves to purify antigenic substances, 
in attempts to determine the particular fractions of the complex 
protein molecule upon which the antigenic properties depend. In 
the course of such work a number of men claim to have obtained a 
truly antigenic substance which no longer gave protein reactions. 
The instance most frequently cited is Jacoby’s81 announcement of 
a protein-free ricin. Jacoby worked with an apparently very impure 
“Ausgangsmaterial” consisting of commercial ricin, which he di- 
gested for five weeks in trypsin solution. At the end of this time he 
obtained a ricin which still possessed the properties of the original* 
castor-bean extract, but no longer gave protein reactions. His “puri- 
fied ricin,” however, was quickly destroyed by further trypsin diges- 
tion, and more recent work by Osborne, Mendel, and Harris82 
appears to have fully refuted Jacoby’s results. They found the 
purified ricin identical with the coagulable albumin of the castor 
bean, and found that tryptic digestion destroys the characteristic 
ricin properties. 
Less easily refuted have been the careful experiments of Ford 83 
upon the active principle of a mushroom (Amanita phalloides) and 
upon that of the poison-ivy plant—(Rhus toxicodendron). These 
substances, he claims, are non-protein. In the case of Amanita 
phalloides Abel and Ford 84 have shown it to be a glucosid, and 
similar structure has been claimed for Rhus by Syme.85 Yet with 
both of these substances Ford has succeeded in producing specific 
antitoxins. Rabe 80 has questioned these with Amanita phalloides. 
He believes that the poison with which Ford worked is not a glucosid, 
but is of protein nature. In the case of Rhus, however, Ford’s con- 
clusions have not, to our knowledge, been challenged. 
With these and a few other less important exceptions, however, 
observers have uniformly concluded that antigenic property and pro- 
tein structure are inseparably associated. All procedures by which 
proteins have been hydrolized into their simpler fractions, chemical 
splitting, tryptic or peptic digestion have in every case resulted in a 
simultaneous loss of protein reaction and antigenic property. 
We will see in a later discussion of relationship of antigenic 
structure to anaphylactic reaction that the possibility has been sug- 
81 Jacoby. Arch. f. exp. Path. u. Pharm., Vol. 46, 1901. 
82 Osborne, Mendel, and Harris. Am. Jour, of Physiol., 1905, Vol. 14. 
83 Ford. Jour. Inf. Bis., Vol. 3, 1906; Vol. 4, 1907. 
84 Abel and Ford. Jour. Biol. Chem., 1907. 
85 Syme. Johns Hopkins Thesis, 1906. 
86 Rabe. Zeitschr. f. exp. Path. u. Therap., Vol. 9, 1911. PHENOMENA FOLLOWING IMMUNIZATION 
113 
gested by Landsteiner and by ns, that the antigenic nature of the 
proteins may be not only a question of chemical structure, but also 
one of molecular size, our own idea being that the antibody mechan- 
ism is a provision by which it is made possible for the body to 
deal with substances that cannot diffuse into the cell substance. 
Many attempts have also been made to show a relation between 
antigenic properties and the lipoid constituents of cells. These en- 
deavors were obviously stimulated by the observation that many li- 
poids are capable of binding antibodies in vitro, and that, in nervous 
tissues, toxin fixation was in some way related to the richness in 
lipoids of these structures. Bang and Forsmann87 accordingly 
treated animals with ether extracts of red blood cells—claiming that 
this resulted in the production of hemolysins. And these results 
have been confirmed by Landsteiner and Dautwitz.88 The latter, 
however, suggest that the hemolysin production may have been in- 
duced, not by the lipoidal substances in solution, but by other anti- 
genic substances which had gone into colloidal suspension in the 
ether extracts. Much similar research on the antigenic nature of 
lipoids has been done, but, after reviewing this very thoroughly, 
Landsteiner comes to the conclusion that no definite proof of the anti- 
genic nature of any pure lipoid has so far been presented. The prob- 
lem is experimentally complicated by the fact that, as Landsteiner 89 
suggests, the antigen may often be present as a lipoid-protein com- 
bination, and as such go into solution or fine emulsion in the organic 
solvents; also the lipoids possess the curious property of altering the 
solubilities of proteins and other substances by their presence. 
Summarizing our present knowledge of the chemical nature of 
antigens, then, we must conclude that, with the exception of Ford’s 
glucosids, no protein-free antigens have been thus far demonstrated. 
In the light of this fact it is all the more remarkable that antigen- 
antibody reactions are specific. For we possess no chemical methods 
by which one variety of protein can be distinguished from another. 
And yet the serum antibodies produced with each species of bacteria 
react with this species only—and the hemolysins, agglutinins, or 
precipitins produced by the injection of bacterial, cellular, or serum 
proteins react respectively only with the particular variety employed 
in their production. This indicates that each of these antigens—of 
almost unlimited number—must possess a chemical structure indi- 
vidually characteristic and different from all the others. It is by 
means of the biological reactions, indeed, that we can detect protein 
in dilutions far beyond the reaction-sensitiveness of chemical tests 
and can distinguish between varieties of protein when the chemical 
87 Bang and Forsmann. llofm. Beitr., 1906; Centralbl. f. Bakt., 40, 1906. 
88 Landsteiner and Dautwitz. Hofm. Beitr., 9, 1907. 
89 Landsteiner. “Wirken Lipoide als Antigene?” Weichardt’s Jahresbe- 
richt, Vol. 6, 1910. 114 
INFECTION AND RESISTANCE 
methods will indicate only protein in general. Our knowledge of 
the chemical constitution of protein has not yet advanced to a point 
at which specificity can be based upon definite variations of chemical 
structure, and the complexity of the problem is such that it does not 
seem likely that we can hope in the near future to attain such knowl- 
edge. We can merely accept it as a fact that the antibody produced 
with one protein differs materially from that produced with another, 
and that this is a definite indication that the antigen in one case must 
be chemically different from that in another. 
The range of such variations is apparently enormous. For each 
variety of bacteria or plant, each species of animal, and to a certain 
extent each individual of the species, possesses certain special anti- 
genic characteristics peculiar to itself. In general there is an under- 
lying antigenic similarity which is peculiar to the species. This is 
true of bacteria and, in the case of animal and vegetable proteins, an 
antibody produced with material from an individual of a certain 
species will react with the protein derived from this species in gen- 
eral. However, that there are also antigenic differences between in- 
dividuals within the same species is indicated by Ehrlich’s experi- 
ments on the antibodies produced by injecting the blood cells of one 
goat into another. And w’e have further indicated that within the 
same animal different organs may possess individual antigenic char- 
acteristics. Added to this we know that certain special organs like 
the testicle, the lens, and some others contain antigens which are 
peculiar to this variety of organ, irrespective of species—a condition 
spoken of as “organ specificity ” Thus an antibody produced by 
injections of the testicular substance of one animal will react with 
testicular protein from many different species—the specificity here 
depending upon the organ and not upon the zoological relationship. 
It is clear, therefore, that there are more different varieties of 
protein, biologically distinguishable, than there are species of living 
beings in nature. As Abderhalden 90 has recently pointed out, this 
is a conception which it is a little difficult to grasp chemically, since 
in breaking up different proteins into their ‘‘building stones” (Bau- 
steine) we encounter again and again the same 20 amino-acids. By 
a simple arithmetical consideration, however, he shows that merely 
by combining these twenty amino-acids in different groupings 
an enormous number of isomeric but varying compounds can be 
formed—even without assuming the additional possibility of 
quantitative variations. He reasons that 3 “Bausteine”—A, B, 
and C—could form 6 different structures, A B C, A C B, B O A, 
B A C, C A B, 0 B A. Similarly 4 could form 26, and finally 20 
could form 2, 432, 902, 008, 176, 640, 000 different compounds.91 
90 Abderhalden. Munch, med. Woch., No. 43, 1913. 
91 We have not repeated the arithmetical labor and take Abderhalden’s 
word for it. PHENOMENA FOLLOWING IMMUNIZATION 
115 
This problem is more fully dealt with in the section dealing with 
antigens. 
Drug Tolerance.—The analogy between the active immuniza- 
tion of animals with the various antigens and certain chemically 
well-defined poisons, alkaloids, etc., is so obvious that it has led 
to much speculation as to a possible similarity in the physiological 
mechanisms of the two phenomena. As a matter of fact the acquired 
tolerance for such substances as morphin, atropin, and other alka- 
loids is not really analogous to the physiological reactions which fol- 
low the treatment of animals with bacterial and other proteins, for 
whatever toxic properties there are in the latter are, as we shall 
see later, rather the results of the interaction of these injected 
substances and the reaction products supplied by the cells and fluids 
of the body. It is at least probable in the light of our modern con- 
ception that such protein antigens are not toxic per se, in the native 
state. This, however, will receive detailed consideration in succeed- 
ing sections. The analogy of drug tolerance, however, to the ac- 
quired immunity against true bacterial toxins and vegetable poisons 
like ricin, crotin, and others is a striking one, since in both classes 
of poisons there is a gradually developed tolerance for substances 
toxic in the native state and often very similar in physiological 
effects (strychnin and tetanus toxin, etc.). In the case of the 
toxins, however, there is a development of immunity by actual neu- 
tralization of the poisonous principle brought about by a specific 
antibody, which circulates in the blood of immunized animals and 
man—the process following, within certain limits, the law of multi- 
ple proportions. In the case of morphin and other alkaloids no 
such neutralizing antibodies have as yet been demonstrated.92 
Whereas toxin immunity is passively transferable from one animal 
to another with the blood serum, and, in vitro, the mixture of the 
toxin with the immune serum brings about a neutralization of the 
poison, no such phenomena have been observed, as a general rule, 
in the case of the alkaloids. We say “as a general rule” since an 
exception is recorded in the observations of Fleisehmann,93 who 
claims to have found antagonistic action to atropin in the blood 
of normal rabbits, this power being absent from the blood of rabbits 
that had thyroid hypertrophies and were, in consequence, atropin- 
susceptible. Other observations of a similar significance have been 
made by Physalix and Contejean 94 on curare, but have not been 
confirmed, and the investigations of all other workers on this sub- 
ject have had negative results. It seems from available evidence that 
92 Hans Meyer and Gottlieb. “Exp. Pharm.,” 2d Ed., Neban & Schwart- 
zenberg, Berlin, 1911, p. 517. 
93 Fleisehmann. Archiv. f. exp. Path. u. Pharm., 62, 1910, cited from 
Meyer and Gottlieb, loc. cit. 
94 Physalix and Contejean. Cited from Meyer and Gottlieb. 116 
INFECTION AND RESISTANCE 
tolerance (immunity) against drugs is due to cellular rather than 
to serum antagonism. In this connection we refer the reader to the 
section on Drug Idiosyncrasies. 
The Origin of Antibodies.—The tissue cell, as the ultimate func- 
tional unit, must, of course, be looked upon as the source from which 
originate the various protective constituents of normal and immune 
sera; and, though perhaps unrecognizable by the coarse tests of 
morphological investigations, it is in the cells that changes must take 
place primarily when the animal body is subjected to any one of the 
processes spoken of as immunization. The exact location of the 
antibody-forming cells and tissues, in spite of much investigation, 
is not at all clear, though many data seem to point to the lymphatic 
organs, the spleen, and the bone marrow as particularly concerned 
with this process. 
Thus Pfeiffer and Marx 95 exsanguinated animals five days after 
injections of dead cholera spirilla and found that at this time bac- 
teriolytic antibodies were more concentrated in the spleen than in 
the blood serum itself. Wassermann’s 96 analogous experiments with 
typhoid bacilli seemed to show a higher antibody content in spleen, 
bone marrow, thymus, and lymph nodes than was present in the 
blood at an early period of immunization. Although these investiga- 
tions, as well as many others of Castellani,97 seem, therefore, to indi- 
cate a particular association of the special lymphatic organs with 
antibody formation,98 extirpation of the spleen 99 before immuniza- 
tion has not prevented animals from responding to injections of bac- 
teria and red blood cells with sharp antibody production. The ex- 
periments of Deutsch,100 in which reduction of antibody formation 
resulted in animals in which splenectomy was practiced three or four 
days after immunization was begun, can hardly be accepted as a con- 
clusion, in the writer’s opinion at least, since any severe operation or 
interference with the normal functions of an animal during the 
severe physiological strain of active immunization would naturally 
lead to a less perfect response. That the resistance of animals and 
man to infection with bacteria is not noticeably diminished by sple- 
nectomy, moreover, has been variously shown. In unpublished 
experiments by the writer splenectomized guinea pigs showed no 
difference from normal animals in regard to their susceptibility to 
tuberculosis. And though these and similar experiments of othei 
workers with various bacteria are not entirely devoid of interest, 
their negative results as a matter of fact have no great significance, 
95 Pfeiffer and Marx. Zeitschr. f. Hyg., Yol. 27, 1898. 
96 Wassermann. Berl. klin. Woch., p. 209, 1898. 
97 Castellani. Zeitschr. f. Hyg., Yol. 37, 1901. 
98 Pfeiffer and Marx. Loc. cit. 
991. Levin. Jour. Med. Res., Vol. 8, 1902. 
,00Deutsch. Ann. de VInst. Past., Vol. 13, 1899. PHENOMENA FOLLOWING IMMUNIZATION 
117 
since our knowledge concerning the true function of the spleen is 
very incomplete, and it is not impossible that on removal of this 
organ other elements of the lymphatic system may take over its 
function in part or as a whole. 
Removal of the spleen has not been an extremely unusual pro- 
cedure in surgery, and there is no evidence to show that patients so 
treated have been abnormally susceptible to infection thereafter. 
Yet, as we have seen, there seems to be an early concentration of 
antibodies in the lymphatic organs in the course of immunization, 
and it may well be that an association between the process and these 
tissues exists which cannot be experimentally demonstrated with 
absolute certainty. 
It is no less likely, however, that similar functions are exerted 
by the cells of other organs. In fact, it is more than probable that 
antibodies may be formed anywhere in the body—and that the local- 
ity of their production is largely dependent upon the locality in which 
the antigen is concentrated. Wassermann and Citron 101 demon- 
strated this by injecting typhoid bacilli into rabbits intraperitoneally, 
intravenously, and intrapleurally, and nine days afterward deter- 
mining the comparative bactericidal strength of blood serum and of 
aleuronat exudates of pleura and peritoneum in each of the three 
animals. Their results showed that the bactericidal titre of the 
intravenously inoculated animal was highest in the blood serum, 
while that of the intraperitoneally and intrapleurally inoculated 
animals was highest in peritoneal and pleural exudates respectively. 
Such experiments point to the possibility of a “local” immunity, 
that is, a production of antibodies directly by the cells with which 
the antigen comes into contact in the most concentrated and direct 
manner. And, indeed, another isolated experiment of the same 
authors, alone successful of a series of similar attempts, would point 
in the same direction. Typhoid bacilli were injected subcutaneously 
into the ear of a rabbit and the ear immediately ligated at its base 
and kept so for several hours. After nine days the bactericidal titre 
of the blood serum was determined and the ear amputated. An 
immediate and rapid drop of antibody contents occurred after the 
amputation—indicating that the chief source of antibody function 
had been removed. More striking examples of the same thing are 
to be seen in the experiments of Romer,102 who instilled abrin into 
a rabbit’s eye and found that the retina of the eye developed an 
antitoxic power against abrin which protected mice against many 
times the fatal dose, while that of the other eye remained practically 
inactive. 
From these facts, as well as from other observations, it is at least 
reasonable to believe that antibody formation is by no means a func- 
101 Wassermann and Citron. Zeitschr. f. Hyg., Yol. 50, 1905. 
102 Romer. Arch. /. Ophthal., 52, 901. 118 
INFECTION AND RESISTANCE 
tion of special organs and that many cells throughout the body may 
take part in the process. It is of especial importance to consider this 
in connection with the possible effects of the treatment of infections 
by means of bacterial vaccines. If the focus of the infection can 
possibly become also a local source of antibody production, then such 
treatment may well seem rationally founded, even in generalized 
acute infections in which no logical basis for such treatment would' 
exist, were the production of antibodies a task for specialized organs 
like spleen and bone marrow only. The therapeutic phases of this 
problem are more extensively considered in a later chapter. 
It is in this fact also that we must seek the explanation of the 
apparent local immunity which occurs in certain infections of the 
skin. Thus it frequently happens that successive crops of boils may 
afflict different parts of a patient’s skin—new ones arising as old 
ones heal, showing that the process of the limitation and healing of 
the infected foci is not due to any increase of generalized resistance, 
but rather to local causes. In the same way, in erysipelas, the process 
extends along the edges while the original central area of infection 
is returning to the normal state, and it rarely occurs in adults that 
the erysipelatous process extends back into the originally infected 
area.103 From these localized laboratories of antibody formation, of 
course, distribution to the circulation probably takes place and the 
complete cure of the patient must await a sufficient concentration of 
these in the body as a whole before further local foci cease to arise* 
That the fixed tissue cells of any part of the body can and do 
take an active part in the local reaction against the invasion of bac- 
teria and other foreign materials is histologically evident. When 
a more or less insoluble foreign body—a thread of lint, paraffin, 
agar-agar, or other material—is deposited in the subcutaneous tissues 
anywhere in the body, and is accompanied by acute infection with 
bacteria, there is a characteristic tissue reaction which results in the 
surrounding of the foreign particle by multinucleated cells spoken 
of as giant cells. In the case of foreign bodies such as those men- 
tioned the process is purely one of local ingestion of the particle 
which later, if the material remains absolutely insoluble, results in 
encapsulation by connective tissue. If soluble, however, there may 
be an eventual digestion of the foreign material by the cell with a 
subsequent degeneration or splitting up of the giant cell and a return 
to normal. This also occurs in the case of such infections as those 
due to yeasts or blastomyces, in which, as the writer has seen, the 
apparent lack of liberation of toxic products gives rise to a purely 
local giant-cell reaction, adjacent tissue cells remaining undegener- 
ated and apparently unaffected. In the case of infection with bac- 
teria like the bacillus of tuberculosis, the leprosy bacillus, that of 
103 In children erysipelas not infrequently returns within a few days over 
a recently healed area PHENOMENA FOLLOWING IMMUNIZATION 
119 
rhinoscleroma, and a few others the purely local picture of giant- 
cell phagocytosis is complicated by secondary reactions arising prob- 
ably from the liberation of toxic products from the living or dead 
invaders which both stimulate specific cell reactions and call forth 
cell degeneration in adjacent tissues, frequently giving the individual 
infection a diagnostically characteristic appearance. CHAPTER V 
TOXIN AND ANTITOXIN 
THE REACTION BETWEEN TOXIN AND ANTITOXIN 
(EHRLICH’S ANALYSIS) 
THE TOXIN-ANTITOXIN REACTION 
When Behring and his collaborators, Kitasato and Wernicke, 
had definitely shown that the cell-free blood serum of animals im- 
munized with tetanus and diphtheria toxins respectively possessed 
the power to protect other animals of the same and different species 
against the poisons, it became of the utmost importance to deter- 
mine, if possible, the mechanism by which the “antitoxic” effect was 
attained. The earlier opinion, expressed by Behring himself, held 
that in all probability the toxin was directly injured or destroyed by 
the action of the antitoxic serum. That this assumption was incor- 
rect was soon demonstrated by the experiments of Roux and Vail- 
lard 1 and by those of Buchner.2 The work of the former investiga- 
tors showed that the mixtures of tetanus toxin and antitoxin, meas- 
ured in such proportions that they were harmless for normal guinea 
pigs, could still be found toxic for animals weakened by preliminary 
inoculation with other bacteria. Buchner claimed in analogous ex- 
periments that similar mixtures, harmless for mice, could still show 
toxicity for guinea pigs. He inferred from this that the nature of 
the cell reactions of different animal species influenced the antitoxic 
effect. Both investigations led the workers to conclude that the pro- 
tective action of antitoxin was not due to a direct effect upon the 
poison but was potent by acting upon the tissue cells of the animal by 
protecting these from subsequent harm by the toxin. Their concep- 
tion implied an indirect protective function on the part of the anti- 
toxin, not due to any direct reaction between it and the poison. 
That this explanation, too, was faulty was made evident by a 
number of investigations which took advantage of the peculiar dif- 
ferences in resistance to temperature between certain toxins and 
their specific antitoxins. 
In 1894 Calmette 3 and Physalix and Bertrand 4 had indepen- 
1 Roux and Vaillard. Ann. de VInst. Past., 1894. 
2 Buchner. Munch, med. Woch., p. 427, 1893. 
3 Calmette. Compt. rend, de la soc. de biol., 1894. 
4 Physalix and Bertrand. Compt. rend, de la soc. de biol., 1894. 
120 TOXIN AND ANTITOXIN 
121 
dently succeeded in obtaining an antitoxin against snake poison. In 
the course of further study of these bodies Calmette 5 determined 
that the venoms of certain varieties of snakes, the naja and cobra, 
would remain potent even when subjected to 100° C. for a very 
short time. In contrast to this the antitoxins to these poisons were 
destroyed at much lower temperatures. How when mixtures of 
the two substances, so proportioned that their injection into ani- 
mals was innocuous, were heated to 68° C. for considerable 
periods, toxic properties again became evident, a demonstration that 
the toxin had not been destroyed, but had remained neutral only in 
the presence of the intact antitoxins. These experiments were con- 
firmed by Wassermann,6 who found that similar conditions pre- 
vailed in the combination between pyocyaneus toxin and antitoxin. 
The filtration experiments of Martin and Cherry 7 are not con- 
vincing since they may be taken as indicating either neutralization 
or toxin destruction. These workers subjected mixtures of snake 
poison and its specific antitoxin to filtration through gelatin filters, 
under pressure. Under the experimental conditions thus estab- 
lished the presumably smaller toxin molecule was allowed to pass 
through the filter while the larger antitoxin molecule was held back. 
They showed that if filtered soon after the ingredients have been put 
together most of the toxin still passes through, but that, as this inter- 
val is prolonged, less and less comes through, presumably because of 
the union of the smaller toxin to the larger antitoxin molecule. The 
chief value of these experiments lies in their proof of the element of 
time as an important factor in the toxin-antitoxin union. 
In his experiments on snake venom just recorded, Calmette in- 
terpreted the restitution of toxicity after the heating of neutral mix- 
tures of cobra neurotoxin and its antitoxin as evidence “qu’il ne 
s’etait pas forme aucune combinaison de ces deux substances ou que 
la combinaison realisee etait, au moins, tres instable.” Later experi- 
ments of Martin and Cherry seemed for a time to contradict this con- 
clusion. Observations by them, analogous to those of Calmette, but 
carried out with the poison of an Australian snake, seemed to indi- 
cate that when the toxin and antitoxin were allowed to remain to- 
gether for a sufficiently long time no restitution of toxicity could be 
obtained by heating. Apparently the application of heat to such mix- 
tures merely prevented the further union of antitoxin with any toxin 
that was not yet bound at the time that the heat was applied. Accord- 
ingly Morgenroth 8 again examined these relations and found that the 
addition of a small amount of hydrochloric acid to mixtures of snake 
poison and the antitoxin resulted in the dissociation of their union. 
5 Calmette. Ann. Past., 1895. 
6 Wassermann. Zeitschr. f. Hyg., 22, 1896. 
7 Martin and Cherry. Proc. of the Royal Soc., Yol. 63, 1898. 
8 Morgenroth. Berl. Klin. Woch., No. 50, 1905, p. 1550. 122 
INFECTION AND RESISTANCE 
To mixtures of the venom lysin and its antitoxin, neutralized and 
even overneutralized so that they were perfectly innocuous to suscep- 
tible animals he added hydrochloric acid until the total concentra- 
tion amounted to N/18. By this method a toxin-HCl modification 
was produced which was dissociated from its union with the anti- 
toxin and was extremely resistant to heat. In such a mixture of 
toxin and antitoxin to which hydrochloric acid had been added, heat- 
ing at 100° C. in a water bath for 30 minutes destroyed the ther- 
molabile antitoxin and, after neutralization, undiminished toxic 
properties could again be demonstrated by animal inoculation. 
These researches and other similar ones of Morgenroth, then, 
form a satisfactory confirmation of the original experiments of Cal- 
mette and seem to show, beyond possibility of contradiction, that the 
inhibition of harmful properties of any true toxin, after mixture 
with its antitoxin, does not depend upon toxin destruction. But 
while Calmette interpreted the facts as pointing toward a failure of 
union of the two substances, Morgenroth’s work is not incompatible 
with the conception of a neutralization of one by the other in the 
chemical sense. These experiments of Morgenroth are of great the- 
oretical importance moreover in that they have shown that dissocia- 
tion of a toxin-antitoxin complex can occur. 
The nature of such neutralizations in regard to quantitative rela- 
tions, speed of action, and relative concentrations, becomes apparent 
partly from experiments like those mentioned above, but more espe- 
cially from those carried out by Ehrlich with ricin and antiricin, ex- 
periments which were primarily planned to demonstrate that the 
reaction between a toxin and its antibody is a direct one, not depend- 
ent upon intervention of the body cells, as at first supposed. 
It had been shown by Kobert and Stillmarck that ricin, the 
powerfully poisonous principle of Ricinus communis (castor oil 
bean) would agglutinate the red blood cells of a number of animals. 
Ehrlich recognized from the beginning how closely analogous the 
neutralization of ricin by antiricin was to that of diphtheria toxin by 
its antitoxin. The former reaction furnished him with a simple 
method of test tube experimentation in that the agglutinating effects 
of ricin upon rabbits’ corpuscles could be directly inhibited by the 
preliminary addition of antiricin. A visible reaction was thus avail- 
able, which, of course, excluded absolutely the participation of the 
tissue cells in the antigen-antibody neutralization, and in which 
careful quantitative measurements were possible. 
Ehrlich 9 determined by means of this method that the neutral- 
ization was accelerated by moderate heat and by concentration of the 
reagents and, most important of all, that the reaction followed 
roughly the law of multiple proportions, characteristics, all of them, 
which were entirely analogous to chemical reactions in general. 
9 Ehrlich. Fortschr. d. Med., Yol. lb, 1897, p. 41. TOXIN AND ANTITOXIN 
123 
When he added 0.3, 0.5, 0.75, 0.1, etc., cubic centimeters of serum 
from a ricin-immune goat to constant quantities of ricin, and then 
added rabbit cells, the hemagglutinating properties of the ricin were 
inhibited in direct proportion to the amount of antiricin mixed with 
it. And his test tube experiments were further found to represent 
with much accuracy the occurrences which took place within the 
animal body. For, similar mixtures injected into mice were toxic 
in direct proportion to the balance of ricin and antiricin established 
in the injected material. 
Although the views of Ehrlich and his followers have great im- 
portance in connection with the union of antigens and their anti- 
bodies in general, these ideas were worked out by him most elab- 
orately in connection with his efforts to arrive at a practicable and 
accurate method of establishing a standard of strength for diphtheria 
antitoxin, and it is essential that we consider this work in detail in 
the following paragraphs. 
Development of Antitoxin Standardization.-^-The earlier at- 
tempts to standardize diphtheria antitoxin by the use of living 
cultures (Roux and Behring) were soon abandoned, since it was 
found that the accurate establishment of fixed lethal doses of the 
culture was not possible. When the facts, just recorded, concerning 
the interaction and quantitative relations of the soluble toxins and 
their respective antitoxins came to light, Behring introduced the 
standardization of the curative sera by the use of toxins, both in the 
case of tetanus and in that of diphtheria. In order to do this con- 
sistently he established for diphtheria poison an arbitrary toxin unit 
which he defined as the amount of any given diphtheria filtrate 
sufficient to cause death in a guinea pig of 250 grams, and, borrowing 
the terms from chemical nomenclature, he designated as a “normal” 
diphtheria poison one which contained 100 such units in one cubic 
centimeter. (D T 17, M250 = diphtheria toxin normal, Meer- 
schweinchen 250 grams.) 
Together with Ehrlich, Behring then established an antitoxin 
unit (I-E, Immunitats Einheit). They designated as a “normal” 
antitoxic serum one “which contained in one cubic centimeter one 
antitoxic unit” (I-E), and state further, “of this serum 0.1 c. c. 
neutralizes 1 c. c. of the Behring normal toxin.” (Conf. Madsen in 
“Kraus u. Levaditi Handbuch,” II, p. 94.) Alterations were subse- 
quently made in this scale of standards and Ehrlich later desig- 
nated as an antitoxin unit a quantity of an antitoxin which 
completely neutralized 100 lethal doses (for guinea pigs of 
250 grams) of a poison at that time in his possession. This 
definition of the antitoxin unit, however, no longer holds good 
and a comprehension of the true meaning of the unit can only be 
obtained by following in detail the methods by which it is measured. 
Since the methods of antitoxin standardization employed at 124 
INFECTION AND RESISTANCE 
present in the United States were worked out by Rosenau10 along 
the lines of Ehrlich’s method, and the standard is based on the 
one introduced by Ehrlich, the antitoxin unit as employed in this 
country is identical with the one spoken of in the following 
paragraphs. 
In measuring the neutralizing value of antitoxin for toxin, then, 
since both substances are chemically unknown and no purely chem- 
ical indicator of neutralization is available, it was necessary to select 
a susceptible animal by means of which excess of toxin, in mixtures 
of the two. could be detected. As the standard test animal guinea 
pigs of 250 grams were chosen, and improve- 
ments in the methods of measurement were 
introduced, in that the toxin and antitoxin, 
instead of being separately injected as here- 
tofore, were mixed, allowed to stand for 15 
to 30 minutes, and then injected together 
subcutaneously. 
By means of this technique Ehrlich set 
out to examine a large number of toxins and 
their antibodies and obtained results which, 
aside from their practical value, have had 
an important influence upon the development 
of the knowledge of antigen-antibody reac- 
tions. These investigations were consider- 
ably complicated by the fact that neither 
the diphtheria toxin nor the antitoxin is very 
stable and deterioration occurs unless special 
methods of preservation are employed. Since the antitoxin, how- 
ever, is much less unstable than the toxin, the former is employed 
in order to preserve the standard, and is preserved in sealed U tubes 
(see Figure) with anhydrous phosphoric acid. Kept in this way, 
in black, light-proof boxes, and at low temperature, it may be pre- 
served for months without appreciable loss of value and may be 
renewed by accurate comparative measurements from time to time. 
This is carried out for the United States, at the present time, by 
the Government Hygienic Laboratories at Washington under the 
U. S. Public Health Service. 
Preservation of the toxin is much more difficult, and it is in 
connection with the investigation of the instability of the toxin that 
Ehrlich gained his first insight into the nature of the reaction. He 
measured, in a number of toxic filtrates, the minimal lethal dose 
for guinea pigs of 250 grams, establishing a time limit for death 
in order to obtain more accurate comparisons. He designated as the 
new M L D or “T” (that is: toxic unit) the quantity of toxin 
Tube for the Preser- 
vation of the Stand- 
ard Antitoxin. 
Taken from Rosenau, U. 
S. Hygienic Labora- 
tory Bulletin, No. 21, 
1905, p. 53. 
10 Rosenau. TJ. S. P. H. <&' Mar. Ilosp. Service, Hygienic Laboratory 
Bull., 21, April, 1905. TOXIN AND ANTITOXIN 
125 
which will kill a guinea pig of the designated weight in from 4 to 5 
days. He then determined for a number of poisons the exact quan- 
tity just neutralized by one antitoxin unit, calling this quantity L0. 
(L meaning Limes or threshold.) 
It is clear that in judging of complete neutralization of a quan- 
tity of toxin by antitoxin, there may be a strong subjective element, 
since any very slight excess of toxin may cause unimportant local 
reactions such as edema or small hemorrhages, which could escape 
the attention of one observer while being noticed and recorded by 
another. In order therefore to exclude definitely all subjective fea- 
tures from such experimentation, Ehrlich now established another 
toxin value—L+ dose (“Limes death”—now, for convenience, writ- 
ten L+)—which eliminated all possible variations of personal per- 
ception. He designated by this symbol that quantity of toxin which 
not only neutralized one antitoxin unit but included enough toxin, 
in excess of this, to give the result of one free toxin unit, that is, to 
cause death in 4 to 5 days in a guinea pig of 250 grams. Since 
the three values just defined form the basis of Ehrlich’s experiments 
as well as that of all practical diphtheria serum standardizations we 
will briefly restate them for the sake of clearness. 
Thus: 
M L D or “T” = the amount of toxin which, subcutaneously injected, 
causes death in a 250-gram guinea pig in from 4 to 5 days. 
Lo = the amount of toxin which is just neutralized by one antitoxin unit 
so that no trace of reaction, local or otherwise, ensues 
and 
L+ = that amount of toxin which will cause death in 4 to 5 days in a 
guinea pig of 250 grams if injected together with one antitoxin unit. 
It will further clarify the meaning of these terms to examine 
experimental protocols which show how these values are determined. 
Thus in the following: 
I. Injections of toxin 
(1) .005 c. c.—guinea pig lives. 
(2) .009 c. c.—guinea pig dies in 6 days. 
(3) .01 c. c.—guinea pig dies in 4 days. 
(4) .02 c. c.—guinea pig dies in 2 days. 
.01 = M L D or T. 
II. 1 Antitoxin unit + .19 toxin = late paralysis. 
1 Antitoxin unit + .20 toxin = sometimes late paralysis and some- 
times acute death. 
1 Antitoxin unit + .21 toxin = death fourth day. 
1 Antitoxin unit + .22 toxin = death in 2 to 3 days. 
.21 = L+ dose. 126 
INFECTION AND RESISTANCE 
III. 1 Antitoxin unit -f- -14 c. c. toxin — no reaction. 
* 1 Antitoxin unit + .15 c. c. toxin = no reaction. 
1 Antitoxin unit + .16 c. c. toxin = slight congestion about point of 
injection, scarcely visible. 
1 Antitoxin unit + .17 c. c. toxin = apparent reaction at site. 
1 Antitoxin unit + .18 c. c. toxin — edema at site. 
Lo = .16.11 
In determining these values with a large number of toxins Ehr- 
lich discovered the curious fact that, although there was a rapid and 
extensive diminution of toxicity in every toxic filtrate in the course 
of time, there was no corresponding alteration in the L0 amount. 
In other words, although more and more of the toxic broth was 
necessary to kill a guinea pig of standard weight in the required 
time, the amount of the same broth which neutralized one antitoxin 
unit remained approximately the same.12 
In seeking an explanation for this apparent paradox, Ehrlich 
concluded that we must assume that the toxin is complexly con- 
structed, consisting of a toxophore and a haptophore group. Assum- 
ing that chemical union between the toxin and the antitoxin (or, in 
disease, the body cell) takes place, it is by means of the haptophore 
group that such union is brought about. The toxophore group, how- 
ever, is the element by which toxic action is exerted after union 
by the haptophore group has been accomplished. It would be 
conceivable, therefore, that in deteriorating in toxicity the toxin 
might undergo alterations in the toxophore group only, its hapto- 
phore group, and, therefore, its antitoxin neutralizing properties 
remaining intact. Such modified toxins, modified only as to the 
toxophore groups, Ehrlich now refers to as “toxoids.” 
In the production of diphtheria toxins for practical purposes it 
has been found advisable to allow them to “season,” that is to stand 
for prolonged periods until they have reached a state of “equilib- 
rium” at which the conversion of toxin to toxoids has been reduced 
11II and III are taken from the article by Rosenau, P. H. d; M. H. S., 
Hyg. Lab. Bulletin, 21, 1905. 
12 This statement plainly contradicts the definition of a toxin unit; i. e., 
the amount which neutralizes 100 toxin units and often leads to confusion 
among students or others who are unfamiliar with this subject. It should 
be borne in mind that, while the definition of an antitoxin unit is the one 
accepted when Ehrlich first arbitrarily established it, the antitoxin unit, 
as at present in use, is really an amount of antitoxin standardized against 
L+ quantities of toxin, this last value again obtained by measurement 
against the original unit. It represents a neutralization value equal to the 
original one, but may protect the guinea pig against 85, 110, 130, etc. 
(variable), toxin units, according to the constitution of the particular toxic 
filtrate employed in the experiment. Indeed, if, in the following pages, the 
reasoning of Ehrlich is consistently adhered to, our definition of an anti- 
toxin unit should be: The amount of antitoxin which contains 200 binding 
affinities for toxin. This will become clearer as the following paragraphs 
are read. TOXIN AND ANTITOXIN 
127 
to a minimum and the change of relationship between L0 and “T” 
or I L D has practically ceased to go on. From the very begin- 
ning of the growth of the culture in the incubator the process of 
toxoid formation has probably occurred, and even freshly prepared 
toxic filtrates therefore are not pure “toxins,” especially since the 
conversion of toxin to toxoid seems to diminish in velocity as time 
goes on. 
Now in spite of the presence of such alteration products, in com- 
paring the values L0 and L+ of any given toxin preparation, one 
would naturally suppose that L+ minus L0 should be equal to one 
M L D, or the quantity just sufficient to kill a guinea pig of 250 
grams in 4 to 5 days. For we have seen that Lo just neutralizes 
one antitoxin unit while L+ is the quantity which, in addition to 
such neutralizing power, has an excess of toxin equal in action to 
one minimal lethal dose. This, however, is not the case. Let us 
illustrate this by a concrete case. One of Ehrlich’s toxins on meas- 
urement showed a minimal lethal dose or M. L D of 0.0025 c. c. 
The L+ dose of this was .25 
while The Lo dose of this was .125 
The difference was .125 or 50 M L D instead of 1 M L D 
as we would suppose. 
Stated in words, this measurement means that after neutralizing 
completely one antitoxin unit with the toxic filtrates, in order to 
obtain death in a guinea pig in 4 days with such a mixture, it was 
necessary to add, beyond the neutralizing quantity, 50 M L D, or 
again as much as was necessary for neutralization. 
This last relation is merely coincidence, since it might have been 
30 or 40 or 60 M L D just as well. The important point is the fact 
that more than llLD was necessary, and by this fact Ehrlich was 
led to resort to an assumption which forms one of the basic princi- 
ples of many of his explanations for serum phenomena, namely, the 
assumption of differences in combining avidity or affinity. 
As applied to the present problem he reasoned as follows: 
It is conceivable that the toxoids resulting from deterioration of 
toxin might possess three different degrees of affinity for the anti- 
toxin. They might have a stronger, an equal, or a lesser affinity than 
the toxin itself. If their affinity for antitoxin were equal to that of 
toxin they would, of course, not influncee the L+ dose itself ; if 
stronger than toxin their influence would be so exerted that toxin 
would be forced out of combination with antitoxin, giving place to 
the toxoid, and the effect would be the opposite from that experi- 
mentally observed. If, however, their affinity for antitoxin were 
weaker than that of toxin each additional toxin unit added to the 
L0 dose would unite with antitoxin, replacing a corresponding quan- 
tity of the toxoid of weaker affinity. In consequence, as far as the 128 
INFECTION AND RESISTANCE 
poisonous properties of the mixture are concerned, the addition of 
toxin would not render the neutral mixture poisonous for guinea 
pigs until the toxoids had been completely displaced from union with 
antitoxin. Finally, after all the antitoxin had been bound to un- 
changed toxin, further addition would then result in the presence of 
free toxin and poisonous properties would again appear in the mix- 
ture. Ehrlich at first spoke of the toxoids possessing weaker affinity 
for antitoxin than the toxin itself as “epitoxoids.” This conception 
can be rendered clear by the following example: 
In the case cited above we had 
T or M L D = 0.0025 c. c. 
L+ = 0.25 c. c. 
L« = 0.125 c. c. 
The difference — 0.125 = 50 M L D 
Supposing that the toxoids (epitoxoids) present in the mixture 
possessed an affinity for antitoxin less than that of toxin, the follow- 
ing conditions might obtain: 
151 toxin-antitoxin + 49 epitoxoid-antitoxin = Lo. 
Add 1 M L D or T and we have: 
152 toxin-antitoxin + 48 epitoxoid-antitoxin + 1 epitoxoid free. 
Add 2 M L D or T and we have: 
153 toxin-antitoxin + 47 epitoxoid-antitoxin + 2 epitoxoid free until, 
finally, adding 50 T, we get: 
200 toxin-antitoxin + 49 epitoxoid free + 1 toxin free := L+!J 
Later experience led Ehrlich to abandon the opinion that the 
epitoxoids were deterioration products of the toxin. He found that 
the relation between L0 and L+ which we have just outlined, was 
demonstrable in the same way, in freshly prepared toxin filtrates, 
in which there had been little time for toxoid formation. He further 
noticed that, even after deterioration had occurred to a considerable 
extent, and the amount necessary to kill a guinea pig had been much 
13 An example identical in significance with the one just given, but some- 
what simpler in its arithmetical conditions, is here added for the sake of 
permitting no possibility of unclearness. This example is taken from Ehr- 
lich’s own work. 
T = .01 c. e. of the toxin bouillon. 
L+ (neutral, of antitoxin unit yet killing 1 pig) = 2.01 c. c. or 201 T. 
Lo (complete neutral, of 1 antitoxin unit) — 1 c. c. or 100 T. 
Difference = 1.01 c. c. or 101 T. 
100 toxin-antitoxin + 100 epitoxoid antitoxin = L»; 
Add 1 T, and we have: 
101 toxin-antitoxin + 99 epitoxoid-antitoxin + 1 epitoxoid free; 
Add 101 T, and we have: 
200 toxin-antitoxin + 100 epitoxoid free + 1 T free — L+. TOXIN AND ANTITOXIN 
129 
increased (the number of fatal doses in L0 constantly decreasing as 
toxoids replaced., toxin), the L+ nevertheless remained unchanged. 
This, he held, could mean one thing only. The elements present in 
toxic broth which possessed a weaker affinity for antitoxin than the 
toxin itself, namely, the epitoxoids, were present from the very begin- 
ning and were probably separate and primary secretion products of 
the diphtheria bacilli, remaining relatively stable and constant as the 
broth was preserved. In order to avoid confusion, therefore, he now 
referred to the “epitoxoids” as “toxons”—to preclude their confu- 
sion with the other toxoids or true toxin deterioration products. 
These toxons possess, according to Ehrlich, a “haptophore” group 
identical with that of the toxin, but have a different toxophore group. 
For there is reason to believe that the toxon, lacking the power of 
causing acute death, gives rise to slow emaciation and paralysis, 
finally killing after a subacute or chronic course. Thus, in the tab- 
ulation just preceding, we have seen that the toxic broth added to 
neutral mixtures of toxin and antitoxin (containing the L0 dose), 
does not give rise to the acutely toxic effect of one MLD or T until 
an amount has been added which considerably exceeds one toxin 
unit. This, we explained, by Ehrlich’s reasoning, on the supposi- 
tion that “epitoxoids” or “toxons” are displaced from their union 
with antitoxin, giving place to toxin and becoming free. Such toxin- 
antitoxin mixtures—in which the amount of toxin broth ranges be- 
tween the L0 and the L+ doses—therefore, contain no free toxin 
units but contain varying amounts of free toxin. An injection into 
guinea pigs is not followed by acute death in these cases, but leads 
with considerable regularity to emaciation, paralysis, and death’ 
after a long incubation period. 
It has been objected to this, as we shall see, that the slow poison- 
ing produced by such mixtures is due, not to a qualitatively differ- 
ent poison but to the presence of minute quantities of free toxin too 
small to produce acute death, yet sufficient to produce this gradual 
injury. This Dreyer and Madsen 14 tried to disprove by experi- 
ments in which they prepared antitoxin-toxin mixtures so balanced 
that they possessed the toxon effect, and of these mixtures injected 
increasing multiples. In no case did they succeed in producing 
acute death even when the amount injected had been multiplied to 
such an extent that free toxin, if present, must have asserted itself. 
The same workers were able to show that the injection of these 
mixtures, in which free toxons were assumed to be present, produced 
immunity against toxin, thus indicating the similarity of the hapto- 
phore group of toxin and toxon—a conception which will become 
more and more clear as we consider the “Side-Chain Theory” which 
Ehrlich evolved as a result of his toxin analysis. 
Ehrlich had thus elicited facts which seemed to him to indicate 
14 Dj-eyer and Madgen. Zeitschr. f. Hy<j,, Vol, 37, 1901, 130 
INFECTION AND RESISTANCE 
the presence of three qualitatively different substances in toxic fil- 
trates of diphtheria cultures. Two of these, the toxin and the toxon, 
were present, he assumed, in freshly prepared filtrates, as indepen- 
dent primary secretion products of the bacilli, the toxin an acute 
poison, the toxon a substance with slower and qualitatively different 
poisonous effects. Both of them, toxin and toxon, possessing similar 
haptophore groups, could unite with antitoxin and neutralize it, but 
the affinity of toxon for antitoxon was weaker than that of toxin. For 
this reason toxin could displace toxon from its union with antitoxin, 
this accounting for the discrepancy between the L+ and the L0 
doses. The third class of substances, the toxoids, were deterioration 
products of toxin, the deterioration implying an alteration in the 
toxophore group only, the haptophore group remaining the same. 
It is plain from this reasoning that Ehrlich’s conception implies 
complete analogy between chemical reactions in general and the 
neutralization of toxin by antitoxin. Accordingly it is but another 
step in the same direction to speculate concerning the actual rela- 
tions of valency existing between the two substances. It seemed to 
Ehrlich that there were many reasons for assuming that the union 
between toxin and antitoxin occurred in proportions of 200 to 1; 
that is, just as the formula for water is H20, that of toxin-antitoxin 
combinations would be “Toxin200Antitoxin.” 
The considerations on which this opinion was based were as fol- 
lows: In examining a large series of toxic filtrates, Ehrlich,15 as 
well as Madsen, had found that the number of toxin units (“T” or 
M L D) necessary to neutralize one antitoxin unit (that is, the 
number of toxin units contained in the L0 dose) corresponded, 
with great regularity, to multiples of 100. Values of 25, 33, 50, 
100, etc., recurred again and again. This indicated that the de- 
terioration of the toxin into toxoids followed a certain regularity of 
progression and seemed to justify the assumption that the absolute 
binding power possessed by antitoxin was represented by a valency 
corresponding to a multiple of 100. Since the number of toxin units 
contained in an L0 dose rarely fell below, and usually above 100, the 
valency could not be less than 100. On the other hand, repeated 
measurements of L0 and L+ doses never showed as many as 200 
toxin units. Ehrlich’s own highest value was 133, and the highest 
ever obtained by any one was a measurement by Madsen of 160. 
How considering the fact that no toxin is “pure” but that, in every 
case, it contains admixtures of toxoid and toxon, the values 133 or 
160 cannot represent all the valencies of an antitoxin unit. They 
represent only that part of the antitoxin unit which is neutralized 
by the “toxin,” as measurable upon guinea pigs, a certain fraction 
of antitoxin being united to toxoid or toxon. It is likely, therefore, 
as Ehrlich reasoned, that, being higher than 100, and in an ob- 
15 Ehrlich. Deutsche med. Woch., No. 38, 1898, Yol. 24. TOXIN AND ANTITOXIN 
131 
viously impure condition approaching but never reaching 200, the 
valency of antitoxin for toxin was just 200. The correctness of this 
surmise seemed rendered more probable by Ehrlich’s further studies, 
since analysis of the ingredients of various toxic filtrates, that is, the 
determination of the relative contents of toxin, toxoids, and toxon, 
appeared to show constantly definite relations to the valency 200. 
The method by which Ehrlich carried out these subsequent stud- 
ies is spoken of as the method of “Partial Absorption.” In prin- 
ciple it represents a reversal of his earlier methods of measurement. 
In these he had titrated various amounts of toxin broth against a 
constant quantity (one unit) of antitoxin, establishing the L+ and 
L0 values. In the method of Partial Absorption, on the other hand, 
he measured varying fractions of an antitoxin unit against a con- 
stant amount of toxin, employing for this a previously determined 
L+ and L0 dose. A measurement carried out in this way is shown 
in the following tabulation in which, at the same time, there is indi- 
cated how many valencies each antitoxin fraction represents, on the 
basis of an assumed total of 200 for each unit.16 
0 antix. unit representing 0 valency + L+ = 85 free T units 
.1 antix. unit representing 20 valencies + L+ = 85 free T units 
.25 antix. unit representing 50 valencies + L+ = 60 free T units 
.8 antix. unit representing 160 valencies + L+ = 10 free T units 
.9 antix. unit representing 180 valencies + L+ = 3.5 free T units 
1.0 antix. unit representing 200 valencies + L+ = 1 free T unit 
It is immediately evident in this table, as it would be evident in 
any other citation of similar measurements, that there is no regu- 
larity in the progress of neutralization; or, in other words, that addi- 
tion of a definite fraction of antitoxin does not remove the arithmet- 
ically corresponding amount of toxic properties from the L+ dose. 
The first 0.1 unit of antitoxin in this table has removed no free 
toxin whatever. The addition of the next 0.15 of an antitoxin unit, 
representing 30 valencies, has removed 25/s5 or 5/n of the total 
toxicity. Throughout the scale there is not the regular progression 
of neutralization, multiple by multiple, which would be expected if 
antitoxin could be titrated against a pure toxin. This, according to 
Ehrlich, is due to the presence of various toxoids which possess vary- 
ing affinities for the antitoxin molecule. The first 0.1 of a unit 
added, in this case, does not diminish the toxicity of the mixture 
because it is bound by “protoxoids” which possess a higher affinity 
for antitoxin than the toxin itself. Next are bound the toxins them- 
selves together with varying amounts of “syntoxoids” which possess 
the same affinity as toxin. Finally there are left the toxons which 
18 This measurement is taken from one cited by Ehrlich in Deutsche med. 
Woch., No. 38, 1898, Yol. 24, and is taken literally except for the first value 
of 1/10 antitoxin unit, which is inserted to illustrate the formation of pro- 
toxoids. 132 
INFECTION AND RESISTANCE 
possess a lesser affinity than toxins or toxoids, and therefore again 
have the discrepancy between the L0 and L+ dose. Ehrlich utilizes 
this method in order to determine the composition of the constituents 
of any given toxic filtrate and expresses the results in the so-called 
“toxin spectra.” 
The construction of these spectra and the principles underlying 
the measurements on which they are based are very clearly illus- 
trated by Madsen,17 from whose article the following type spectra 
are taken:18 
Toxin Spectrum After Madsen, Ann. de VInst. Past., Yol. 13, 1899, p. 57. 
This figure represents the diphtheria filtrate composed of 50 
valencies of protoxoid, 100 toxin and 50 toxon equivalents. The 
measurements in this case may be represented by the following tabu- 
lation : 
Lo + 1 antitox. unit = 200 valencies = 0 lethal dose 
Lo + .75 antitox. unit = 150 valencies — 0 lethal dose 
Lo + .25 antitox. unit = 50 valencies = 100 lethal doses 
Lo + 0 antitox. unit = 0 valency = 100 lethal doses 
The following diagram, also from Madsen, represents the same 
poison after it had deteriorated to % its toxic power. L0, therefore 
would contain only 50 toxic doses. 
After Madsen, Ibid., p. 578. 
The measurements corresponding to this table are as follows: 
Lo + 1. antitox. unit = 200 valencies — 0 lethal dose 
Lo + .75 antitox. unit — 150 valencies — 0 lethal dose 
Lo + .745 antitox. unit = 149 valencies — 0 lethal dose 
Lo + .74 antitox. unit — 148 valencies — 1 lethal dose 
etc., until 
Lo + .25 antitox. unit = 50 valencies — 50 lethal doses 
Lo + 0 antitox. unit — 0 valency = 50 lethal doses 
17 Madsen. Ann. Past., Yol. 13, 1899, p. 576. 
18 We have chosen to illustrate the principles of the toxin spectrum from 
the article of Madsen rather than from Ehrlich’s original work,, because the 
former presents this difficult phase of the subject in a hypothetical toxin of 
extremely simple structure. Some of Ehrlich’s spectra constructed from 
actual measurements may be found in Deutsche med. Woch., No. 38, 1898. TOXIN AND ANTITOXIN 
133 
The following spectrum, the third in Madsen’s article, represents 
the same toxin described in the preceding spectrum but at a later 
period, at which considerable further deterioration had taken place. 
The L0 dose now contained but 30 M L 1) or, in other words, the 
amount of toxin contained in the L0 dose was sufficient to kill 30 
guinea pigs only. 
After Madsen, Ibid., p. 579. 
Madsen’s description of the method in which this spectrum is 
constructed is the following: 
Lo + 20%oo of an antitoxin unit kills 0 guinea pig 
L# + 15%oo of an antitoxin unit kills 0 guinea pig 
Lo + 141/&oo of an antitoxin unit kills 0 guinea pig 
Lo + 14%oo of an antitoxin unit kills 1 guinea pig 
Lo + 13%oo of an antitoxin unit kills 2 guinea pigs 
Lo + 12%oo of an antitoxin unit kills 3 guinea pigs 
Lo + 10%oo of an antitoxin unit kills 5 guinea pigs 
Lo + °%oo of an antitoxin unit kills 5 guinea pigs 
Lo + 9%oo of an antitoxin unit kills 6 guinea pigs 
Lo 4* 9%oo of an antitoxin unit kills 6 guinea pigs 
Lo + 9%oo of an antitoxin unit kills 7 guinea pigs 
Lo -j- 9%oo of an antitoxin unit kills 10 guinea pigs 
Lo + 5%00 of an antitoxin unit kills 30 guinea pigs 
Lo + bioo of an antitoxin unit kills 30 guinea pigs 
The amount of toxon has remained the same in spite of deteriora- 
tion. As less and less antitoxin is added, between the values of 1,,%oo 
and 10%oo of an antitoxin unit, there are now liberated only 5 fatal 
doses of the toxin. It is in this zone that deterioration has taken 
place, since, in the preceding spectrum, the difference between the 
addition of 15%oo and 10%oo of an antitoxin unit represented 25 
fatal doses for guinea pigs. When in this last spectrum the amount 
of antitoxin is gradually reduced from 100 valencies to 50 valencies 
25 fatal doses are liberated, a quantity corresponding to the similar 
zone in the preceding spectrum. Thus in this particular zone of the 
spectrum no change has taken place. I lie same is true of the pro- 
toxoid zone. 
It is unnecessary to cite a larger number of such measurements 
in this place, since the ones given sufficiently illustrate the methods 
and the conclusions drawn from them. As a result of such experi- 
ments Ehrlich concludes: 
I. That the diphtheria bacillus produces primarily two kinds 
of substances (a) toxin, (b) toxon, both of which bind the antibody. 134 
INFECTION AND RESISTANCE 
II. The toxins (and perhaps also the toxons) may deteriorate 
and be modified into secondary substances (toxoids) which may be 
distinguished by their different degrees of affinity for antitoxin. 
III. This classification does not exhaust all possible complica- 
tions, since each subdivision of toxin consists apparently of equal 
parts of two different modifications which are similar to each other 
in their relation to antitoxin but differ in varying resistance to in- 
fluences of deterioration. A more complete analysis of these condi- 
tions may be found, together with a series of illustrative spectra, in 
Ehrlich’s article in the Deutsche meet. Wochenschr., Sept., 1898, 
which has been quoted above. 
The complex deductions arrived at by Ehrlich are largely de- 
pendent, as we have seen, upon strict adherence to the analogy be- 
tween the toxin-antitoxin reactions and those occurring between 
strong acids and strong bases. In such cases there is a complete 
reaction, in which chemical change ceases only when there has been 
a complete neutralization of one by the other. If, for instance, we 
mix molecular equivalent amounts of H2S04 and NaOH, an ap- 
parently complete change into Na2SO and H20 occurs: 
H2SO4 + 2 NaOH = NasSCU + 2 ILO 
The reverse process does not seem to take place, and if traces 
of uncombined H2S04 and NaOH are present, as may be theoret- 
ically assumed, they are so slight in amount that they are not de- 
monstrable. There are, however, many chemical reactions in which 
the process is not a complete one, in that the chemical change does 
not proceed until the reagents are completely used up. Reaction in 
these cases ceases when an equilibrium is reached at which there are 
present definite amounts of the reaction products and of the original 
substances at the same time.19 
This occurs when a weak acid is added to a weak base. In such 
cases the reaction is incomplete and reversible and, together with the 
neutralization products, both free acid and free base may be present. 
The conditions are best explained by citing an example of a reversi- 
ble reaction which is commonly given in text-books of physical chem- 
istry, namely, the reaction between ethyl-alcohol and acetic acid. 
(Our citation is taken from Philip’s “Physical Chemistry,” London, 
Arnold, 1910) : “When one gram mol. of ethyl alcohol is added to 
one gram mol. of acetic acid, a reaction takes place which results in 
the formation of ethyl acetate and water; the reaction, however, is 
incomplete and stops at an equilibrium point at which the reaction 
mixture contains % gram mol. alcohol, % gram mol. acid, % gram 
mol. ethyl acetate, and % gram mol. water. If, on the other hand, 
19 See Cohn. “Vortrage f. Artze iiber Physik. Chem.,” Engelman, Leip- 
zig, 1901. TOXIN AND ANTITOXIN 
135 
1 gram mol. of ethyl acetate is mixed with 1 gram mol. of water, a 
reaction sets in which results in the formation of ethyl alcohol and 
acetic acid. This change likewise stops in equilibrium at a point at 
which the composition of the reaction mixture is the same as that 
already stated.” The reaction is thus reversible and may be written: 
C*H.OH + CHsCOOH ?=* CttCOOCiH. + ILO 
Another example somewhat simpler and more easily brought into 
analogy with the toxin-antitoxin reaction is that of the dissociation 
of phosphorus pentachlorid into phosphorus trichlorid and chlorin 
(see Alexander Smith, “General Chemistry,” Century Company, 
1ST. Y., 1911, p. 181). 
Here the reaction takes place: 
PCh PCb + Cl* 
Chemical equilibrium is reached when the reaction speed is the same 
in both directions, and there will be present, at equilibrium, PC13, 
Cl2, and PC15. Now the “Law of Mass Action” (Guldberg & 
Waage) states that reaction goes on at a velocity proportionate to 
the concentration of the reacting molecules. It is plain, therefore, 
that at the point at which the reaction takes place with equal veloci- 
ties in both directions, that is, at the equilibrium point, a very defi- 
nite relation of molecular concentrations must obtain, and this rela- 
tion can be expressed as a formula. For the example given above 
this may be written as follows: 
Cone. PCb X Cone. Cl „ . , .. 
“ K (constant) 
This formula is expressed in words by Alexander Smith as follows: 
“If we change the amount of the pentachlorid placed in the vessel, 
or if we use amounts of chlorin and trichlorid which are not equiv- 
alent, the numerical value at equilibrium of each concentration will, 
of course, be different, but the product of the concentrations of tri- 
chlorid and chlorin, divided by the concentration of the pentachlorid, 
will always give the same numerical value for (K), the constant, at 
the same temperature.” 
Now to return to the application of these facts to the neutraliza- 
tion of toxin by antitoxin, if the reaction is one analogous to that of 
a strong acid and alkali, as cited above in the case of H2S04 and 
NaOH, we would expect a complete neutralization of one by the 
other, multiple for multiple, and the explanation of Ehrlich based 
on the assumption of different toxin constituents, of varying affini- 
ties, and different pharmacological effects, is the only one which will 
account for the experimental results. Assuming, however, that the 136 
INFECTION AND RESISTANCE 
reaction is one analogous to that taking place between a weak acid 
and a weak base—such as boric acid and ammonia—we have an en- 
tirely different state of affairs. For here the reaction goes on to a 
point of equilibrium, and in mixtures containing equivalent amounts 
of the weak acid and the base there will be present the reaction prod- 
ucts and also small amounts of unbound free acid and free base. 
And according to the law of “Mass Action,” the quantities of free 
acid and base present will depend entirely on the masses of the 
reagents put together. Thus, for each particular mixture of the two, 
different quantities of the original substances will be found uncom- 
Curve Representing the Neutralization of Tetanolysin by Different 
Quantities of Antitoxin. 
Taken from Arrhenius, “Immunochemistry,” Macmillan, 1907, p. 175. 
bined, and, furthermore, the gradual addition of one to the other 
will not have a neutralizing value in exact proportion to the amount 
added. Were the toxin-antitoxin reaction analogous to such chemical 
systems, then we could assume that every mixture of the two sub- 
stances, whatever the relative amounts, would contain not only the 
united toxin-antitoxin molecule, but also varying amounts of disso- 
ciated free toxin and free antitoxin, the quantities of each depending, 
according to the law of mass action, upon the molecular concentra- 
tions obtaining in the individual mixture. This, indeed, is the 
conception of toxin-antitoxin union formulated by Arrhenius and 
Madsen. 
Arrhenius and Madsen,20 bearing in mind these conditions, 
made comparative studies of the neutralization of tetanolysin by its 
20 Arrhenius and Madsen. Zeitschr. f. physik. Chem., 44, 1903, and 
Festschrift Kopenhagen Serum Instit., 1902. Arrhenius. “Immunochem- 
istry,” Macmillan, N. Y., 1907. TOXIN AND ANTITOXIN 
137 
antilysin on the one hand, and that of ammonia by boric acid on the 
other. Ammonia, like most bases, is a hemolytic agent, while boric 
acid, unlike stronger acids, has no hemolyzing properties. For this 
reason, in mixtures of the two, the toxicity is proportional to the 
concentration of free ammonia (though, as Arrhenius states, “a cor- 
rection must be made for the lowering action of the ammonium salt, 
as indicated by experiments on this action”). Because the reaction 
between boric acid and ammonia is reversible, that is, the salt formed 
is dissociated by the hydrolytic effect of the water, there is always 
present a slight amount of free ammonia even if the largest possible 
quantities (to saturation) of boric acid are added. (See Arrhenius, 
“Immunochem.,” p. 174.) The curve of toxicity indeed descends 
as more boric acid is added, but never reaches 0. 
By a modification of the formula expressing the law of Mass 
Action, Arrhenius and Madsen could calculate the amount of free 
ammonia present in a series of mixtures in which increasing quanti- 
ties of boric acid were added to a constant quantity of ammonia, and 
the values so obtained corresponded with much accuracy to those re- 
sulting from measurements of toxicity upon red blood cells. The fol- 
lowing table taken from Arrhenius and Madsen illustrates this: 
Toxicity (q) of 0.1 N. NIL (1 Equivalent) within Equivalents of Boric 
Acid. (Taken from Arrhenius, loc. cit. p. 176.) 
Equivalents of boric 
acid added 
Quantity of free 
ammonia—i. e., 
toxicity—observed 
q = Ammonia 
-toxicity calculated 
from formula 
Aq obs. 
0 
100 
(100) 
0.17 
85 
79 
15 
0.33 
69 
64 
16 
0.67 
43 
42 
26: 2 = 13 
1 
25 
27 
18:2= 9 
1.33 
20 
18 
5:2 1 
1.67 
13 
13 
7:2 U.5 
2 
10 
10 
3:2 J 
Here, in the last column, there is indicated the proportion of 
toxicity which is neutralized by the successive addition of Yo of an 
equivalent of boric acid. The first additions lower it to a degree 
proportionate to the amount of acid added; the next additions neu- 
tralize it to a much slighter degree, and, as further additions are 
made, each successive one possesses progressively less relative neu- 
tralizing power than the preceding. 
This, it is plain, is closely analogous to the phenomena observed 
by Ehrlich in his “Partial Absorption” method, and Arrhenius com 
eludes that the two phenomena, toxin-antitoxin and boric acid-am- 
monia neutralization, are closely analogous. His point of view is 138 
INFECTION AND RESISTANCE 
further strengthened by his experiments with tetanolysin and its 
specific antibody, in which he constructed a curve similar to that 
given for boric acid, derived a formula and found that the observed 
and the calculated values closely coincided for various mixtures of 
the two. He claims, in consequence, that the phenomena observed by 
Ehrlich should not be interpreted as due to “partial toxins”—toxoids 
or toxons, but dependent rather upon the presence of varying quan- 
tities of free toxin dissociated from union with antitoxin because of 
the reversibility of the union. 
The opinions of Arrhenius and Madsen are not generally ac- 
cepted. It is in the first place doubtful whether substances like toxin 
and antitoxin, which, as far as we know their chemical nature at all, 
belong to the class of substances spoken of as colloids, can be re- 
garded as subject to the laws of Mass Action in their reactions. 
Nernst21 has criticized Arrhenius’ deductions chiefly on the 
basis of their assumption of the reversibility of the union of toxin 
and antitoxin, since reversible reactions between colloids, though not 
at all inconceivable, have so far not been definitely shown. Further- 
more, as Nernst states, if complete reversibility of such reactions 
were possible it would be hard to understand how antitoxin can pro- 
tect the animal against the actions of toxin. 
These objections to the dissociation ideas of Arrhenius and Mad- 
sen do not, in our opinion, hold good any longer. Recent work 
upon the immunization of animals with partially and completely 
neutralized toxin-antitoxin mixtures, the work upon human im- 
munization in diphtheria with similar preparations, and the increas- 
ing knowledge we are obtaining about the manner in which proteins 
may unite with other substances at different hydrogen ion concen- 
trations, seems to prove that in regard to dissociation, proteins may 
follow laws not fundamentally different from those prevailing with 
other substances. Therefore, to refute the opinions of Arrhenius 
and Madsen entirely on the objection that dissociation does not 
take place with substances like toxin and antitoxin, is neither jus- 
tified nor sound. This matter is referred to again below in connec- 
tion with the analogy to enzyme reactions. 
Another point of view concerning the toxin-antitoxin union 
which has been gaining ground especially through the work of 
Landsteiner and his pupils, is that of Bordet.22 Bordet expresses his 
views in the following way: 
I. When one mixes with a certain quantity of toxin an amount 
of antitoxin which is insufficient to produce a complete neutraliza- 
tion, the molecules of antitoxin are not taken up by a definite frac- 
tion of the toxin molecules, satisfying the affinities of these entirely 
21 Cited from Landsteiner in “Kolle u. Wassermann Handbuch,” 2d Ed., 
Yol. 5. 
22 Bordet. Ann. de I’Inst. Past., Yol. 17, 1903. TOXIN AND ANTITOXIN 
139 
while other units remain intact; on the contrary, the antitoxin mole- 
cules distribute themselves equally upon all the toxin molecules pres- 
ent, and these are therefore, all of them, partially saturated, and 
lose proportionately a part of their initial toxicity. One could say 
that there is an attenuation of the toxin since there is a formation of 
a less poisonous complex. 
II. The symptoms of poisoning produced by such a complex in- 
jected into animals or placed in contact with sensitive cells cannot 
be identical with those which would be produced by a fully saturated 
mixture of toxin and antitoxin, or by intact toxin. 
III. Between these two extremes, free toxin and entirely neu- 
tralized toxin, one can imagine many transitions, progressive stages 
of attenuation. Every time that one mixes toxin and antitoxin in 
the same way one attains the same degree of attenuation. 
Briefly put, this means that Bordet estimates toxin-antitoxin 
combinations of different degrees of toxicity as representing differ- 
ent stages in the completeness of the saturation of the individual 
toxin units. When 10 parts of toxin are added to 1 part of anti- 
toxin, the result, according to him, would not be such that 1 part is 
neutralized by 1 part of antitoxin, leaving 9 parts of toxin free. He 
assumes rather that each unit of toxin is attenuated by the absorp- 
tion of 0.1 of a part of antitoxin. He compares this process to the 
action of iodin upon starch. Starch can absorb variable quantities 
of iodin and, according to the amount taken up, is colored slightly 
or deeply blue. This mode of action is common to most staining 
processes. The substance that is stained fixes varying quantities of 
coloring matter and the coloring matter does not limit itself to a 
definite fraction of the substance stained but distributes itself 
equally to the material, coloring it slightly or deeply, in its entirety, 
according to the relative amount of color added. We will see later 
that there are many reasons for regarding other antigen-antibody 
combinations as following similar laws of proportion. 
Bordet and others speak of this point of view as the “Adsorption 
Theory,” and Biltz, in studying this point of view by physical 
methods, comes to the conclusion that the observed figures of the 
quantitative relations between toxin and antitoxin in the process of 
neutralization are fairly consistent with the values to be expected if 
the process were actually an absorption phenomenon. 
Bordet 23 supports this point of view not only by test tube ex- 
periments, but also by general observations in animals. He cites, for 
instance, the observation that it takes more tetanus toxin to kill a 
guinea pig than a mouse, but if one takes into consideration the 
difference in weight between the two animals, it is found that the 
guinea pig is in reality more sensitive. How, if a mixture of toxin 
and antitoxin is made of such strength that a mouse can still sur- 
23 Bordet, “Traite de l’lmmunite,” Masson, Paris, 1920, p. 530. 140 
INFECTION AND RESISTANCE 
vive it, the same mixture may kill a guinea pig. Such an effect 
cannot be explained, as Bordet rightly argues, by traces of free toxin, 
but might well be explained by the fact that the toxin was suffi- 
ciently neutralized in its individual units to have no effect on the 
mouse, but not sufficiently so to permit poisoning of the guinea pig. 
According to Bordet, then, the quantitative factor, as well as the 
qualitative, that is, the degree of neutralization in the condition of 
each particular toxin unit must be considered in all experiments 
in which toxin and antitoxin are combined. 
Bordet’s conception of antigen-antibody union in this case can 
be brought out by other antibody reactions rather more easily than 
with toxin-antitoxin mixtures, and it is relatively easy to prove 
that cellular antigens, such as red blood cells and bacteria, can be 
sensitized with varying degrees of intensity, unit by unit, according 
to the concentration of the antibody solution with which they are 
brought in contact. Moreover, a very striking example of this was 
recently demonstrated in our laboratory in experiments on pneu- 
mococcus agglutination by J. T. Parker. When pneumococcus sus- 
pensions were brought in contact with antiserum and immediately 
shaken for a short time, agglutination occurred promptly. When 
the mixture was left unshaken, agglutination proceeded much more 
slowly. When shaking was delayed for three or four minutes after 
the first contact, the hastening effect of the shaking was no longer 
apparent. These observations seemed to us most easily explained as 
due to the immediate distribution of the agglutinating antibodies to 
the entire mass of the suspension in the cases where shaking was 
practiced. 
One of the most important arguments in favor of Bordet’s ideas 
as against the conception of Arrhenius and Madsen, and the estab- 
lishment of a definite equilibrium, are the various experiments in 
which the reversibility of antigen-antibody reactions has been shown' 
to be a very slight one only, and that, although an equilibrium is 
reached between the two, it is not affected by dilution and time 
factors in the same way in which the ordinary mass action phe- 
nomena take place. For instance, Madsen, while he has noticed that 
dilution can increase the toxicity of mixtures of toxin and antitoxin, 
has shown that this is only true of freshly mixed, but not of old 
preparations. Similarly, Madsen and Walbum have shown that 
diphtheria toxin and antitoxin diffuse into gelatin with different 
speeds; but that the difference in diffusion of mixtures is only no- 
ticed when the experiment is done with fresh toxin-antitoxin prepa- 
rations. If the mixtures have stood for some time, the neutralized 
solutions no longer dissociate. 
Again, the long train of analogies between antigen and antibody 
reactions in the case of precipitins, for instance, the necessity for 
proportionality of reacting substances, zone reactions, etc., which are TOXIN AND ANTITOXIN 
141 
discussed in another part of the hook (see chapter on Precipitins) 
supports the idea that these unions take place more like adsorption 
phenomena than like the chemical reaction between inorganic sub- 
stances. While all this is true, we must nevertheless, remember that 
in view of the more recent advances in our knowledge of colloidal 
reactions, such as those being made by Loeb at the present time, may 
in the end show that, fundamentally, these reactions are not as dif- 
ferent from each other as they now appear.24 
A curious occurrence which seems to bring the toxin-antitoxin 
reactions close to colloidal reactions in general is that which is 
known as the “Danysz 25 Effect” or as the “Bordet 26-Danysz Phe- 
nomenon.” Danysz discovered that when ricin or diphtheria toxin 
were brought into contact with their homologous antibodies the de- 
gree of neutralization depended upon the manner in which the two 
were put together. When the toxin was added to the antitoxin in 
two fractions, a considerable time being allowed to elapse between 
the additions, the final mixture was much more toxic than when the 
total amount was added at once. In other words, although both 
mixtures contained exactly the same quantities of the two reacting 
substances, nevertheless the amount of toxin left free varied in the 
two cases, according to the speed with which they had been put to- 
gether. This was confirmed in 1904 by von Dungern 27 for diph- 
theria toxin, and Craw 28 was able to observe it in the case of mega- 
theriolysin and its antilysins. 
A good deal of light may be hoped for in regard to the nature of 
antigen-antibody unions from investigations upon the nature of 
enzyme reactions. It was largely due to the researches of Duclaux,29 
Bayliss,30 and their followers that the combination of an enzyme 
with a substrate was looked upon as taking place by adsorption. 
24 For a more complete discussion of this problem see Bordet, loc. cit. 
25 Danysz. Ann. de I’Inst. Past., Vol. 16, 1902. 
26 Bordet. Ann. de VInst. Past., Vol. 17, 1903. 
27 Von Dungern interpreted this in the sense of Ehrlich, by assuming 
it to be due to what he calls “epitoxonoids.” This epitoxonoid he assumes 
to be a constituent of toxic broth, which has still less affinity for antitoxin 
than the toxon. It can combine with diphtheria antitoxin, but not until all 
the true toxin is bound. However, when it is once united with diphtheria 
antitoxin it is not very easily displaced from the union, especially when 
a considerable time has elapsed since the union. Therefore, he thinks, when 
the toxin is added to the antitoxin in two fractions, this epitoxonoid is 
bound and keeps the toxin, which is added later, out of combination. 
Whereas if the toxic broth is added as a whole, it is the epitoxonoid which 
is left unbound. This explanation of von Dungern’s may be looked upon 
as an ingenious refinement of the reasoning introduced by Ehrlich into this 
field. Von Dungern. Deutsche med. Woch., 30, 1904. 
28 Craw. Jour. Hyg., Vol. 7, 1907. 
29 Duelaus. “Chemic Biol.,” Paris, 1883. 
80 Bayliss. Proc. Royal Soc. London, Series B, 84, 1911, 90. Bayliss. 
“The Nature of Enzyme Action, Monographs of Biochemistry,” London, 1914. 142 
INFECTION AND RESISTANCE 
Northrop 31 has recently studied the reaction of trypsin and pepsin 
with its substrate in detail, and has, in our opinion, thrown con- 
siderable light upon the processes. In his pepsin studies he found 
that the apparent divergence of pepsin action from the results pre- 
dicted from the law of mass action, can be quantitatively explained 
by the fact that the enzyme in solution combines with some of the 
substances32 produced by the digestion, .and maintains an equi- 
librium with this substance which obeys the ordinary laws of mass 
action. He bases his assumption that the pepsin combines with 
such substances as peptone (using the word “peptone” broadly, as 
signifying substances in solution with which the pepsin combines but 
does not hydrolyze) upon the experiments of Pekelharing and 
Ringer 33 who found that while pure pepsin solutions showed no 
iso-electric point when tested between two electrodes, the direction 
of migration of the pepsin was changed upon the addition of pep- 
tone at a Ph of 3, which is the iso-electric point of the added pep- 
tone. By such a conception Northrop explains the experimental 
fact that the rate of digestion of proteins by pepsin is not pro- 
portional to the total concentration of the pepsin. As the pepsin 
digests the protein, peptone-like substances are formed with which 
pepsin goes into combination and with which it then maintains an 
equilibrium following mass action laws. Since it is only the un- 
combined or dissociated pepsin which affects the further hydrolysis, 
the curves experimentally obtained may be readily explained. In 
adding increasing amounts of peptone to pepsin solution, as a result 
of this manner of combination, the first amounts added inactivate 
more pepsin than the later additions. We need hardly call attention 
to the fact that, as Northrop points out, this is in very striking 
analogy to the manner in which antitoxin and toxin react in 
Ehrlich’s experiments. 
In similar studies on trypsin action, Northrop confirms his pep- 
sin studies in general. Here the process in its relative quantitative 
relations goes on as follows: At the beginning of the experiment, the 
trypsin solution contains undigested protein, and as the trypsin acts 
upon this, trypsin-inhibiting substances are formed. These will nat- 
urally increase during the experiment, and the observed quantitative 
results will consist in, first, a reversible inactivation of the trypsin 
due to the presence of the inhibiting substance, and, secondly, an 
irreversible destruction of free trypsin. The curve of trypsin activ- 
ity resulting from this is one in which there is, at first, a very rapid 
drop of activity, which gradually becomes slower. In studying the 
effect of inhibiting solutions on the rate of digestion by trypsin, 
31 Northrop, J. H. Jour. Gen. Physiol., 2, 1920, 471, and 4, 1922, 227, 
245, 261. 
32 These substances are not, as once supposed, the amino-acids. 
33 Pekelharing and Ringer. Zeit. f. Phys. Chem., 75, 1911, 282. TOXIN AND ANTITOXIN 
143 
Northrop made another very interesting observation which is in very 
close analogy to the Dansyz phenomenon. When the inhibiting so- 
lution and trypsin were mixed, the act of mixing and the time of 
standing had no effect at ordinary temperatures. If the experiment, 
however, was done at 38°, the results were different. The solution 
in which the inhibitor (antitoxin) was added at the beginning, was 
much more active at the end of the experiment than one in which 
the inhibitor was added in fractions. In commenting upon this, 
Northrop states that the main discrepancy in the analogy is the fact 
that in the trypsin experiment the result is more marked if a rela- 
tively small quantity of trypsin is added in the beginning of the 
experiment, whereas, in the Dansyz it is necessary to add an excess. 
This he believes is due to the fact that in the trypsin experiment 
(toxin) the trypsin is changed during the experiment, while in the 
Dansyz phenomenon it is probably the free antitoxin (inhibitor) 
which is altered. 
These experiments of Northrop do not, of course, solve the 
question of antigen-antibody unions, but they do serve to bring the 
analogy of toxin-antitoxin relations much closer to laws governing 
the union of enzyme wdth substrate. Moreover, they show that 
enzyme is actually used up in its reactions, just as toxin is used 
up in its reactions with antitoxin, and that equilibrium following 
the laws of mass action may be a definite factor in the quantitative 
relations governing the reactions. It is not at all impossible that 
the general laws which govern reactions between trypsin and its 
inhibiting substances are similar to those which govern the toxin- 
antitoxin reaction. Moreover, while it is a dangerous subject upon 
which to theorize, it is yet not utterly impossible that the toxins are 
closely analogous to enzymes, and that they produce in their action 
upon cells, products of injury which, passing into the circulation, 
become the specific inhibitors of the toxin or the antitoxin. We 
do not wish to dignify this with the label of a theory, but in sub- 
jects as vague as the origin and biological meaning of antitoxins, 
we must grasp at every straw that may suggest experimental paths 
for enlightenment. 
THE SIDE-CHAIN THEORY 
Mechanism of Antibody Formation 
The discovery of antitoxins in the blood serum of toxin-immune 
animals by Behring and his collaborators furnished a point of new 
departure for the investigation of the phenomena of immunity, and 
Ehrlich’s work upon the nature of the reaction between toxin and 
antitoxin, both in vitro and in the animal body, firmly established 
that the protective effect of the latter was one of direct neutral iza- 144 
INFECTION AND RESISTANCE 
tion, and not, as at first supposed, one of toxin destruction or of 
indirect influence through the mediation of the body cell. As we 
have seen, moreover, it was quickly noted that these reactions were 
strictly specific in that an antitoxin produced with any one of the 
known toxins reacted solely with this one to the exclusion of all 
others. All these facts were of the utmost practical importance and 
gave hope of ultimate extensive therapeutic application, a hope 
which has, in part, been realized. 
The physiological mechanism by which these phenomena were 
brought about, however, was, and is, to a great extent still, a mys- 
tery, and a most extensive and painstaking series of researches has 
occupied itself with its elucidation. 
When we consider the invariable production of a specific anti- 
toxin in response to the treatment of an animal with a toxin it is 
but natural that Buchner and others should have at first assumed 
that the antitoxin is, in each case, a product obtained by the action 
of the body tissues from the toxin itself. While difficult to refute 
at a time when little was known of the laws of antitoxin production 
and of quantitative relationships, such an assumption is entirely un- 
tenable in the light of more recent knowledge. We now know that 
such a simple conversion of toxin into antitoxin cannot explain the 
phenomenon because the amount of antitoxin incited in the immu- 
nized animal is out of all proportion great in comparison with the 
amount of toxin injected. Thus Knorr 34 has found that 100,000 
units of antitoxin may be produced by the injection of the toxin 
equivalent of one unit. Moreover the discovery by Salomonsen and 
Madsen35 that pilocarpin injections will increase the amount of 
antitoxin produced by an animal distinctly pointed to the likelihood 
of the participation of the general physiological activities of an 
immunized subject in the production of antibodies. Unquestionable 
proof of this was also brought by the experiments of Roux and Vail- 
lard,36 in which antitoxin production in immunized animals con- 
tinued even after the entire volume of blood had been removed by 
repeated bleeding. This observation points distinctly to the direct 
secretion of antibodies by the tissue cell, in the nature of what has 
been termed by Koux37 an “internal secretion.” And it is this 
activity of the body cell in the production of antibodies which forms 
the fundamental premise from which the now classical “Side-Chain 
Theory” of Ehrlich takes its departure. 
In order to approach this theory logically it will be of advantage 
to consider briefly the general subject of the assimilation of food- 
stuffs and other substances distributed by the circulation to the cells 
34 Knorr. Munch, med. Woch., 1898, pp. 321, 302. 
35 Salomonsen and Madsen. Ann. de I’Inst. Past., Yol. 12, 1898. 
36 Roux and Vaillard. Ann. de VInst. Past., Yol. 7, 1893. 
37 Roux. Ref. in Semaine Medicale, 1899. TOXIN AND ANTITOXIN 
145 
of the animal body. Eor, as Ehrlich has expressed it, “The Reac- 
tions of Immunity, after all, represent only a repetition of the 
processes of normal metabolism, and their apparently wonderful 
adjustment to new conditions is only another phase of ‘Uralter 
Protoplasma Weisheit.’ ” 38 It is impossible to conceive the nutri- 
tion of body cells without assuming that the assimilable nutritive 
substances come into physical and, eventually, chemical relationship 
with the protoplasm of the nourished cell. Considering the large 
variety of substances which may thus be brought into contact with 
cells in the course of normal and abnormal metabolism, the body cell, 
chemically considered as a complex of enormous molecules, must 
possess a correspondingly great variety of atom groups, by means 
of which it can unite with these substances to assimilate them and 
make them a part of its own protoplasm. In order to enter into 
similar relationship with toxins and other antigens, then, it is only 
logical to suppose that the cell, in the same way, unites chemi- 
cally with the antigenic substance, and either assimilates it without 
sustaining harm, as in the case of non-poisonous complexes, or is in- 
jured in the process, as in the case of the poisons. 
The living cell, from this point of view, is conceived as consist- 
ing of a central chemical nucleus, the “Leistungskern,” more or less 
stable, in that the specialized tissue function is dependent upon it, 
and a manifold variety of “side chains,” or atom groups by means 
of which it can enter into relationship with the nutritive and other 
materials carried to it by the body fluids. The latter term, “side 
chains,” is taken from the nomenclature of chemistry, and, although 
the analogy is a loose one, it serves satisfactorily to elucidate Ehr- 
lich’s meaning. Thus we may conceive the “Leistungskern” as the 
central carbon ring of any compound of the Benzol series, as, for 
instance, in salicylic acid in which the hydrogen atoms, the hydroxyl, 
OH 
I 
C 
/ \ 
H—C C—COOH 
I I 
H—C C—H 
\ / 
C 
Salicylic acid 
H 
OH 
I 
C 
I | CO2CH3 
\/ 
Methyl salicylate 
and the acid radicles represent “side chains. ’ By means of the lat- 
ter the compound can enter into relation with other substances, as, 
for instance, with CH3 in the formation of methyl salicylate. 
Graphically, though this analogy formulates Ehrlich’s fundamental 
38 Ehrlich. Introduction to “Gesammelten Arbeiten,” Berlin, Hirsch- 
wald, 1904. 146 
INFECTION AND RESISTANCE 
conception, it must not be taken as too literally representing the 
existing conditions, since, in actual metabolic interchange, there is 
an infinite variety of possible “side-chain” groups; for we are deal- 
ing with an enormous number of assimilable substances, most of 
them of chemically unknown constitution. The cell, therefore, is 
looked upon as an active chemical complex, retaining its own pe- 
culiar functional characteristics by reason of the “Leistungskern,” 
but constantly getting rid of waste products and entering into new 
union with extraneous materials by virtue of its “side chains.” 
These side chains, because of their “receiving” function, are spoken 
of by Ehrlich as cell “receptors.” 
That the chemical structure of certain bodies determines their 
ability to enter into relation with cell derivatives such as enzymes 
is, of course, a fact well established by experiment and explains the 
specific action of bacterial and other ferments upon certain sub- 
stances to the exclusion of others. Thus Pasteur noted the fact that 
bacterial ferments could decompose dextrorotatory tartaric acid 
while they did not affect the levorotatory variety, and Emil 
Eischer 39 showed that only those carbohydrates possessing 6 and 9 
carbon atoms were subject to fermentation by yeasts, and of these 
only the ones belonging to the “d” series, observations which, by 
demonstrating the relationship between these active agents of extra- 
cellular digestion, and the stereochemical configuration of the mole- 
cules acted upon, lend much support to the logic of Ehrlich’s con- 
tentions. 
Moreover, the recent experiments upon the growth of tissues in 
plasma outside of the animal body in which cartilage cells produce 
cartilage, kidney cells, etc., have shown that, given the same nutri- 
tive materials, the cells themselves must command a certain selective 
power in the choice of these materials, which can only depend upon 
a specific element in the structure of the cell receptors. As Eischer 
has expressed it for fermentation, the ferment must possess an atom 
group which fits into some group of the fermentable substance as a 
“key does into a lock,” an analogy which is equally applicable to 
Ehrlich’s conception of the relation of the “side chain” to a nutri- 
tive molecule. 
Now the toxins and other antigens are, all of them, so far as we 
know, complex chemical substances, derivatives of animal and vege- 
table cells, and, for this reason, should have much in common with 
the materials available for nutrition. It is not strange, therefore, 
that, coming into contact with the cells of the body during the acci- 
dents of disease or other abnormal conditions, they should find re- 
ceptors by means of which they can combine with the cell. Under 
the ordinary conditions of nutrition a suitable particle taken up by 
the cell in this way would be assimilated and the receptor either 
39 See Oppenheimer, “Die Fermente,” Yol. 1. The Structure op Cell-Receptors and Immune Bodies, According to Ehrlich’s Conception. 
(After Aschoff.) 
147 148 
INFECTION AND RESISTANCE 
freed for further use or regenerated for the further absorption of 
similar substances, by virtue of a mechanism delicately coordinated 
to the needs of cell-nutrition. In the case of the absorption of sub- 
stances belonging to the class of antigens, however, foreign proteins 
difficult of assimilation, or of toxins even directly harmful, the re- 
ceptors occupied by these substances are rendered useless to the cell, 
and, if the cell continues to live, must be regenerated. If the degree 
of poisoning or the amount of other antigen introduced has been 
extremely slight, this regeneration may possibly take place, as in 
the course of nutritive processes, without further disturbance. If, 
however, the amounts of antigen are greater than this, or are repeat- 
edly thrust upon the cell, the process of regeneration may be not only 
sufficient to compensate for the loss of the eliminated receptors, but 
may follow the general law of overcompensation, formulated by Wei- 
gert, and receptors of the variety occupied by the antigen are pro- 
duced in excessive number. 
Here again Ehrlich has called analogy to his aid, and has taken 
his conception of “overcompensation” from the well-known phe- 
nomena of pathological anatomy where, for instance, in the restora- 
tion of cellular elements after injury, there is often an overpro- 
duction of granulation tissue, far beyond the needs of simple healing. 
Thus the restitution of cell receptors, if sufficiently stimulated 
by large quantities or repeated administration of the antigen, far 
exceeds the quantity normal to the cell, and may proceed to such a 
degree that the cell, becoming as it were “top-heavy” with these 
elements, sloughs them off into the surrounding lymph and blood, 
where they circulate as free receptors. These free receptors then, 
having specific affinity and combining power for the antigen which 
incited their production, unite with subsequently introduced antigen 
in the blood stream, diverting it from the cells themselves, and, in 
the case of the variety of antigens spoken of as toxins, this union 
with the free receptors in the blood stream would serve to protect 
the cells from harm, exerting thereby an antitoxic action. 
The antibodies appearing in the blood of immunized animals, 
therefore, represent atom complexes, normally parts of the body cells 
and concerned in the metabolic processes, but now produced in ex- 
cess and extruded into the body fluids under the influence of the 
stimulation of immunization. The very substances, as Behring has 
put it, which make possible the poisoning of the cell by the toxins be- 
come protective when, detached from the cell, they circulate in the 
blood. Thus the theory, beside explaining the causes leading to 
antibody formation, offers a plausible reason for the relatively strict 
specificity observed in antibody-antigen reactions. 
Formulated in direct connection with the investigations upon 
toxins and antitoxins, the side-chain theory has been extended by 
Ehrlich and his associates to all known phases of antibody-antigen TOXIN AND ANTITOXIN 
149 
reactions. The differences in the nature and complexity of various 
antigens would naturally necessitate variation in the receptors capa- 
ble of assimilating them, and these receptors, appearing subsequently 
in the blood as antibodies, must, of necessity, differ from each 
other. On this basis Ehrlich has conceived of three main varieties or 
“orders” of receptors or “haptines,” as he calls them. Of these the 
simplest are those of the first order which attach to the toxins, and 
by over-regeneration appear in the blood stream as antitoxins. Those 
of the second order, adapted to the assimilation of more formidable 
protein molecules, are, of necessity, of greater structural com- 
plexity, appearing in immunized animals as the agglutinins and 
precipitins, while those of the third order, dependent upon the co- 
operation of alexin or complement, for proper functionation, appear 
as the cytotoxins or lysins. The detailed structure of these various 
haptines will be discussed in connection with other considerations 
dealing with their special reactions. 
Limiting ourselves, for the present, to a broad consideration of 
the theory as a whole, it may be briefly recapitulated as follows: 
Toxins or other antigens, in order to exert any influence upon the 
animal body, must enter into chemical relationship with the cells. 
This they do by virtue of union with chemical units or atom groups 
of the cells, spoken of as “side chains.” These side chains or re- 
ceptors, thrown out of function by this union, and necessary for the 
metabolic processes of the cell, are regenerated, and under the influ- 
ence of repetition of this process are produced in excess, to such a 
degree that they are eventually thrown off bv the cells and enter the 
circulation as antibodies. Thus far the theory, comparing the union 
of antigen with cells to the processes of nutrition, is eminently log- 
ical and likely, necessitating the assumption of over-regeneration 
as the only criterion not directly amenable to experimental proof. 
That the antigen can be bound by the body cells has been vari- 
ously shown in a large number of investigations, some of which have 
been reviewed in our section on the action of bacterial poisons. We 
have there seen that Donitz demonstrated the rapid disappearance 
of tetanus and diphtheria toxins from the circulation of susceptible 
animals, and that conversely Metchnikoff showed that the poison 
may persist unabsorbed and unchanged for weeks and months in the 
blood of such insusceptible animals as the turtle and the lizard, 
facts which furnish indirect evidence of the absorption of the toxins 
by the body cells. More direct evidence has, of course, been possible 
in the test tube experiments with hemolytic and other cell poisons 
where a directly specific combination between antigen and antibody 
has been easily demonstrable. Thus, in his earlier experiments 
with spider poison, Sachs was able to show that rabbit erythrocytes, 
which are sensitive to the poison, could absorb it out of solution, 
while dog and other corpuscles, which were insusceptible to the 150 
INFECTION AND RESISTANCE 
poison, did not bind or absorb it. This can be easily demonstrated 
for many antigens and antibodies and may be accepted as a fact. 
This point established, and repeatedly confirmed, and the origin 
of antitoxins from the cells of the body having been rendered likely 
by the experiments of Salomonsen and Madsen, and by those of 
Roux and Vaillard just cited, it would follow, by the theory of 
Ehrlich, that we should find the site of antibody production in the 
very cells which possessed specific affinity (receptors) for the 
antigen. This question has been variously investigated, chiefly in 
the case of the toxins and antitoxins, since this phase of the subject 
is most easily amenable to experiment. It will be remembered also 
that Wassermann and Takaki discovered that emulsions of the tissue 
of the central nervous system of rabbits and guinea pigs, shown 
by Meyer and Ransom and others to be the special points of attack 
for tetanus toxin, possessed the power of neutralizing this poison 
in vitro, while emulsions of spleen, kidney and other organs had no 
such effect. They assumed from this that the poison was fixed by 
cell receptors, antecedents of antitoxin in the sense of Ehrlich. 
Kempner40 made similar observations with botulinus toxin and 
further confirmation has been derived from experiments like those 
of Blumenthal,41 who found that the toxin was neutralized by the 
brain tissue of susceptible animals but showed conversely that the 
brain substance of the chicken, an animal but slightly susceptible 
to tetanus, possessed little or no neutralizing power. Similar re- 
sults were obtained by Metchnikoff in the cases just cited. 
The great importance of these experiments lies not only in show- 
ing that body cells may absorb the toxins, but that there is direct 
relationship between the susceptibility of tissue and the toxin-bind- 
ing properties. Furthermore the facts demonstrated by Metchnikoff 
that no antitoxin was produced by those animals (turtle, lizard) in 
which the tissues had no power of fixing poison and are consequently 
insusceptible, furnish evidence in favor of Ehrlich’s view. 
It becomes of great importance, therefore, to determine whether 
in the case of the fixation of tetanus toxin by the brain cells the 
union between cell and toxin is a specific and chemical one com- 
parable in every way to the union of toxin with antitoxin. 
Metchnikoff, in spite of his results in the experiments just cited, 
objected to this interpretation on the ground that although the brain 
emulsion of a guinea pig neutralized tetanus toxin in vitro, the in- 
jection of the toxin into such an animal, subdurally, produced the 
disease. This can hardly be regarded as a valid argument against 
Wassermann’s interpretation, since the very premises of the Ehrlich 
theory require that these neutralizing elements, when still attached 
40 Kempner and Poliak. Deutsche med. Woch., 1897, No. 23, p. 505. 
Kempner and Shepilewsky. Zeitschr. f. Hyg., Yol. 36, 1901, p. 1. 
41 Blumenthal. Deutsche med. Woch.} 1898, No. 12, p. 185. TOXIN AND ANTITOXIN 
151 
to the living cell, as “sessile” receptors, are the cause of the poison- 
ing, since they serve to “unlock” the cell to the entrance of the toxin. 
Similar objections on the part of Metclmikoff42 were based on some 
of his own experiments, as well as on those of Courmont 43 and 
Doyen, in which it was found that the poison disappears but slowly 
(in 2 to 3 months) from the circulation of frogs, and the brain cells 
show hardly any toxin neutralization in vitro, whereas these animals 
can be rendered tetanic if they are warmed to 25° to 30° C. Further 
work, however, by these authors as well as by Morgenroth 44 has 
satisfactorily cleared up this difficulty. As a matter of fact, tetanus 
poison disappears more rapidly (that is, is bound by the cells more 
rapidly) from the circulation of frogs, if the frogs are warmed to 
30° C. or more. Furthermore, if the toxin is injected into these 
animals, and they are kept at low temperatures, no disease results, 
but if they are then warmed up to the temperature stated, they grad- 
ually succumb to the disease. Morgenroth has shown that the ap- 
parently anomalous behavior of frogs in this respect is actually a 
question of temperature. At low temperatures the poison is bound, 
though with extreme slowness, but the toxophore group of the toxin 
does not functionate. When the animals are warmed, not only does 
the binding proceed more rapidly, but the toxophore group becomes 
active. He thus not only has answered Metchnikoff’s objections to 
Ehrlich’s theory on this ground, but has furnished an additional in- 
direct confirmation of the dual constitution of toxin, that is, its 
constitution of a haptophore and a toxophore atom group, suggested 
by Ehrlich in his diphtheria-toxin analysis. 
There is apparently, then, a strong absorption of tetanus toxin 
by the brain and nervous tissue of all animals which are susceptible 
to the poison, an absorption which amounts, as we have seen, to neu- 
tralization, the brain emulsion acting like antitoxin when mixed 
with the toxin before injection, as in Wassermann’s and Takaki’s 
experiments. 
A serious objection has been brought, however, to the assumption 
that this binding can be identified in its nature with the similar bind- 
ing of toxin by antitoxin, and a number of authors have claimed that 
the binding by the brain is not a binding by specific receptors, but 
an accidental property due to the presence of some fortuitous fixing 
substance in the central nervous system. Besredka 45 showed, for 
instance, that the brain of susceptible animals could bind much more 
toxin than it could actually neutralize, and that, if antitoxin was 
added to a brain emulsion previously saturated with the toxin, the 
toxin is removed from its combination with the brain cells and these 
42 Metchnikoff. Ann. de I’Inst. Past., Vol. 12, 1898. 
43 Courmont and Doyen. Arch, de Physiol., 1893. Courmont and Doyen. 
Compt. rend, de la soc. de biol., 1893. 
44 Morgenroth. Arch, internat. de Pharm., Vol. 7, 1900, pp. 265-272. 
45 Besredka. Ann. de VInst. Past., Vol. 17, 1903, p. 138. 152 
INFECTION AND RESISTANCE 
again regain their original absorbing property. These experiments 
would seem to point to a difference, especially in regard to affinity 
and firmness of union between the nature of the combination between 
toxin and brain emulsion on the one hand, and toxin and antitoxin 
on the other. This, of course, would prove a serious obstacle to the 
interpretation of the binding of toxin by susceptible cells in the sense 
of Ehrlich, as depending as it were upon union with specific recep- 
tors, or, as they might be termed, “sessile” antitoxin. Moreover, to 
strengthen such objections to this point of view, the work of Land- 
steiner and v. Eisler 46 has brought out the fact that extraction of 
brain tissue with ether materially reduces its toxin-binding powers 
by removing fatty or lipoidal substances, such as cholesterin and 
lecithin. And it has indeed been confirmed that lipoids can possess, 
in many instances, binding properties not only for toxins but for 
other forms of antibodies. On the basis that at least a part of the 
toxin absorption by brain emulsions depends upon such lipoidal fixa- 
tion, the results of Besredka are readily explained, but were this the 
sole cause of toxin fixation by these tissues it would indeed be diffi- 
cult to interpret the phenomenon, with Wassermann and Takaki, in 
support of Ehrlich’s theory. For, without going into further refine- 
ments, the fact of the probable proteid, certainly not lipoidal, nature 
of the antitoxins, discussed in a previous section, would alone serve 
to distinguish the two modes of toxin fixation. 
However, a number of facts have been ascertained which show 
that, although the lipoids play some part in the antitoxic action of 
brain cells, they do not by any means account for the entire process. 
In the first place it is found that the heating of brain emulsions al- 
most completely removes their power to bind the toxin, while no 
such reduction of the fixative property follows the heating of lipoids 
like cholesterin or lecithin. The experiments of Marie and Tif- 
feneau 47 have done much to clear up the confusion regarding this 
point. They determined that the “lipoidal binding” constituted only 
about one-tenth of the total binding power of the brain emulsions, 
by showing in the first place that only one-tenth of the total was left 
after heating, and that all but one-tenth could be destroyed by sub- 
jecting the tissue to the action of proteolytic enzymes. It appears 
from this that a large part, at any rate, of the toxin fixation of the 
brain tissues is dependent upon substances of an albuminous nature, 
a smaller but definite part being dependent upon fixation by lipoids, 
a phenomenon entirely apart from the former in underlying princi- 
ples. This would, it seems, both justify the original interpretation 
of Wassermann and still explain the apparently contradictory results 
of Besredka and others. 
48 Landsteiner and v. Eisler. Centralbl. f. Bakt., Yol. 39, 1905. 
47 Marie and Tiffeneau. Ann. de I’Inst. Past., Vol. 22, pp. 289 and 644, 
1908. CHAPTER VI 
THE BACTERICIDAL PROPERTIES OF BLOOD SERUM, 
CYTOLYSIS, AND SENSITIZATION 
In spite of the profound physiological alteration of the animal 
body which is implied by the acquisition of immunity against any 
particular infection, we have seen that no anatomical or histological 
changes in the organs and tissues accompany such alteration. The 
same is true of the difference between animals of different species, 
in which the most marked variation in resistance against any given 
infection is inexplicable on the basis of structural or microscopic 
characteristics in the organs. We have mentioned briefly the at- 
tempts that have been made to discover chemical and physical changes 
or differences to account for such conditions and have seen that the 
attention of investigators was soon attracted to the blood. 
A possible relationship between the blood and the defence of the 
body against infection had been foreshadowed by observations made 
long before the days of bacteriological knowledge. As early as 1792, 
John Hunter, in his “Treatise on the Blood, Inflammation and Gun- 
shot Wounds/’ had noted that the blood did not decompose as readily 
as other putrescible material, and a century later, during the period 
of great interest in the living nature of fermentation and putrefac- 
tion, Traube (1874) expressed the opinion that blood could destroy 
bacteria. Similar observations were made by Lister and by Groh- 
man 1 but no experimental work aimed at this point was carried on 
until 1886, when the subject was taken up by Hut tall,2 von Fodor,3 
and Fliigge, and a little later by Buchner.4 These authors, working 
with defibrinated blood, peptone blood, and blood serum, showed that 
such substances all exerted a definitely measurable destructive influ- 
ence upon bacteria, and Hut-tall, later confirmed by Buchner, further 
found that this bactericidal power was weakened on standing, and 
could be rapidly destroyed by heating to 60° C. 
Their method of procedure consisted in the planting of controlled 
amounts of various bacteria in measured quantities of blood and, 
after several hours at 37° C., pouring plates and thus determining 
the numbers of surviving organisms. The fact of bactericidal power 
1 Grohman. Quoted from Adami, “Principles of Pathology,” Vol. 1, p. 
497. 
2 Nuttall. Zeitschr. f. Hyg., 4, 1888. 
3 Yon Fodor. Deutsche med. Woch., 1887. 
4 Buchner. Centralbl. f. Bakt., Yol. 5, 1889. 
153 154 
INFECTION AND RESISTANCE 
established, there was, of course, much early difference of opinion 
as to the mechanism responsible for the destruction of the bacteria, 
and a number of simple explanations were suggested which, though 
entirely refuted at the present time, still possess considerable inter- 
est in showing the stages of development through which the concep- 
tions of the mechanism of immunity have progressed. 
These early theories were formulated chiefly upon the under- 
lying thought that the animal body was primarily passive in its rela- 
tion to the invading micro-organisms, and that the disappearance of 
bacteria in the body fluids was due to the existence of a chemically 
or physically unfavorable environment which prevented their multi- 
plication and therefore induced gradual mortality among them. 
Thus Billroth 5 believed that bacteria could thrive in the body only 
after a preceding putrefactive change had prepared a favorable pab- 
ulum. Others attempted to discover a relation between the degree 
of alkalinity of the blood serum and the destruction of bacteria. 
This argument was soon refuted by the experiments of Buchner, who 
showed conclusively that the bactericidal power of serum was not 
reduced by the neutralization of its natural alkalinity with weak 
acetic acid. 
Another theory which has been kept alive until the present day by 
Baumgarten,6 and in favor of which much has been written by Fischer, is 
the so-called “Osmotic” explanation. The basis of this conception is the 
observation that vegetable and other cells, which are in themselves delicate 
osmotic systems, undergo changes when they are placed into fluids of dif- 
ferent osmotic tension.7 Thus, of course, cells of all kinds may be destroyed 
by being placed in distilled water on the one hand, or in hypertonic salt solu- 
tion on the other. The point of view of Baumgarten, as explained in a 
recent edition of his “Text-book of Bacteriology,” is the following: The bac- 
terial (or blood) cell, like all cells, is surrounded hv a semi-permeable mem- 
brane. Under ordinary conditions, this membrane permits the passage of 
certain substances which must enter and leave the cell in the course of normal 
metabolism. When the bacteria are placed in a specific bacteriolytic serum 
there is a chemical union between the antibody and the cell membrane, and 
the latter is, in consequence, injured. The result of the injury is that now 
the cell becomes permeable for salts and other substances to which it was 
impermeable before, and there are consequent swelling and increased intra- 
cellular pressure. This, in turn, brings about the extrusion from the cell of 
proteins and other ordinarily non-diffusible substances, and destruction of 
the cell results. This explanation is practically an adaptation of the earlier 
more primitive osmotic theories to the facts subsequently discovered. It 
stands in direct contradiction to the prevailing’ opinion that the process of 
bacteriolysis and cytolysis in general is an enzymotic process, brought about 
by the injury of the cell by specific substances comparable to digestive fer- 
ments. Interesting though the suggestion of Baumgarten is, it can hardly 
5 Billroth. Quoted from Sauerbeck, “Die Krise in der Immunitatsforseh.,” 
Klinkhardt, Leipzig, 1909. 
6 Baumgarten. “Lehrbuch der pathogenen Mikroorg.,” Hirzel, Leipzig, 
1911. 
7 See also Pfeiffer’s “Pflauzen Physiologie.” BACTERICIDAL PROPERTIES OF BLOOD SERUM 
155 
receive more than casual attention given it for the sake of completeness, 
since careful experimental work by von Lingelsheim 8 has shown definitely 
that altered salt contents of serum do not exercise the effect upon bacteriolysis 
which we would be entitled to expect from Baumgarten’s reasoning. 
In explanation of the natural immunity possessed by many animals 
against various infections, Baumgarten has offered another explanation 
which, like the preceding, we may classify, in agreement with Sauerbeck,9 
with the “passive” theories. This theory, which he calls his “Assimilation 
Theory,” assumes that the bacteria do not find suitable food material in 
the tissues and fluids of certain animals, and, since bacteria do not have to 
be killed to be eliminated, but may be checked merely by their inability to 
grow and multiply, they must soon succumb in surroundings in which they 
find no suitable foodstuffs. This point of view approaches somewhat the 
earlier exhaustion theory of Pasteur, which has been mentioned in another 
place.10 
In contrast to these “Passive” theories of immunity are the now 
prevailing and well-founded opinions that the resistance of the ani- 
mal body against bacterial invasion is not a mere fortuitous result 
of chemical and physical conditions encountered by the infectious 
agents, but is rather the result of the struggle against the invasion 
by active forces of the body cells and fluids. The part played by the 
cells had already been emphasized by Metchnikoff and his school 
when the discovery of the bactericidal power of the normal blood 
was made. The study of the antibacterial powers of the blood now 
introduced a new element which became the basis of the so-called 
“humoral” theories. In the prolonged controversies waged, with 
great astuteness and experimental skill, between the adherents of 
these two schools, most of the facts which we possess regarding im- 
munity were discovered, and it is only within recent years that we 
have obtained information which has made possible a correlation 
between these two main paths of thought. 
The humoral theory was conceived by Buchner, as the first 
important theoretical result of Nuttall’s discovery. Buchner, as 
we have seen, confirmed the observations of Nuttall both as to 
the primary fact of the bactericidal power of the fresh normal 
blood and as to the unstable nature of this bactericidal property. 
He looked upon the antibacterial power as depending upon a con- 
stituent of the fresh blood plasma, which he named “alexin” (pro- 
tective substance), and which he believed to be comparable to a 
proteolytic enzyme. The action of this alexin was conceived as 
potent against all bacteria equally, without showing specific selection 
of various species to any great extent. The analogy to ferment 
action was formulated by Buchner because of the heat sensitiveness 
and the instability of the bactericidal substance on standing; and he 
8 Yon Lingelsheim. Zeitschr. f. Hyg., Yol. 37, 1901. 
9 Sauerbeck. “Die Krise in der Immunitatsforschung,” Kiinkhardt, 
Leipzig, 1909. 
10 The influence of foodstuffs, temperature, and other environmental con- 
ditions upon natural immunity has been discussed in an earlier section. 156 
INFECTION AND RESISTANCE 
suggested that this alexin might possibly be a product of the tissue or 
blood cells, possibly leukocytic in origin. 
Buchner found that the action of the ferment-like alexin upon 
bacteria was most marked at the temperature of the body, and that 
it was capable of destroying bacteria in the subcutaneous tissues and 
the serous cavities of the animal body, without the aid or coopera- 
tion of cellular elements. He inferred that there was a direct rela- 
tion between the potency of the alexin and resistance against 
infection. 
The Pfeiffer Phenomenon.—The next great step in the under- 
standing of the bactericidal processes was now made by Pfeiffer as 
a consequence of studies upon the nature of cholera immunity. 
Pfeiffer 11 found that the injection of cholera spirilla into the 
peritoneal cavity of a guinea pig which had recovered from a previous 
cholera infection was followed by a rapid destruction of the bacteria. 
If small quantities of exudate were taken out of the peritoneum 
at varying intervals after the injection, a granular change and 
swelling of the bacteria were noticed, followed, soon after, by com- 
plete dissolution and disappearance. Such animals would recover 
from doses of bacteria which, in control animals of the same weight, 
resulted in death. He further found that the phenomenon was 
specific, in that the dissolution of cholera organisms only occurred 
in the cholera-immune animals, other bacteria being unaffected. In 
other words, the guinea pig had acquired a specific antibacterial 
power, expressed by the process of “bacteriolysis,” a property pos- 
sessed to only a very slight extent by the peritoneal exudate of a 
normal animal. It was the next logical step to determine whether 
the bacteriolytic power could be transferred to the peritoneal cavity 
of a normal animal by injecting, together with the bacteria, a small 
amount of the serum of such an immune animal. This was indeed 
found to be the case and, although such immune serum, like normal 
serum, is deprived of its in vitro bactericidal power on heating, 
Pfeiffer found, in his intraperitoneal experiments, that heated serum 
is quite as effectual as fresh immune serum in transferring passive 
immunity to a normal guinea pig. We may summarize the important 
harvest of facts obtained from these experiments of Pfeiffer in the 
following statements: 
1. Rapid dissolution of cholera spirilla takes place in the 
peritoneal cavity of a cholera-immune guinea pig. Similar lysis 
takes place not at all, or only to a slight extent, in the peritoneum 
of a normal pig. In consequence of the lysis the immune pig will 
survive the injection of quantities of bacteria which invariably kill 
normal animals of the same weight. 
2. The protection obtained in this way is specific. 
11 Pfeiffer. Zeitschr. f. Hyg., Vol. 18, 1894; also Yols. 19 and 20. Pfeif- 
fer and Isaefif. Deutsche med. Woch., No. 18, 1894. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
157 
3. The protection may be transferred from an immune to a 
normal guinea pig, by injecting a little immune serum together with 
the bacteria into the peritoneum of the normal animal. In a normal 
animal so treated lysis is in every way similar to that observed in 
the immune pig. 
4. The transfer of the lytic power and consequent immunity 
can be brought about not only by means of fresh immune serum but 
by heated serum as well, although the latter has lost all its alexic 
power because of the heating. 
Of the phases of this “Pfeiffer phenomenon” the one most diffi- 
cult to understand, in the light of the knowledge of that time, was 
the transference of the lytic property with the heated serum. Pfeiffer 
very naturally took his experiments to signify that the actual de- 
struction of bacteria in the animal body could take place entirely 
without the phagocytic participation of the body cells, a view in 
sharp contrast to that of the Metchnikoff school, and based upon his 
observation of the complete extracellular disintegration of the spi- 
rilla in the peritoneal exudate. He assumed, however, that there was 
an indirect participation on the part of the cells. The observation 
that heated serum, inactive outside the body, was efficient when in- 
troduced into the peritoneum, persuaded him that the cooperation 
of the living tissues was a necessary factor, and he assumed a pos- 
sible activation by substances derived from the endothelial cells lin- 
ing the peritoneal cavity. In the same way he explained his failure 
to observe actual bacterial dissolution in hang-drop preparations, 
even when fresh serum was used in the experiment. 
It will be interesting to examine a protocol of an experiment such 
as those carried out in the performance of the Pfeiffer phenomenon 
in order to make the actual occurrences entirely clear. In such 
experiments the quantity of bacteria used must be chosen with some 
regard to the virulence and toxicity of the particular culture em- 
ployed, since, as we shall see, protection of animals by bactericidal 
or bacteriolytic sera does not follow the law of multiple proportions 
as in the case of the protection against toxins by antitoxins. While 
the dose of bacteria chosen should be considerably above the minimal 
lethal dose for an animal of the weight used, it should nevertheless 
be remembered that the bactericidal serum does not possess antitoxic 
properties against the poisons liberated or produced as the bacteria 
undergo dissolution, and at best the protection by bacteriolysis is 
limited to a very definite maximum of bacteria, beyond which no 
further increase of serum quantity will avail. The following table 
will illustrate an experiment of this kind in which, in a series of 
guinea pigs, the bacteriolytic protective power (titre) is determined 
by comparative tests.12 
12 For extensive discussion of the technique of such tests see Boehme in 
Kraus u. Levaditi Handbuch, etc., Vol. 2, p. 366. The scheme of presenta- 158 
INFECTION AND RESISTANCE 
PFEIFFER PHENOMENON 
Weight 
of 
guinea pig 
Dose of bacteria* 
cholera spirilla 
Amount of 
inactivated 
immune serum 
Result 
(1) 215 gm. 
2 mg. 
0.1 c. c. in 1 c. c. 
salt solution. 
Complete dissolution in less 
than 1 hour. Lives. 
(2) 230 gm. 
2 mg. 
0.05 c. c. 
About the same as first. 
(3) 200 gm. 
2 mg. 
0.01 c. c. 
Somewhat slower than in other 
two; a few unchanged spi- 
rilla after 1 hr. Final disso- 
lution. Pig lives. 
(4) 245 gm. 
2 mg. 
0.005 c. c. 
Similar to (3) but complete dis- 
solution in 2 hrs. Pig lives. 
(5) 220 gm. 
Normal 
control 
2 mg. 
0.001 c. c. 
After 30 min. the spirilla seem 
to have begun to multiply. 
Dies with innumerable active 
spirilla in peritoneum. 
(6) 210 gm. 
2 mg. 
0.1 c. c. normal 
inactive rab- 
bit serum. 
Very slight lysis at the begin- 
ning. Soon rapid multipli- 
cation. Dies. 
* The bacteria may be measured for such an experiment by standard loopfuls (1 loop 
being equal to 2 milligrams), or by volume in emulsion with salt solution. 
Pfeiffer has established a system of standardization for the meas- 
urement of sera by this technique, lie speaks of one immunity unit 
as the smallest amount of such a serum which is capable of causing 
complete dissolution of 2 milligrams of culture material 13 (of a 
standard culture) and saving the life of the animal. The unit of 
the serum in the preceding test would accordingly be 0.005 c. c., and 
the titre of the serum, expressed in Pfeiffer’s language, would be 
200 units to the cubic centimeter. Owing to the great variation in 
the virulence and toxicity of different strains of the same organism, 
and also because of the difficulties opposed to the visible dissolution 
of many bacteria, which may be killed by the serum without show- 
ing much evidence of solution, the practical application of Pfeiffer’s 
standardization is not universally possible. In doing experiments 
by this technique, whatever their purpose may be, accurate adjust- 
ment of bacterial amounts and preliminary studies of virulence must 
be made in order that the tests may be of real value and, failing 
visible lysis, the death of the animals must be taken as the indicator 
tion of our example is taken from that used by him. See also Pfeiffer, 
Zeitschr. f. Hyg., Yol. 19, 1895, p. 77. . 
13 The standard “loop” used in many laboratories for the rough meas- 
urement of quantities of bacteria from agar cultures takes up approximately 
2 milligrams of the material. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
159 
of the titration. Comparisons of results obtained with two different 
cultures of the same species are consequently of value only when the 
minimal lethal dose of each and its toxicity have been studied before 
the final tests are made. 
The cardinal points of Pfeiffer’s phenomenon were rapidly con- 
firmed, but his assumption that the process could take place only 
within the animal body was soon corrected by both Metchnikoff 14 
and Bordet.15 Both of these investigators succeeded in producing 
extracellular lysis of cholera spirilla in hang-drop preparations. The 
former produced the phenomenon by adding to the hang-drop prep- 
arations small quantities of extracts of leukocytes, and thus at- 
tempted to correlate Pfeiffer’s observations with his own opinions 
regarding the importance of the leukocytes in bacterial destruction. 
The latter, however, subjected the phenomenon of bacteriolysis, both 
in vivo and in vitro, to a careful analysis and obtained results which 
definitely disproved the necessity of cellular intervention in this 
phenomenon, and furnished facts regarding the process which stand 
uncontradicted to the present day. Upon the basis of these our 
modern views of the mechanism of cytolysis in general are founded. 
Mechanism of the Bacteriolytic and Hemolytic Process.—Bordet 
showed that the bacteriolytic properties of immune serum are in- 
deed destroyed by heating to from 50° to 60° C. If, however, tc 
such a heated immune serum there is added a small quantity of 
fresh normal serum, the bacteriolytic power is restored with undi- 
minished vigor. He recognized in consequence that there were two 
distinct serum elements necessary for the process. Fresh normal 
serum by itself had very slight or no bacteriolytic power. Fresh 
immune serum had powerful and rapid effects. Heated immune 
serum had lost its power completely, but this was restored to it by 
the addition of the fresh normal serum. He noted, furthermore, 
that the specific nature of the bacteriolysis by the immune serum was 
unchanged after it had been inactivated by heat and reactivated sub- 
sequently by the normal serum. The inference was plain. Immu- 
nization of an animal incites the production, in the blood of this 
animal, of a “preventive” substance, which is moderately resistant 
to heat, and which is specific for the bacteria employed in the 
immunization. This substance cannot act upon the bacteria alone, 
however, but depends for its effective functionation upon the co- 
operation of another substance present universally in normal serum, 
the “bactericidal” substance, which is non-specific, corresponds to 
Buchner’s alexin, and is apparently not increased by the process of 
immunization. These are the fundamental facts revealed by the 
early studies of Bordet, and they are stated in the present connec- 
tion merely as experimental facts, without further elaboration of 
14 Metchnikoff. Ann. de VInst. Past., Vol. 9, 1895. 
15 Bordet. Ann. de VInst. Past., Yol. 13, 1899. 160 
INFECTION AND RESISTANCE 
the later theoretical interpretation placed upon them by Bordet 
himself and by Ehrlich and his followers. 
In the course of these studies Bordet 16 had used the immune 
serum produced in a goat by injection of cholera spirilla. As normal 
serum he had used guinea-pig serum, and the latter frequently con- 
tained a few blood corpuscles. He noticed that these corpuscles were 
frequently clumped in the goat serum and correlated this with the 
similar clumping (agglutination) of cholera organisms which he 
had noticed in this and other sera. In his incidental observation of 
the phenomenon of agglutination he had concluded that the living 
nature of the bacteria had no importance as far as their agglutina- 
tion was concerned, dead organisms being as readily agglutinated as 
living. 
Reasoning from this similarity between blood cells and bacteria 
in their behavior in serum, it occurred to him that the phenomena 
both of agglutination and of lysis might be expressions of general 
biological laws, not limited to bacteria. Accordingly he injected 
rabbit blood into guinea pigs, and examined the serum of animals 
so treated for its action upon rabbit corpuscles, in vitro. He found 
that the sera of “blood-immune” animals had acquired not only in- 
creased agglutinative power against the corpuscles injected, but had 
also acquired specific “hemolytic” powers, that is, the property of 
causing a solution of hemoglobin out of the red cells. (For the 
process of serum hemolysis does not consist of a complete dissolution 
of the red corpuscles, but rather in the liberation of the hemoglobin 
from the cell stromata.) The latter (shadow forms) can be recovered 
undisintegrated by the centrifugation of hemolyzed blood. The 
process, like that of bacteriolysis, was specific in that the hemolytic 
power was lost if the serum was heated to from 50° to 60° C., but 
could be restored undiminished by the addition of a little fresh nor- 
mal serum, in itself possessing no hemolytic properties for the given 
species of cell. The specificity of the phenomenon again was seen 
to reside entirely in the heat-stable factor, the heat-sensitive or 
“alexin” factor being non-specific, and not increased during the 
process of immunization. 
Observations related to those of Bordet concerning hemolysis 
were made independently, in the same year, by Belfanti and Car- 
bone, who had observed that the serum of animals treated with blood 
cells of another species became toxic for this species, and extensive 
confirmation of the phenomenon of hemolysis was obtained, in the 
year following, by the work of von Dungern, and by that of Land- 
steiner. 
After Bordet had thus established the important fact that hemol- 
16 See Bordet’s own account in a “Resume of Immunity”; “Studies in 
Immunity,” Bordet, collected and translated by Gay, Wiley & Son, N, Y., 
1909. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
161 
ysis was in every way analogous to bacteriolysis in that, like bac- 
teriolytic sera, hemolytic sera could be inactivated by heat, but 
reactivated by the addition of small quantities of fresh normal 
serum, Ehrlich and Morgenroth 17 undertook an elaborate study of 
the mechanism of hemolytic phenomena, hoping thereby to elucidate 
the mechanism of lysis in general. For it is obvious that hemolysis 
lends itself far more easily to experimentation than does bacteriol- 
ysis, and, as wre shall see, experiments on hemolysis can be made 
with a considerable degree of accuracy. Ehrlich and Morgenroth 
approached the investigation of the hemolysins from the point of 
view of the side-chain theory, formulated by Ehrlich in connection 
with his work on the toxins. According to this theory, it will be 
remembered, the hemolytic substances in the sera of animals treated 
with blood corpuscles represent the receptors or side chains of tissue 
cells. These receptors were originally integral chemical elements of 
the body cells, by means of which the cell became united to the 
injected erythrocyte (or bacterial) protein. Since union with the 
foreign substance blocked these receptors or side chains, thereby 
rendering them useless, they had been regenerated and, under the 
influence of immunization, regenerated in excess, cast off by 
the cell, and wrere now free in the blood stream as hemolysins (or 
bacteriolysins). 
If this conception of the process was the correct one, Ehrlich 
and Morgenroth argued, the hemolytic substances of any immune 
hemolytic serum should possess specific chemical affinity, “hapto- 
phore groups,” as they expressed it, for the blood cells which had 
been used in the immunization. 
In order to show this, they inactivated at 56° C., by the method 
of Bordet, a goat serum which was hemolytic for beef blood, left it 
in contact with beef blood corpuscles for 15 minutes at 40° C., and 
then separated the cells from the supernatant fluid by centrifuga- 
tion. To the blood cells they then added a little normal goat serum 
(by itself not hemolytic for beef blood) and found that complete 
hemolysis occurred. The subsequent addition of normal goat serum 
and beef blood cells to the supernatant fluid, however, resulted in no 
change. 
In the following diagram we have tried to represent this basic 
experiment, giving the facts only of the experiment without using 
any of the usual symbols which imply agreement with a theory. 
EXPERIMENT TO SHOW THAT THE ANTIGEN (iN THIS CASE RED BLOOD CELLS 
ABSORBS THE SPECIFIC HEAT STABLE ANTIBODY OUT OF THE IMMUNE SERUM. 
4 c. c. of 5 per cent, emulsion of washed beef blood. 
1 c. c. of inactivated blood serum of a goat treated with beef 
blood. 
In a test tube 
17 Ehrlich and Morgenroth. Berl, klin, Woch.} Nos. 1, 21, and 22, 1900, 162 
INFECTION AND RESISTANCE 
These substances are left together at 37.5° C. for one hour and 
then centrifugalized into: 
I 
Sediment of Corpuscles.—To this are 
added 4 c. c. salt solution and 0.8 
c. c. fresh normal goat serum, by 
itself not hemolytic for beef cor- 
puscles. 
Result = Complete hemolysis. 
II 
Supernatant Fluid Containing the 
Serum and Salt Solution.—To this 
are added washed beef corpuscles 
and 0.8 c. c. fresh normal goat 
serum. 
Result = No hemolysis. 
Summarized, together with the facts we have already outlined, 
this basic experiment has the following significance: the fresh serum 
of the goat, previously injected (“immunized”) with the beef blood, 
possessed the property of dissolving the hemoglobin out of beef cor- 
puscles, viz., hemolyzing them. Heating this serum to 56° C. for 
20 minutes, as Bordet has shown, deprives the serum of all hemolytic 
power, i. e., inactivates it. The addition of a little fresh goat serum, 
in itself inactive, completely reactivates the hemolytic properties of 
the heated immune serum. So far, as we have already seen, this 
shows that hemolysis is a dual process in which a heat-sensitive 
and a heat-stable substance cooperate, neither of them capable of 
producing lysis by itself. The lieat-sensitive ingredient, correspond- 
ing to Buchner’s “alexin,” is present in normal serum, and, as Bor- 
det 18 and von Dungern 19 had shown, is not increased in the process 
of immunization, and is apparently not specific. The heat-stable 
substance, therefore specific and increased in immunization, must 
represent the receptors, overproduced and cast off into the circula- 
tion. And, as Ehrlich and Morgenroth have now shown in the ex- 
periment just described, this heat-stable element is actually bound 
to the red corpuscles, and renders them susceptible to the action of 
the heat-sensitive substance in the normal goat serum. And further- 
more, in attaching to this heat-stable element, the blood cells have 
removed it from the solution. For we have seen, in the experiment, 
that addition of corpuscles and normal serum to the supernatant fluid 
resulted in no hemolysis, showing that the third necessary element, 
originally in the mixture, had been carried down with the red cells. 
In these and other experiments then, it was shown that only the 
lieat-stable substances could be fixed by the red cells, and this even 
at temperatures at or about 0° C. (a fact which indicates the strong 
affinity between the two substances), while the heat-sensitive 
“alexin,” which Ehrlich now’ called “complementcould not attach 
directly to the red cells. For if such complement, in the form of 
fresh serum, was added to washed red blood cells, and the mixture 
after standing at 40° C. for some time was centrifugalized, the com- 
18 Bordet. Ann. de VInst. Past., Yol. 12, 1898. 
*9 y. Dungern. Miinch, med. Woch., No. 20, 1900, p. 677, BACTERICIDAL PROPERTIES OF BLOOD SERUM 
163 
plement remained in the supernatant fluid, as could be easily shown 
by an experiment such as the one represented in the following 
protocol. 
EXPERIMENT TO SHOW THAT COMPLEMENT OR ALEXIN IS NOT ABSORBED BY 
UNSENSITIZED CELLS 
Mixed in a test tube 
'4 c. c. of 5 per cent, emulsion of washed beef blood. 
0.8 c. c. of fresh normal goat serum (alexin or comple- 
ment), not, by itself, hemolytic for beef blood. 
These substances are left together at 37.5° C. for one hour, then 
centrifugalized into: 
I 
Sediment of Cells.—To this is added 
inactivated serum of immune goat 
which would cause hemolysis if 
alexin were present. 
Result = No hemolysis. 
II 
Supernatant Fluid (salt solution and 
serum).—To this is added washed 
beef blood and inactivated serum 
of immune goat containing heat 
stable element. 
Result — Complete hemolysis. 
Although, therefore, the red cells bind the thermostable specific 
antibody of the immune serum and not the complement or alexin, 
it 'was shown both by Bordet and by Ehrlich and his collaborators 
that the red cells, after absorption of the thermostable substance, 
when exposed to the action of the complement, were not only disin- 
tegrated by hemolysis but, in the process, fixed or attached the com- 
plement, so that this was no longer available for further activation 
of other sensitized cells. 
The fact that the alexin or complement is used up during proc- 
esses of lysis, as first described by Bordet, Ehrlich, and others, has 
recently been made the subject of repeated investigation, since this 
is out of keeping with the general enzyme or fermentlike nature of 
complement indicated by many of its other properties. 
Muir,20 who studied the conditions thoroughly, comes to the con- 
clusion that the complement is in truth used up in hemolysis, but 
that it does not always disappear completely, this depending upon 
the relative amount of sensitizer or amboceptor present. (He con- 
firms the quantitative ratios between the two substances found by 
Morgenroth and Sachs in hemolytic reactions, a subject discussed 
by us in another place.) 
Liefmann and Cohn,21 in a more recent publication, have come 
to different conclusions. They believe that the disappearance of free 
complement from hemolytic complexes is not due to its chemical 
20 Muir. Lancet, Yol. 2, 1903, p. 446. 
21 Liefmann and Cohn. Zeitschr. f. Immynitatsf orsch. Or.. Vol. 8, p. 58, 
1911, 164 
INFECTION AND RESISTANCE 
union with the sensitized cells in the process of hemolysis, but is due 
rather 
(1) to a fixation by the products of hemolysis (stromata, etc.) 
after the reaction is accomplished, 
(2) to dilution, and 
(3) to weakening because of prolonged preservation in dilute 
solution at 37° C.22 
Theoretically this is of considerable importance if confirmed, 
since it would bear out strongly the conception of complement as a 
true enzyme or ferment. From the point of view of the practical 
utilization of complement fixation for various purposes it makes 
little difference, since here the disappearance of complement is the 
essential thing, irrespective of whether this occurs in the course of 
its activity or because of fixation by the products of its own action. 
We now have the basic principle of hemolysis; facts which can 
easily be shown to hold good for bacteriolysis and for the bacteri- 
cidal processes even when no actual solution takes place. Briefly 
reviewed, these facts are as follows: The antigen (blood cells, bac- 
terial cells, etc.) undergoes hemolysis or bacteriolysis when acted 
upon by two factors, one a thermostable substance, specific and in- 
creased during immunization, the other a thermosensitive substance 
present in fresh serum, not increased 23 by immunization of the ani- 
mal with the antigen and not specific. The specific thermostable 
substance becomes united with or fixed to the antigen regardless of 
the presence or absence of the thermosensitive alexin or comple- 
ment, and with such avidity that the union takes place even at 0° C. 
The alexin or complement, however, cannot enter into relation with 
the antigen unless this has been rendered susceptible to it by attach- 
ment to the thermostable specific substance. When this has taken 
place, union with complement occurs, but only at temperatures above 
0° C. (the speed and completeness of the union increasing as the 
temperature approaches 40° C.), and the result of the union is lysis 
or, in the case of bacteria not easily soluble, the bactericidal effect. 
The “Amboceptor” Conception.—It appears from the preceding 
that the thermostable hemolytic antibody must, of necessity, unite 
with the red cell before the complement or alexin can exert its action 
upon it. Ehrlich conceives this process as a mediation on the 
part of the heat-stable substance between the antigen and the alexin 
or complement. The heat-stable body, which he calls “amboceptor,” 
because of its assumed mode of action, possesses two combining 
groups—one the “cytophile,” by means of which it is anchored to 
the sensitive cell, the other the “complementophile,” by means of 
which it exerts affinity for the complement. The original cell re- 
22 In the ordinary dilution used in Wassermann tests, the unit of comple- 
ment employed may deteriorate entirely within several hours at 40° C. 
23 Bordet. Ann. de I’Inst. Past., Yol. 12, 1898, Confirmed by v. Dun- 
gern, Munch, med, Woch,, No. 20, 1900, BACTERICIDAL PROPERTIES OF BLOOD SERUM 
165 
ceptor, from which such an “amboceptor” takes its origin, is one 
which not only can combine with the antigenic substance offered for 
assimilation, but which also possesses another atom group by means 
of which it can enlist the aid of the digestive ferment of the blood, 
the alexin or complement. Cast off into the blood stream, as a re- 
sult of overregeneration, it now appears as a “double” receptor, 
which can form a link between the antigen and the complement or 
alexin. 
Ehrlich’s conception of the relationship of antigen, amboceptor, and comple- 
ment in the bactericidal and hemolytic process. In A the receptor is still 
a part of the body cell, in B it has been overproduced, and is free in 
the circulating blood. 
Schematic Representations of a Receptor of the Third Order. 
In his general scheme of diagrammatic representation of these 
processes Ehrlich refers to the “amboceptors” as “haptines” of the 
third order. 
Now it is quite plain, from the extreme specificity which results 
when an animal is immunized with any given variety of blood cells 
or bacteria, that there must be as great a variety of such amboceptors 
as there are different antigens, and indeed an animal immunized 
with two or more antigens may simultaneously contain in its blood 
serum a corresponding number of separately recognizable ambo- 
ceptors. 
This assumption of the multiplicity of “amboceptors” in the 
same serum is, of course, forced upon us by the fact of specificity, 
and the frequently repeated observation that the same serum may 
contain heat-stable lytic antibodies against a variety of antigens, each 
antigen absorbing out of such a serum that antibody only which 
specifically reacts with it. This fact has, of course, never been 
denied, and it is a frequent misunderstanding of the views of Bor- 
det, which will be discussed directly, to assume that he has combated 
the “multiplicity of amboceptor” in the sense just outlined. Ehrlich 
and Morgenroth, however, have expressed themselves in favor of the 166 
INFECTION AND RESISTANCE 
conception of a multiplicity of “amboceptor” not only in this sense, 
but as occurring in response to immunization with one and the 
same antigen. 
Ehrlich and Morgenroth 24 assume that any cellular antigen, 
blood or bacterial cell, substances of great complexity of chemical 
structure, must necessarily be possessed of a large number of differ- 
ent side chains or receptors. When immunization is practiced with 
such cells a correspondingly varying number of different ambocep- 
tors must result. They found, for instance, that when rabbits are 
immunized with ox blood, the resulting antiserum is capable of pro- 
ducing hemolysis not only of ox blood but of goat’s blood as well, 
though to a lesser degree. They conclude from this that the hemo- 
lytic action of the serum must be referred to the presence of at least 
two kinds of amboceptor, especially since repeated experiments with 
different anti-ox-blood sera showed that there was no regularity in 
the proportions of hemolysins for ox and goat blood, respectively. 
This opinion they further fortify by showing that exposure of the 
serum to ox blood deprives it of all its hemolysins, both those for 
ox and those for goat’s blood, whereas absorption with goat’s blood 
alone removes the specific goat’s blood hemolysins only. They trans- 
late their understanding of the conditions to graphic form by the 
following diagram: 25 
If ox blood is in- 
jected, a and (3 receptors 
being present, a and |3 
amboceptors are formed, 
and ox blood can conse- 
quently anchor both am- 
boceptors. The presence 
of (3 receptors in goat’s 
blood also explains the 
modern hemolysis of this 
blood by the antiserum, 
but lacking the a recep- 
tors which, in this case, 
represent the larger pro- 
portion, these blood cells 
cannot remove all the 
amboceptor for ox blood 
out of the serum. The 
example given, of course, 
represents the simplest assumed case, and Ehrlich and Morgenroth 
believe that the same blood or bacterial cells may possess an entire 
series of such receptors, some of them being dominant for the given 
Schematic Representation of Ehrlich and 
Morgenroth’s Conception of the Com- 
plex Structure of an Antigen. 
(After Ehrlich and Morgenroth. Berl. klin. 
Woch., Yol. 38, 1901.) 
24 Ehrlich and Morgenroth. Berl. klin. Woch., Nos. 21 and 22, 1901. 
25 Ehrlich. “Gesammelte Arbeiten,” p. 147. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
167 
species, others being merely secondary or “partial,” in varying 
proportions. 
If we grant the fundamental premises of Ehrlich respecting the 
“double receptor” or “amboceptor” nature of the specific antibody 
and its mediation between antigen and complement by means of a 
cytophile and a complementophile receptor, certain logical conse- 
quences of this conception suggest themselves, which, in their many 
ramifications, have been the subject of much investigation. And 
although many phases of these researches are no longer commonly 
accepted, some, indeed, being untenable in the light of more recent 
discoveries, the influence of this work upon the development of 
immunology has been so important that it must be briefly reviewed 
in order that controversial questions may be given the critical 
discussion which they merit. 
The comparison of the action of hemolytic sera with that of fer- 
ments, and the possibility of producing antiferments by the injection 
of the ferments into animals, obviously suggests a similar induction 
of antihemolysins by the treatment of animals with lysins. This, 
we have seen, was the method employed by Ehrlich and Morgenroth 
in their studies of the causes of the failure of autolysin formation 
in goats. They extended this work with the purpose of ascertaining 
whether or not there were differences in the structure of the cyto- 
phile groups of the various amboceptors formed when various ani- 
mals were injected with any given species of red blood cells. After 
obtaining a strong hemolytic serum by injecting ox blood into a 
rabbit, they treated a goat with the inactivated serum of this rabbit. 
The result was that the serum of the goat so treated, when mixed 
with ox blood cells and the hemolytic serum, prevented the sensiti- 
zation of the cells by the hemolysin. They then measured the neu- 
tralizing power of such an “anti-amboceptor” or “anti-sensitizer” 
against a variety of hemolytic sera produced with ox blood in differ- 
ent animals and found that, while this “anti-amboceptor” neutralized 
the hemolytic action of an antiserum produced in rabbits, it had but 
an indifferent or entirely ineffective neutralizing power upon sim- 
ilar ox blood hemolysins derived from goats, geese, dogs, rats, or 
guinea pigs. They concluded from this that, although these various 
lysins had been produced in the different animals by the injection of 
the same antigen, viz., ox blood, and possessed affinity for the ox 
blood in consequence, they must necessarily differ from each other in 
some way, since they were not equally neutralized by the same anti- 
lysin. It seemed to them that the difference in such cases must 
depend upon variations in the structure of the cytophile group of 
the amboceptor, a conclusion which they based upon the foregoing 
experiments and sought to support by the following reasoning*. 
When an animal is treated with sensitizers or amboceptors, they 
reasoned, these bodies react with the tissue cells by means of the. 168 
INFECTION AND RESISTANCE 
cell-receptors. These receptors are then overproduced and extended 
into the circulation as free atom-groups. 
They now act as “anti-amboceptor,” free in the serum, but are 
in structure merely overproduced cell receptors, identical with those 
which originally united on the cell with the injected amboceptor. 
Ehrlich and Morgenroth,26 therefore, believed that the neutral- 
ization of the amboceptor by the anti lysin depended upon a union of 
the latter with the “cytophile” group of the former, preventing its 
subsequent union with the red cells. And 
since one and the same antilysin did not thus 
invalidate the action of all the amboceptors 
for ox blood (derived from different ani- 
mals), they concluded that these “ambocep- 
tors” must possess different “cytophile 
groups.” 
That this conclusion of Ehrlich and Mor- 
genroth is not correct seems to follow the 
subsequent work of Bordet.27 He demon- 
strated that it is not necessary to inject ani- 
mals with specific hemolytic sera in order to 
obtain antilytic sera, but that the same object 
may be attained by injecting animals with 
the normal serum of an untreated animal. 
Moreover, if an “antisensitizing” serum so 
produced was added to corpuscles which had 
already absorbed “amboceptor,” it prevented 
the subsequent union of these sensitized cells 
with alexin or complement. From this it be- 
comes clear that, in the first place, the anti- 
sensitizer or anti-amboceptor cannot be 
identical with the cell receptors of the 
corpuscles, and, further, that the inhibition 
of the hemolysis which such an antisensitizer 
exerts, cannot be due to union with the 
“cytophile” group. This both contradicts the Ehrlich conception of 
the mechanism of “anti-amboceptors” and invalidates his argument, 
in this instance, in favor of the plurality of the amboceptors pro- 
duced by the injection. 
Bordet’s experiments were later confirmed by Ehrlich and 
Sachs,28 who admit the error of the former “anticytophile” interpre- 
tation of Ehrlich and Morgenroth’s experiments, but they still main- 
tain that Bordet’s experiments do not disprove the conception of 
an “amboceptor” or “Zwischenkorper” of Ehrlich. They claim that 
Schematic Represen- 
tation op Ehrlich 
AND MORGENROTH’S 
Conception of the 
Neutralization of 
a Hemolytic 
Serum by Antily- 
sin or Antiambo- 
ceptor, Reacting 
WITH THE CYTO- 
phile Group. (Ehr- 
lich and Morgen- 
roth, loc. cit.) 
This conception, as we 
shall see, has be- 
come untenable,. 
26 Ehrlich and Morgenroth. Berl. klin. Woch., No. 22, 1901, p. 600. 
27 Bordet. Ann. de VInst. Past., Yol. 18, 1904, p. 593. 
28 Ehrlich and Sachs. Berl. klin. Woch., No. 19, 1905. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
169 
Bordet’s results merely prove that the anti-amboceptor or anti- 
sensitizer is “anticomplementophile” instead of “anticytophile.” 
The principles involved we will discuss in another place in con- 
nection with Moreschi’s analysis of the “anticomplements.” How- 
ever this may be, we may conclude that Ehrlich and Morgenrotil’s 
differentiation of amboceptors or sensitizers by the cytophile group 
is no longer valid. 
The studies of Bordet on the antisensitizers (anti-amboceptor) 
had important results apart from their refutation of Ehrlich and 
Morgenroth’s opinion. In addition to showing that such antisensi- 
tizer did not represent cell receptors identical with those that an- 
chored the sensitizer (amboceptor) to the red blood cells, his experi- 
ments revealed the fact that such an antisensitizer neutralizes 
unspecifically various specific sensitizers as well as normal anti- 
bodies in the serum of the same animal; and this showed that there 
is no necessity of assuming a variety of specific antisensitizers, as 
had been done by the German workers. 
As regards the multiplicity of amboceptor or sensitizer, however, 
though the proof of this, by means of anti-amboceptors, has had to 
be abandoned, as we have seen, there is still a great deal of evidence 
advanced in favor of such an assumption. The chief support for 
such an opinion is found in the “group reactions” among bacteria, 
similar to those observed for blood cells by Ehrlich and Morgenroth, 
and described above (see page 166). For it is frequently observed 
that the antibodies produced by immunization with one species of 
bacteria may have a certain though lesser degree of action upon 
other related forms, these in turn absorbing only a part of the ambo- 
ceptor out of the serum, while the species originally used for im- 
munization takes out all the amboceptor present. Considering the 
great chemical complexity of the bacterial or tissue cells, moreover, 
we may well expect such multiplicity. And it is, indeed, entirely 
reasonable to suppose that a structure as complex as the bacterial 
cell may contain a number of antigens and consequently give rise 
to a number of sensitizers which differ in that each is specific for its 
particular antigen only. This is merely a restatement of the phe- 
nomenon of specificity and has, as a matter of fact, no modifying 
influence on the general principles involved. 
Multiplicity or Singleness of Alexin.—From the point of view 
of a general understanding of the processes of immunity, however, 
the question of multiplicity of sensitizer is not so fundamentally 
important as is the similar controversy which has been waged re- 
garding the unity or multiplicity of alexin or complement. Here 
again there has been some misconception as to the meaning of those 
who maintain the unity of alexin. Neither Bordet, nor anyone 
else familiar with experimental conditions, has ever maintained 
that the alexins of different animals were functionally identical. 170 
INFECTION AND RESISTANCE 
It is a well-known fact that tlie fresh blood sera of various animal 
species differ from each other considerably in their power to activate 
bactericidal or hemolytic systems. In regard to hemolysis, fresh 
guinea-pig serum is very powerful in activating many sensitized 
blood-cell complexes, but weak in activating sensitized guinea-pig 
corpuscles. Often one finds that the alexin of an animal is entirely 
impotent or but weakly capable of producing hemolysis of the sensi- 
tized cells of its own species, though this is not a general rule. 
Again, even without such species relationship, a given alexin may 
be very weak for certain complexes and strong for others. The 
alexin of horse blood can even be fixed to sensitized cells 29 without 
producing much, if any, hemolysis.30 An alexin which may be strong 
for a given hemolytic complex may be weak for certain bactericidal 
complexes, or vice versa. Thus there is a large mass of evidence 
which shows that no two alexins are exactly alike, though the 
difference between them can, of course, be defined functionally only. 
The difference between the opinions of Ehrlich and his school 
on the one hand, and the followers of Bordet, on the other, revolves 
not about this point, upon which all agree, but about the question of 
whether one and the same serum may contain more than one alexin 
or complement. Ehrlich and Morgenroth 31 and Ehrlich and 
Sachs 32 have brought forward evidence from which they deduce the 
existence of a number of different alexins or complements for hemo- 
lytic complexes in the same serum. The earlier experiments of 
Ehrlich and Morgenroth on this question were carried out by means 
of the filtration of normal goat serum through Pukall filters; 33 in 
these it appeared that the serum which passed through the filters was 
complementary for sensitized guinea-pig cells, while that part which 
had, in the original serum, hemolyzed sensitized rabbit cells was left 
behind. Similar differentiation of complement they later based 
upon experiments with anticomplementary sera which, they showed, 
did not equally neutralize all the complementary functions of a 
serum. 
In support of their contention Heisser 34 described two comple- 
mentary substances in rabbit serum, the one active for bactericidal 
complexes, the other for hemolytic, and similar experimental evi- 
dence has been brought forward by Wassermann 35 for guinea-pig 
and by Wechsberg 36 for goat serum. 
29 For the sake of clearness it may be repeated here that by sensitized 
cells we mean cells which have absorbed specific “amboceptor” or “sensitizer,” 
and have thereby become amenable to the action of alexin or complement. 
30 Browning. Wien. klin. Woch., No. 15, 1906. 
31 Ehrlich and Morgenroth. Berl. klin. Woch., No. 31, 1900. 
32 Ehrlich and Sachs. Berl. klin. Woch., No. 21, 1902. 
33 Sachs. Berl. klin. Woch., Nos. 9 and 10, 1902. 
34 Neisser. Deutsche med. Woch., 1900, p. 790. 
35 Wassermann. Zeitschr. f. Hyg., 37, 1901. 
36 Wechsberg. Zeitschr. f. Hyg., Vof. 39, 1902. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
171 
The evidence advanced by these writers is based chiefly on ex- 
periments in which it was found that a normal serum which pos- 
sessed both bactericidal and hemolytic powers could be deprived of 
the complement for one or the other of these activities only, by ab- 
sorption with the respective cells. In addition to this, Ehrlich and 
Morgenroth, Ehrlich and Sachs,37 Wendelstadt,38 and others, claimed 
to have differentiated various complements in the same serum by 
careful heating, by the action of weak acids or alkalis, or such' 
methods as the digestion of sera by papain. 
As a rule, these experiments have been carried out with normally 
hemolytic serum and unsensitized cells, though in certain cases 
Ehrlich has employed sensitized cells; but whenever this was done 
exposure to complement for purposes of absorption has been for 
much briefer periods than when normal serum was used. This point 
is significant when we come to consider the objections to the inter- 
pretation of the preceding experiment in favor of a plurality of com- 
plement, objections raised chiefly by Wilde 39 and by Bordet. 
Wilde refuted particularly the experiments of Neisser, who 
claimed that the absorption of fresh rabbit serum with anthrax 
bacilli deprived this serum only of its bactericidal but not of its 
hemolytic complement. Wilde showed that, if a sufficient excess of 
anthrax bacilli (or in given cases of typhoid bacilli or cholera spi- 
rilla) were added, both bactericidal and hemolytic complement could 
be absorbed from normal serum. He concludes that there is actually 
only one alexin present, but that the red cells and anthrax bacilli 
differ in their susceptibility to this alexin (or, in other words, that 
the sensitization of these cells by the normal serum is unequal, a 
conclusion which seems rational in view of the fact, now well known, 
that one and the same complement may differ greatly in the degree 
of its activity upon different sensitized complexes. 
Bordet has analyzed the conditions in a similar way. He found 
that absorption of normal serum with unsensitized cells rarely de- 
prived this serum of all of its alexin, even when these cells were used 
in considerable amounts. This he attributed to the feeble sensitiza- 
tion of the cells. If, however, strongly sensitized cells were added 
to such a normal serum, all the alexin would be taken up. He refers 
the phenomenon of specific alexin absorption, observed by previous 
workers, to insufficiency in the perfection of sensitization on the 
part of the cells used in the preliminary exposure; and subsequent 
work with complement fixation seems to bear him out. 
Most of these arguments, though they seem to us perfectly valid 
in the light of the experimental facts, have been answered by Ehrlich 
and his school by the assumption of the existence of so-called “poly- 
37 Ehrlich and Sachs. Berl. klin. Woch., Nos. 14 and 15, 1902. 
38 Wendelstadt. Centralbl. f. Bakt.} I, Yol. 31, 1902. 
39 Wilde. Habilitations Schrift, Munich, 1901. Also Berl. klin. Woch., 
No. 34, 1901. 172 
INFECTION AND RESISTANCE 
ceptors.” Ehrlich now admits that the amboceptors cannot be shown 
to differ from each other. However, he does not believe that differ- 
ences in the intensity of sensitization ex- 
plain variation in the functional efficiency 
of different complements upon sensitized 
cell complexes, nor does he accept, for proof 
of this, the fact that complement may be 
entirely absorbed out of a serum by a com- 
plex, even though the complement may be 
comparatively inefficient as an activator in 
the given case. He assumes that the sensi- 
tizer or “amboceptor” may possess a num- 
ber of complementophile groups (polycep- 
tors), by means of whirdi a nnmW nf 
different comple 
ments may become 
active in the giver 
case. Thus, al 
though such t 
p o 1 y c e p tor, o1 
course, is capable 
of uniting with the 
complement which 
activates the domi- 
nant complement, 
it is capable also 
of union with a 
number of other 
3omplements which 
have slight or no 
functional action 
whatever—the non-dominant complements. 
This opinion is rendered diagrammatic by 
Ehrlich and Marshall40 in the following 
way: 
If one carefully considers the reasons 
advanced for the assumption of the exist- 
ence of such polyceptors it does not seem 
that they are sufficiently forcible to lead one 
to desert the much simpler explanation of 
Bordet. 
Belated to the problems discussed in 
connection with the production of “anti- 
amboceptors” or “antisensitizers” are those which have arisen re- 
garding the existence of “anticomplement” or “antialexins.” 
Ehrlich and Morgen- 
roth’s Conception 
of the Action of 
Anticomplement. 
A. Scheme of Hemolysis. 
B. Action of Anticomple- 
ment upon Hemoly- 
sin. 
b. = blood cell, c. = com- 
plement, i. = immune 
body, a. = anticom- 
plement. 
The complementoids are 
not included in "the 
scheme, since in this 
case they are without 
influence. 
Polyceptor According 
to Ehrlich and 
Marshall 
(a) Receptor of the Cell. 
(b) Haptophore Group 
of the Amboceptor. 
(c) Dominant Comple- 
ment. 
(d) Secondary Comple- 
ments. 
Complementophile groups 
of the Amboceptor: 
(1) for the Dominant 
Complement. 
(2) for the Secondary 
Complement. 
(After Ehrlich and Mar- 
shall, Berl. klin. 
Woch., No. 25, 1902.) 
40 Ehrlich and Marshall. Berl. klin. Woch., No. 25, 1902. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
173 
Ehrlich and Morgenroth claimed that, by the injection of active 
horse serum into a goat, they had obtained substances in the goat 
serum which neutralized horse complement. They believed that 
the “anticomplements” thus produced neutralized the comple- 
ment by uniting with its haptophore group, thus preventing its 
combination with the “complementophile group” of the amboceptor. 
This was their conclusion because they found that the “anticomple- 
mentary” serum exerted no protective influence upon sensitized cells, 
when these were exposed to the serum and then removed, but that 
it protected against hemolysis when added to the cells together 
with the complement. There was apparently no union of the pro- 
tective substance with the “complementophile” group of the ambo- 
ceptor, but the protecting substance did act in direct antagonism to 
the complement itself. 
From the fact that similar anticomplements could be produced 
when inactivated serum was injected into animals, they concluded 
that, on inactivation, there was not a complete destruction of the 
complement, but that during the process of heating the zymophore 
group of the complement only was injured, the “haptophore group,” 
by means of which union to the tissue elements would take place, 
and through which, therefore, specific antibody production would be 
incited, remained intact. Such altered complement they speak ot 
as “complementoid.” 
Bordet has made similar observations upon the production of anti- 
alexins by the injection into animals both of active and of inactive 
serum, but in the light of further researches, which will be discussed 
in connection with the problems of alexin-fixation, chiefly those of 
Moreschi and of Gay, we are forced to the conclusion that the ex- 
istence of true anticomplements is by no means certain, and that 
the older evidence in their favor is found to be unconvincing at the 
present time. 
In the preceding paragraphs we have emphasized the conceptions 
of the cytolytic phenomena formulated by Ehrlich and his followers, 
and although we have brought out, whenever possible, the objections 
of other investigators to many of these opinions, we have not yet 
followed out in a systematic manner the reasoning of any of Ehrlich’s 
opponents. In opposition to the views of his school the leading 
position has been taken by Bordet, who, after all, furnished in his 
investigations the fundamental facts which have led to a comprehen- 
sion of the cytolytic processes. In explaining Bordet’s views we can 
do no better than to follow out his own exposition as set forth in his 
article, “A General Kesume of Immunity,” 41 published with a collec- 
tion of his papers. He expresses himself, in substance, as follows: 
That the antigen, in the form of bacteria, blood cells, or cells of 
41 “Studies in Immunity,” by Bordet and collaborators. Gay, Wiley & 
Sons, N. Y,, 1909. 174 
INFECTION AND RESISTANCE 
any other nature, meets in the body of the treated animal a “recep- 
tor” complex with which it unites is, of course, plain and agreed to 
by everyone. That the antibody produced by the tissues in response 
to such union of antigen with receptor is a direct product of the cells 
containing the receptors is likely. It is by no means certain, how- 
ever, or, at any rate, it has never been experimentally demonstrated, 
that, as Ehrlich maintains, the antibody is identical with the original 
receptor by which the antigen was fixed or anchored to the tissue 
cell. It might be assumed with equal justice that the cells of the 
immunized animal could 
build up a new substance, 
not identical with the re- 
ceptors, in consequence of 
stimulation by the anti- 
gen. It is also by no 
means certain whether the 
injected antigen reacts 
with the body cells them- 
selves or with the normal 
antibodies which we know 
to exist in many cases. 
Thus the blood serum of 
goats may normally often 
contain hemolysins against 
rabbit corpuscles. Is it 
not reasonable to suppose 
that possibly these may 
furnish the point of at- 
tachment and the source of further antibody production when 
rabbit cells are injected into goats ? In criticism of Ehrlich’s as- 
sumption of the mode of action of heat-stable lytic antibody, Bordet 
very justly maintains that no proof whatever exists of the “ambo- 
ceptor” nature of this substance. All that is certain is that the 
stable substance must unite with the antigen before the alexin or 
complement can exert its action upon it or be fixed by it. There is 
no entirely valid proof of the existence in this antibody of a “com- 
plementophile” and a “cytophile” group, and no satisfactory 
instance has been observed in which alexin has united with a heat- 
stable antibody which has not previously been united with an anti- 
gen.42 All that has been shown is that the antigen, together with 
its specific antibody, forms a complex which has an avidity for 
alexin, a complex which is “endowed with properties of absorption 
for complement which neither of its constituents alone possesses.” 
Bordet speaks of the “amboceptors,” therefore, as “sensitizers,” 
Schematic Representation of Bordet’s 
View Concerning the Inability of 
Complement to Unite with Either 
Antigen or Sensitizer Alone and Its 
Ability to be Fixed by the Complex 
Formed When the Antigen is Sensi- 
tized. 
Compare this figure with that representing 
Ehrlich’s conception of the same process. 
42 Refer also to the discussion of the conglutinins at the end of this 
chapter. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
175 
meaning by this that the antigen, by union with its antibody, is sensi- 
tized to the action of the alexin. The term “sensitizers” in no way, 
therefore, implies a preconceived notion, experimentally unproved, 
of the mode of action or structure of the sensitizer. Since we have 
graphically explained Ehrlich’s opinions, a similar diagrammatic 
representation may be permitted of Bordet’s opinion of the same 
process of union of antigen and heat-stable antibody with the con- 
sequent development of alexin-fixing property. 
In this diagram the ability to absorb or unite with complement 
becomes evident only after a complex has been formed by the union 
of the two elements, antigen and antibody. The diagram must not 
be assumed to mean that the notch into which the complement fits 
symbolized necessarily an “atom group,” but merely expresses the 
idea of “ability to absorb alexin,” not assuming that this ability is 
either chemical affinity by means of a definite atom group or a mere 
physical change of molecular equilibrium permitting a specific com- 
plement absorption.43 
“Zone” Phenomena.—It will be seen from the preceding that 
the controversy between Ehrlich’s “amboceptor” conception and the 
“sensitization” idea of Bordet turns largely upon the existence of a 
so-called complementophile group of the thermostable antibody. For 
if it were the case that this antibody possessed an atom group which 
permitted it to unite with alexin, independent of previous union 
with antigen, it would go far to support Ehrlich’s view. One of 
the strongest arguments brought into the field in favor of such a» 
occurrence by Ehrlich’s followers is the phenomenon of Heisser and 
Wechsberg, which is usually spoken of as “complement deviation” 
(Komplement Ablenkung). 
In order to make the conditions underlying this phenomenon 
clear, it will be of advantage to consider for a moment the methods 
of determining quantitatively the amount of bactericidal antibody 
(sensitizer amboceptor) in any given immune serum, since it was in 
working with such titrations that ISTeisser and Wechsberg made their 
observations. 
In carrying out such measurements, it is customary to add in 
series, to constant amounts of bacteria, varying amounts of inacti- 
vated antiserum and constant amounts of complement or alexin. 
43 The diagram on page 174, though possibly not expressing with absolute 
accuracy the idea of sensitization, was devised because it will remove what 
seem to the writer frequent misconceptions of Bordet’s views. Statements 
are found in the literature which imply (Ehrlich, “Kraus und Levaditi 
Handbuch,” Vol. 1, p. 8) that Bordet assumes “dass das Komplement direkt 
an die Zelle angreift,” and deny that there is experimental evidence to sup- 
port this. It is perfectly true that there is no evidence to show such 
“direktes Angreifen” upon the unaltered cell, but there is evidence that 
this union takes place after the cell has absorbed the antibody, and no satis- 
factory evidence to show that the thermostable body is an intermediary, 
that is; forms a link as conceived in the amboceptor idea. 176 
INFECTION AND RESISTANCE 
These mixtures are set away in the thermostat for 3 to 4 hours, are 
then mixed with agar and plates are poured. The colonies which 
develop will give an indication of the number of bacteria killed in 
each mixture when compared with similar plates poured from tubes 
in which the same original amounts of bacteria had been mixed with 
alexin alone. The following table will exemplify such a test: 
Typhoid bacilli 
Typhoid 
antiserum 
inactive 
Alexin 
Result in colonies 
after 3 hours’ 
incubation 
Constant quantity 
.1 C. C. 
.07 C. C. 
Many thousand 
Constant quantity 
.01 c. c. 
.07 C. C. 
Many thousand 
Constant quantity 
.005 c. c. 
.07 C. C. 
150 colonies 
Constant quantity 
.001 c. c. 
.07 c. c. 
200 colonies 
Constant quantity 
.0005 c. c. 
.07 c. c. 
800 colonies 
Control I, constant quantity. . 
Control II, constant quantity.. 
.07 c. c. 
Many thousand 
Many thousand 
In this table it is noticeable that, although there has been con- 
siderable bactericidal action in the mixtures in which 0.005, 0.001, 
and 0.0005 c. c. of antiserum were used, the mixtures in which as 
much as 0.1 and 0.01 c. c. were present, and in which one would nat- 
urally expect a still greater antibacterial action, the contrary oc- 
curred. 
This surprising and curious phenomenon, showing that an ex- 
cess of antibody could actually be harmful to the bactericidal 
complex, was explained by Neisser and Wechsberg by the following 
reasoning. In tests like the one given above a limited amount of 
bacteria and alexin has been mixed with the enormous amount of 
antibody represented in the immune serum. Although bacteria can 
absorb more of this antibody than is necessary for their solution or 
destruction, nevertheless the higher concentration given in the table 
will contain quantities of “amboceptor” so far in excess of the 
amount that can be absorbed that much of it must remain free in 
the fluid. How this amboceptor, possessing a complementophile 
group, is able to anchor complement or alexin as wTell as that which 
has become united with the bacteria. In consequence, there being 
only a limited amount of complement, some of this is deviated from 
the amboceptor-antigen complexes by the free amboceptor, and is, in 
consequence, ineffective so far as bactericidal action is concerned. 
In the higher dilutions of the antiserum, in which no such excess is 
present, the complement will be concentrated upon the “attached” 
or “anchored” amboceptor, and greater efficiency will result. 
As to the accuracy of the observations of Heisser and Wechsberg 
there can be no question, and everyone who has occasion to carry out 
bactericidal tests with any frequency is sure to meet with the phe- 
nomenon again and again. But their explanation, which involves BACTERICIDAL PROPERTIES OF BLOOD SERUM 
177 
the assumption of union between free sensitizer or amboceptor, and 
alexin or complement, without the participation of antigen, cannot 
be accepted since, search as we may, through the extensive experi- 
mentation that this problem has inspired, there is no instance on 
record in which indisputable evidence of such an occurrence has 
been advanced. On the contrary, there is a mass of satisfactory 
evidence available which indicates clearly that amboceptor or sensi- 
tizer alone cannot absorb alexin, and the Neisser-Wechsberg expla- 
nation seems consequently to be merely an interesting and cleverly 
conceived but improbable possibility. 
What, then, is the explanation of the diminution of bactericidal 
elfect in the presence of an excess of sensitizer? We will see that, in 
the study of agglutinin and precipitin reactions, phenomena exactly 
analogous to the Neisser-Wechsberg effect have been noticed, in the 
case of the agglutinins, the so-called “pro-agglutinoid” zone being a 
case in point. For these phenomena, as well as for that of 
Neisser and Wechsberg, explanations have been advanced by the 
Ehrlich school, similar in principle in that they all depend upon 
more or less arbitrary assumptions regarding affinity between the 
reacting bodies. Such explanations, though not outside the realm of 
possibility, have, however, lost much force since it has been recog- 
nized that the reactions between serum antibodies and their antigens, 
in general, take place according to laws far more closely analogous 
to those governing reactions between colloids than to those governing 
chemical reactions in which the laws of definite proportions can be 
applied. And, indeed, the reacting substances in antigen-antibody 
complexes are, beyond doubt, of the nature of colloids. Now, in 
many precipitations resulting when two colloids are mixed, an ex- 
cess of one or the other factor will completely inhibit the occurrence 
of the precipitation; the reaction taking place only when definite 
proportions between the reacting bodies are present. The occurrence 
of such inhibition zones, due to an excessive concentration of one 
reagent, can be shown for agglutination and precipitation, exactly 
as it can in ordinary colloidal reactions, and it is more than likely 
that the Neisser-Wechsberg phenomenon is merely an example of a 
similar phenomenon. 
This inhibiting effect upon reactions produced by excess of either 
antigen or antibody has been particularly noticeable in complement 
or alexin fixation reactions which are considered in a separate sec- 
tion below. In this place, however, it is important to discuss the 
so-called zone phenomena as observed in researches done upon this 
occurrence in connection with alexin fixation experiments. Dean 44 
has extensively investigated the matter in parallel experiments in 
which precipitation reactions and complement fixation reactions 
were compared. He found that an optimum precipitation or an 
44 Dean. Zeit. f. Immunitats., 13, 1912, 84. 178 
INFECTION AND RESISTANCE 
optimum complement fixation could be obtained only when very 
definite relative proportions of antigen and antibody were used. If 
either of the reacting substances was increased from the amount 
in the optimum mixture, both reactions were diminished in in- 
tensity. Within certain limits, if both were diminished propor- 
tionately, intense reactions could still be obtained. Work of entirely 
■similar significance has recently been done in our laboratory by 
J. T. Parker with residue antigens such as those described at the 
end of the complement fixation section. These investigations def- 
initely contradict the explanation of Neisser and Wechsberg and 
bring such reactions into very close analogy with what we generally 
speak of as colloidal reactions. They will be taken up again in our 
discussion of the precipitation reaction in connection with which 
they have been most carefully studied. 
Looked at from this point of view, far from supporting the sup- 
position of a separate complementophile group and therefore of the 
“amboceptor” nature of the heat-stable lytic antibody, the Neisser- 
Wechsberg phenomenon indeed becomes rather a strong argument 
in favor of Bordet’s views, and against those of Ehrlich. For, by 
introducing the analogy between the lytic and bactericidal processes 
with colloidal reactions, it takes away much force from the supposi- 
tion that antigen-sensitizer alexin reactions take place according to 
laws of definite proportion, an idea which still underlies, though 
somewhat loosely, many of the more important views of antigen- 
antibody reactions as conceived on the basis of the analysis made 
by Ehrlich and his school. 
Gay has suggested also that the Neisser-Wechsberg phenomenon 
may well be explicable on the basis of the fixation of complement by 
precipitates. In a succeeding section we will discuss the fixation of 
alexin, which occurs when a dissolved protein is brought together 
with its specific antiserum. It is not impossible that this may occur 
when bacterial emulsions, from which a small amount of bacterial 
protein may well go into solution, are brought together with anti- 
serum in concentration. Under such conditions a reaction might 
readily occur which would lead to the fixation of alexin and its con- 
sequent deviation from the sensitized bacteria.45 
Quantitative Relationship of Sensitizer and Alexin.—Of all 
explanations considered, therefore, that of Heisser and Wechsberg 
seems to be the least likely. It would seem to us that Bordet’s 
interpretation of these facts is borne out indirectly by certain ex- 
periments of Morgenroth and Sachs46 themselves, in which the 
mutual quantitative relations between complement and “amboceptor” 
were studied. In these experiments it was shown that the more 
45 In this connection read also sections on zone phenomena in chapters on 
Alexin fixation and on precipitation and the section on the union of antigen 
with antibody. 
46 Morgenroth and Sachs. Berl. klin. Woch. No. 35, 1902. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
179 
highly cells were sensitized, the smaller was the quantity of comple- 
ment which was needed for their hemolysis, and vice versa, the less 
the sensitization (the smaller the quantity of amboceptor) the more 
complement was necessary to produce the same result. The fol- 
lowing extract from one of their protocols will illustrate this: 
BEEF BLOOD CELLS 5%, 1 C. C., AXTIBEEF GOAT SERUM, GUINEA-PIG COMPLEMENT 
Amount of 
amboceptor 
Relative amount of 
amboceptor 
Amount of 
complement for 
complete hemolysis 
.05 
1 
.008 
2 
4 
.0025 
.4 
8 
.0014 
A similar relation may be observed by all who have occasion to 
work with hemolytic reactions. In the present connection this seems 
to bear out Bordet’s interpretation, since, knowing the differences in 
functional efficiency of various complements for different hemolytic 
and bactericidal complexes, we could well expect that insufficient 
sensitization of a red cell or bacterial antigen, not particularly amen- 
able to the complement employed, might fail to absorb it completely 
out of the serum, thus giving a negative result which would simulate 
complete lack of affinity. 
This research of Morgenroth and Sachs seems further of funda- 
mental importance in its contradiction of the regularly progressive 
quantitative relations which strict adherence to the “amboceptor” 
idea would seem to impose. 
The quantitative relations here outlined have been diagrammat- 
icallv represented by Noguchi as follows: 
The essential point of difference between the opinions of Ehrlich 
and Bordet concerning the processes of hemolysis and bacteriolysis 
lies, as wTe have seen, in the conception of the union of alexin or com- 
plement with amboceptor or sensitizer. Although Ehrlich and his 
followers admit that the union of complement with amboceptor does 
not usually occur unless the amboceptor has previously united with 
the antigen, they still maintain that this may occasionally take place 
in the case of special complexes in which the complement may di- 
rectly unite with free amboceptor. This, we have seen, is the basis 
of the Neisser-Wechsberg conception of complement—“Ablenkung” 
or deviation, and of other ramifications of this theory. Bordet, on 
the other hand, consistently holds that alexin or complement is at- 
tached only by the complex antigen-sensitizer (antigen-amboceptor). 
The Conglutinin Effect.—In the controversy which this differ- 
ence aroused, an observation was reported by Ehrlich and Sachs,47 
which seemed to represent, as they themselves express it, an “Ex- 
47 Ehrlich and Sachs. Berl. klin. Woch., No. 21, 1902. 180 
INFECTION AND RESISTANCE 
perimentum Crucis” proving Ehrlich’s contention of the intermediary 
function of the amboceptor in contrast to Bordet’s “sensitization” 
idea. The facts, as they record them, are as follows: When fresh 
horse serum is added to guinea pig corpuscles, slight hemolysis re- 
sults. When inactivated ox serum alone is added to such corpuscles, 
Noguchi’s Diagram Illustrating the Quantitative Relations Between 
Antigen, Amboceptor and Complement 
(Taken from Noguchi,“Serum Diagnosis of Syphilis, ” Lippincott, Philadelphia, 
1910.) 
of course no hemolysis results. If the corpuscles are, on the other 
hand, exposed to the action of the inactive ox serum, together with 
fresh horse serum, very active hemolysis is brought about. Appar- 
ently the ox serum sensitizes (or furnishes amboceptor to) the 
guinea pig corpuscles, rendering them amenable to the action of 
the complement in the fresh horse serum. In other words, inacti- 
vated ox serum can be reactivated by the addition of fresh horse BACTERICIDAL PROPERTIES OF BLOOD SERUM 
181 
serum. From this one would expect that if the guinea pig cells 
were exposed to inactive ox serum, then separated from the serum 
by centrifugalization and fresh horse serum subsequently added, 
hemolysis would ensue. However, this was not the case. When 
the cells were so treated it was found that they had not been sensi- 
tized, and, what is more, it could be shown that the ox serum so 
employed had lost no/ie of its ability to produce strong hemolysis 
when added to another complex of cells and fresh horse serum. 
Ehrlich and Sachs concluded that this experiment definitely showed 
the ability of the amboceptor in the ox serum to unite with alexin 
independently. The relation to the cell occurred only after the 
union of the amboceptor in the ox serum and the complement in 
the horse serum had been established, and if their interpretation 
is correct, of course, it constitutes strong evidence against the general 
principle of “sensitization” as conceived by Bordet. 
This apparent inability of the corpuscles to absorb amboceptor 
independently out of the inactivated ox serum, and the fact that 
hemolysis results only if the corpuscles, ox serum, and fresh horse 
serum are all simultaneously present, are extraordinary and not at 
all in keeping with the preceding work of Ehrlich and Morgenroth, 
and indeed with experience of these phenomena in general. It is log- 
ical therefore to examine more closely the peculiar conditions main- 
tained in these experiments before applying the reasoning deduced 
from obviously different phenomena to their explanation. 
Bordet and Gay 48 accordingly studied the Ehrlich-Sachs phe- 
nomenon carefully and obtained results which confirmed the experi- 
mental data of these writers but cast much doubt upon the validity 
of their conclusions. 
In going over the experiments of Ehrlich and Sachs, Bordet and 
Gay made an observation which had apparently escaped the atten- 
tion of the former investigators. Heated bovine serum has but a 
slight agglutinating power for guinea pig corpuscles. Fresh horse 
serum agglutinates them only slightly and slowly. On the other 
hand a mixture of the two sera agglutinates them very rapidly and 
completely. The bovine serum apparently possessed an accelerating 
or fortifying influence both upon the weakly active normal hemoly- 
sins and agglutinins in the horse serum. Bordet and Gay conse- 
quently suspected that this property might be due to an undescribed 
substance, peculiar to the bovine serum. To eliminate the uncertain 
elements obtaining in experiments in which normal sensitizer is 
used they now experimented with guinea pig corpuscles, anti-guinea 
pig sensitizer (from a rabbit immunized with guinea pig blood cells) 
and guinea pig alexin. 
They found that sensitized guinea pig cells are hemolyzed by 
guinea pig alexin very slowly and imperfectly, as is often the case 
48 Bordet and Gay. Ann. de I’Inst. Past., Yol. 20, 1906, p. 467. 182 
INFECTION AND RESISTANCE 
when the alexin comes from the same animal species as the cells. 
When heated bovine serum was added to the complex of sensitized 
cells and alexin, rapid agglutination and hemolysis resulted. Their 
experiments may be tabulated as follows: 
1. Cells + guinea pig alexin + heated bovine serum = no agglutination; 
very slight hemolysis on next day. 
2. Cells + sensitizer + heated bovine serum = slight agglutination; no 
hemolysis. 
3. Cells + sensitizer + alexin + bovine serum — powerful agglutination 
and complete hemolysis in 10 minutes. 
4. Cells + sensitizer + alexin = very slight agglutination and incomplete 
hemolysis in 30 minutes. 
5. Cells + sensitizer = slight agglutination; no hemolysis. 
In tube (1) the slight hemolysis was due to the small amount of 
normal sensitizer present in the bovine serum, and the slight agglu- 
tination in tube (5) is referable to the agglutinating power of the 
sensitizer. In tube (3) we see the powerfully accelerating effects 
exerted both upon agglutination and hemolysis when bovine serum 
acts upon sensitized corpuscles in the presence of alexin. 
Bordet and Gay’s interpretation of the Ehrlich-Sachs phenom- 
enon, in the light of these new experiments then, is, in their own 
words, as follows: “When guinea pig corpuscles are added to a 
mixture of the two sera they are affected by the sensitizer of the 
horse serum and, to a certain extent, by the sensitizer in the heated 
bovine serum. This second sensitizer is, however, superfluous. Its 
presence is by no means necessary for the experiment. When this 
sensitization is effected the corpuscles are then in condition to fix 
the horse alexin. This alexin, however, has only slight hemolytic 
power. But once the Corpuscles have become sensitized and laden 
with alexin they are modified in their properties of molecular ad- 
hesion to such an extent that they become able to attract a colloidal 
substance of bovine serum, which unites with them. The adhesion 
of this new substance produces two results: it causes the blood 
corpuscles to be more easily destroyed by alexin and also agglutinates 
them energetically. Consequently, a powerful clumping, followed 
by hemolysis, is observed.” 
Bordet and Gay, therefore, assume that the action of the bovine 
serum is due to a new substance which they speak of as “bovine 
colloid.” This substance resists heating to 56° C., is probably al- 
buminous, and has the property of uniting with cells that are laden 
with sensitizer and alexin, but remains free in the presence of nor- 
mal or merely sensitized cells. 
They fortify this opinion by showing experimentally that thp 
“colloid” is removed from bovine serum by absorption with sensi- 
tized bovine corpuscles which have been treated with horse alexin.49 
49 Browning (Wien. klin. Woch., 1906) had shown that horse alexin may 
be absorbed by sensitized beef cells without causing hemolysis. BACTERICIDAL PROPERTIES OF BLOOD SERUM 
183 
Bordet and Streng50 later studied this “colloid” more thor- 
oughly and have suggested for it the name “conglutinm.” Streng 51 
later showed that the agglutinating action of this substance could be 
shown not only for sensitized and “alexinized” red blood cells, but 
also for similarly treated bacteria, and that conglutinins were pres- 
ent not only in bovine serum, but in that of goats, sheep, antelopes, 
and a number of other herbivores, but apparently absent in cats, 
dogs, guinea pigs, and birds. 
The body described by these workers as conglutinin is probably 
identical with a similar heat-stable serum component reported by 
Manwaring 52 and called by him “auxilysin.” 
50 Bordet and Streng. Centralbl. f. Bakt., Orig. Yol. 49, 1909. 
“ Streng. Zeitschr. f. Immunitiitsforscli., Orig. Yol. 2, 1909, p. 415. 
52 Manwaring. Centralbl. f. Bakt., 1906; Orig. Yol. 42. CHAPTER VII 
FURTHER DEVELOPMENT OF OUR KNOWLEDGE CON- 
CERNING COMPLEMENT OR ALEXIN. COMPLE- 
MENT FIXATION 
Origin of Alexin in Leukocytes.—It will be remembered that 
Buchner in his first studies upon the “alexin” compared its action 
to that of an enzyme or ferment, and suggested that the source of 
this substance might possibly be found in the white blood cells. 
This thought was very obviously suggested by the observation that 
bacteria were destroyed within the white blood cells, after phago- 
cytosis, by a process analogous in many ways to that by which they 
were destroyed by the serum constituents. Hankin,1 in an elaborate 
study dealing with the problem, maintained the leukocytic origin 
of alexin on the basis of the observation that increased bactericidal 
properties closely followed upon the heels of periods of leukocytosis, 
lie assigned the particular property of alexin production to the 
eosinophile cells, proposing for them the designation “alexocytes.” 
Further study, however, has not justified such an association with 
the eosinophiles, and Hankin’s opinion has not been experimentally 
upheld. 
After Ilankin the problem occupied the attention of a number 
of other investigators, and many of them succeeded in showing that 
there was, indeed, an increased bactericidal power in exudates rich 
in leukocytes, and further that bactericidal substances could be 
directly extracted from leukocytic emulsions. We refer particularly 
to the early work of Denys and Havet,2 of Hahn,3 of Van de Velde,4 
and others, studies which will be described in our chapter on 
phagocytosis. This work was done before the complex nature of the 
bactericidal constituents of serum had been demonstrated and before 
the work of Schattenfroh and others had shown that the bactericidal 
substances extracted from leukocytes were of a nature quite distinct 
from the active elements of the serum, and were independent of the 
participation of alexin. Although these earlier investigations cannot 
properly be regarded, therefore, as proving the leukocytic origin of 
1 Hankin. Centralbl. f. Bakt., Yol. 12, 1892. 
2 Denys and Havet. La Cellule, Vol. 10, 1894. 
3 Halm. Archiv f. Hyg., Yol. 25, 1895. 
4 Van de Velde. La Cellule, Vol. 10, 1894. 
184 FURTHER DEVELOPMENT OF KNOWLEDGE 
185 
alexin, Metchnikoff and his school have nevertheless adhered to this 
conception for various additional reasons. 
Metchnikoff distinguishes between two kinds of alexin—the 
microcytase, which is the bactericidal complement or alexin, and 
is supposed to originate from the microphages or polynuclear leu- 
kocytes, and the macrocytase, which represents the hemolytic and 
cytolytic alexin or complement, and originates from the mononuclear 
cells or macrophages. As in the case of the bactericidal alexin, ex- 
traction methods have been employed to demonstrate that the hemo- 
lytic alexin took its origin in the macrophages, and at MetchnikofFs 
suggestion, Tarassewitch 5 prepared hemolytic substances by extract- 
ing spleen tissue and other “macrophagic organs” in various ways. 
Here again the identity of the hemolytic extracts with serum hemoly- 
sins has been placed in doubt. Korschun and Morgenroth6 have 
shown that the hemolytic organ extracts were heat stable and alcohol 
soluble; Donath and Landsteiner,7 and others, have obtained similar 
results. It would be quite thankless to review the extensive literature 
which has accumulated upon this point. It would seem, in summariz- 
ing it, that no definite proof of the presence of true, active alexin, 
either hemolytic or bactericidal, within the leukocyte or mononuclear 
cells has been brought by methods of extraction, and the apparently 
positive results reported by earlier observers are adequately explained 
by the discovery of the heat-stable and non-reactivable bactericidal 
and hemolytic substances in extracts of such cells by Schattenfroh, 
Korschun and Morgenroth, and many others. It appears, moreover, 
from these investigations that probably the intracellular substances 
by which the digestion of ingested bacteria or blood cells is brought 
about are of a nature entirely distinct from that of the serum anti- 
bodies and alexins. A very ingenious demonstration of this is found 
in an experiment first made by Neufeld. Neufeld 8 allowed leu- 
kocytes to take up highly sensitized red cells. Instead of undergoing 
prompt hemolysis, as they would if small amounts of alexin had been 
added, they were slowly broken up without hemolysis, fragments of 
hemoglobin remaining after complete morphological disintegration 
of the erythrocytes. At no time were intraphagocytic “shadow” 
forms observed. 
The failure to extract alexins from dead leukocytes does not, 
however, preclude the possibility of the secretion of alexins by living 
leukocytes. This point is one which is, of course, much more diffi- 
cult to investigate directly. Indirectly the increased bactericidal 
properties of exudates rich in leukocytes, as found by Denys and 
Havet, would point in this direction. However, even this is not 
5 Tarassewitch. Cited from Metchnikoff. 
6 Korschun and Morgenroth. Berl. klin. Woch., No. 37, 1902. 
7 Donath and Landsteiner. Wien. klin. Rundschau, Yol. 40, 1902. 
8 Neufeld. Arb. a. d. kais. Gesundheitsamt, Yol. 28, 1908, p. 125. 186 
INFECTION AND RESISTANCE 
conclusive, since at the time when these investigations were carried 
out no discrimination was made between the bactericidal serum sub- 
stances and those other “endolysins” which might well have been 
extracted from the accumulated white blood cells. The writer some 
years ago attempted to approach this problem directly by keeping 
leukocytes alive in inactivated serum and in Ringer’s solution at 
37.5° C. for several days in the hope that, after 48 hours, alexin, 
hemolytic or bactericidal, might appear in these fluids. The experi- 
ments were entirely negative, but were regarded as inconclusive, since 
it was impossible to determine accurately how long, or in what pro- 
portion, the leukocytes had remained alive. 
One of the basic premises of Metchnikoff’s theory on the nature 
of alexin consists in the conception that alexin is not found in the 
circulating blood plasma, but appears only when there has been leu- 
kocytic injury, as in the clotting of blood or in the “phagolysis” 
which, as we have seen in the chapter on phagocytosis, usually occurs 
after foreign substances have been injected into the peritoneum, pre- 
ceding a local accumulation of leukocytes. This point of view seems 
to be rendered improbable because of the rapid hemolysis which 
occurs when we inject sensitized red blood cells into the circulation 
of an animal, but we might here, too, assume a preliminary injury 
to white blood cells resulting from the intravenous injection of 
foreign material. 
Much less likely to be accompanied by cell injury is the method 
of obtaining blood serum by creating an area of artificial edema bv 
ligating a limb—or, as in Metchnikoff’s 9 experiments, the ear of 
a rabbit. And, indeed, in edema fluids so obtained little or no alexin 
is ordinarily found. This fact has been interpreted in favor of 
Metchnikoff’s views, as has also the curious absence of alexin in the 
aqueous humor of the anterior chamber of the eye.10 In this fluid 
nO alexin is present under normal conditions, but if puncture is prac- 
ticed, and the fluid again taken after a period of three or four hours, 
alexin is now found, probably, according to Metchnikoff’s school, be- 
cause of the coincident entrance of leukocytes into this space. It is 
conceivable, however, that the aqueous humor may be free from 
alexin for other reasons than the absence of leukocytes; and an injury 
which is followed by the invasion of leukocytes is pretty sure to be 
followed also by the entrance of the fluid elements of the blood; i.e., 
alexin. 
Is there Alexin in the Circulating Blood?—Much experimental 
work has been done in which it has been attempted to demonstrate 
directly that the blood plasma contains no complement or alexin. 
The most important investigation of this kind is that carried out 
9 Metchnikoff. Ann. de VInst. Past., Yol. 9, 1895. Bordet. Ann. de VInst. 
Past., Yol. 9, 1895. 
10 Metchnikoff. Loc. cit. Mesnil. Ann. de VInst. Past., Yol. 10. FURTHER DEVELOPMENT OF KNOWLEDGE 
187 
by Gengou 11 in 1901. It was Gengou’s primary purpose to obtain 
the plasma of mammals in such a way that no cell injury would 
occur. This he accomplished by special methods in which coagula- 
tion was avoided without the addition of foreign anticoagulants like 
hirudin, etc. His technique was, in essence, as follows: He took 
the blood directly through a paraffined cannula into tubes that had 
been coated with paraffin, and centrifugalized it at low temperatures 
until cell free. This plasma, taken from the paraffin tubes, quickly 
clotted, and the material with which the experiments were done 
actually consisted of blood serum. Upon examining the serum so 
obtained, he found that it exerted practically no bactericidal action. 
As a result of this investigation he claims to have demonstrated 
the truth of Metclmikoff’s contention that the circulating blood 
plasma contains no alexin. 
If borne out, it is true that Gengou’s results would very power- 
fully support this theory, and for this reason a large number of ex- 
periments have been made since then, with the same end in view. 
In all such investigations the technical procedures are extremely 
difficult and, as Addis 12 has recently said, in our opinion quite cor- 
rectly, it would be impossible to carry out bacteriolytic or hemolytic 
experiments with mammalian paraffin plasma without obtaining 
coagulation, and for this reason most of the writers who have re- 
peated Gengou’s experiments have worked, as he did, not with 
plasma, but with serum. Falloise,13 following Gengou’s method 
exactly, obtained results diametrically opposed to those of Gengou; 
Schneider,14 also with the same technique, failed to confirm Gengou’s 
results; Herman,15 on the other hand, confirms Gengou. 
In order to overcome the technical difficulties encountered in 
working with mammalian plasma a number of writers have more 
recently experimented with bird blood, which, as is well known, 
coagulates much less easily than does mammalian blood. Hewlett,16 
who worked with goose plasma and peptone plasma, could not con- 
firm Gengou’s results. Lambotte,17 examining the plasma of chick- 
ens, found no difference between the serum and plasma in their 
contents of bactericidal alexin, as measured against cholera spirilla. 
Von Dungern, working with fish plasma, obtained similarly negative 
results, and recently Addis, in a careful comparative study of 
chicken plasma, found no evidence of differences between plasma 
and serum in either the bactericidal or the hemolytic alexin. As far 
as we can tell at present, therefore, we cannot accept, as conclusively 
11 Gengou. Ann. de l’Inst. Past., Yol. 15, 1901. 
12 Addis. Jour. Inf. Dis., Yol. 10, 1912. 
“Falloise. Bull, de VAcad. Roy. de Med., 1905, p. 230. 
14 Schneider. Archiv f. Hyg., 1908, Vol. 65, p. 305. 
15 Herman. Bull, de VAcad. Roy. de Med., 1904, p. 157. 
16 Hewlett. Archiv f. exp. Path. u. Pharmk., 1903, Vol. 49, p. 307. 
"Lambotte. Centralbl. f. Bakt., I, Orig., 1903, Vol. 34, p. 453. 188 
INFECTION AND RESISTANCE 
proven, the contention that the circulating plasma contains no alexin. 
Nevertheless the Metchnikoif school have not been discouraged by 
the various contradictions of Gengou’s work, found in the experiments 
we have enumerated, because they are not satisfied that the technique 
of other workers has conclusively excluded cell injury. Owing to 
the great difficulties of investigations of this kind, when carried out 
with mammalian blood, it is not impossible that they are justified in 
this, but nevertheless the assumption of the absence of alexin in the 
plasma finds so many objections in other observations that the bur- 
den of proof would certain rest with Gengou and his supporters. 
Not the least important of these objections, it seems to us, is based 
on the very simple experiment of injecting bacteria into the veins of 
a living animal and finding a very rapid and active phagocytosis. 
And considering the very probable participation of alexin in the 
opsonic functions this would seem to point strongly toward the pres- 
ence of these substances in the circulating blood. The evidence also 
furnished by the recent developments of our understanding of an- 
aphylaxis would further tend to strengthen our belief in the presence 
of alexin or complement in the normal circulation. For, in the 
process, as we shall see in a later chapter, complement plays an 
important role. When 3 per cent, salt solution is administered (as 
in Friedberger’s experiments), and the action of complement is 
thereby inhibited, anaphylactic shock may be greatly diminished. 
It has also been claimed, chiefly by Walker 18 and by Henderson 
Smith,19 that, as serum stands upon the clot it at first gains in alexin 
or complement contents, an occurrence which they attribute to the 
liberation of alexin from the leukocytes. This observation has not 
been universally borne out and, even were it unquestionable, it might 
be dependent upon any one of the numerous factors involved in the 
complicated process of coagulation rather than upon leukocytic 
changes only. 
Alexin and Glandular Organs.—The failure to obtain definite 
proof of the origin of alexin from the white blood cells has led to 
search for the source of these substances in various organs. An in- 
teresting series of investigations on this subject are those of Mile. 
Louise Fassin,20 who believes that she has found reasons for defi- 
nitely associating the thyroid gland with alexin production. She 
found that the subcutaneous injection of thyroid extract into dogs 
and rabbits was followed by a rapid increase of alexin, both hemo- 
lytic and bactericidal, and that the same thing was true when 
thyroid substance was administered by mouth. When the thyroid 
gland was removed from rabbits a reduction of alexin resulted. 
Although important, these researches do not necessarily prove that 
18 Walker. Jour. Hyg., Yol. 3, 1903. 
19 Smith. Proc. Roy. Soc., Series B, Yol. 79, 1906. 
20 Louise Fassin. C. R. de Soc. Biol., Yol. 62, 1907. FURTHER DEVELOPMENT OF KNOWLEDGE 
189 
the thyroid can be looked upon as a source of alexin, and, indeed, 
Fassin gives experimental results without drawing any very sweep- 
ing conclusions. It might well be that the thyroid secretion is 
simply concerned in stimulating the production of alexin from 
another source. Marbe 21 has similarly associated the thyroid gland 
with the production of opsonins, which, when we consider the 
probable identity of alexin and normal opsonin, may be taken as 
a confirmation of Fassin’s work. 
Of great interest also are the series of investigations which asso- 
ciate the liver with the production of alexin. The basis of such 
investigations is found in the observations made by Morgenroth and 
Ehrlich 22 that there is a diminished production of complement or 
alexin in dogs subjected to phosphorus poisoning, with consequent 
degeneration of the liver. The first investigator to study this ques- 
tion experimentally was Nolf.23 ISTolf tried to approach it by extir- 
pating the liver in dogs, and found that his results were unreliable 
by this method. He then experimented with rabbits and found that 
when the liver was extirpated in these animals and the vena cava 
anastomosed with the portal vein (Eck fistula) the animals would 
survive for three or four hours. This period, though short, was suffi- 
cient to show definite changes in the blood. Taken just before death 
it differed from that taken just before the operation in a number of 
important respects. There was relative incoagulability, there was 
autohemolysis, and with these there occurred an extreme fall of 
alexin or complement. Serious objections may be brought against 
Nolf’s experiments. In the first place the operation as performed by 
him results in shock and injury so profound that rapid death ensues, 
conditions under which not only thef complement-producing functions 
but all functions, secretory and otherwise, are reduced. Muller 24 ob- 
jects to Nolf’s experiments chiefly for the reason that he did not pre- 
vent the absorption of toxic substances from the intestine, materials 
which could now enter the general circulation without any longer be- 
ing neutralized by the liver functions. Muller, for this reason, re- 
peated Half’s work but, by a complicated technique, temporarily shut 
off the intestinal circulation in addition to extirpation of the liver. 
He found, in agreement with Holf, that exclusion of the liver from 
the circulation resulted in the prompt diminution of complement or 
alexin. In all such experiments, however, the very profound shock 
which necessarily occurs in the animals wrould seem to us to vitiate 
the results. Moreover, Liefmann 25 has repeated Muller’s experi- 
ments without being able to obtain the same results. Hot satisfied 
with these experiments, however, Liefmann experimented on frogs; 
21 Marbe. C. R. de la Soc. Biol., Yols. 64, et seq., 1908-1909. 
22 Morgenroth and Ehrlich. In Ehrlich’s “Gesammelte Arb.,” etc. 
23 Nolf. Bull, de VAcad. de Science de Belg., 1908. 
24 Muller. Centralbl. f. Bakt., Yol. 57, 1911. 
25 Liefmann. Weichhart’s Jahresbericht, Yol. 8, 1912, p. 155. 190 
INFECTION AND RESISTANCE 
in whom, as Friedberger has shown, extirpation of the liver is not 
so rapidly fatal as in warm-blooded animals. He removed the livers 
of frogs in a number of cases and, although his animals lived about 
a week, there was no definite diminution of the hemolytic properties 
of the serum. It seems, therefore, that the origin of alexin in the 
body is by no means settled and requires a considerable amount of 
further investigation. 
Nature of Alexin.—Equally unsatisfactory have been the at- 
tempts to define the chemical nature of the complement or alexin. 
In the investigations dealing with the hemolytic action of cobra 
venom it seemed at first as though a clue to this problem had been 
found. Flexner and Noguchi 20 made the interesting observation 
that cobra poison alone does not hemolyze the blood cells of certain 
animals, namely those of cattle, goats, or sheep, if these cells are 
washed entirely free of serum. This seemed to suggest that the 
serum of these animals contained some activating substance. It also 
seemed to indicate that the cells of other animals, which were easily 
hemolyzed, even when entirely freed of serum, might contain such 
an activating substance within themselves. The behavior of this 
activating substance toward snake-venom hemolysis was therefore 
very similar to the action of complement, except in one important 
respect, namely, as Calmette 27 showed, almost all sera were ren- 
dered more efficient for the activation of snake venom when heated 
to 65° C., whereas complementary properties of sera for other hemo- 
lyzing complexes are, of course, destroyed at 56° C. Eyes,28 on 
further studying these phenomena, extracted the red blood cells of 
rabbits and other animals whose cells were hemolyzed by snake 
venom alone, by shaking them up with distilled water, and showed 
that, with these extracts, he could activate the venom against ox, 
goat, and sheep corpuscles, cells which were not ordinarily hemolyzed 
by the venom without the addition of serum. Similar activation of 
the venom with extracts of the ox, goat, or sheep corpuscles was not 
possible. He concluded from this that the blood cells of the rabbit, 
dog, guinea pig, and man possessed an “endocomplement” for the 
snake venom; that is, a complementary substance contained within 
the cells, while in the other species it was found in the activating 
serum only. 
The thermostability of such venom “complements” encouraged 
him to attempt their isolation, and he found that they were ether- 
soluble, indicating their lipoidal nature; and, finally, after several 
negative attempts with activation by other lipoids, he determined 
that lecithin, added to the corpuscles and the snake venom, brought 
26 Flexner and Noguchi. Jour. Exp. Med., Yol. 6, 1902; TJniv. Pa. Med. 
Bull., 1902 and 1903. 
27 Calmette. C. R. de VAcad. des Sciences, p. 134, 1902. 
28 Kyes. Berl. klin. Woch., Nos. 38 and 39, 1902. Kyes and Sachs. 
Berl. klin. Woch., Nos. 2-4, 1903. FURTHER DEVELOPMENT OF KNOWLEDGE 
191 
about a rapid hemolysis. This seemed to explain both why heated 
serum could activate the venom in some cases, and why some varie- 
ties of blood cells could be hemolyzed without serum, since lecithin 
is a substance widely distributed both in the liuids and cells of the 
animal body. His further studies seemed to show that, by proper 
chemical manipulation (bringing together cobra poison with lecithin 
in chloroform solution), he could produce a combination of the two 
which he called “cobra lecithid,” a substance which apparently “acti- 
vated” cobra venom. He conceived it as the “amboceptor-comple- 
ment” complex of the cobra hemolysin, which acted hemolytically 
upon all varieties of blood cells. 
These researches of Kyes aroused much interest, chiefly because 
they seemed to furnish an example of a chemically definable com- 
plement, lipoidal in its constitution. Recent researches by Von 
Dungern and Coca,29 however, seem to prove that, while Kyes’ ex- 
perimental facts were perfectly accurate, his conclusions do not seem 
to have been warranted. Von Dungern and Coca showed that the 
cobra venom contains a lipoid-splitting ferment which acts upon the 
lecithin, liberating substances from it which hemolyze in the same 
way as do many other non-specific substances. The cobra-lecithid, 
according to this, would represent merely a lecithin derivative which 
happens to have hemolytic action without any specific relationship to 
the hemolytic properties of the venom itself. Thus, even in this case, 
unfortunately, we are not in possession of facts which bring us 
nearer to a chemical understanding of the complementary substances 
of alexins. 
In the further development of attempts to define alexin or com- 
plement chemically, two further researches are of importance, 
namely, those of Von Liebermann 30 and of Noguchi. In both in- 
vestigations it is suggested that the alexin may consist of a combina- 
tion of soaps and proteins. Noguchi 31 showed that the hemolytic 
organ-extracts described by various observers were soaps, a possi- 
bility which had been previously considered by Sachs and Kyes.3 
Noguchi further established analogies between his soaps and com- 
plement as follows: Sensitized blood cells are hemolyzed by mix- 
tures of soaps and inactivated guinea pig serum, while normal ery- 
throcytes are not hemolyzed by similar mixtures; furthermore, like 
normal complement, such serum-soap mixtures are inactivated by 
prolonged preservation and by heating at 56° C. Objections were 
soon made to the findings of both Noguchi and Von Liebermann by 
Hecker,33 whose experiments seemed to show that when sensitized 
blood cells were thoroughly washed free of serum soaps did not have 
29 Von Dungrem and Coca. Munch, med. JVoch., 1907, p. 2317. 
30 yon Liebermann. Biochem. Zeitschr., Vol. 4, 1907. 
31 Biochem. Zeitschr., Vol. 0, 190/. 
32 Sachs and Kves. Berl. Bin. Woch., 2-4, 1903. 
33 Hecker. Arh. a. d. konig. Inst. f. exp. Ther., Helt 3, 190/. 192 
INFECTION AND RESISTANCE 
this hemolyzing action, and Friedemann and Sachs 34 claimed that 
they were unable in any case to inactivate the hemolytic serum soap 
mixtures by heating to 56° C. These writers, as well as others, 
attribute Noguchi’s results to the fact that the sera which he used to 
produce his “artificial complement,” i. e., his serum soap mixtures, 
were heated to 50 to 51° C. only, a fact which would justify doubt of 
complete inactivation. Knaffl-Lenz 35 has more recently carried out 
experiments on the same question. Ilis results seem to show that the 
hemolytic action exerted by fatty acids or soaps is a phenomenon 
quite incomparable to true complement action, and that these hemoly- 
sins are heat stable, remaining unchanged by heating at 56° C. We 
have referred in a number of places to the analogy between alexins 
and ferments or enzymes. The chief objection to this conception 
formerly brought forward was based upon the fact that the comple- 
ment or alexin, unlike an enzyme, was used up during its reactions, 
and that a definite quantitative relationship existed between the 
alexin and the amount of cells or bacteria upon which it could act. 
Recent experiments by Kiss 36 seem to show that this quantitative re- 
lationship is not as strict and regular as was formerly supposed. He 
showed that the action of complement depends very largely upon its 
concentration. For instance, to cite his work directly: “0.05 com- 
plement is sufficient to hemolyze completely a definite quantity of 
sensitized blood cells if the experiment is done in a total volume of 
5 c. c. 0.02 c. c. of complement gives absolutely no hemolysis in a 
similar volume. When, however, the total volume is reduced to 2.5 
c. c., then 0.02 c. c. of the complement begins to act, and it produces 
complete hemolysis if the total volume is reduced to 1.25.” In far- 
ther developing this observation he showed that, if sufficiently con- 
centrated, a very small amount of complement can act upon an 
extremely large amount of red blood cells, an amount incomparably 
larger than those acted upon in more dilute solutions. These obser- 
vations would tend to strengthen considerably the conception of the 
ferment nature of alexins in general. Moreover, investigations, such 
as those of Northrup discussed in the Section dealing with the union 
of toxin and antitoxin, have considerably changed the older ideas in 
which it was assumed that enzymes were not used up or—more 
properly speaking—inhibited in the course of their action. 
Kiss’ observations are furthermore in agreement with the inves- 
tigations of Liefmann and Cohn,37 whose work we have mentioned 
in the preceding chapter on Cytolysis. These writers assert that the 
fixation of complement during hemolysis is not due to its chemical 
union with the sensitized cells, but is due to fixation by the end 
34Friedemann and Sachs. Biochem. Zeitschr., Yol. 12, 1908. 
35 Knaffl-Lenz. Biochem. Zeitschr., Yol. 20, 1909. 
36 Kiss. Zeitschr. f. Imm., Yol. 3, 1909. 
37 Liefmann and Cohn. Zeitschr. f. Imm., Yol. 8, 1911. FURTHER DEVELOPMENT OF KNOWLEDGE 
193 
products of the reaction; in other words, by the stromata of the red 
cells and possibly by other substances given up by these cells. A 
further factor contributing to the disappearance of complement in 
such reactions is, they claim, its rapid deterioration at 37° to 40° C., 
when diluted. If they are right, these considerations also remove 
important objections to the conception of complement as a ferment. 
It is clear, therefore, that although we have gained much detailed 
information regarding the functional activity of the complement or 
alexin, and may assume, in a general way, that its action is similar 
to, if not identical with, that of an enzyme, we are nevertheless still 
very much in the dark concerning its chemical nature. The same 
thing may be said in regard to its physical characteristics. One 
method of investigating the physical properties of complement has 
been that of filtration. It may be remembered that one of Ehrlich 
and MorgenrotlTs 38 arguments in favor of the multiplicity of com- 
plement was the fact that, when goat serum was filtered through a 
Pukal candle, the complement which was active upon rabbit corpus- 
cles was retained, while that which acted upon guinea-pig cells 
passed through. Immune bodies or amboceptor always passed 
through. 
Vedder,39 in similar experiments upon bactericidal complements, 
claims to have been able, in the same way, to separate the comple- 
ments acting upon different bacteria. The problem has been more 
recently investigated by Muir and Browning.40 Their conclusions 
are briefly as follows: In the early stages of filtration through a 
Berkefeldt filter complement is often completely held back. After 
continued filtration it begins to pass through. If the complement is 
inactivated by the addition of hypertonic salt solution (5 per cent.), 
it passes through, and the filtrate can be reactivated by dilution to 
isotonicity. Sensitizer or amboceptor always passes through. Just 
how these experiments are to be interpreted is a little obscure. The 
fact that the addition of salt renders the complement capable of 
passing through the filter would seem to indicate that its original 
inability to permeate ‘did not depend upon the size of the molecule. 
On the other hand, it is also possible that the addition of salt to the 
complement may increase its dispersion in such a way that the indi- 
vidual particles are rendered smaller. This, however, is purely 
speculative, and we are at a loss for a fully satisfactory explanation 
of the results of Muir and Browning. We have repeated some of the 
experiments of Muir and Browning and, in substance, confirmed 
their results. It is our opinion that new filters remove complement 
by adsorption, just as this is accomplished when complement is 
Shaken up with kaolin or other finely suspended material. 
38 Morgenroth and Ehrlich. Ehrlich’s “Gasammelte Arbeiten,” etc. 
39 Yedder. Jour. Med. Res., Yol. 9, 1903. 
40 Muir arid Browning. Jour. Path, and Bad., Vol. 13, 1909. 194 
INFECTION AND RESISTANCE 
Inactivation of Alexin by Heat.—While, of course, it is one of 
the basic attributes of alexin that 56° C. for one-half hour destroys 
it, the process by which this moderate amount of heating inactivates 
is not any more definitely understood than is the effect of similar or 
slightly greater degrees of heat upon toxins or enzymes. One of the 
earlier ideas was that possibly heating to these moderate tempera- 
tures, while not causing visible coagulation, nevertheless produced 
a condition of diminished dispersion, the proteins in which the com- 
plement was contained being in an entirely different state of col- 
loidal equilibrium, with possible surface tension changes. As a 
matter of fact, a purely physical interpretation of inactivation was 
encouraged by the fact that other physical methods, such as pro- 
longed shaking, allowing to stand without heat, and the addition of 
such things as neutral salts, could lead to definite, permanent, or 
temporary inactivation of the alexin. It is not unlikely, however, 
that it will be necessary to change these ideas in view of our more 
recent knowledge concerning the possibilities of chemical union of 
proteins with the ions of salts, of acids and of bases, and relation- 
ships between salt contents, on the one hand, and hydrogen ion con- 
centration and temperatures at which inactivation occurs will have 
to be taken into consideration. 
Effects of Salt Concentration on Alexin.—That the addition of 
salts of various kinds in quantities greater than isotonicity (or more 
than the equivalent of 0.85 = 0.9 per cent. NaCI)41 exerts a pro- 
found action upon the activity of complement is well known. 
Nolf 42 noted this in 1900, and the problem has been studied since 
that time by many investigators. Yon Lingelsheim,43 who studied 
it in connection with his work on the refutation of the “osmotic” 
theories of immunity, showed that increasing the salt contents of 
serum (KN03, NaCl, K2HP03, etc.) progressively diminished its 
bactericidal power. Hektoen and Ruediger44 also, after a very 
thorough study of this phenomenon, conclude that the action of the 
salts in such cases is exerted upon the alexin or complement and not 
upon the heat-stable sensitizers, and that it probably depends upon 
“physicochemical” causes. However, the manner in which such 
salt-inactivation is brought about is, to a great extent, obscure. There 
is no visible precipitation from serum after the addition of salts 
sufficient in quantity to weaken its action. Nothing is, as far as we 
can tell, removed from solution, and yet there is temporary inactiva- 
tion which, at the same time, renders the complement filtrable, facts 
from which we can only surmise some physical alteration. 
41 Alexin can be preserved in the refrigerator for long periods if hyper- 
tonic salt solution (15 to 25%) is added. It will again become active if iso- 
tonicity is restored with distilled water. 
42 Nolf. Ann. Past., Yol. 14, 1900. 
43 Y. Lingelsheim. Zeitschr. f. Hyg., Yol. 37, 1901. 
44 Hektoen and Ruediger. Jour, of Inf. Dis., Yol. 1, 1904. FURTHER DEVELOPMENT OF KNOWLEDGE 
195 
Inactivation of the complement also follows the removal of salts, 
but here the process is accompanied by a definite chemical change in 
that the serum globulins are precipitated. 
Action of Acid and Alkali on Alexin.—It is well known that any 
considerable amount of acid or alkali added to fresh serum perma- 
nently removes the alexin function. Brooks 45 has determined that 
hydrogen ion concentrations high enough to transform serum pro- 
teins from the cation to the anion condition, that is, passed the iso- 
electric point, permanently inactivate it. Coulter46 studied the 
activity of alexin at various hydrogen ion concentrations, both with 
and without the influence of heat. He took into consideration the 
complex structure of complement into albumin and globulin frac- 
tions. He found that the inactivation which complement undergoes 
when heated in distilled water dilutions is closely related to the 
properties of the euglobulin fraction, since the destruction is least at 
the reaction at which euglobulin is least soluble. The euglobulin 
sediment from serum washed in water and brought into solution 
again by the addition of NaOH at a Ph of 7.4, becomes least soluble 
on the addition of HC1, between Ph 5.1 and 5.7, and is iso-electric 
at about Ph 5. When examined in whole serum, however, the euglob- 
ulin has its iso-electric point at Ph 6.2 to 6.4, at which point 
Coulter found it existed not as a pure euglobulin, but as a compound 
of some other substance in the serum. At this reaction the destruction 
which complement undergoes on being heated, Coulter thinks de- 
pends on the relative preservation of the mid-piece function at this 
point. When salt is added, the destruction by heat increases as 
rapidly with the acidity as it does in the absence of salt, but on the 
alkalin side of this point the NaCI protects the destruction, according 
to Coulter, probably by a depression in the ionization of the 
euglobulin. 
Alexin Splitting.—Studies of this process have led to important 
modifications in our conception of the nature of alexin, since they 
have shown that this body, formerly assumed to be single and homo- 
geneous, may be subdivided into at least two component parts by a 
number of experimental procedures. Ferrata was the first one to 
point this out as a consequence of investigations undertaken by him 
primarily with the purpose of determining the nature of the influ- 
ence of salts upon hemolytic processes. Older studies of Buchner 
and Orthenberger 47 had shown that bactericidal action was inhibited 
when salts were removed from the medium, but the causes under- 
lying such inhibition had not been made clear. Ferrata 48 found, in 
the first place, that the absence of salts exerted no effect upon the 
45 Brooks. Jour. Gen. Physiol., 3, 1921, 185. 
48 Coulter. Jour. Gen. Physiol., 3, 1921, 771. 
47 Buchner and Orthenberger. Archiv f. Hyg., Vol. 10, 1890. 
48 Ferrata, Perl, klin. Woch., 1907, No, 13, 196 
INFECTION AND RESISTANCE 
mechanism of sensitization, but that amboceptor or sensitizer became 
attached to the cellular elements as readily when salts were absent 
as when the reagents were suspended in normal salt solution. It was 
a natural inference, therefore, that the failure of hemolysis, which 
he observed in salt-free media (analogous to the similar experiences 
of Buchner in the case of bacteriolysis), must be attributed to fail- 
ure of functionation on the part of the complement. On further 
investigation he obtained a very simple explanation. Ferrata re- 
moved the salts from his sera by dialyzing for twenty-four hours 
against distilled water. In this process, of course, there is a pre- 
cipitation of the globulins while the water-soluble albumins remain 
in solution. The former may be redissolved in normal salt solution 
and the latter rendered isotonic by the addition 
of calculated amounts of concentrated salt. In 
this way the original serum components are di- 
vided into two parts, neither of which, as 
Ferrata found, is alone capable of producing 
hemolysis of sensitized cells. In order to ob- 
tain the complementary action possessed by the 
original serum it is necessary to combine the 
two. This principle discovered by Ferrata is 
probably responsible also for the results ob- 
tained by Sachs and Teruuchi,49 who likewise 
noted the destruction of the complementary 
function in sera diluted with distilled water, 
but attributed this, in their publication, to the 
action of a complement-destroying ferment, 
which is assumed to be active in salt-free media 
only. 
In his first experiments Ferrata reported that the precipitated 
globulin fraction was thermostable, the thermolability of complement 
being due entirely to the unprecipitated albumin fraction. The 
work of Ferrata was soon continued, however, by a number of other 
workers, who confirmed the essential fact of the partition of the 
complement but modified and considerably extended the original 
observations. Brand 50 found that both fractions were equally ther- 
molabile, and that the globulin sediment, after being redissolved in 
salt solution, could not be preserved in an active condition for more 
than a few hours. Preserved in distilled water or as sediment, it 
may retain its activity for several days, but dissolved in salt solution 
it becomes inactive within 3 to 4 hours, at room temperature. 
Michaelis and Skwirsky 51 have since shown that the globulin frac- 
tion, thermolabile when free, is unaffected by a temperature of 56° 
Conception of Com- 
plement - Split- 
ting as First 
SUGGES TED BY 
Brand. 
49 Sachs and Teruuchi. Berl. klin. Woch., 1907, Nos. 16, 17, and 19, 
“Brand. Berl. klin. Woch., 1907, No. 34. 
51 Michaelis and Skwirsky. Zeitschr. f. 1mm., Yol, 4, 1910, FURTHER DEVELOPMENT OF KNOWLEDGE 
197 
C. after it has become attached to sensitized cells. Brand further 
studied the relationship of the two fractions to the sensitized cells 
and found that the globulin fraction may attach directly to such 
antigen-antibody complexes, but that the albumin fraction cannot be 
bound in this way unless the globulin fraction has been previously 
attached. For this reason he has referred to the former as the “end- 
piece” and the latter globulin sediment as the “mid-piece,” assuming, 
on the basis of the conception of Ehrlich, that the globulin fraction 
serves to establish a link between the sensitized cell and the end-piece 
analogous to that formed by the “amboceptor” between the cell and 
the whole complement. It is possible, therefore, to treat sensitized 
cells with mid-piece in such a way that they are thereafter susceptible 
to hemolysis by the end-piece alone. Such cell-sensitizer-mid-piece 
combinations have been spoken of by Miehaelis as “persensitized” 
cells. 
Tsurusaki 52 confirmed the findings of Brand as to the thermol- 
ability of both “mid-piece” and “end-piece,” but was unable to 
separate the complement of normal hemolysins into the two compo- 
nents in the same way, since he found that hemolytic power was, in 
such cases, completely destroyed after twenty-four hours of dialysis. 
It seems to us not impossible that the natural deterioration of alexic 
power which takes place during such periods of time, at temperatures 
of from 16° to 20° C., may easily be held accountable for this, since 
the very feeble sensitization of cells likely to take place in normal 
hemolysin complexes would require a correspondingly larger amount 
of alexin for activation. 
We have mentioned that the so-called “mid-piece” undergoes a 
rapid change when dissolved in salt solution and, after 3 or 4 hours, 
may lose its ability to induce hemolysis when added to sensitized 
cells together with end-piece. Although Ilecker 53 was able to con- 
firm this, he nevertheless showed that this fact does not imply a 
destruction of the mid-piece. For when such apparently inactive 
“mid-piece” was added separately to sensitized cells, and end-piece 
was subsequently allowed to act upon the complex, hemolysis re- 
sulted. This seems to show that the “mid-piece” undergoes a change 
on standing in salt solution which does not alter its ability to com- 
bine with the sensitized cells, but which subjects it to inhibition of 
such union when end-piece is present. It is also a peculiar fact, evi- 
dent in many of our own experiments, that when “mid-piece” and 
“end-piece” are first mixed and then added to sensitized cells the ef- 
fect in hemolysis is less powerful than when the “mid-piece” is added 
to the cells first, and the “end-piece” later. This effect is so 
instantaneous that, if, in a series of experiments in which combina- 
52 Tsurusaki. Biochem. Zeitschr., Vol. 10, 1908. 
53 Hecker. Arb. a. d. konig. Inst. f. exp. Ther., Frankfurt a/M., Heft 
3, 1907, See also Guggenheimer. Zeitschr. f. 1mm., Yol. 8, 1911. 198 
INFECTION AND RESISTANCE 
tions of mid- and end-piece are used, the end-piece is run into the 
tubes containing the cells just before the mid-piece is added instead 
of the other way round, hemolysis is inhibited. 
In working with dialysis, also, we have regularly had an experi- 
ence which may explain the difficulties which many other investiga- 
tors have had in such experiments. The globulin precipitate, which 
fell out after dialysis of 24 hours or more, almost without exception, 
retained moderate or slight hemolytic properties, which could not be 
removed until the precipitate had been dissolved in salt solution and 
reprecipitated with distilled water two or three times. This would 
imply that a minute amount of the end-piece, carried down in pre- 
cipitation, must suffice to activate the mid-piece and would seem 
to point to the fact that in whole serum the two fractions are present 
as a complex and not separately. This question has been much dis- 
cussed and many facts have been brought out on both sides. Hecker 
showed that the combination of mid-piece with the sensitized cells 
can take place at a temperature of 0° C., while that of end-piece 
with the “persensitized” cells requires a considerably higher tem- 
perature. The bearing this fact may have upon similar earlier ex- 
periments of Ehrlich and Morgenroth upon the thermal conditions 
governing the union of amboceptor and complement with antigen is 
self-evident. In the present connection, however, the fact that the 
two fractions may be separately absorbed out of the serum by sensi- 
tized cells at 0° C. would suggest the probability of their being sepa- 
rate in the whole blood. No crucial experiment has so far been 
possible, and there is not enough evidence on either side as yet to 
justify a definite opinion. However, the experiments of Michaelis 
and Skwirsky and later ones of Skwirsky alone have much indirect 
bearing on this question, though final interpretation is as yet im- 
possible. Michaelis and Skwirsky,54 after determining that an acid 
reaction inhibits the hemolysis of sensitized blood cells, found that 
under such conditions “mid-piece” alone is bound, but that “end- 
piece” or the albumin fraction is left unbound. They recommend 
the use of strongly sensitized cells in an acid medium as a method 
of obtaining free “end-piece” from serum. 
Skwirsky 55 subsequently found that during the ordinary Wasser- 
mann reaction the complex of syphilitic serum and antigen binds 
the mid-piece only. If the Wassermann reaction has been strongly 
positive; that is if there has been absolutely no hemolysis, and we 
remove the supernatant fluid by centrifugation, active end-piece can 
be demonstrated in it by the addition of persensitized cells. Bron- 
fenbrenner and Noguchi have also studied this phenomenon, but do 
not believe that Skwirsky’s experiments prove that end-piece is free 
in such “fixation” supernatant fluids. These supernatant fluids, ac- 
54 Michaelis and Skwirsky. Zeitschr. f. 1mm., Yol. 4, 1910. 
65 Skwirsky. Zeitschr. f. 1mm., Vol. 5, 1910. FURTHER DEVELOPMENT OF KNOWLEDGE 
199 
cording to them, differ from all other “end-pieces” in that they are 
active upon persensitized sheep corpuscles only, but not upon other 
cells. An explanation for this is lacking. 
There is much that is confusing in the facts so far revealed about 
the two component parts of the alexin. The most difficult fact to 
explain is the peculiar inactivation of the mid-piece in salt solution, 
which prevents its functionation in the simultaneous presence of 
end-piece, but does not seem to interfere with its ability to combine 
with the sensitized cells. As was to be expected, explanation for 
this has been sought by the Ehrlich school in changes of affinity. 
Sachs suggests that the mid-piece, by its preservation in salt solu- 
tion, has lost its avidity for the sensitized cells and has gained in 
avidity for the end-piece, an alteration which therefore prevents its 
union with the cells. The same idea was suggested by Hecker him- 
self. It is a little difficult to reconcile this explanation, however, 
with the fact that whole serum can be preserved and remain active 
in its complementary function for a number of days, mid-piece and 
end-piece being present together, in a medium which, as far as salt 
contents are concerned, is isotonic with the salt solution in which 
mid-piece deteriorates so rapidly when alone. 
That there is, after all, much similarity between the alexins of 
different animals is evident from the fact that, as Marks and other 
have shown, the end-piece of one animal may activate the mid-piece 
of another species. It appears also from experiments like those of 
Bitz and Sachs 56 that an animal may possess a mid-piece for certain 
sensitized cell complexes without possessing a corresponding end- 
piece. Thus they found that the serum of mice contained a mid- 
piece but not an end-piece active for the hemolysis of sensitized 
guinea-pig corpuscles. 
Much that has been found out about the so-called globulin por- 
tion, moreover, tends to engender doubt as to the wisdom of applying 
to these complement fractions the terms “mid-piece” and “end- 
piece,” an objection which is based upon reasons similar to those 
which prevent Bordet from accepting the term amboceptor. For so 
little is actually known concerning the mechanism of complement 
functionation, that it seems unwise to establish on a firm basis a 
preconceived idea of the mechanism by adapting the terminology to 
a theory. The most confusing feature of the problem lies in the 
surprising quantitative relations which seem to exist in the reactions 
of the two fractions. Thus Liefmann and Cohn 57 claim that in the 
presence of moderately sensitized cells no measurable amount of 
the so-called mid-piece or globulin fraction is bound, that is, removed 
from solution; and yet, when both fractions are added to such cells, 
rapid and complete hemolysis results. In the presence of heavily 
56 Ritz and Sachs. Zeitschr. f. 1mm., Vol. 14, 1912. 
57 Liefmann and Cohn. Zeitschr. f. 1mm., Yol. 7, 1910. 200 
INFECTION AND RESISTANCE 
sensitized cells (20 to 50 units) a small quantity only is removed. 
N evertheless tliis fraction has had a demonstrable effect on the cells, 
since it has rendered them amenable to the action of the albumin 
fraction. In all such experiments, therefore, as Liefmann justly 
points out, the degree of sensitization must be taken into considera- 
tion before conclusions are formulated. It is curious also that a 
slight excess of the globulin fraction may prevent complement action 
completely. In experiments cited by Marks 58 it appears that the 
most ineffective complement is obtained when “mid-piece” and 
“end-piece” are added to the sensitized cells in proportions of 1 to 1. 
If the proportion of “mid-piece” is increased two or threefold over 
that of “end-piece,” hemolysis is inhibited. This, however, is true 
only when the two fractions are simultaneously added to the sensi- 
tized cells. When the sensitized cells are exposed to the excessive 
quantity of the “mid-piece” separately, and “end-piece” added later, 
the effect is one of stronger hemolysis than when smaller amounts 
are used. It is thus seen that the relations between the complement 
fractions in hemolysis are very involved. All that we can be sure 
of is that there are at least two separable parts, that one of these 
acts directly upon the sensitized cells, forming a so-called persensi- 
tized complex and rendering them amenable to the subsequent action 
of the unprecipitated albumin fraction. 
The many difficulties encountered in the interpretation of the 
confusing phenomena observed in connection with this problem have, 
very naturally, led to a corresponding multiplicity of opinion. Most 
observers at present incline to the opinion that the globulin and 
albumin portions of fresh serum, separated by Ferrata’s or any other 
of several common methods, represent actually two complement frac- 
tions. This is not, however, accepted by all workers. Bronfenbren- 
ner and Noguchi 59 believe that the entire active complement is con- 
tained in the albumin fraction or so-called “end-piece.” They hold 
that “complement-splitting” by dialysis or other methods is an inac- 
tivation of end-piece by change of reaction. In their experiments 
they were able to restore the functional activity of end-piece by the 
adjustment of reaction, either with acid or alkali, respectively, or by 
the addition of amphoteric substances. The mid-piece activates, they 
believe, by reason of its amphoteric nature and consequently adjusts 
any excessive acidity or alkalinity of the medium. They were able 
to substitute for mid-piece indifferent amphoteric substances such as 
alanin. Liefmann 60 has been unable to confirm the experiments of 
Bronfenbrenner and Noguchi, and believes that their results were 
caused by incomplete splitting of the complement. Incidental to a 
study of normal opsonins the writer has also repeated the experi- 
58 Marks. Zeitschr. f. 1mm., Vols. 8 and 11, 1911. 
59 Bronfenbrenner and Noguchi. Jour. Exp. Med., Yol. 15, 1912. 
60 Liefmann. Weichhardt’s Jahresbericht, Yol. 8, 1912. FURTHER DEVELOPMENT OF KNOWLEDGE 
201 
ments of Bronfenbrenner and Noguchi without being able to confirm 
them.61 
The method of Ferrata for the separation of the two parts of the 
complement is successful only if dialysis is very thorough and suffi- 
ciently prolonged to lead to complete precipitation of the globulins. 
Neufeld and Haendel 62 have had difficulty in thus separating the 
fractions, and the writer has noticed similar failures but has always 
been able to obtain eventual separation by sufficient prolongation of 
the dialysis. Because of the occasional difficulties and because of 
the time-consuming and inconvenient nature of the method other 
means of separation have been devised. The one used with success 
by many workers has been that introduced by Sachs and Altmann,63 
namely, precipitation of the sera with weak hydrochloric acid, 
t° T3<e Biefmann has separated the components by precipitation of 
the globulins by the introduction of C02. In carrying out this 
method, Fraenkel64 has found it advantageous to dilute the serum 
ten times with distilled water, then allowing the C02 to flow in at 
low temperatures. It is likely that any of the usual methods of 
globulin separation will serve for complement partition. The salt- 
ing out methods are, however, extremely inconvenient because of the 
necessity for prolonged dialysis subsequently necessary to remove 
the salts. 
Spontaneous Reactivation of Alexin.—The inactivation of com- 
plement or alexin by the addition of salts or by splitting is, very 
apparently, a temporary inactivation in which prompt restitution 
can be practiced by bringing back original conditions either by dilu- 
tion to isotonicity or by reconstruction of the divided substance, 
respectively. Heating to 56° C., the simplest and most commonly 
employed method of inactivations was, until of late, regarded as an 
irreversible process, the complement being irretrievably destroyed 
in the procedure. Gramenitski65 has recently carried out experi- 
ments which seem to show that this opinion is erroneous. His experi- 
ments were suggested by the fact, observed by Bach and Chodat,66 
that certain oxydases and diastases may spontaneously regain some 
of their activity after inactivation by heat. His work with com- 
plement indicated a similar gradual return to an active condition 
after moderate heating. The great theoretical importance of this 
observation will justify our insertion of one of Gramenitski’s 
protocols. 
Experiment 1. Complement 10 times diluted was heated to 56° 
61 Zinsser and Cary. Jour. Exp. Med., Yol. 19, 1914. 
62 Neufeld and Haendel. Arb. a. d. kais. Gesund., 1908. 
63 Sachs and Altmann. Cited from Sachs in “Kolle u. Wassermann Hand- 
buch.,” Vol. 2, p. 877. 
64 Fraenkel. Zeitschr. f. 1mm., I, Yol. 8, 1911. 
65 Gramenitski. Biochem. Zeits., Vol. 38, 1912. 
66 Bach and Chodat. Cited from Gramenitski, loc. cit., p. 511. 202 
INFECTION AND RESISTANCE 
C. for 7 minutes. It was then tested against sensitized beef blood 
at varying intervals as follows: 
Time after heating at which test 
was made 
Quantity of hemoglobin gone into solution after 
% 10 min. 
% 20 min. 
% 30 min. 
% 40 min. 
Immediately after heating 
0 
20 
40 
70 
V/2 hour 
0 
30 
60 
80 
24 hours 
20 
70 
80 
100 
48 hours 
10 
40 
70 
In other experiments in which heating was more prolonged a 
similar regeneration was observed, though not as pronounced as in 
the one cited above. The largest amount of restored complement 
seemed to be present after about 24 hours. After this gradual de- 
terioration again ensued. It is quite impossible to offer an adequate 
explanation for this at the present time. Gramenitski 67 acknowl- 
edges this, but permits himself certain speculations which we repeat 
in nearly his own language, since there is much in them which seems 
to us reasonable. The complement, as indeed all other active serum 
constituents, must be looked upon as colloidal in nature. When heat 
is applied to such substances alterations occur which gradually lead 
to coagulation. As this occurs there is an aggregation of particles 
and a consequent diminution of surface tension. This last point has 
been experimentally demonstrated by Traube,68 who has regularly 
found a fall of surface tension as serum was heated to 56° C. And 
of greatest interest in this connection is the further determination by 
Traube that a gradual restoration of the surface tension takes place 
as the serum is allowed to stand. It is not inconceivable, therefore, 
that the inactivation of complement by heat may depend upon an 
alteration of its colloidal state, i. e., an aggregation of the particles, 
which, if not carried too far, may be reversible and followed by a 
gradual dispersion as the serum is kept 24 hours. On the same 
grounds the gradual deterioration of complement on standing may be 
compared to the slow settling out of colloidal suspensions which 
eventually results in spontaneous precipitation, a process which 
occurs not only in chemically well-defined colloids, but is often ob- 
served in sera. Bechold has referred to this as “das Altern Kol- 
loidaler Losungen.” 
Brooks has confirmed the observations of Gramenitski. He 
showed that after partial thermoinactivation, alexin recovers part of 
its hemolytic power, the recovery taking place more rapidly at 37° 
than at 7°. This recovery of power may restore at least one-third 
of its hemolytic action. 
67 Gramenitski. Loc. cit., p. 504. 
68 Traube. Zeitschr. f. Imm., Yol. 9, 1911, and BiocTiem. Zeitschr., 1908. FURTHER DEVELOPMENT OF KNOWLEDGE 
203 
Inactivation by Shaking.—Of great interest, furthermore, in 
connection with the physical properties of complement is the dis- 
covery made by Jacoby and Schiitze 69 that complement can be in- 
activated by shaking. This astonishing observation has been con- 
firmed by Zeissler,70 Noguchi and Bronfenbrenner,71 Ritz,72 and 
others . It appears, according to these observers, that guinea-pig 
serum, when subjected to active shaking, can eventually be robbed 
thereby of its activating properties. The success of such experiments 
depends somewhat upon the concentration of the serum, and is best 
observed in a dilution of 1 part to 10 parts of salt solution. Under 
such conditions complete inactivation may be observed within 20 to 
25 minutes. Between the inactivation of complement by heat and 
that which results from shaking, there are certain similarities which 
seem to strengthen the opinion regarding the nature of heat inactiva- 
tion which we have cited above. For it has been variously shown 
that prolonged shaking of protein solutions, like heating, gradually 
leads to coagulation. It would be important to determine whether 
or not the inactivation by shaking, like that produced by heat, is 
accompanied by a fall of surface tension. 
The Photoinactivation of Alexin.—There seems to be no question 
about the fact that the radiation of fresh serum with ultra-violet light 
to a slight extent inactivates alexin action. In this the temperature 
at which the radiation is done plays an important role. Brooks 73 
particularly has studied this. By a very delicate method of alexin 
titration, he determined the actual inactivation by light radiation 
without possibility of doubt. The explanations for this which he 
gives cannot be gone into here since they involve complex considera- 
tions of photo-chemical reactions which we are not competent to 
appraise. We refer the reader to Brooks’ original papers, the refer- 
ences of which are given. 
ALEXIN OR COMPLEMENT FIXATION 
The controversy regarding the multiplicity of alexin and the 
existence of a “complementophile group” cannot, of course, be re- 
garded as closed, however much we may lean toward the acceptation 
of Bordet’s point of view, since German experimenters of eminence 
still adhere to the Ehrlich interpretations. Moreover, it is, of course, 
extremely difficult to disprove such an assumption as that of the 
“polyceptor” conception of the complementophile group. However, 
69 Jacoby and Schiitze. Zeitsehr. f. 1mm., Yol. 4, 1910. 
70 Zeissler. Berl. klin. Woch., No. 52, 1909. 
71 Noguchi and Bronfenbrenner. Jour, of Exp. Med., Yol. 13, 1911. 
72 Ritz. Zeitsehr. f. 1mm., Vol. 15, 1912. 
73 Brooks. Jour. Med. Res., 41, 1919, 411, and Jour. Gen. Physiol., 3, 
1921, 169 & 185. 204 
INFECTION AND RESISTANCE 
we may safely assert that the functional unity of complement (and, 
after all, that is all that Bordet has maintained) is being upheld by 
the constantly increasing evidence in its favor which is being fur- 
nished by the practical and experimental application of the phe- 
nomenon of “alexin fixation” described, in 1901, by Bordet and 
Gengou.74 It will be well to bear in mind that this phenomenon 
should be strictly distinguished from the so-called “complement de- 
viation” (“Ablenkung”), described by Neisser and Wechsberg. The 
latter was advanced as an explanation of the inactivity of bacteri- 
cidal sera when used in too great concentration, as described in an- 
other place (p. 176) (.Neisser and Wechsberg phenomenon), and has 
been variously utilized as support for the assertion that alexin can 
unite with unattached sensitizer. It is regarded by most observers, 
moreover, as untenable in the light of later investigation. In spite of 
this, the term “Komplement-Ablenkung” has been employed by a 
number of German writers (see Citron, Vol. 2, “Kraus und Levaditi 
Handbuch”) as synonymous with “fixation” in the sense of Bordet 
and Gengou. 
The phenomenon of Bordet and Gengou, briefly described, is 
nothing more than an experimental utilization of the fact which we 
have discussed at length, that alexin is fixed by antigen and antibody 
after union, but by neither alone. 
The condition, as observed by them, may be best described by 
submitting the protocol of the first experiment detailed in their 
communication: 
An emulsion of a 24-liour slant of plague bacilli was used as 
antigen, heated antiplague horse serum represented the antibody, 
and fresh guinea-pig serum was used as alexin. A series of tubes 
was then prepared as follows: 
1. Alexin + plague bacilli + inactivated antiplague serum. 
2. Alexin + plague bacilli + inactivated normal horse serum 
3. Alexin + inactivated antiplague serum. 
4. Alexin + inactivated normal horse serum. 
5. Plague bacilli + inactivated antiplague serum. 
6. Plague bacilli and normal horse serum. 
These mixtures were left together for 5 hours and, at the end of 
this time, sensitized rabbit corpuscles were added to each tube. The 
result showed hemolysis in all the tubes except “1,” in which there 
were plague bacilli, antiplague serum, and alexin, and in tubes 5 and 
6, which had contained no alexin from the beginning.75 
It was plain, therefore, that the bacilli when specifically sensi- 
tized had become capable of absorbing alexin and preventing its sub- 
74 Bordet and Gengou. Ann. de VInst. Past., 1901, Yol. 15, p. 289. 
75 We will see later that unsensitized bacteria in emulsion will non-specifi- 
eally fix small amounts of complement. FURTHER DEVELOPMENT OF KNOWLEDGE 
205 
sequent action upon the sensitized erythrocytes. That the occurrence 
was not exceptional was shown by the fact that, in the same series, 
similar results were obtained with anthrax, typhoid, and proteus 
bacilli, and their respective antisera. 
Schematized in accordance with the conceptions of Ehrlich our 
diagram would be as follows: 
Complement Fixation Schematized According to Ehrlich’s Views. 
If the antibody in I is present then complement is fixed by the antigen-anti- 
body complex, and is no longer free to act upon the hemolytic complex II. 
In the same way antigen I could be determined if a known antibody I 
were used. For, in the absence of either of these parts of the complex I 
complement would remain unfixed and free to act on complex II. 
We represent the phenomenon graphically in the symbols of Ehr- 
lich merely because they facilitate clearness of exposition. 
In the presence of both parts of Complex I the alexin is held and 
is no longer available for Complex II. If either of the reacting 
parts, antigen or sensitizer, of Complex I are lacking the alexin is 
left unfixed and free to react with Complex II. 
With this technique Bordet and Gengou were able to demonstrate, 
by indirect experiment, the presence of specific sensitizers in the 
sera of animals immunized with various bacteria, a fact which was, 
of course, surmised but had been amenable to proof heretofore only 
in the case of bacteria like the spirillum of cholera in which lysis 
under the influence of immune serum and alexin could be directly 
observed under the microscope. The practical possibilities of their 
method were, of course, immediately apparent. By the use of a 
known antigen specific sensitizers can be demonstrated in this way, 
and, vice versa, in the presence of a known antibody, the method 
will serve to identify the nature of a doubtful antigen. Thus bac- 
terial differentiation can be carried out by adding to the suspected 
bacteria, in emulsion, a small quantity of a known antiserum and 
alexin, and determining whether or not the alexin has become fixed. 206 
INFECTION AND RESISTANCE 
And, conversely, Bordet and Gengou 76 have more recently utilized 
the method in support of their claim of the specific etiological im- 
portance of the bacillus isolated by them from whooping cough, by 
showing that the serum of children suffering from this disease 
formed a specific alexin-fixing complex when treated with the 
bacillus. 
The phenomenon of Bordet and Gengou thus found rapid prac- 
tical application in the diagnosis of a number of infectious diseases, 
and has, of course, attained great clinical importance in the diag- 
nosis of syphilis in the form of the “Wassermann” reaction and its 
many modifications. Before discussing these practical features in 
greater detail, however, it will be useful to discuss more particularly 
the many important theoretical considerations which have followed 
in the train of the complement-fixation phenomena. 
Alexin-fixation with Non-cellular Antigens.—A year after the 
publication of Bordet and Gengou’s paper Gengou 77 made another 
fundamentally important observation by showing that complement 
or alexin fixation was not limited to the complexes of cellular antigens 
and their antibodies, but that the sera of animals immunized with 
dissolved proteins (animal sera, etc.), when brought together with 
their specific antigens, likewise formed combinations which fixed 
alexin. Thus egg-white or dog serum, brought together with “anti- 
egg-white” or “anti-dog” rabbit serum, respectively, strongly fixed 
alexin, whereas neither the antigenic substances nor the antisera 
exerted such fixation alone. The interpretation put upon this by 
Gengou was the following: “In sera obtained by injecting rabbits 
with large doses of cow’s milk, etc., there are, in addition to the 
precipitins of Bordet and Tschistovitch, substances analogous to the 
sensitizers described by Bordet in bacteriolytic and hemolytic sera, 
and later found in the majority of antimicrobial sera.” The impor- 
tant point in this interpretation is that Gengou conceived the exist- 
ence of antiprotein sensitizers, in addition to the precipitins, formed 
as a response to immunization with amorphous protein. Moreschi 78 
soon confirmed Gengou’s experimental determinations, and Ueisser 
and Sachs'79 took the further logical step of applying this knowledge 
to the determination of proteins for forensic purposes. This, too, 
we will further discuss when we speak of the practical features of 
these phenomena. It thus appears that the fixation of alexin is a 
generalized property of all mixtures in which an antigen is brought 
into contact with its specific antibody, whether the antigen is in the 
form of the whole bacterial or other cell, or in that of a dissolved 
protein, animal serum, or egg-white, etc. 
76 Bordet and Gengou. Ann. de VInst. Past., 1906. 
77 Gengou. Ann. de VInst. Past., 16, 1902. 
78 Moresehi. Perl. klin. Woch., No. 37, 1905. 
79Neisser and Sachs. Berl. klin. Woch., No. 44, 1905. FURTHER DEVELOPMENT OF KNOWLEDGE 
207 
The observation of Gengou, though for a time insufficiently 
valued, has had a profound influence upon the subsequent under- 
standing of serum reactions. The fundamental importance of 
this work was not fully recognized until his studies had found 
logical continuation in the investigations of Gay 80 and in those of 
Moreschi. 
Moreschi 81 studied the antihemolytic properties possessed by the 
serum of a rabbit which had been treated with normal goat serum. 
He found that such a serum had distinct anticomplementary powers 
when it was added to a hemolytic system of ox blood sensitizer (ob- 
tained against ox blood from rabbits), and goat complement. With 
such a hemolytic system, however, there was anticomplementary ac- 
tion only against goat complement and not against rabbit or guinea- 
pig complement. If, however, he used a hemolytic system in which 
the amboceptor or hemolytic sensitizer employed was one obtained 
from a goat, the serum was anticomplementary for all complements 
which were used. Moreschi concluded from this that the apparent 
anticomplementary action of the serum could not be interpreted as 
the action of a specific anticomplement in the sense of Ehrlich, but 
that it resulted from the reaction which took place as the consequence 
of union of the antibody in the anti-goat rabbit serum and goat pro- 
tein, which was introduced into the tubes, in the first case with the 
complement, and in the second with the amboceptor. He proved his 
contention by obtaining similar universal anticomplementary action 
when he added a little normal goat serum to the tubes set up as above 
described. It is plain, therefore, that anticomplementary action can 
be explained in observed cases by the simple consideration of the phe- 
nomenon of Gengou. Similar findings were later recorded by Muir 
and Martin,82 and it may well be doubted, as a result of these and 
other researches, whether we are at all justified in assuming the ex- 
istence of anticomplements. 
Fixation of Alexin by Precipitates.—The work of Gay, pub- 
lished independently in the same year as that of Moreschi, has, in a 
general way, the same significance, but Gay recognized the relation 
of the conditions observed by him to the precipitin reaction, a feature 
absent from both the original study of Gengou and the work of 
Moreschi. Gay noticed that an inactivated hemolytic immune serum, 
left for some time in contact with its specific cells, and then separated 
from them by centrifugation, would often possess anticomplementary 
or anti-alexic properties. He further noted that after such a serum 
had been freed from the cells by a short centrifugation, if it was again 
vigorously centrifugalized, a slight, cloudy sediment would appear 
80 Gay. Centrcdbl. f. Bakt. I Orig. Yol. 93, 1905, p. 603. 
81 Moreschi. Berl. klin. Woch., 1905, No. 37, ibid., No. 4, 1906. 
82 Muir and Martin. Jour. Hyg., Yol. 6, 1906. See also Muir’s “Studies 
on Immunity,” Froude, London, 1909, 208 
INFECTION AND RESISTANCE 
at the bottom of the tubes. If this sediment was removed the serum 
lost its alexin-fixing properties. He recognized that the precipitate 
formed in these tubes was a specific precipitate resulting from the 
union of a precipitinogen and its antibody. The reaction was due 
entirely to the fact that insufficient washing of the cells used in pro- 
ducing the hemolysin, gave rise to the formation of precipitin against 
the serum of the animal from which the cells had been taken, and 
subsequently insufficient washing of the cells of this same species 
employed in the tests furnished enough antigen to give a precipitin 
reaction in the tubes in which the inactivated hemolytic (and pre- 
cipitating) serum was mixed with the cells. Subsequently, numerous 
investigations 83 have shown Gay’s interpretation to be correct, and 
we may now accept it as a fact that precipitates formed by the union 
of specific antigen with its antibody possess the power of fixing alexin 
and that, in a general way, this fixation is proportionate in energy 
to the amount of precipitate which is formed. 
Gay utilized his results primarily to contradict certain assertions 
of Pfeiffer and Priedberger concerning antibacteriolytic substances 
supposed to occur in normal sera. These authors had found that, if 
normal sera possessing no “antagonistic” properties in the first place 
were left in contact with certain bacteria, they acquired antibac- 
teriolytic properties for these particular bacteria. Thus normal 
inactive rabbit serum, left in contact with typhoid bacilli, and again 
separated from the bacteria, now prevented the lysis of sensitized 
typhoid bacilli if tested by the intraperitoneal method spoken of as 
the Pfeiffer reaction. Sachs 84 applied these observations to analo- 
gous hemolytic reactions and obtained similar results. He found 
that, if normal, inactive rabbit serum was left in contact with sheep 
or guinea pig corpuscles, it acquired the property of preventing the 
hemolysis of these corpuscles if, later, it was brought together with 
them in the presence of specific hemolysin and alexin. Gay now 
showed by experiment that Sachs’ method was referable to insuffi- 
cient washing of the corpuscles. When, in the first contact, the rab- 
bit serum was exposed to the sheep corpuscles, a certain amount of 
sheep serum adherent to the cells was carried over into the rabbit 
serum. This sheep antigen later reacted with the antisheep pre- 
cipitin present in the hemolytic immune serum and, in this way, 
fixed alexin and prevented hemolysis. 
It seems that the analysis of Gay is correct, and that Sachs’ 
conclusion as well as those of Pfeiffer and Priedberger, by analogy, 
cannot be taken as demonstrating the existing of specific anticom- 
plemcnts or anti-amboceptors. Gay has further offered the same 
mechanism as an explanation of the Peisser-Wechsberg phenomenon, 
which has been discussed in another place. 
83 Dean. Zeitschr. f. Imm., I, Yol. 13, 1912. 
84 Sachs, Deut, med, Woch., 1905, No, 18, FURTHER DEVELOPMENT OF KNOWLEDGE 
209 
To summarize, then, we have learned that there are a number of 
varieties of specific alexin absorption or fixation processes, one that 
is exerted by cells treated with specific sensitizer, be they blood or 
bacterial, the other that which occurs when unformed protein is 
brought into contact with its specific antiserum. The latter has been 
correlated with the precipitin reaction, in that it has been found that, 
whenever a specific precipitate is formed in such reactions, it is this 
precipitate on which the fixation depends. On the other hand, it is 
necessary to note that the formation of a precipitate is by no means 
necessary for the fixation, for, as is well known, if a series of pre- 
cipitin tubes are set up, in each successive one of which the amount 
of antigen is diminished, a degree of dilution will soon be reached 
at which no visible precipitate will occur, but which nevertheless 
will show alexin fixation. The following is an illustration of such 
an experiment: 
Sheep serum + antisheep serum 
Precipitate 
Fixation of 0.5 c. c. 
guinea pig complement 
0.5 c. c. (1:20) + 0.5 c. c 
+ 
Complete 
0.5 c. c. (1:50) 0.5 c. c 
+ 4* 
Complete 
0.5 c. c. (1:100) 0.5 c. c 
+ + + 
Complete 
0.5 c. c. (1:200) 0.5 c. c 
+ + + 
Complete 
0.5 c. c. (1:500) 0.5 c. c 
+ + + 
Complete 
0.5 c. c. (1:1,000) 0.5 c. c 
+ 
Complete 
0.5 c. c. (1:2,000) 0.5 c. c 
+ 
Complete 
0.5 «. c. (1:5,000) 0.5 c. c 
Partial 
0.5 c. c. (1:10,000) 0.5 c. c 
Partial 
0.5 c. c. 0.5 c. c 
None 
From such experiments it follows moreover that the fixation of 
alexin, carefully titrated, is a more delicate method of determining 
the presence of an antigen or, vice versa, of an antibody than is the 
observation of a visible precipitate, a fact which has been made use 
of, as we have mentioned, by Neisser and Sachs and others for 
forensic antigen determinations. 
It should also be remembered that, if to such a precipitate there 
is added an excess of the antigen, the precipitate may be partially 
dissolved, and this dissolved precipitate, as Gay 85 has shown, may 
possess fixation properties. This, too, accounts for the fact, observed 
by a number of workers, that if, in a series of precipitin tests the 
supernatant fluids and the washed precipitates are separately exam- 
ined for alexin fixation, the fixation properties reside entirely in the 
precipitates except in those tubes in which a considerable excess of 
antigen was used and in which, as in tubes 1 and 2 of the preceding 
85 Gay. Univ. of Cal. Public, in Pathol., Vol, 2, No, 1, 1911. 210 
INFECTION AND RESISTANCE 
protocol, the precipitates were relatively slight. The subject, though 
involved, is worthy of detailed consideration in this place since it 
seems to us to have an important bearing on certain theoretical con- 
ceptions which will be taken up below. 
The important question now arises: what is the nature of the 
alexin fixation by the complexes formed by unformed proteins with 
their antibodies and, more especially, what is the nature of the 
alexin fixation exerted by specific precipitates? There have been 
much experimentation and speculation concerning this, and a number 
of different views are held. Gengou assumed, as we have seen, that 
this fixation, as studied by him, was entirely analogous to the fixation 
by sensitized bacterial or blood cells. He expressed the belief that 
treatment of an animal with an unformed protein produced not only 
specific precipitins but also specific sensitizers, analogous to those 
produced in response to treatment with bacterial or other cells. He 
noticed the parallelism between the quantity of the precipitate 
formed and the alexin fixation, but did not associate the two proc- 
esses. 
His conception of specific antiprotein sensitizers was accepted by 
a number of workers, and Wassermann and Bruck,86 Friedberger 87 
and several others brought out the facts that actual precipitate forma- 
tion is not a necessary criterion of fixation. Thus the last-named 
writer showed that the precipitating power of a serum may be de- 
stroyed by moderate heat without a corresponding destruction of its 
fixing property. A similar independence of the precipitation from 
the complement-fixing property, in the presence of an antigen, has 
been observed by Muir and Martin.88 
Gay,89 also, though he was the first definitely to associate 
precipitin formation with the alexin-fixing property and, indeed, 
determined a rough parallelism between the amount of precipitate 
and the degree of alexin fixation, has nevertheless recently declared 
himself in favor of the assumption of the presence in protein anti- 
sera of two antibodies, the alexin-fixing lysins and the precipi- 
tins. This he does on the basis of certain experiments from which 
he concludes that the antigen-antibody complex which fixes alexin 
is distinct from the precipitin-precipitinogen complex, but is usually 
“brought down in its formation in such a way as to simulate fixation 
by the precipitate.” Nicolle 90 goes even further than this in declar- 
ing that the “coagulins” or precipitins are “anticorps bons,” which 
prevent the action of the albuminolysin upon the antigen, thereby 
inhibiting the liberation of poisonous cleavage products. 
88 Wassermann and Bruck. Mediz. Kl., 1905, Yol. 1, No. 55. 
87 Friedberger. Deut. med. Woch., 1906, No. 15. 
88 Muir and Martin. Jour. Uyg., Yol. 6, 1906. 
89 Gay. Loc. cit. Also TJniv. of Cal. Publ. in Pathol., Vol. 2, No. 1, 1911, 
90 Nicolle. Ref. in Bull, de Vlnst, Past,, Vol, 5, 1907, FURTHER DEVELOPMENT OF KNOWLEDGE 
211 
It seems to the writer91 that the assumption of a separation 
between the precipitin and the albuminolysin is a needlessly compli- 
cated interpretation of the phenomena. 
Without going into further complicated detail, it would seem to 
us 92 to be justified that we look upon the so-called precipitins not as 
separate antibodies but as identical with so-called albuminolysins. 
They unite with the antigen, producing an alexin-fixing complex. 
Since both reacting bodies are colloidal in nature, they precipitate 
each other in the test tube, but, following the laws governing other 
mutually precipitating colloids, they do so only when brought to- 
gether in concentrations which lie within definite zones of relative 
proportions. The visible precipitation would seem, therefore, to be 
merely a secondary phenomenon, the essential one being the union of 
an antigen with a sensitizer by which it is rendered amenable to the 
action of the alexin. This would enable us to comprehend also the 
experiments of Friedberger, discussed in the section on anaphylaxis, 
in which it was shown that the action of alexin upon precipitates 
gives rise to the formation of toxic bodies just as this occurs when 
alexin acts upon sensitized cells. 
Dean,93 who has lately analyzed the relation between precipita- 
tion and alexin fixation on the basis of extensive experimentation, 
comes to the conclusion that the proportions of antigen and antibody 
which are favorable for rapid and complete precipitation do not 
favor the most complete alexin fixation. He states that the two reac- 
tions do not run a parallel course but believes that this does not mean 
that they are necessarily distinct phenomena. He says: “They rep- 
resent two phases of the same reaction ... a flocculent precipitate 
represents the final stage of a change which can be recognized in its 
earliest and incomplete stage by means of a complement fixation.” 
Our view differs from this only in that we believe that the pre- 
cipitation is merely a secondary, colloidal phenomenon, which may, 
or may not, coincide with the phase of greatest alexin fixation, ac- 
cording to other fortuitous conditions which may favor or retard 
flocculation. Indeed, if our view he accepted, rapid compact pre- 
cipitation may possibly be assumed to interfere with alexin fixation 
in that it would inhibit perfect contact of the alexin with the antigen- 
antibody complexes. 
Dean finds that the alexin fixation in mixtures of antigen and 
antibody is quantitatively determined very largely by the character 
of the precipitation. If antigen and antibody are mixed in such 
proportions that a heavy precipitate is rapidly formed, there is either 
a very slight or no fixation of alexin, and there is, consequently, no 
91 Zinsser. Jour. Exp. Med., Yol. 15, 1912. Also Proc. of Soc. of Exp. 
Biol, and Med., April, 1913. 
92 Zinsser. Jour. Exp. Med., Sept., 1913. 
93 Dean. Zeitschr. f. Imm., Yol. 13, 1912. 212 
INFECTION AND RESISTANCE 
relationship whatever between the amount of the precipitate and 
the amount of alexin fixed. If, on the other hand, antigen and 
antibody are mixed in such proportions that the turbidity develops 
slowly, as a very fine and gradual flocculation, large amounts of 
alexin may be fixed. The relationship is such a delicate one that if, 
in a series of experiments, varying amounts of antigen are mixed 
with a constant amount of antibody, two definite points may be 
found, at one of which there is the heaviest precipitation, and at 
the other the most intense alexin fixation. The amount at which 
the alexin fixation is most effective is usually the one at which there 
is materially less antigen than at the other. The necessity for 
optimum relative amounts for both reactions has been briefly touched 
upon in a discussion of the Neisser-Wechsberg phenomenon a few 
pages above. In cases in which an excess of antigen produces in- 
complete precipitation, no alexin fixation can be expected, according 
to Dean. In experiments in which antiserum and antigen are mixed 
in optimum relative amounts, the degree of alexin fixation is pro- 
portionate to the amount of antiserum used. Dean believes that 
alexin fixation probably occurs during the earliest stages of the pre- 
cipitation, when the particles of precipitate are extremely small (per- 
haps because of the high surface tension), and that after the stage at 
which a definite turbidity is observed, very little alexin fixation takes 
place. Therefore, in order to obtain the maximum amount of alexin 
fixation, the alexin should be present in the tubes at the time when 
antigen meets antibody. Subsequent addition usually results in less 
fixation. 
/These considerations of Dean which have been confirmed gen- 
erally in many of their aspects, are of the utmost importance to all 
workers with complement fixation, since neglecting them may reverse 
the results of experimental procedures. 
Another view of the mechanism of alexin fixation is that which 
has been advanced by Neufekl and Haendel.94 
These workers have found that sensitized cholera spirilla will fix hemolytic 
complement at 0° C., whereas the same bacteria at 37° C. will fix both the 
hemolytic and the bactericidal complement. They conclude from this that 
the fixation at 37° C. was brought about by virtue of the bactericidal ambo- 
ceptor, whereas at 0° C. fixation was brought about by an antibody which 
is distinct from amboceptor or sensitizer. They believe from this and other 
observations, which we cannot consider in detail, that alexin fixation may be 
brought about by a special fixing antibody, the “Bordetscher Antikorper,” 
which is not identical with any of the other known antibodies.95 
Non-specific Alexin Fixation.—In all experiments which deal 
with alexin fixation by specific antigen-antibody complexes it is of the 
84 Neufeld and Haendel. Arb. a. d. kais. Gesund., Yol. 28, 1908. 
95 As will be seen in a later section we are in entire disagreement with 
this view and believe that Dean’s conception is the correct one. FURTHER DEVELOPMENT OF KNOWLEDGE 
213 
greatest importance that we should guard against the errors easily 
introduced by fortuitous non-specific antihemolytic agencies. Thus 
there are a number of factors which will interfere with the functiona- 
tion of alexin upon a sensitized antigen, either by direct non-specific 
absorption of the alexin itself or by producing physical conditions 
in the presence of which alexin cannot act. 
Thus many animal tissue cells, in emulsion, will absorb alexin, 
and the same property may be possessed by tissue extracts. Von 
Dungern 96 was the first to call attention to this, and his observations 
have been variously confirmed. Muir 97 showed that the stromata of 
hemolyzed red blood cells exert strong anticomplementary action, 
and that this is due to a firm union with the complement. It is not un- 
likely that the action of cells in this respect is referable to their 
lipoidal contents. This suggestion was first made by Landsteiner 
and von Eisler,98 who found that the petroleum-ether extracts of red 
blood cells possessed strong anticomplementary action which, to a 
limited extent, was specific toward the particular corpuscles from 
which the etracts had been made. Similar observations have been 
made by Noguchi,99 who speaks of the substance he extracts as “pro- 
tectin.” In general, the protective action of the lipoidal extracts 
seems to depend largely upon cholesterin, and, since this substance is 
present to some extent in many tissues, their antihemolytic action is 
easily understood. In another section we have discussed the similar 
neutralizing action of lipoidal substances upon poisons of various 
kinds (saponin, tetanolysin, and snake poison, but, as we have noted 
there, the neutralizing properties of the extracts do not, as a rule, 
equal those of the whole tissues.100 It is not unlikely that in such 
cases as Landsteiner suggests the potent agent is not the lipoid itself 
but rather a lipoid-protein combination, a class of substances of which 
we know very little, but the importance of which, in many phases of 
serum reactions, seems assured. 
We have already mentioned that yeast cells may absorb alexin. 
And it has been found by Wilde 101 and others that almost all bac- 
teria in emulsion may possess varying degrees of alexin-fixing prop- 
erties even though unsensitized. There seems to be no regularity 
either qualitatively or quantitatively in regard to this, but the fixa- 
tion is usually sufficiently marked to render the use of whole bacteria 
unreliable for specific fixation experiments. For this reason, as we 
will see, bacterial extracts must be used in such work unless careful 
quantitative controls are made. Upon what this fixation depends it 
is difficult to determine. It may be that it is purely non-specific and 
96 Yon Dungern. Munch, med. Woch., Nos. 20 and 28, 1900. 
97 Muir. “Studies in Immunity,” London, Vol. 19. 
98 Landsteiner and von Eisler. Wien. Min. Woch., No. 24, 1904. 
99 Noguchi. Jour. Exp. Med., Vol. 8, 1906, p. 726. 
100 See also Ivar Bang, “Biochemie der Lipoide.” 
101 Wilde. Berl. Min. Woch., 1901, Yol. 38, and Archiv f. Ilyg., 39, 1902. 214 
INFECTION AND RESISTANCE 
due to absorption of the fine emulsion of the bacteria comparable to 
that observed on the part of kaolin or quartz sand emulsions, or, pos- 
sibly fixation by such bacterial emulsions may occur because of the 
small amounts of normal sensitizer almost always present in the 
serum employed as alexin. 
Apart from the lipoids, a number of other substances have been 
found to fix alexin and exert consequent antihemolytic action. Thus 
Landsteiner and Stankovic,102 and Landsteiner and von Eisler 103 
describe the anti-alexic action of various proteins coagulated or pre- 
cipitated. They refer this action not to particular chemical struc- 
ture but to the colloidal state, since they obtained similar antilytic 
action with such inorganic emulsions as quartz sand and kaolin 
(aluminium-orthosilicate). Since anticomplementary action has, 
moreover, been noted in the case of a large number of extracts of 
such materials as wool, leather, etc., it is clear that the methods of 
alexin fixation, as applied to the forensic differentiation of blood, 
must be carefully controlled with this point in view.104 
Among the most practically important non-specific agencies 
which fix alexin there are some which appear under certain condi- 
tions in normal serum. Noguchi105 has found that serum will often 
develop anticomplementary properties as a consequence of heating 
during the process of inactivation. On more detailed investigation 
he determined that the anticomplementary action increased as the 
serum was heated to about 90° C. Above this temperature it is de- 
stroyed. He refers this property to the serum lipoids, since he was 
able to remove it by extraction with ether, the ether extract possessing 
the same anticomplementary power as the original serum before 
extraction. 
Neisser and Doring 106 have noticed anti-alexic or anticomple- 
mentary properties of human sera which were destroyed on heating, 
and which they associate with disease of the kidneys, since they 
noted it in sera of uremic patients. Browning and McKenzie 107 
have observed a similar heat-sensitive anti-alexic action on the part 
of normal serum, and the subject has been studied by Zinsser and 
Johnston.108 It was found that all normal sera will develop anti- 
alexic properties on preservation at room temperature within a few 
days, and more slowly but no less regularly in the ice chest. This 
anti-alexin is destroyed on heating to 56° C., and may be precipitated 
out with the globulins of the serum. There appeared in these studies 
102 Landsteiner and Stankovic. Centralbl. f. Bakt., 1906, Yols. 41 and 42. 
103 Landsteiner and von Eisler. Wien. klin. Woch., 1904, No. 24. 
104 Uhlenhuth. Deut. med. Woch., 1906, Nos. 31 and 51, and Centralbl. f. 
Bakt., 1906, I, Ref., Yol. 38. 
105 Noguchi. Jour. Exp. hied., Vol. 8, 1906, p. 726. 
106Neisser and Doring. Berl. klin. Woch., 1901, No. 22. 
107 Browning and McKenzie. Jour. Path, and Bad., Vol. 13, 1909. 
108 Zinsser and Johnston. Jour. Exp. Med., Yol. 13, 1911. FURTHER DEVELOPMENT OF KNOWLEDGE 
215 
no particular association between the anti-alexic property and 
nephritis. 
The action of alexin upon sensitized cells may be prevented, also, 
by physical or chemical conditions without actual fixation or binding 
of the alexin. We refer to the effects of the addition of salts, prob- 
lems which have been considered above. CHAPTER VIII 
PRACTICAL APPLICATIONS OF THE COMPLEMENT- 
FIXATION METHOD 
THE WASSERMANN REACTION 
The principle of specific alexin fixation has been practically 
utilized in the diagnosis of disease and in the forensic determination 
of the nature of spots of blood or other protein material. 
Soon after Bordet and Gengou’s experiments Wassermann and 
Brack 1 showed that bacterial extracts could be successfully substi- 
tuted for whole bacteria in these reactions. Citron,2 too, made simi- 
lar observations, and, indeed, we now know that the use of bacterial 
extracts is more suitable for these experiments than are emulsions 
of whole bacteria, since, as we have mentioned above, bacterial emul- 
sions may often fix small amounts of complement of themselves 
(without specific sensitization), thereby confusing the results of the 
reaction. 
On the basis of their experience with bacterial extracts Wasser- 
mann and Brack 3 then determined that complement fixation could 
be carried out in tuberculosis when the various tuberculin prepara- 
tions were used as antigen.4 These investigations fell into the period 
during which active research upon the Spirochceta pallida in syphilis 
was going on, and it occurred to Wassermann that the technique of 
complement or alexin fixation might be utilized in the diagnosis of 
syphilis. Together with Neisser and Brack 5 he subjected this idea 
to experimental test. The publication of their first results appeared 
in 1906. They used in their experiments the syphilitic monkeys 
which were being observed in Neisser’s clinic. Their method con- 
sisted in mixing inactivated serum from syphilis-inoculated monkeys 
with organ extracts, serum, etc., of syphilitic human beings, and 
adding a small amount of fresh guinea pig complement. After these 
materials had been together for a certain time, sensitized red blood 
1 Wassermann and Bruek. Med. Klinik, Vol. 55, 1905. 
2 Citron. Centralbl. f. Bakt., Vol. 41, 1906. 
3 Wassermann and Bruck. Deut. med. Woch., No. 12, 1906. 
4 Complement fixation in tuberculosis is not yet on a practical or reliable 
basis. Recent claims of Besredka (Ann. Past., 1913) for his new antigen 
promise a successful technique, but no extensive confirmation has followed up 
to the present time. 
5 Wassermann, A. Neisser, and Bruck. Deut. med, Woch,, No. 19, 1906. 
216 PRACTICAL APPLICATIONS OF METHOD 
217 
cells were added. If the complement was bound during the first 
exposure no hemolysis resulted and the reaction was regarded 
as positive. From their results they drew the following con- 
clusions : 
1. Immune serum from monkeys, produced by treatment with 
syphilitic material, will sensitize syphilitic material from human 
beings or monkeys, so that an alexin-fixing complex is formed. 
2. Complement fixation results only when the syphilitic immune 
serum of monkeys is added to similar material from men or 
monkeys, but not when added to organ extracts of normal men or 
monkeys. 
3. Normal monkey serum has no such action. 
They concluded that their results justified them in assuming a 
specific fixation due to specific antisyphilitic immune bodies in the 
blood of the treated monkeys. They excluded experimentally the 
possibility of fixation by a precipitin reaction resulting from the 
treatment of the monkeys with human material. It might well have 
happened that precipitins against human protein appearing in the 
serum of the treated monkeys might subsequently react with the 
human protein material used as antigen, a complement-fixing com- 
plex resulting. This, however, was excluded by the fact that they 
obtained positive reactions only when the human material was ob- 
tained from luetic lesions. 
The same authors, with Schucht,6 very soon after this, extended 
their method to the diagnosis of syphilis in human beings. The 
same thing had been done shortly before their publication appeared 
by Detre 7 on a smaller material. By these and many other investi- 
gations it was very soon shown that syphilis may be reliably diag- 
nosed by complement fixation when extracts of the syphilitic organs, 
employed as antigen, are mixed with the inactivated serum of syphi- 
litic individuals. It was incidentally shown by Wassermann and 
Plaut 8 that the reaction could be obtained not only with blood serum 
but also with spinal fluid in paralytic cases. 
It was generally assumed, at this time, that the reaction in syph- 
ilis depended, as in the case of other infections, upon the presence in 
the syphilitic serum of specific antibodies. For it seemed reasonable 
to suppose that the specific antigen obtained in the extracts was de- 
rived from the extraction of large numbers of spirochetes demon- 
strable in the extracted organs. 
This, of course, is the most logical and simple theoretical concep- 
tion of the reaction, and is justified on the basis of analogy. Un- 
fortunately, however, it was soon found by a number of workers, 
6 Wassermann, Neisser, Bruck, and Schucht. Zeitschr. f. Hyg., Yol. 55, 
1906. 
7 Detre. Wien. kl. Woch., Yol. 19, No. 21, 1906. 
8 Wassermann and Plaut. Deut. med. Woch., No. 44, 1906. 218 
INFECTION AND RESISTANCE 
Marie and Levaditi,9 Weygant, Kraus and Volk, Landsteiner, 
Miiller, and Pdtzl,10 and others that antigens perfectly capable of fix- 
ing complement in the presence of syphilitic serum could be pro- 
duced from normal organs.11 
Theoretically it must be admitted that we are very much in the 
dark at present. The fact, now entirely unquestionable, that the 
sera of syphilitic patients will give fixation with antigens derived 
from extracts of normal organs, as well as from those of syphilitic 
organs, seems to throw doubt upon the simple specific antigen-anti- 
body conception at first held. 
In order to understand the questions involved in the theories of 
the Wassermann reaction as at present conceived it will be necessary 
to consider the types of antigen which are now employed. 
Wassermann’s original method of antigen preparation consisted 
in using the liver or spleen of a congenitally syphilitic fetus. The 
organs were finely divided and emulsified in 4 to 6 parts of normal 
salt solution. This mixture was shaken for 24 hours, centrifugal- 
ized, and the clear supernatant fluid used as antigen. Later the 
specific organ substances were extracted by Porges and Meier 12 in 
five times the volume of absolute alcohol for 24 hours. This alco- 
holic extract was evaporated in vacuo and the residue taken up in 
salt solution and shaken until an even suspension resulted. 
After it had been discovered that normal organ extracts could 
serve as antigen as well as the extracts of syphilitic organs, Land- 
steiner, Porges and Meier, and others, introduced antigens produced 
by alcoholic extraction of normal organs of animals and of man. 
Landsteiner introduced the alcoholic extract of normal guinea pig 
organs, especially extracts of the heart and liver, and Weil and 
Braun 13 made use of extracts of normal human organs. There are 
various methods of preparing extracts for this purpose. We may 
mention, to illustrate these methods, the one suggested, first, we be- 
lieve, by ISToguchi, a procedure which is applicable to the extraction 
of normal human organs (spleen), beef hearts, and guinea pig hearts. 
The finely divided or triturated organ substance is shaken up with 
five times its weight of absolute alcohol and allowed to stand in the 
incubator at 37.5° C., for from 5 to 7 days. At the end of this time 
it is filtered through cheesecloth and then through coarse paper, and 
9 Marie and Levaditi. Cited from McIntosh and Fildes’ “Syphilis.” 
Longmans & Co., 1911, p. 94. 
10 Landsteiner, Muller, and Potzl. Wien. kl. Woch., Yol. 20, 1907. 
11 An extensive historical review of the development of the Wassermann 
reaction is found in the book of Boas, “Die Wassermannsche Reaktion,” 
Karger, Berlin, 1911. Since these earlier publications have appeared the 
literature of the Wassermann reaction has become very extensive. It is 
enumerated more fully than we can afford space for here in the book of 
Noguchi (“Serum Diagnosis of Syphilis”) and that of Boas, mentioned above. 
12Porges and Meier. Berl. kl. Woch., No. 15, 1908. 
13 Weil and Braun. Berl. kl. Woch., No. 49, 1907. PRACTICAL APPLICATIONS OF METHOD 
219 
the filtrate placed in a large crystallizing dish in which it is evapo- 
rated at room temperature with the aid of an electric fan. A gummy 
yellow residue is left, which is then taken up in as small a quantity 
of ether as possible. This ether solution is then precipitated with 
4 times its volume of acetone, in consequence of which there is a pro- 
fuse precipitation of coarse white flakes. This acetone-insoluble 
substance, which is at first white, later yellowish, in color, is the 
stock antigen. A little of this is taken up in a very small quantity 
of ether, and this ethereal solution is shaken up in salt solution until 
the ether has evaporated or the material has gone into very fine col- 
loidal suspension in the salt solution. This is the antigen ready to 
be used. 
It is immediately evident that these antigenic substances must 
consist very largely of lipoidal extractives of the organ substances, 
and it has been found that such antigen contains sodium oleate, leci- 
thin and cholesterin. Indeed, Porges and Meier have claimed that a 
1 per cent, solution of commercial lecithin may be used with success. 
Browning and Cruikshank 14 have found further that the addition 
of small amounts of cholesterin to syphilitic antigen very largely 
increases its specifically diagnostic value, and this idea has since 
been utilized more especially by Sachs,15 Walker and Swift,16 and 
others. Sachs, especially, has obtained excellent antigens in the 
following way: 1 gram of moist guinea pig heart substance was 
extracted with 5 c. c. of alcohol and left at room temperature for 
twelve hours or in the ice box for two days; it was then filtered and 
0.5 to 1 per cent, of cholesterin was added; frequently the alcohol 
extract had to be diluted two or three times before use. Sachs and 
Rondoni 17 have also recommended artificial mixtures of lipoids con- 
taining sodium oleate, lecithin, and oleic acid. 
The fact that cholesterin added to alcoholic organ extracts in- 
creases the antigenic value of these for the Wassermann reaction is 
all the more curious inasmuch as cholesterin alone has practically no 
antigenic action. Walker and Swift have recommended an antigen 
in which alcoholic extracts of human or guinea pig hearts were made 
up to 0.4 per cent, of cholesterin, 0.4 per cent, having been found by 
comparative test to be the most favorable concentration. Cholesterin- 
liver extracts or even alcoholic extracts of syphilitic livers without 
cholesterin were found to be inferior in specific antigenic value to 
0.4 per cent, cholesterin-heart antigens. From the experience of 
many investigators it now seems unquestionable that additions of 
cholesterin increase the delicacy of the reaction in that more cases 
react positively with such an antigen than with the uncholesterinized 
14 Browning and Cruikshank. Jour, of Path, and Bad., Yol. 16, 1911. 
15 Sachs. Berl. hi. Woch., No. 46, 1911. 
16 Walker and Swift. Jour, of Exp. Med., Yol. 18, 1913. 
17 Sachs and Rondoni. Zeitschr. f. 1mm., Yol. 1, 1909. 220 
INFECTION AND RESISTANCE 
preparations. The experience of Hopkins and Zimmermann, how- 
ever, would indicate that great caution must be exercised when the 
reaction is done in this way, since occasional positive results are 
obtained with cases clinically not syphilitic. These workers believe 
that cholesterinized antigen is extremely useful, but advise its use 
only parallel with the ordinary lipoidal antigens and together with 
careful study of the clinical aspects of the case. 
The fact that these antigens are non-specific in origin naturally 
necessitates careful determination of their usefulness before they 
are used. Before any antigen can be regarded as reliable, therefore, 
a titration must be carried out in the following way: Two series of 
tubes are prepared, in the first of which antigen and complement are 
added to normal serum, and in the second the same substances are 
added to known syphilitic serum. The antigen must, of course, be 
such that in no test tube does it cause alexin fixation in the presence 
of normal serum, but, in the quantities used, it must give fixation 
regularly with syphilitic serum. An example of such a titration 
may be tabulated as follows: 
Example of Antigen Titration 
Antigen by Landsteiner’s method: normal guinea pig heart 
freed from fat and ground up in a mortar. To each gram is added 
5 c. c. of absolute ethyl alcohol and the mixture allowed to extract 
at 60° C. for 12 hours (or several days at 37.5° C.). It is then fil- 
tered through paper. The following titration is then carried out: 
• A 
Tube 1 
Tube 2 
Tube 3 
Tube 4 
Tube 5 
Normal serum 
0.2 
0.2 
0.2 
0.2 
*_ 
Antigen 
0.05 
0.1 
0.2 
0.8 
0.6 
Alexin 
0.1 
0.1 
0.1 
0.1 
0.1 
B 
Tube 1 
Tube 2 
Tube 3 
Tube 4 
Syphilitic serum 
0.2 
0.2 
0.2 
0.2 
Antigen 
0.05 
0.1 
0.2 
0.3 
Alexin 
0.1 
0.1 
0.1 
0.1 
The volume in all of these tubes is brought to 3 c. c. with isotonic 
salt solution. After one hour at 37.5° C., sensitized red cells are 
added to each tube.18 If the antigen is suitable in that it does not 
18 Tube “5” is the antigen control which shows that the antigen in large 
amounts is neither anticomplementary nor hemolytic by itself. It is well, in 
addition, also to test out various amounts of the antigen and alexin, without PRACTICAL APPLICATIONS OF METHOD 
221 
fix alexin by itself or in the presence of normal serum, hemolysis 
will result in all of the tubes of series A. If it is suitable in that it 
fixes in the presence of syphilitic serum, the tubes in series B will 
show no hemolysis; if there is slight hemolysis in B 1, it is inferred 
that 0.05 c. c. of the antigen is insufficient, and the smallest amount 
(0.1 c. c.), which completely fixes 0.1 c. c. of alexin in the presence 
of the positive serum, is the quantity used. Again the antigen may 
be able to cause hemolysis by itself if used in too large amounts. If 
this is the case in tube B 4, then this antigen is suitable only in 
amounts varying between 0.1 c. c. and 0.2 c. c. 
The titration is done with varying quantities because too little 
antigen might fail in fixing the alexin, even if the serum were posi- 
tively syphilitic, whereas too much antigen might possess alexin-fix- 
ing properties in itself, even in the presence of normal serum, or 
possibly without any serum at all, an attribute which is not uncom- 
monly possessed by lipoidal extracts. 
It is thus seen that Wassermann reactions can be carried out 
with antigens which do not contain extracts of syphilitic lesions or 
of the micro-organisms which give rise to syphilis. This fact alone 
would exclude the possibility of considering the fixation of comple- 
ment as at present carried out in the Wassermann reaction as being 
due to a specific antigen-antibody union. 
This conclusion is strengthened by the recent discovery that a 
specific antigen prepared from cultures of Spirochceta pallida cannot 
be successfully used in diagnostic Wassermann tests. The first in- 
vestigations of this kind were made by Schereschewsky,19 who used 
as antigen extracts of mixed cultures in which the spirochete was 
present; his results were inconclusive. Noguchi 20 later investigated 
this phase of the problem, preparing his antigens by the extraction of 
pure cultures and of syphilitic rabbit testicles in which the spirochetes 
were very profuse. He found that positive tests with such an antigen 
were obtained only in isolated cases of prolonged syphilis which had 
been thoroughly treated, and that the ordinary Wassermann reaction, 
as obtained in active cases, is not due to antibodies which combine 
specifically with the pallida antigen. Craig and Nichols 21 also have 
found that cases of untreated syphilis which gave positive reactions 
with syphilitic liver extracts gave absolutely negative results when 
culture antigens were used. 
From these results also we may infer that the Wassermann reaction does 
not represent a fixation of alexin by the union of a specific syphilitic antigen 
either normal or syphilitic serum, to determine the largest amount of antigen 
which, by itself, is devoid of the actions mentioned above. 
19 Schereschewsky. Deut. med. Woch., 1909, p. 1653. 
2° Noguchi. Jour. A. M. A., Vol. 58, 1912. 
21 Craig and Nichols. Jour, of Exp. Med., Yol. 16, 1912. 222 
INFECTION AND RESISTANCE 
with antibodies found against the Spirochceta pallida. Noguchi concludes 
that it is caused by “lipotropic” substances in the sera of syphilitic human 
beings; a conclusion which is justified by the fact that the antigens used, 
all of them, contain large quantities of lipoids. It must be acknowledged, 
however, that we have no definite information concerning the nature of the 
reaction beyond this. Schmidt 22 believes that it is a colloidal reaction, and 
depends upon the union of the serum globulins with the extract colloids in 
the antigen. In normal serum such a union is prevented by the albumins 
which act as a sort of protective colloid. In syphilitic serum the globulins 
are increased quantitatively or are changed qualitatively in the degree of 
their dispersion, or possibly in both characteristics. He regards the serum 
globulins in the Wassermann reaction as directly uniting with the extract 
colloid. 
Levaditi and Yamanouchi23 also conclude that the Wassermann reaction 
depends upon the union of two colloidal substances—one a non-proteid 
constituent of syphilitic serum (cholesterin derivatives or fatty acids), the 
other the lipoidal constituents of the antigen. Like others they found that 
the active substances in the antigenic extracts are non-protein and alcohol 
soluble. 
It is interesting to note, moreover, that Porges and Meier24 observed 
actual precipitation when syphilitic serum was added to lecithin emulsions. 
In consequence, attempts have been made to make the diagnosis of syphilis 
by direct precipitation of syphilitic serum by such emulsions of lecithin 
and of sodium glycocholate (Merck). The results of these investigations as 
well as those of Klausner,25 who claims that syphilitic sera are more easily 
precipitated by distilled water than are normal sera, have led to no diag- 
nostically reliable results, but they have seemed to show that the serum 
globulins are probably more plentiful and more easily precipitated out of 
syphilitic than out of normal sera. 
The inference of many workers, therefore, has been that the Wassermann 
reaction is primarily due to the precipitation of (probably) globulin by the 
lipoidal colloids of the antigen, the resulting precipitate being capable of 
absorbing alexin. Jacobsthal26 has examined mixtures of syphilitic serum 
and antigen by the ultramicroscopic method, and claims that precipitates 
are always present even when they are not macroscopically visible. Bergel,27 
who has recently suggested the importance of specific lipase production as a 
cause of hemolysis, suggests that the Wassermann reaction is due to fixation 
exerted by the products of the action of a specific lipase formed in the 
syphilitic body against “lues-lipoids.” This theory is open to objections 
similar to those mentioned above, namely, that the antigen need not neces- 
sarily be a lues-lipoid, but may be derived from normal organs. Other 
theories have been brought forward by Bruck, Weil, Braun, Manwaring, and 
more recently by Rabinowitch.28 The data supporting most of these theories 
are, as yet, too speculative to justify our discussion of them at any length. 
The only fact which seems established with any reasonable certainty is the 
independence of the Wassermann test from a specific antigen-antibody re- 
action in the usual sense. 
Another suggestion which has been made, particularly by Weil and 
22 Schmidt. Zeitschr. f. Hyg., Yol. 69, 1911. 
23 Levaditi and Yamanouchi. C. R. de la Soc. de Biol., 1907, Yol. 63, 
p. 740. 
24 Porges and Meier. Berl. kl. Woch., No. 15, 1908. 
25 Klausner. Wien. kl. Woch., No. 7, 1908. 
26 Jacobsthal. Munch, med. Woch., 1910. 
27 Bergel. Zeitschr. f. 1mm., Vol. 17, 1913. 
28 Rabinowitch. Centralbl. f. Bakt., Orig., 1914. PRACTICAL APPLICATIONS OF METHOD 
223 
Brown,29 is that the antibody involved in syphilis is an auto-antibody, re- 
sulting probably from the destruction of tissue by the syphilitic infection, 
with the liberation of lipoid-protein complexes against which the body, itself, 
then reacts with antibody formation. Such lipoid-protein substances, they 
suppose, are present in small amounts in normal tissues, but are enormously 
increased in the course of syphilitic infection. There is not much experi- 
mental or theoretical basis for the opinion of Weil and Brown. 
While, therefore, we are still considerably in the dark concerning the true 
nature of the Wassermann reaction, we may state with safety that there 
is present in the syphilitic serum a substance which leads to the formation 
of a precipitate when brought into contact with properly prepared alcoholic 
extracts of normal tissues. The so-called antigenic substances, then, are 
probably lipoidal in nature, or, at any rate, represent lipoid-protein com- 
plexes. The precipitate formed, probably because of its physical properties, 
is now capable of fixing alexin. 
Although the Wassermann reaction is thus apparently not based 
on those principles in the investigation of which it was discovered, 
its practical diagnostic value is not therefore diminished. For its 
proper performance any of the methods of antigen preparation con- 
sidered above may be employed, provided that the usefulness of the 
preparation utilized is carefully controlled in each case as indicated. 
Since, of course, a hemolytic system is used in such tests as an in- 
dicator, it is necessary also to titrate sensitizer and alexin. 
From what has been said in another place concerning the quanti- 
tative relations of alexin and amboceptor or sensitizer (see reference 
to work of Morgenroth and Sachs, p. 163), it is evident that the use 
of too strongly sensitized cells might result in hemolysis, if a slight 
fraction of alexin were left unbound by a weak syphilis reaction. 
Conversely the use of too large a quantity of alexin would result in 
hemolysis, since, even if the amount of syphilitic fixation were con- 
siderable, a sufficient excess of alexin might remain. The use of 
uniform amounts of fresh guinea pig serum in each case does not 
control this adequately, for different specimens of guinea pig serum 
may vary considerably in alexin content. In consequence, titra- 
tions of both sensitizer and alexin should be made. For practical 
purposes it is quite enough to titrate the hemolytic sensitizer every 
few weeks and use a stated amount in successive reactions. The 
alexin or complement can then be titrated individually for each set 
of reactions. Examples of such preliminary titrations follow: 
Titration of Hemolytic Amboceptor or Sensitizer 
Rabbit injected 3 times at 5-day intervals with washed sheep 
corpuscles . . ., 3, 4, and 5 c. c., and bled 10 days after the last 
injection.30 
29 Weil and Brown. Berl. klin. Woch., 44, 1907, 1570. Cited from Bruck, 
“Serodiagnose der Syphilis,” p. 618. 
30 In immunizing animals with blood cells for this or any other purpose 224 
INFECTION AND RESISTANCE 
This serum is inactivated at 56° C. for 20 minutes. 
Washed sheep corpuscles 
5% emulsion 
in salt solution 
Sensitizer 
Fresh 
K- P- 
serum 
Hemolysis 
1 
1 C. C 
0.01 
0.1 
+ + + 
2 
1 c. c 
0.005 
0.1 
+ + + 
3 
1 c. c 
0.003 
0.1 
+ + + 
4 
1 c. c 
0.001 
0.1 
+ + + 
5 
1 c. c 
0.0005 
0.1 
+ + 
6 
1 c. c 
0.0002 
0.1 
=fc 
7 
1 c. c 
0.1 
8 
1 c. c 
salt sol. 
In this case 0.001 c. c. still causes complete hemolysis of 1 c. c. 
of a 5 per cent, emulsion of sheep cells (volumetric measurement of 
cells sedimented in the centrifuge), and this amount ( fVuu c. c.) 
is called the “hemolytic unit” of sensitizer; two units are then used 
in the reactions. 
Against these cells alexin can, in each case, be titrated as follows: 
Alexin Titration: 
Fresh Guinea Pig Serum Pipetted from Clot 
Red cells 
5% emulsion 
Sensitizer 
as above 
determined 
Guinea pig 
serum 
Hemolysis 
1 
1 C. C. 
2 units (.002) 
0.1 c. c. 
+ + + 
2 
1 c. c. 
2 units (.002) 
0.05 c. c. 
+ + + 
3 
1 c. c. 
2 units (.002) 
0.025 c. c. 
=4= 
4 
1 c. c. 
2 units (.002) 
0.01 c. c. 
The smallest amount of alexin which completely hemolyzes the 
red cells (0.05 in this case) is the amount used. Since it is easier 
to measure larger volumes with accuracy, the alexin is diluted 1 to 
10 in salt solution before use. A typical Wassermann reaction can 
then be carried out as follows: 
it is necessary to wash the cells very carefully in salt solution. Unless this is 
done blood serum or plasma will be injected with them and the treated animal 
will respond by the formation not only of hemolysin but of precipitins for 
the serum proteins as well. When a subsequent hemolytic test is carried out, 
a precipitin reaction between the precipitin in the antiserum and serum ad- 
hering to the corpuscles will follow, and this, as we have seen, will fix alexin, 
obscuring other reactions which may be under observation. PRACTICAL APPLICATIONS OF METHOD 
225 
Scheme for Wassermann Test 
ADAPTED TO ORIGINAL WASSERMANN SYSTEM AFTER SCHEME OF NOGUCHI 
Test with 
unknown serum 
Test with known 
positive syphilitic 
serum 
Test with known 
negative normal 
serum 
Test without serum 
to control efficiency 
of hemolytic system 
Serum .2 c. c. 
6 Serum .2 c. c. 
Serum .2 c. c. 
d 
© 
+ 
+ 
+ 
gf 
O Complement 
O Complement 
O Complement 
O Complement 
£ 53 
.1 c. c. 
| .1 c. c. 
.1 c. c. 
.1 c. c. 
’© 2 
08 2 
+ 
1 + 
+ 
+ 
Salt sol. 
•3 Salt sol. 
Salt sol. 
Salt sol. 
9 
3 c. c. 
° 3 c. c. 
3 c. c. 
3 c. c. 
2. 
H 4. 
6. 
8. 
Serum .2 c. c. 
Serum .2 c. c. 
Serum .2 c. c. 
+ 
+ 
+ 
O Complement 
O Complement 
O Complement 
O Complement 
.1 c. c. 
6 .1 C. C. 
.1 c. c. 
.1 c. c. 
+ 
“ + 
+ 
+ 
d 
* S) 
Antigen 
Antigen 
Antigen 
Antigen 
g*“ 
(required amount 
a 
in 1 c. c. salt sol.) 
o 
2-5 
fcrg 
+ 
a + 
+ 
+ 
Salt sol. 
o Salt sol. 
Salt sol. 
Salt sol. 
2 c. c. 
2 c. c. 
2 c. c. 
2 c. c. 
1. 
3. 
5. 
7. 
O = Test tube. 
Place in water bath at 40° C. for one hour, then add to all tubes red blood 
cells and amboceptor. These are previously mixed so that 2 c. c. contains the 
equivalents of 1 c. c. of a 5 per cent, emulsion of sheep corpuscles and 2 units of 
amboceptor. Again expose to 40° C. If the serum tested is positive, tubes 1 
and 3 should show no hemolysis, all the other tubes showing complete hemolysis 
in one hour. 
Since many human sera normally contain small amounts of antisheep 
sensitizer, it is the habit of many workers to add the sheep corpuscles, without 
the sensitizer or amboceptor, and incubate for a half-hour. If, at the end of this 
time, no hemolysis has occurred either in the front or the back row, then 
amboceptor may be added. This technique avoids the possible error introduced 
by an excess of amboceptor, a condition which easily occurs when any large 
amount is normally present in the serum and in addition to this 2 units are 
added as in the test described above. 
The above represents the typical “Wassermann” as at present 
carried out in most laboratories. It may be carried out just as well 
and with greater economy of material by using one-half the amounts 
throughout. It is evident that the performance of the reaction calls 
for experience of serum technique, and knowledge of such reactions, 
so that fortuitous irregularities may be intelligently controlled. It is 
our opinion that the performance of routine Wassermann tests by 
workers without a thorough knowledge of the fundamental facts of 226 
INFECTION AND RESISTANCE 
serum phenomena is worse than useless in that insufficient attention 
to special conditions and to details may easily result in a positive re- 
action when syphilis is not present, and vice versa. 
Kecently Archibald McNeil and others have exposed the mix- 
tures of complement, antigen, and patient’s serum at refrigerator 
temperature for a number of hours instead of in the water bath or 
thermostat at 37.5° C., before adding the sensitized cells. It is a 
curious fact, which has not yet been satisfactorily explained, that 
such a procedure increases the delicacy of the reaction. It may be 
that, when the tubes containing the antigen, patient’s serum, and 
alexin are left at incubator temperature, partial alexin fixation only 
can take place during the brief period of 30 minutes to one hour, 
which is usually employed. More prolonged exposure at this tempera- 
ture would not be advisable on account of deterioration of the alexin. 
On the other hand, at ordinary ice-box temperatures of about 8° to 
10° C., the exposure can be continued for as long as 10 hours without 
extensive complement deterioration, and meanwhile more complete 
fixation can occur. This so-called ice-box method has become a 
regular routine in many Wassermann laboratories. In our own 
department, it has been employed for a good many years in parallel 
with the water-bath tests, each serum being used for a water-bath 
reaction in which the cholesterinated antigen is used, as well as in 
an ice-box test with an ordinary alcoholic antigen without cholesterin. 
Whether or not it will be advisable to continue such duplication after 
further development of the direct precipitation methods of Sachs 
and Georgi, and of Kahn, remains to be seen. 
Many modifications of the Wassermann test have been suggested. 
Probably the most important is that of Noguchi. The chief justifi- 
cation for this modification is the fact that many normal human sera 
contain hemolysins for sheep corpuscles. For this reason many 
workers carry out the ordinary Wassermann technique without add- 
ing antisheep sensitizer or amboceptor until they have first observed 
whether or not the tested serum (in the “back row,” without antigen) 
will not hemolyze the corpuscles without such an addition, adding 
the sensitizer only when this does not take place. This is advisable 
since the presence of any considerable amount of normal antisheep 
sensitizer in the human serum which is being examined (if added to 
the amount used in the ordinary reaction, 2 units) may so increase 
the total quantity that hemolysis will result even after most of the 
alexin has been fixed. Noguchi excludes this uncertainty by avoid- 
ing the use of the “sheep cell-antisheep sensitizer” system entirely, 
substituting a hemolytic complex consisting of human cells and anti- 
human sensitizer, produced by injecting washed human corpuscles 
into rabbits. 
His technique may be best illustrated in the following tabula- 
tion : PRACTICAL APPLICATIONS OF METHOD 
227 
Reagents 
1. Sensitizer prepared by injecting washed human blood corpuscles into 
rabbits. 
2. 1 per cent, emulsion of washed human blood cells. 
3. Alexin—fresh guinea pig serum diluted with one and one-half volumes 
of salt solution, 40 per cent. 
The reaction is performed in the following way: 
Noguchi’s Method of Complement Fixation for the Serum Diagnosis of Syphilis 
Set for diagnosis 
Test with the serum in 
question 
Positive control set 
Test with a positive syphi- 
litic serum 
Negative control set 
Test with a normal serum 
St 
o 
M 
a. Unknown ser- 
um, 1 drop* 
b. Complement, 
a. 'Positive syph. 
serum, 1 drop* 
b. Complement, 
a. "Normal serum, 
1 drop* 
b. Complement, 
c 
S 
a 
<x> 
u 
p 
to 
3 
£ 
S3 
U 
03 
2 units 
2 units 
2 units 
u 
o 
(S 
O c. Corpuscle 
O c. Corpuscle susp., 
O c. Corpuscle 
o 
A 
a 
A ai 
susp., 1 c. c. 
1 c. c. 
susp., 1 c. c. 
t- 
o 
a 
03 
N 3 
u. >-» 
•S £ 
a. Unknown ser- 
a. 'Positive syph. 
a. "Normal serum, 
d 
s 
is 
um, 1 drop* 
serum, 1 drop* 
b. Complement, 
1 drop* 
ro 
■■§■§ 
CO - 
o 
b. Complement, 
b. Complement, 
cl 
3 § 
2 units 
2 units 
2 units 
fl 
O 05 
a M 
£ 
O c. Corpuscle 
O c. Corpuscle susp., 
O c. Corpuscle 
03 
a *■ 
•2f2 
•2 3 
5 S3 
susp., 1 c. c. 
+Antigen 
1 c. c. 
susp., 1 c. c. 
-(-Antigen 
P 
•2 ® 
35 
-(-Antigen 
fl 
h-t 
T3 p 
< 
a 
* When working with inactivated serum 4 drops (0.08 c. c.) should be em- 
ployed. With cerebrospinal fluid, 0.2 c. c. (not inactivated) is used. 
(Taken from Noguchi’s “Serum Diagnosis of Syphilis,” Lippincott, 1910, 
p. 57.) 
Bauer 31 has introduced a modification in which he utilizes the 
presence of normal sheep sensitizer in many human sera. He per- 
forms his tests without the addition of antisheep sensitizer at 
first, adding this only to those tubes in which controls have shown 
that no normal sensitizer is present. Stern,32 on the other hand, 
utilizes the alexin normally present in human serum. The syphi- 
litic serum to be tested is, therefore, not inactivated, and the 
sheep cells are more heavily sensitized (9 to 12 units). It seems 
to us that this method is objectionable chiefly because of the 
anticomplementary action which develops in most normal human 
sera if kept for a short time, and which can he removed only by 
inactivation. 
Other modifications of the Wassermann reaction are those of 
31 Bauer. Semaine Medicare, 28, 1908. 
32 Stern. Zeitschr. f. I mm., Yol. 1, 1909., 228 
INFECTION AND RESISTANCE 
Jacobaeus 33 and of Wechselman.34 It seems, however, that, as the 
reaction is gaining in importance in clinical diagnosis, most labora- 
tories are adhering to the original system used by Wassermann and 
his associates, except for the substitution of the non-specific lipoidal 
antigens for the originally employed organ extracts. 
The value of the Wassermann test in the diagnosis of the various 
stages of syphilis is a problem which can be approached only by 
careful statistical analysis of the results obtained. This has been 
done by various investigators, and some of the results have been 
tabulated in the books of Noguchi, of Boas, and of McIntosh and 
Fildes. The figures we cite are those largely taken from Boas, as 
summarized in F. C. Wood’s “Chemical and Microscopical Diag- 
nosis” (D. Appleton & Co., 1911), pp. 706 et seq. 
Primary syphilis, 974 cases, 56.5 per cent, positive. 
The reaction may appear before the primary sore, but this is 
very rare. Usually it is positive in from 5 to 6 weeks after infection. 
Secondary syphilis, 2,762 cases, 88 per cent, positive. In untreated 
cases they are stated to be 100 per cent, positive. 
Tertiary syphilis, 830 cases, 80 per cent, positive. 
Tabes, 360 cases, 70 per cent, positive. 
Dementia paralytica, 95 to 100 per cent, positive. 
The tabulation on the following page, taken directly from Boas, 
will give a comprehensive summary of this phase of the problem. 
Since the reaction is not a specific antigen-antibody union but 
depends on some substance liberated or produced by reason of the 
syphilitic injection, it is not out of question that other infections 
may give rise to a “positive Wassermann.” And this, indeed, is 
the case. It was claimed for a time that a positive reaction may 
be obtained in tuberculosis, but this has been refuted by subsequent 
experience, and the earlier positive results probably depended upon 
faulty technique. There can be little doubt, however, that occasional 
positive reactions are obtained in cases of leprosy, scarlet fever, 
malaria, and trypanosoma infections. 
The spinal fluid may'be used instead of the blood serum in cases 
of syphilis of the central nervous system, but even here, as Citron 35 
has shown, the results with blood serum are more frequently positive 
than those done with the spinal fluid itself. In isolated cases posi- 
tive reactions have been obtained with ascitic fluids, pleural and 
pericardial exudates. Bab 36 reports a case of positive reaction in 
the milk of a syphilitic mother. Serum obtained at autopsy is not 
33 Jacobaeus. Zeitschr. f. Imm., Vol. 8, 1911. 
34 Wechselman. Zeitschr. f. Imm,., Yol. 3, 1909. 
35 Citron. Deut. med. Woch., 1907, No. 29, p. 1165. 
36.Bab. Munch, med. Woch., Vol. 46, 1907. PRACTICAL APPLICATIONS OF METHOD 
229 
Table Compiled by Boas, loc. cit., p. 138. 
Stage of disease 
Number of 
cases 
Positive 
reaction 
Negative 
reaction 
Control cases (not syphilitic) 
1,064 
1 
(scarlatina) 
1,063 
Induration 
76 
56 
20 
Secondary 
Early untreated 
269 
269 
0 
Recurrent after treatment 
199 
187 
12 
Tertiary 
No treatment of early tertiary mani- 
festations 
63 
63 
0 
Treatment 
20 
16 
4 
Latent syphilis 
Within 3 yrs. after infection 
243 
89 
154 
After 3 yrs 
111 
44 
87 
Tabes 
Untreated 
17 
17 
0 
Treated 
26 
11 
15 
Dementia paralytica 
0 
Serum 
139 
139 
Spinal fluid 
67 
61 
6 
Congenital 
54 
54 
0 
With symptoms 
Without symptoms 
10 
7 
3 
suitable for the reaction, since this, for unknown reasons, may often 
give a positive reaction in non-syphilitic cases. 
Direct Precipitation Methods in the Diagnosis of Syphilis.—In 
view of what we have said about the fixation of alexin by specific 
precipitates, it is interesting to know that the probable physical basis 
of the reaction consists in the actual formation of a precipitate by 
the union of the so-called antigenic substance and the reacting sub- 
stance in the syphilitic serum. We have already mentioned that 
Jacobsthal in 1910 showed that a microscopic coagulation of particles 
could be detected in mixtures of syphilitic serum and lipoidal anti- 
gens under the ultra-microscope. This observation was confirmed 
by a considerable number of workers, and has borne fruit in a num- 
ber of diagnostic tests in which syphilitic serum and antigen extracts 
produce macroscopically visible precipitates. The best known of 
these is that spoken of as the Sachs-Georgi reaction. The following 
description is one prepared for us by our associate, Dr. Frederic 
Parker, Jr., for our “Textbook of Bacteriology,” 37 on the basis of 
comparative investigations made with this reaction and the Wasser- 
mann reaction upon a series of positive and negative cases. 
37 Hiss and Zinsser. “Textbook of Bacteriology,” Fifth Edition, D. 
Appleton & Co.; N. Y.; 1922. 230 
INFECTION AND RESISTANCE 
Sachs-Georgi Reaction for Syphilis (Direct Precipitation). 
Preparation of Extract.—A beef heart is freed from fat and endo- 
cardium, cut up finely and ground in a mortar. It is then shaken 
with 5 volumes of 95 per cent, alcohol and a few glass beads in a 
shaking machine for 5 hours, allowed to stand at room temperature 
over night, filtered through ordinary filter paper next morning, then 
placed in the ice-box for at least two days, when it is again filtered 
through ordinary filter paper, and is now ready for use. It must 
first be titrated against a standard extract on a number of sera to 
determine the optimum dilution and cholesterinization. For this, 
the alcoholic extract is diluted with 1, 2, and 3 parts of alcohol, and 
to fractions of each of these dilutions, 0.3, 0.45, 0.6 and 0.75 per 
cent, of a 1 per cent, alcoholic solution of cholesterin is added. These 
various portions are then diluted with 5 parts of saline as described 
below, and set up against a standard extract. At least two such ex- 
tracts should be used in each test. 
Extract Dilution.—The alcoholic cholesterinized extract is diluted 
with 5 parts of saline as follows: The required amount of extract is 
placed in an Erlenmeyer flask; to it is rapidly added from a burette 
an equal volume of saline. It is shaken gently and allowed to stand 
10 minutes, then the remaining 4 volumes of saline are rapidly run 
in. It is again shaken and is ready for use. 
Serum.—Serum should be as fresh as possible. Three or four 
days is not too old. A slight degree of hemolysis does not interfere. 
Before use in the test, it is heated for y2 hour at 55° to 56° C. and 
should not be used sooner than 3 hours after heating. Sachs and 
Georgi 38 recommend that spinal fluids should be used undiluted in 
amounts of 1 c. c. and 0.5 c. c. 
Saline.—0.85 per cent, sodium chlorid in distilled water. Should 
be sterile and as fresh as possible. 
Test.—0.1 c. c. serum is diluted with 0.9 c. c. saline and to this is 
added 0.5 c. c. extract dilution. On each serum a control should be 
set up consisting of 0.1 c. c. serum 4" 0.9 c. c. saline -f- 0.5 c. c. of 95 
per cent, alcohol, diluted 1:6 with saline. Each extract dilution 
should be controlled by a tube containing 0.5 c. c. extract dilution 
-f 1 c. c. saline. 
The tubes are thoroughly shaken and placed in the incubator at 
37.5° C. for 20 hours. A preliminary reading may now be made; 
then the tubes are placed at 14° and 18° C., or in the ice-box for 20 
hours, and the final and decisive readings are taken. 
The reactions present appearances similar to macroscopic bacterial 
agglutination or precipitin reactions, and are read with the naked eye, 
the positives showing varying amounts of precipitates and the nega- 
tives remaining opalescent as at the beginning of the test. Sus- 
38 Sachs-Georgi. Med. Klinik, No. 33, 1918, 805, and Parker and Haigh, 
Archiv. Dermat, and Syphilol., 4, 1921, 67, PRACTICAL APPLICATIONS OF METHOD 
231 
picious tests are centrifuged at moderate speed for a few minutes, 
and are proved positive or negative by the fact that in the positives 
after centrifuging a few definite white compact flocculi can be shaken 
from the bottom of the tube, whereas the negatives, at most, show 
a slight grayish sediment which disperses on shaking. The serum 
controls should remain clear. If a precipitate does occur, the serum 
is unsuitable and another specimen must be obtained. The extract 
controls should remain diffusely opalescent, and should show no 
precipitate. 
The Meinecke Reaction.—The Meinecke reaction is one of the 
many precipitation reactions that have been described of late years 
for syphilis, since Jacobsthal’s observation of a precipitate in the 
Wassermann reaction. The Meinecke reaction differs from the 
others in that it adds to the precipitation reaction the question of the 
solubility of precipitates in salt solution. When certain concentra- 
tions of sodium chlorid are added to the ordinary Wassermann anti- 
gen, it is flaked out and remains precipitated. Many substances in 
human serum are prevented from flaking out by the presence of salt, 
such, for instance, as the globulins, and if flocculated by other means 
are redissolved by the addition of salt solution. Meinecke 39 claims 
that if normal and syphilitic sera are precipitated, respectively, with 
dilute solutions of a lipoidal alcoholic antigen in distilled water, the 
precipitate formed by the syphilitic serum is less soluble than is that 
formed from normal sera. The Kahn Reaction 39a—The recent modi- 
fication of the Sachs-Georgi reaction of Kahn has been found satis- 
factory by Miss Rockstraw in our laboratory. The essential point 
in the technique is the antigen preparation which consists in the 
extraction of dried heart muscle with alcohol, in the ice-box for 9 
days and 1 day at room temperature. This avoids the larger lipoid 
contents of the antigen when incubator extraction is used. This is 
used both with and without cholesterin, diluted with salt solution. 
There are many smaller technical points which must be attended 
to for success, and for these we refer to Kahn’s own publications. 
COMPLEMENT OR ALEXIN FIXATION AS A METHOD OF 
DETERMINING THE NATURE OF UNKNOWN PROTEIN 
FORENSIC ALEXIN FIXATION TESTS 
Our preliminary discussions of the principles underlying alexin 
or complement fixation have revealed that alexin is hound not only 
by sensitized cells but also by the specific precipitates formed when 
an unformed protein antigen is mixed with its specific antiserum. 
This discovery, made by Gengou, was attributed by him, it will be 
89 Meinecke. Berl. klin. Woch., 54, 1017, 613, also, 55, 1918, 83. 
**“ Kahn, Archiv. Dermat. and Syphilol., 5, 1922, 570, 734, and 6, 1922, 334. 232 
INFECTION AND RESISTANCE 
remembered, to the presence of “albuminolysins,” or protein sensi- 
tizers, antibodies which have been by many observers regarded as 
separate from the precipitins, but which we believe, for stated rea- 
sons (see p. 211), to be very probably identical with the precipitating 
antibodies or precipitins. However this may be, when a dissolved' 
antigen is mixed with its antiserum alexin fixation is exerted by the 
complex, and this, even when the reacting quantities, antigen and 
antibody, are so small that visible precipitation will not take place. 
For this reason, it is plain, it should be possible by means of com- 
plement fixation to detect amounts of a foreign protein too small to 
be demonstrable by direct precipitation with an antiserum. 
The method has, therefore, been suggested chiefly by Neisser and 
Sachs 40 for the forensic determination of unknown proteins, as an 
adjuvant to, and improvement upon, the forensic precipitin test. Our 
discussion of the principles involved in the introductory paragraphs 
of this chapter will render unnecessary an extensive discussion of 
the reasoning upon which this reaction is based. It is well to remind 
the reader, however, of the facts which we have discussed regarding 
the quantitative proportions which govern the occurrence of precipi- 
tation when an antigen, say human serum, is mixed with its antibody, 
in this case antihuman rabbit serum. The actual precipitation may 
be absent either when an excess of the antigen is used or when the 
antigen is present in too small a quantity. Thus a given quantity 
of the antiserum may precipitate strongly dilutions of the antigen 
ranging from 1-50 to 1-10,000. No precipitation or, at least, a very 
slight one only may occur when concentrations stronger than 1-50 
are used and when the dilution is greater than 1-10,000. Neverthe- 
less, in both cases, alexin fixation may be exerted by the complex 
although no precipitation takes place. As Gay 41 has shown, com- 
plement fixation may be exerted even when a formed precipitate has 
been redispersed by the subsequent addition of more antigen. The 
importance of the forensic reaction of Neisser and Sachs, however, 
lies chiefly in its application to the detection of quantities of un- 
known protein too small to be detected by precipitin reactions. 
The tests are carried out by mixing a dilution of unknown pro- 
tein with given quantities of antiserum, adding small quantities of 
alexin (quantities determined best by previous alexin titration as 
indicated in our section on the Wassermann reaction) ; these reagents 
are left together for a given time at 37.5° C., and then sensitized cells 
are added to determine whether or not the alexin has been bound. 
The table on the following page, taken directly from the article 
of Neisser and Sachs, loc. cit., will not only illustrate the method of 
carrying out the reactions but will also give an indication of their 
extreme delicacy. 
40 Neisser and Sachs. Berl. klin. Woch., Yol. 42, No. 44, 1905, p. 1388. 
41 Gay. Univ. of Cal., “Publications in Pathology,” 1912. PRACTICAL APPLICATIONS OF METHOD 
233 
0,1 human antiserum -f 0.05 complement and variable amounts of different 
normal sera (brought to 1 c. c. volume with salt solution); the mixtures kept 
1 hour at room temperature. Then added 1 e, c. 5 per cent, washed beef 
blood + 0.0015 c. c. amboceptor and left 1-2 hours at 37° C. 
The results are as follows: 
TaMe Taken from Neisser and Sachs, loc, cit., p. 1388 
Amounts 
of 
Hemolysis on addition of serum of: 
normal 
serum 
Man 
, Monkey 
Rat 
Pig 
Goat 
Rabbit 
Ox 
Horse 
0.01 
0 
0 
0.001 
0 
0 
0.0001 
0 
moderate 
com- 
com- 
com- 
com- 
com- 
com- 
0.00001 
slight 
complete 
' plete 
’ plete 
’ plete 
' plete 
'plete 
r plete 
0.000001 
complete 
complete 
0 
complete 
complete 
It will be seen that 0.00001 c. c. of the normal human serum still 
gave almost complete complement fixation of 0.05 c. c. of complement 
in the presence of 0.1 c. c. of the antihuman serum. The table also 
shows that this reaction follows a general law of relative specificity 
so often noted in other reactions, namely that, of all the animals 
tested, the serum of monkeys alone gave reactions with the human 
antiserum; and this in quantities as small as 0.001 cubic centimeter. 
The forensic complement fixation reaction of Neisser and Sachs 
is both theoretically and practically valid. Its extensive use in many 
investigations for theoretical purposes has well established its reli- 
ability. However, it is more complicated and requires much more 
experimental training and care than does the simpler precipitin test, 
and it will rarely occur that an unknown protein is available in 
quantities too small to permit of successful precipitation. 
THE USE OF COMPLEMENT FIXATION TESTS IN THE DIAG- 
NOSIS OF MALIGNANT NEOPLASMS 
A great many attempts have been made to establish a method of comple- 
ment fixation by which a diagnosis of malignant tumors could be made. It 
had been hoped that the substance of malignant tumors might contain a 
form of protein or protein lipoid combination which might represent sub- 
stances specific for such tumors, and might therefore functionate as a specific 
antigen. On this basis it might be possible that the serum of tumor patients 
would contain a specific antibody which could react with a specific antigen 
in tumor extracts, with the resulting formation of an alexin-fixing complex. 
No experimental facts have so far justified our assumption of the pres- 
ence of either specific antigen in tumor extracts, or that of a specific anti- 
body in the serum of such patients. However, we have seen that the Wasser- 
mann reaction is a perfectly useful clinically diagnostic method, in spite of 234 
INFECTION AND RESISTANCE 
the fact that the antigen need not be specific, and the purely empirical basis 
on which the syphilis reaction is at present based has justified extensive at- 
tempts to establish an analogous empirical method for tumor diagnosis. 
The literature on this question is confusing. A number of observers using 
antigens variously prepared from tumor substances have reported favorable 
results. Simon and Thomas 42 report many positive reactions, as do Sanpietro 
and Tesa43 and a number of others. Clowes44 has carried out a reaction 
on sarcoma rats and obtained positive reactions in animals in which the 
tumors were small, negative ones when the tumor had grown to a large size. 
Ranzi, on the other hand, obtained negative results throughout. Ranzi45 
found that normal serum would often give complement fixation with carci- 
noma extracts, also that many tumor extracts and sera of tumor patients 
inhibited complement by themselves. The reactions were so irregular that 
he assumed them to be without value. Recently the subject has been very 
thoroughly investigated by v. Dungern.46 
Yon Dungern claims to have finally evolved a method by which the diag- 
nosis of malignant disease can be made with reasonable accuracy. Like 
the Wassermann reaction his method is purely empirical. He admits that 
probably it is not a specific antibody determination and depends rather 
upon the presence of pathological products of metabolism in the sera of 
tumor patients. The reliability of his method depends upon the observation 
of a number of details which he has determined empirically. 
He obtains his antigen in a purely non-specific manner, using, as just 
stated, for this reaction acetone extracts of human blood cells. We take the 
description of the reaction entirely from his own article in “Weichhardt’s 
Jahresbericht.” The antigen is prepared in the following way: Blood is 
taken from a vein, preferably from a paralytic patient, since v. Dungern 
claims that individual specimens of blood vary, and he has had the best 
results with that of paralytic cases. Clotting is prevented by sodium oxalate 
and the blood cells are thoroughly washed in the centrifuge. To the sedi- 
ment are added 19 volumes of pure acetone (Merck). This is allowed to 
stand three days at room temperature and is occasionally shaken during this 
time. It is then filtered, the acetone evaporated in the incubator at 37° C., 
and the residue taken up in 96 per cent, alcohol. This alcoholic extract is 
diluted before use with four parts of salt solution. Of this final preparation 
0.8 c. c. is used in the individual test. 
Particular precautions must also be taken in the handling of the serum 
of the patient. In his earliest tests v. Dungern determined that the inactiva- 
tion of the tumor sera greatly diminishes their specific fixation properties, 
and for this reason he at first advised that the serum be used unheated. He 
has found recently that the best results are obtained when the serum is heated 
to 54° C., together with a little sodium hydrate solution. He handles the 
blood in the following way: After being taken from the patient it is allowed 
to stand 1 to 2 days in the refrigerator; just before use he adds two parts of 
an /o NaOH solution with one part of serum and heats it for half an hour 
at 54° C. As it is important that the sodium hydrate should contain no 
sodium carbonate, he advises the use of the Kahlbaum preparation. In set- 
ting up the test he uses graded quantities of the mixture corresponding to 
42 Simon and Thomas. Jour. Exp. Med., Yol. 10, 1908. 
43 Sanpietro and Tesa. Cited from v. Dungern in “Weichhardt’s Jahres- 
bericht,” etc., Yol. 8, 1912, p. 163. 
44 Clowes. Jour. A. M. A., 1909, Vol. 52. 
45 Ranzi. Wien. klin. Woch., 1906, p. 1552. 
46 V. Dungern. Munch, med. Woch., Nos. 2, 20, 52, 1912; Berl. klin. 
Woch., 1913, “Weichhardt’s Jahresbericht,” Yol. 8, 1912, p. 163. PRACTICAL APPLICATIONS OF METHOD 
235 
0.2, 0.1, 0.05, and 0.025 c. c. of the original serum. To each of these quan- 
tities he adds the stated quantity, 0.8 c. c. antigen preparation described 
above, and the 0.05 guinea pig complement. Controls must be set up with the 
antigen alone and with the patient’s serum alone to prevent error from 
independent fixation by these substances. These reactions are allowed to 
stand three hours at room temperature, and then one cubic centimeter of a 
5 per cent, solution of beef blood sensitized with two units of hemolytic 
serum is added (as in the Wassermann reaction). It is important to use a 
strongly sensitizing serum, so that not too much of the hemolytic rabbit 
serum must be added to the tubes. Experiments done in this way with 
normal sera usually result in complete hemolysis within one hour, although 
in certain other diseases, i. e., tuberculosis and syphilis, slight inhibition 
may result. However, fixation with the patient’s serum in quantities of 
0.1 c. c. or less is, according to von Dungern, fairly specific for malignant 
tumors, since normal sera treated in the way described usually do not cause 
fixation in quantities of less than 0.2 c. c. and, in syphilis and tuberculosis, 
if fixation is at all present, it is usually not evident in quantities less than 
0.1 c. c. 
With a reaction so carried out von Dungern has examined 244 cases. 
The following tabulation states his results: 
Malignant tumor of 
No. of cases 
Reaction positive* 
Pharynx 
3 
3 
Esophagus 
6 
6 
Stomach 
15 
11 
Rectum 
14 
10 
Larynx 
2 
2 
Tongue 
5 
5 
Bladder 
1 
1 
Breast 
22 
22 
Uterus 
10 
10 
Skin 
8 
7 
Ethmoid bone 
1 
1 
Upper maxilla 
1 
1 
* Taken from von Dungern, "Weichhardt’s Jahresbericht,” Vol. 8, 1912, p. 174. 
We report von Dungern’s results exactly as he states them in his last 
summary, since his well-known experimental ability necessitates serious con- 
sideration of all of his work. We may say, however, that a survey of the 
entire literature of complement fixation in the diagnosis of malignant tumors 
does not yet justify our acceptation of this method as of anything like 
the established value which the similar method has attained in syphilis. 
Complement Fixation in Glanders 
The diagnosis of glanders by the mallein test and by agglutina- 
tion has been recently reenforced by the method of complement fixa- 
tion. In carrying out these tests the method of preparation of the 
antigen is of the greatest importance. The directions which we give 
are those employed in the Diagnostic Laboratory of the Hew York 
Department of Health, under the immediate supervision of Dr. 
McNeil and Miss Olmstead, from whom we have our information. 236 
Infection And resistance 
The particular strain of glanders bacilli employed seems to be of 
little importance. The organisms are grown on 1.0 per cent, acid 
glycerin-potato-agar. This stock culture is transplanted every other 
day. From it cultures are planted upon salt-free veal peptone agar. 
It is of the greatest importance that this medium shall be neutral to 
phenolphthalein. After twenty-four hours in the incubator the 
growth is washed off with distilled water, which also should be neu- 
tral, and the emulsion heated for from four to six hours at 80° C. in 
a water bath. It is then filtered through a Buchner filter simply to 
facilitate subsequent filtration through a Berkefeld “N” or “V” 
filter. After filtration this antigen is again sterilized for one hour 
.at 80° C. and then on two successive days at 56° C. for one-half hour. 
The fixation tests carried out with these antigens have yielded 
■ excellent results as reported by Dr. McNeil47 at the Hew York 
• Serological Society. 
It is unnecessary to give further directions as to the technique of 
"this reaction, since it is simply that of complement fixations in gen- 
eral, the chief difficulty being that of antigen preparation. 
Complement Fixation in Gonorrheal Infections 
There are certain conditions following gonococcus infection of 
the genito-urinary tract which are not easily distinguished from a 
number of other tests unless the organisms can be cultivated or a 
specific serum reaction can be applied. Most important of these are 
gonorrheal rheumatism, salpingitis, and endocarditis. Complement 
fixation with the sera of such patients, and an antigen produced from 
gonococci, has been employed by many observers during recent years, 
and promises to be of great value. 
Here, too, the production of the antigen is the only feature of 
the reaction which has offered difficulties. Since the researches of 
Torrey have shown that not all races of gonococcus are antigenically 
alike, it seems necessary to produce a polyvalent antigen. At the 
Hew York Department of Health at present the antigen is prepared 
by using the ten Torrey strains. Stock cultures are carried on neu- 
tral veal agar and cultures are planted upon salt-free veal agar. 
Twenty-four-hour growths, washed off in neutral distilled water, are 
kept in a water bath at 56° C. for two hours, and then filtered 
through, first, a Buchner and then a Berkefeld filter. They are then 
sterilized for one hour. 
Complement Fixation in Tuberculosis 
Attempts to apply complement fixation to the diagnosis of tuber- 
culosis have been made by as many as thirty or more investigators 
47 McNeil, Archibald. N. Y. Serological Soe. Meeting, April 4, 1914. PRACTICAL APPLICATIONS OF METHOD 
237 
with varying results. Wassermann and his collaborators attempted 
it before they succeeded in developing the Wassermann reaction in 
syphilis. Recently, intensive work has been done on the subject by 
Besredka,48 Petrol!,49 Craig,50 Bronfenbrenner,51 and Miller and 
Zinsser.52 Results have warranted the application of the reaction to 
clinical tuberculosis, although the actual degree of usefulness of the 
reaction must still await the multiplication of cases tested. The diffi- 
culty has of course consisted in the development of a suitable antigen. 
The antigen of Petroft' has consisted of a filtrate of potato broth. 
Besredka has used a filtrate of cultures made upon egg broth. Craig 
has used Besredka’s antigen and suspensions of ground bacilli. Mil- 
ler and Zinsser 52 have employed an antigen made by triturating liv- 
ing and dead bacilli with crystals of table salt, then adding distilled 
water to isotonicity. With all of the antigens favorable results have 
been obtained. Our own results seem to check up with clinical 
diagnoses in over 80 per cent, of the cases and so far have appeared 
to give practically no positive results in negative cases and indicated 
only active tuberculosis and not healed lesions. 
The reactions are carried out by methods entirely analogous to 
those employed in the Wassermann reaction, but careful antigen 
titrations must be done. 
The Preparation of Bacterial Antigens for Alexin Fixation 
Reactions.—We have mentioned only a few of the more commonly 
used alexin fixation reactions with bacteria. Alexin fixation can be 
used for experimental and diagnostic purposes in the case of almost 
any species of bacteria. The technical difficulties have involved 
chiefly the production of a suitable antigen, since many bacterial 
suspensions and extracts are non-specifically anticomplementary, or, 
in other words, these bacterial preparations fix alexin by themselves 
without previous sensitization. The reactions, for this reason, have 
necessitated a very careful titration of the maximum amounts of 
alexin fixed by the unsensitized antigen in order that this might be 
allowed for in the final reaction. Such a procedure, when the anti- 
complementary action of the suspension is at all marked, involves 
delicate manipulation and titrations, and great care in interpreta- 
tion. 
We have recently made studies of bacterial antigens which are 
mentioned in other parts of this book; but it may be useful at this 
point to call attention to the practical advantages for all kinds of 
bacterial alexin fixations represented in the use of our so-called 
48 Besredka. Zeitschr. f. Immunity 1914. 
49 Petroff. 
50 Craig. Am. Jour. Med. Sc., Dec., 1915, p. 781. 
51 Bronfenbrenner. Arch. Int. Med., 1914, No. 6, 786. Zeitschr. f. 
Immunit., 1914, XXIII, 2, 21. 
52 Miller and Zinsser. Proc. Soc. Exper. Biol, and Med., 1916, 134. Jour. 
Lab. and Clinical Med., Yol. I, p. 817. 238 
INFECTION AND RESISTANCE 
residue antigens. These substances which are antigenic in the sense 
that they unite with antibody, giving both clear precipitations and 
alexin fixations with specific sera and almost entirely devoid of anti- 
complementary action, consist in the residue left when alkalin ex- 
tracts of bacteria are boiled for a short time with acid, filtered clear 
and neutralized. The antigens may be further purified by precipita- 
tion in the neutral residues with ten volumes of alcohol, and resolu- 
tion in normal salt solution. The technique for the preparation of 
these antigens from different bacteria varies. 
Tubercle Bacilli.—This antigen devised by us 52a is being used 
extensively by Petroff at the present time. The result of his 
investigations has not yet been published. In the case of tubercle 
bacilli, the antigens are made by shaking pulverized tubercle bacilli, 
about 50 milligrams in 100 c. c. of slightly alkaline salt solution. 
After about three hours’ shaking, a massive precipitate is obtained 
by acidifying with acetic acid, added drop by drop. This is filtered 
off and the filtrate boiled for from 2 to 5 minutes over a free flame, 
still in the acid condition. The liquid is allowed to cool and the 
slight cloud which forms, is filtered off by passing through a Berkfeld. 
The filtrate is neutralized to about Ph 7, and the precipitate of phos- 
phates which forms as the neutral point is approached, again removed 
by filtration. The water-clear residue which gives a slight precipita- 
tion when mixed with about 10 volumes of alcohol can be used 
directly, and titrated for alexin fixations without further manipula- 
tion, or may be precipitated with 10 volumes of alcohol and taken 
up in any desired concentration with sterile salt solution. In this 
form it can be kept a long time and can be boiled repeatedly without 
losing its antigenic properties. 
Influenza Bacilli.52b—Influenza bacilli are grown on chocolate 
agar in large pie plates, or any moderately large surface, washed off 
with salt solution and shaken in neutral salt solution for about two 
hours, then filtered and treated as above. The influenza bacillus 
antigen is much less stable than the tubercle bacillus preparation. 
Pneumococci.—Pneumococci are washed off agar surfaces, 
shaken a short time, not more than an hour is necessary, in slightly 
alkaline salt solution, acidified, boiled, etc., and in other ways treated 
as described for the tubercle bacillus above. The pneumococcus anti- 
gen is extremely stable. 
Meningococcus.—Meningococcus antigen is prepared just ex- 
actly as for pneumococci, only in this case, as in the influenza bacillus, 
great care must be taken to preserve the final residue in a slightly 
acid condition and to apply no heat except when the antigen is slightly 
acid. If alkaline boiling weakens and eventually destroys it. 
Staphylococci.—Staphylococcus residue antigen may be pre- 
Baa Zinsser. Jour. Exp. Med., Yol. 34, 1921, p. 495. 
Bab Zinsser and Parker. Jour. Exp. Med., Yol. 37, 1923, p. 275. PRACTICAL APPLICATIONS OF METHOD 
239 
pared in the same way, except that with the staphylococci it is neces- 
sary to collect dried bacteria, grind them and extract the pulverized 
staphylococci as the tubercle bacillus. They are subsequently treated 
like tubercle bacillus powder. 
The only organisms, so far worked with, with which we have not 
succeeded in obtaining satisfactory amounts of residue antigen are 
typhoid bacilli. CHAPTER IX 
THE PHENOMENON OF AGGLUTINATION 
When bacteria are added to homologous immune serum there 
occurs a peculiar agglomeration of the individual micro-organisms 
into small clumps. The phenomenon is so general and so easily 
observed that it is not surprising that it was noticed and reported by 
a number of workers during the period of active investigation upon 
serum reactions which preceded and followed the discovery of the 
Pfeiffer phenomenon. Thus, in the years from 1891 to 1895, 
Metchnikoff,1 Charrin and Roger,2 Isaeff and Ivanoff,3 Washburn,4 
and several other workers made this observation with a variety of 
bacteria and immune sera. But all of these observers failed to follow 
up or analyze the process they incidentally noticed in the course of 
other investigations. A thorough study of the phenomenon was not 
made until 1896, when Gruber and Durham,5 in Vienna, in the 
course of their studies upon bacteriolytic reactions with colon bacilli 
and cholera spirilla, again noticed the agglutination of these bacteria 
in homologous immune sera, recognized the specificity of the reaction, 
and called attention to its apparent independence of other previously 
studied serum phenomena. 
The process known as agglutination consists in the following 
train of occurrences. If we add to an even emulsion of bacteria a 
small amount of homologous immune serum the micro-organisms 
may be seen to collect rapidly in groups or masses, with a resultant 
clearing of the fluid in which they have been suspended. The clumps 
of bacteria gather in flakes which, not unlike flakes of snow, sink 
to the bottom of the test tube. The speed and completeness with 
which this phenomenon occurs depend, as we shall see, upon the 
agglutinating strength and other qualities of the serum which is em- 
ployed, but the essential process of clumping is alike in all cases. 
There are a large number of different methods by which this 
occurrence can be observed, each one particularly adapted to some 
special purpose for which the reaction is carried out. Gruber and 
Durham, who were investigating the properties of bacteriolysins 
1 Metchnikoff. Ann. de Vlnst. Past., 1892. 
2 Charrin and Roger. C. R. de la Soc. de Biol., 1889. 
3 Isaeff and Ivanoff. Zeitschr. f. Hyg., Vol. 17, 1894. 
4 Washburn. Jour. Path, and Bact., 1896, p. 228. 
6 Gruber and Durham. Munch, med. Woch., 1896. 
240 THE PHENOMENON OF AGGLUTINATION 
241 
when they observed agglutination, naturally recognized the specific 
nature of the reaction and proposed to make use of it for the purpose 
of bacterial differentiation and species determination. For this pur- 
pose, which has become one of the most important of the practical 
applications of the agglutination reaction, the phenomenon is best 
observed by the so-called “macroscopic method,” in which a series of 
serum dilutions are mixed, in small test tubes, with equal volumes 
of emulsions of the bacteria. Thus, if we wish to determine the 
nature of an unknown bacil- 
lus, suspected of belonging to 
the typhoid bacillus group, by 
this method, we may deter- 
mine its agglutination in the 
serum of an animal im- 
munized with a known strain 
of typhoid. The tubes are in- 
cubated after the mixtures 
have been made, and the 
agglutination which has taken 
place in the various tubes is 
recognized by a clearing up 
of the fluid and the flaking 
of the bacteria after from one 
to three hours. The test tube 
method has the advantage of 
permitting the use of larger 
quantities of reagents than can be used in the other methods de- 
scribed below, and therefore more exact quantitative measurements 
can be made. 
Although this method for the determination of bacteria has found 
universal application, it is probably most frequently employed at the 
present time for the rapid identification of colonies of doubtful 
typhoid or dysentery, incident to the isolation of these organisms for 
stools by such methods of plating as those of Conradi-Drigalski, of 
Endo, or of Hiss. The suspicious colonies can thus be fished directly 
to an agar slant, and the cultures, when developed, emulsified and 
identified by agglutination. The advantages of such a method for 
the determination of the biological interrelationship of the organisms 
of a given group, like, for instance, that of the dysentery bacilli, are 
obvious. 
An ingenious use of this reaction was also made by Shiga when 
he determined, among various bacteria isolated from the stools of 
dysentery cases, the particular one which was specifically aggluti- 
nated by the patient’s serum, thus discovering the dysentery bacillus 
.which bears his name. 
.Within a few months, after the publication of Gruber and Dur- 
Microscopic Agglutination. 242 
INFECTION AND RESISTANCE 
ham’s work, Widal and, apparently independently of him, Griin- 
baum,*5 by a process of reasoning the converse of that detailed above, 
applied the reaction to the diagnosis of infectious disease. 
It is obvious that a human being or an animal infected with a 
given variety of bacteria may develop agglutinating properties 
against these bacteria. It is of great value, therefore, to determine 
the agglutinating power of the serum of a patient for the bacteria 
which are known to cause the disease suspected in the particular 
case in which a diagnosis is desired. This method has become a rou- 
tine measure in the early diagnosis of typhoid fever under the name 
of “Widal” or “Gruber-Widal” reaction and, since the quantities of 
serum which can easily be obtained from a patient are usually small, 
it is convenient to carry out the reaction by the microscopic method. 
This consists in mixing serum and bacterial emulsion in hang-drop 
preparations and observing them with the microscope. An excellent 
method, also, is the so-called Proeseher 7 method in which serum and 
bacterial emulsion are mixed in small watch-glasses or salt cellars. 
Proeseher used this method extensively in the study of staphylococcus 
agglutinations. The mixtures in the salt cellars were set away at 
37° C. for two hours, and then observed with a magnification of 60 
to 70 diameters. 
Close observation of the occurrence under the higher power of a 
microscope shows that the bacteria, if motile, lose their motility, if 
non-motile the Brownian motion is arrested. They are then rapidly 
gathered in small clumps, isolated individuals between these clumps 
being gradually drawn into them, until finally the fluid between the 
masses is entirely clear. This complete clearing, of course, happens 
only when there is not an excess of bacteria, for, like other serum re- 
actions, this phenomenon is a quantitative one in which definite pro- 
portions must be maintained. 
Clinically the most frequent use of the agglutination reaction is 
in the diagnosis of typhoid fever. The technique used for this test 
is, in the large majority of cases, the microscopic hang-drop method. 
In Germany the Proeseher method is sometimes used, and the micro- 
scopic method with dead organisms, as first introduced by Ficker, is 
also not uncommon at the present day. 
Since the serum of normal human beings very often contains 
moderate agglutinating powers for the typhoid bacillus, the diag- 
nostic value of the reaction in this disease depends upon the elimina- 
tion of this error by sufficient dilution. If dilutions of the serum of 
from 1-40 to 1-60 are used diagnostic errors on this point are 
avoided, since the normal agglutinating power of human beings is 
never such that typhoid bacilli will be clumped by it in these dilu- 
tions within one hour. Prompt clumping in serum dilutions of 1-20 
6 Griinbaum. Lancet, 1896, Yol. 2. 
7 Proeseher. Centralbl. f. Bakt., Yol. 34, 1903. THE PHENOMENON OF AGGLUTINATION 
243 
is fairly reliable, but does not absolutely exclude an unusually high 
normal agglutinating power. In carrying out tests clinically dilu- 
tions of 1-20, 1-40, and 1-80 are usually made and observed for one 
hour. From such tests diagnosis can be made without danger of 
error. In rare cases of icterus the agglutinating power for typhoid 
bacilli may be increased. Just what is the cause of this is not cer- 
tain; Wood 8 reports cases in which agglutination of 1-40 was pres- 
ent with slight jaundice (hepatic cirrhosis). On the other hand he 
has frequently failed to notice agglutination in other cases of intense 
jaundice. It is not impossible, as Wood suggests, that the occasional 
presence of unusual agglutinating power in individuals with jaun- 
dice has some relation to the frequent persistence of typhoid bacilli 
in the gall bladder. 
Occasionally it will be noticed that dilutions of the patient’s 
serum of 1-5 to 1-20 fail to agglutinate, while higher dilutions will 
give positive tests. This is referable to the so-called “pro-agglu- 
tinoid zone,” the principles underlying which we shall discuss in an- 
other place. 
The Widal test in typhoid cases rarely appears before the end 
of the first week, and, in the majority of cases, is present before the 
end of the second week. It may proceed for months, although Wood 
states that he has seen it disappear at the end of three to six weeks. 
In paratyphoid fever the diagnosis can often be made by agglu- 
tination, and in dysentery, as we have seen, the fact that the pa- 
tient’s blood agglutinated the bacteria was one of the important facts 
utilized by Shiga in his discovery of the organism which bears his 
name. 
In pneumonia agglutination of the pneumococcus, isolated from 
the patient’s sputum by sera prepared by immunization with various 
types of pneumococci, has become of considerable importance clin- 
ically, since Heufeld and Haendel and, in this country, Cole, Dochez, 
and Gillespie have determined that there are a number of different 
types of this micro-organism. The use of pneumococcus serum in the 
disease will be of value only if a serum is used which has been pro- 
duced with an organism of the same type as the one infecting the 
patient. Therefore, determinations of the type by highly potent 
agglutinating sera give a finger-point to the variety of serum to be 
used. Whatever may be the eventual outcome of the serum treat- 
ment in pneumonia, no results whatever can be expected, according 
to our present knowledge, unless such determinations are made. The 
technique of agglutinations in pneumococcus work is facilitated by 
growing mass cultures of organisms, as advised by Hiss, in flasks of 
glucose broth containing 1 per cent, calcium carbonate, but ordinarily 
no difficulty is encountered and no special methods are necessary. 
8 Wood. “Chemical and Microscopical Diagnosis,” D. Appleton & Co., 
p. 242. 244 
INFECTION AND RESISTANCE 
The same method of growing micro-organisms is useful in the 
case of streptococcus agglutinations, since the insoluble calcium car- 
bonate, if thoroughly shaken, breaks the chains of streptococci and 
thereby facilitates judgment as to the reaction. 
Agglutination reactions have been of considerable usefulness also 
in the diagnosis of glanders in horses. The early work on this sub- 
ject was done chiefly by Macfadyean,9 and the reaction has been 
particularly studied by Wladimiroff.10 Since the serum of normal 
horses will often agglutinate glanders bacilli in dilutions of as much 
as 1-500, Wladimiroff advises making the positive diagnosis on dilu- 
tions only higher than 1-1,000, since he states that normal horses may 
occasionally reach an agglutination titre of 1-1,000. The same writer 
states, moreover, that glanders bacilli are subject to great variations 
in agglutinability, and that for this reason the choice of a suitable 
strain is of great importance. 
The motility of bacteria has absolutely no relation to the reac- 
tion, and their agglutination is entirely passive. 
Some of the earlier investigators of agglutination associated the 
reaction with alteration in the flagellar mechanism of the micro- 
organisms. It is now well known, however, that non-motile, as well 
as motile, bacteria are subject to the phenomenon, and that no visible 
change in the appearance or arrangement of flagella accompanies the 
clumping. Although this is the case, observation of the motility of 
such organisms as the bacillus of typhoid fever, while subjected to 
the action of agglutinating serum, may be of great value in aiding 
in the determination of the degree of completeness with which the 
reaction is taking place. 
Agglutination, furthermore, does not lead to the death of the 
bacteria. Of course, whenever the reaction is carried out in un- 
heated serum the concomitant effects of the bactericidal substances 
bring about bacterial death. Agglutination does not, however, de- 
pend upon the cooperation of alexin, and serum may be inactivated 
without interference with its power of agglutination. In such heated 
serum clumping takes place without bactericidal effects, and, more 
than this, the bacteria may grow, if exposed to proper temperature 
conditions, when suspended in the serum. In fact, it is of consider- 
able interest to carry out the reaction in this way, for the bacteria 
growing in agglutinating serum form long convoluted threads and 
skeins even when in ordinary culture they habitually occur as sep- 
arate individuals. Thus colon bacilli, typhoid bacilli, pneumococci, 
cholera spirilla, and other organisms, which ordinarily grow as free 
single cells, or, at most, in chains of two or three, if kept in the 
incubator for ten to twelve hours together with homologous serum, 
will grow in long, delicate chains, like those of streptococci. This 
0 Macfadyean. Jour. Comparative Path, and Ther., Vol. 9, 1896. 
10 Wladimiroff. “Kolle u. Wassermann Handbuch,” 2d Ed., Yol. 5. THE PHENOMENON OF AGGLUTINATION 
245 
form of reaction has been especially studied by Pfaundler,11 who 
attributed particularly delicate specificity to it. However, the 
“Thread Reaction,” as it is sometimes called, may be regarded as 
merely another manifestation of the phenomenon of agglutination 
and subject to the same laws and limitations of specificity which 
apply to other methods. 
1 he purely passive role played by the bacteria in agglutination 
is best shown by the fact that dead bacteria, killed in various ways, 
are specifically clumped just as are the living cultures.12 On this 
fact depends the method spoken of as “Ricker’s Reaction,” in which 
emulsions of typhoid bacilli, killed by formaldehyd or carbolic acid 
(distributed commercially), are agglutinated in small test tubes by 
the serum of typhoid patients. The original method of Ricker is 
said to be a proprietary secret; however, a number of other methods 
which attain the same purpose are in use in various places. Volk 13 
describes the method used in Vienna, and states that there carbolic 
acid is used to kill the cultures. Similar to this is the method de- 
scribed by J. H. Borden,14 who proceeds as follows: 
The bacilli are grown on agar slants in large tubes for 24 hours. 
They are then washed from the medium with a sterile mixture of 
salt solution 450 parts, glycerin 50 parts, and 95 per cent, carbolic 
acid 2.5 parts. After this solution has been kept a week it becomes 
translucent and by this time the bacilli are dead. The preparation 
is then ready for use and can be kept a long time in dark bottles in a 
cool place. Borden very carefully controlled this bacterial emulsion 
with positive and negative typhoid sera and found it reliable. The 
great advantage of all these methods, of course, consists in the possi- 
bility of furnishing the general practitioner with materials for clini- 
cal agglutination tests in which the necessity of preserving and sus- 
pending living cultures is eliminated. 
The facts which we have just considered tend to show that agglu- 
tination is not a vital phenomenon 15 dependent in any way upon 
the living nature of the bacterial cell, but, like other phenomena of 
antigen-antibody reactions, a purely chemical or physical process in 
which the substance of the bacterial cell enters specifically into rela- 
tion with the agglutinating factor of the serum. In uniformity with 
other analogous reactions the antigenic substance is here spoken of 
as “agglutinogen,” the antibody as “agglutinin.” 
Agglutinogen.—The agglutinogen, or agglutinin-inducing sub- 
stance in the bacteria is apparently an inherent part of the bacterial 
protein, and agglutinins may be produced in animals by injection 
11 Pfaundler. Wien. klin. Woch., 1898, and Centralbl. f. Bakt., I, Vcl 23 
1898. 
12 Bordet. Ann. Past., Yol. 10, 1896. 
13 Volk. “Kravis und Levaditi Handbuch,” Yol. 2. 
14 Borden. Medical News, N. Y., Mar., 1903. 
15 Bordet. Ann. Past., Yol. 10, 1896. 246 
INFECTION AND RESISTANCE 
not only of living and dead whole bacteria, but by bacterial extracts, 
prepared in various ways. And, furthermore, just as the agglutinins 
of serum are absorbed out of a serum by the whole bacteria, they may 
be neutralized by the dissolved bacterial extracts. 
Just what the nature of the agglutinogen is has been a matter 
of prolonged controversy, Pick 16 and others claiming that it is pos- 
sible to obtain an agglutinogen by alcohol precipitation from old 
bacterial cultures which, upon further purification, can be found to 
give none of the usual protein reactions (Biuret, Millon). It is by 
no means certain, however, that Pick’s results are correct. That the 
agglutinogen is, to a certain extent, subject to dialysis has been 
claimed because of experiments in which specific agglutinins have 
appeared in the sera of animals into whose peritoneal cavities celloi- 
din sacs, filled with bacteria, have been placed.17 
There has been a great deal of discussion regarding the possible localiza- 
tion of the agglutinogen of bacteria in the ectoplasmic layers of the cells, 
and especially in the flagellar substance. We have seen that, as a matter 
of fact, nonmotile bacteria are subject to the phenomena of agglutination 
just as are the motile forms, but numerous attempts were made during the 
earlier stages of our knowledge of this reaction to demonstrate that changes 
in ectoplasm and flagella accompanied the actual agglutination. Gruber18 
himself held the opinion for a time that the clumping was due to an ectoplas- 
mic swelling which rendered the bacteria sticky, causing them to hold to- 
gether after chance approximation. He soon gave up this idea himself, 
but a similar theory was for some time upheld by Nicolle 10 and others. 
Malvoz 20 in 1897 devised an ingenious method by which he believed that 
he could show that the agglutination of bacteria depended upon their ecto- 
plasmic substances. He passed the typhoid emulsion through Chamberland 
filters and, when the bacilli had been caught upon the filters, he subjected 
them to prolonged washing. The bacilli, now removed from the filter by 
passing fluid through in the opposite direction, were no longer motile or 
agglutinable either by formalin, safranin, or other chemical agents, nor 
by agglutinating sera. Dineur,21 repeating the experiments of Malvoz, came 
to the same conclusions. He decided that in agglutination the bacteria be- 
came entangled with each other by means of the flagella. Harrison,22 in 
later studies working under Tavel, attempted to dissolve out the ectoplasmic 
layers of bacteria with pyocyanase, and from his experiments also came to 
the conclusion that the agglutinogen was contained in the external layers. 
Similar results were obtained by De Rossi.23 
16 Pick. “Hofmeister’s Beitrage,” 1901, 1902. 
17 This would be in keeping with Pick’s work just referred to, and should 
be subjected to the same criticism before final acceptance. For a more de- 
tailed discussion of these conditions the reader is referred to the article by 
Paltauf, “Kolle u. Wassermann Handbuch,” Yol. 4, part 1. 
18 Gruber. Munch, med. Woch., 1896. 
19 Nicolle. Ann. de I’Inst. Past., 1898. 
20 Malvoz. Ann. de VInst. Past., Yol. 11, 1897. 
21 Dineur. Bull, de I’Acad. de Med. de Beige, 1898, cited from Smith and 
Reagh. 
22 Harrison. Centralbl. f. Baht., Yol. 30,1, Orig. 1901. 
23 De Rossi. Centralbl. f. Baht., I, Vols. 36 and 40. THE PHENOMENON OF AGGLUTINATION 
247 
Further studies on the same problem are those of Smith and Reagh.24 
These investigators worked with two strains of bacilli, both of which they 
regarded as belonging to the hog-cholera group, though the one was motile 
and the other nonmotile. When rabbits were immunized with the nonmotile 
bacillus an agglutinin was obtained which acted upon this bacillus differently 
and less powerfully than did the agglutinin produced with the motile one. 
Contact with the nonmotile bacillus did not deprive the serum produced 
with the flagellated organism of the agglutinins for the latter. They con- 
clude that two agglutinins were involved—one incited by the ectoplasm and 
flagellar substance, the other by the bacterial cell body proper. Rehns as 
well as Paltauf have criticized these results by questioning the species identity 
of the two bacilli employed in the experiments, referring the phenomenon 
to the occurrence of group agglutination. 
As a matter of fact our present knowledge of agglutination no 
longer justifies the association of agglutination with flagella. ISTon- 
motile as well as motile bacteria are readily agglutinated, and we 
have much evidence which will be discussed presently which shows 
that the agglutination reaction is governed by many of the laws 
which obtain in colloidal flocculations. This, however, does not ex- 
clude the possibility that it is the ectoplasmic zone chiefly which 
takes part in the reaction. Furthermore, loss of motility, which 
always accompanies agglutination when a motile organism is under 
observation, is an extremely valuable aid in guiding us in our judg- 
ment of incomplete reactions. 
That changes may be brought about in bacterial agglutinogen by 
various methods of treatment has been shown by a number of workers, 
although the fundamental principles underlying such changes are not 
at all clear. 
Joos 25 was the first to study agglutination with particular refer- 
ence to the effects upon the reaction of heating both the antigen and 
the antibody. On the basis of extensive and complicated experiments 
upon the agglutinin produced in horses by immunization with heated 
and unheated typhoid bacilli, he drew the conclusion that agglu- 
tinogen (agglutinin-inducing substance) in bacteria was not a single 
element but consisted of at least two definite parts of which he speaks 
as a and 5-agglutinogen, is a constituent of the 
living bacteria, and is easily destroyed at 60° to 62° C. The 5-agglu- 
tinogen is also present in normal bacilli, but is more heat-resistant 
and will withstand 60° to 62° C. for several hours. The injection of 
living unheated bacilli then induces the formation of both a and 
5-agglutinin, which have respectively a particular affinity for a and 
5-agglutinogens. The injection of heated bacilli, on the other hand, 
induces the formation only of 5-agglutinin and not a trace of a.-agglu- 
tinin. The a-agglutinin is considerably heat-resistant, resisting 60° 
to 62° C., whereas the 5-agglutinin loses its agglutinating property 
24 Smith and Reagh. Jour. Med. Res., Yol. 10, 1903. 
25 Joos. Centralbl. f. Bakt., Yol. 33, 1903. 248 
INFECTION AND RESISTANCE 
when heated to GO0 C. The a-agglutinin is entirely incapable of unit- 
ing with B-agglutinogen. However, i?-agglutinogen can combine or 
react with both the a, and B constituents of the bacilli. For this 
reason Paltauf has spoken of agglutinin produced with the heated 
bacteria as “umfanglicher.” This is a point of great interest, and 
if Joos is right is, of course, of considerable practical importance. 
However one may look upon these experiments, as well as the 
similar ones of other workers, it seems established that heating bac- 
teria leaves them still capable of inciting agglutinins powerfully 
and rapidly, perhaps of an “umfanglicher” nature than those pro- 
duced with the native cells. 
Heating bacteria may also alter their agglutinability. Thus, ac- 
cording to Eisenberg and Volk,26 heated above 65° C. the bacteria 
no longer agglutinate in the presence of specific immune serum, but 
still absorb agglutinin. Eisenberg and Volk, therefore, distinguished 
between a heat-sensitive constituent of the antigen, which is particu- 
larly associated with the clumping, whereas the thermostable sub- 
stance represents the haptophore or combining portion. It seems 
simpler, in this case also, to assume a change in the colloidal stability 
of the bacteria by heating than to seek it in a differentiation into 
combining and agglutinable parts of the same antigen. 
The points raised by Joos’ work have been followed up particu- 
larly by Kraus and Joachim27 and by Scheller.28 Scheller sum- 
marizes the results of his work as follows: First, in agreement with 
Joos he found that immune sera obtained by injection of bacteria 
modified by heat vary considerably from each other. Secondly, im- 
munization with living typhoid bacilli produces sera which agglu- 
tinate living bacilli very highly and less highly bacilli heated to 60° 
C. The titre of agglutinating serum is altered very little toward 
living bacilli after heating to 60° to 62° C., but is sometimes dimin- 
ished toward bacteria that have been heated. Bacilli that have been 
heated to 100° C. but slightly agglutinate unheated serum. Sera 
produced by the injection of typhoid bacilli heated to 60° to 62° C. 
agglutinate with both living and heated bacilli. Very important 
furthermore in Scheller’s work are the determinations that typhoid 
bacilli which have been heated absorb agglutinins out of the sera 
more actively than do the unheated bacteria, and that the highest 
agglutinin titres can be obtained by agglutination with bacilli that 
have been heated to 60° C. The analogy of Scheller’s results with 
similar work done in connection with the precipitin reaction is strik- 
ing and will be referred to in another place. 
Alterations in the agglutinability of bacteria may also occur 
spontaneously, without previous heating, as in the preceding experi- 
26 Eisenberg and Yolk. Zeitschr. f. Hyg., Vol. 40, 1902. 
27 Kraus and Joachim. Centralbl. f. Bakt., I, Vol. 36, 1904. 
28 Scheller. Centralbl. f. Bakt., Vol. 36, 1904, and Vol. 38, 1905. THE PHENOMENON OF AGGLUTINATION 
249 
merits. It has been frequently noticed that typhoid bacilli, recently 
cultivated out of the human body, will not agglutinate in sera which 
have high agglutinating power for strains kept for some time on 
laboratory media. Much investigation has been focused upon the de- 
termination of the cause for this, and although many explanations 
have been suggested no satisfactory solution has been found. Most 
workers who have attempted to attack this problem have based their 
reasoning upon the receptor conception of Ehrlich and have assumed 
that such inagglutinable bacteria are characterized by a diminished 
equipment in “receptors.” Such strains have been especially well 
studied in the case of typhoid bacilli and cholera spirilla. Inagglu- 
tinable typhoid bacilli have been cultivated by many investigators 
from the spleen, gall bladder, and urine of typhoid patients, and 
many of these, when studied for prolonged periods, have been found 
to regain normal agglutinability after several generations of culti- 
vation upon artificial media. Apparently some alteration of the 
bacilli had taken place in the presence of the body fluids (immune 
serum) which affected their sensitiveness to the agglutinins, i. e* 
their ability to unite with or absorb this antibody. The phenomenon 
involves an important principle, emphasized some years ago by Pro- 
fessor Welch, namely, that the bacteria may acquire a quasi- 
immunity against the attacking forces of the body, a property which 
may be responsible for the increase of virulence noted when some 
bacteria are repeatedly passed through the bodies of animals, and, 
indeed, alterations of virulence signify biologically a process of 
adaptation on the part of the bacteria just as increased immunity 
indicates a similar process on the part of the invaded subject. 
This lessened susceptibility to antibodies is noticeable not only in 
strains cultivated from the body in disease, but can be produced arti- 
ficially by cultivating the bacteria in inactivated homologous immune 
serum. This has been accomplished by Walker 29 especially, and by 
Muller,30 with both typhoid bacilli and cholera spirilla cultivated 
upon broth mixed with serum. Such strains not only increase in 
virulence but lose in both agglutinability and susceptibility to bac- 
tericidal effects. Sacqueppee 31 obtained similar results by keeping 
the organisms in collodium sacs in the peritoneal cavity, and Bail 32 
found similar inagglutinability of typhoid bacilli taken from the 
peritoneal exudates of guinea pigs dead of infection. 
Gay and Clay pole in 1913 33 produced the carrier state in rabbits 
with typhoid bacilli, and found that the organisms reisolated from 
the rabbits did not agglutinate in a serum which agglutinated stock 
29 Walker. Jour. Path, and Bact., Yol. 8, 1902. 
30 Muller. Munch, med. Woch., 1903. 
31 Sacqueppee. Ann. Past., Yol. 4, 1901. 
32 Bail. Archiv. f. Hyg., Vol. 42. 
33 Gay and Claypole. Archiv. Intern. Med., 12, 1913, 614. 250 
INFECTION AND RESISTANCE 
cultures in dilutions as high as 1:20,000. The writer, many years 
ago, noticed similar inagglutinability in a typhoid bacillus isolated 
from a case of gall stones which developed some 17 years after 
typhoid fever. A curious by-product of Gay and Claypole’s observa- 
tion is the fact that they claim that the blood and bile cultures which 
were not agglutinated by ordinary anti-serum were easily aggluti- 
nated by a serum produced by the immunization of rabbits with cul- 
tures grown on blood agar media. These observations, however, 
were not confirmed in an extensive investigation on this point by 
Bull and Pritchett.34 The artificial production of inagglutinability 
by cultivation on immune sera was again extensively worked out in 
1921 by Morishima 35 in our laboratory. Morishima’s conclusions 
are as follows: Typhoid bacilli grown upon normal serum do not 
become inagglutinable. Cultivated continuously upon specific im- 
mune serum, they at first become inagglutinable, but if such cultiva- 
tion is persisted in for two weeks or longer, eventually these strains 
again become agglutinable. In some cases this return to normal 
does not occur until the 72nd day; usually, however, it occurs sooner 
than this. The inagglutinability of the typhoid bacillus is accom- 
panied by inability to absorb agglutinin. The development of capsu- 
lar material does not seem to be concerned in the inagglutinability, 
since treatment by the Porges method, that is, washing with weak 
acid, does not render the inagglutinable strains agglutinable. 
Zinsser and Dwyer 36 have noticed similar inagglutinability in 
typhoid bacilli recovered from the peritoneal cavities of guinea pigs 
injected with anaphylatoxin and bacteria. The anaphylatoxin in 
these cases possessed distinct aggressive action, and the conditions 
here were probably very similar to those observed by Bail. 
There are various possible explanations, the most prevalent ones 
all representing variations of the opinion that such inagglutinable 
strains possess an inadequate receptor apparatus. Cole 37 advances 
this because he found these cultures possessed less power to absorb 
agglutinin than others, and, injected into animals, produced sera 
which were not highly agglutinating for the injected strain. Some 
of Cole’s experiments show clearly the variable agglutinability dis- 
played by different strains of the same species. Thus the agglutina- 
tion in the same serum. 
Of strain E = 1:8,000 
Of strain H = 1:7,000 
Of strain I = 1:4,500 
Of strain W = 1:4,500 
Of strain C = 1:4,000 
34 Bull and Pritchett. Jour. Exper. Med., 24, 1916, 55. 
35 Morishima. Jour. Bacter., 6, 1921, 275. 
86 Zinsser and Dwyer. Proc. Soc. for Exp. Biol, and Med., Feb., 1914. 
37 Cole. Zeitschr. f. Hyg., Yol. 46, 1904. THE PHENOMENON OF AGGLUTINATION 
251 
The difference here between E and C actually amounted to a 
relation of 1 to 2. A rabbit immunized with strain I furnished a 
serum which agglutinated strain E more powerfully than I itself. 
Muller’s experiments have the same general significance. It has 
also been suggested that the inagglutinable bacteria, especially those 
from the peritoneal exudate, which Bail found were neither agglu- 
tinable nor absorbed agglutinin, may have taken up altered agglu- 
tinin or agglutinoid. We will have occasion to recur to this problem 
in connection with our discussions of the capsulated bacteria and of 
virulence. The explanations given above do not seem on the whole 
satisfactory, and the problem is an exceedingly complex one. It has 
been found indeed that the acquired resistance of bacteria against 
agglutinins is not at all unique, and that acquired resistance against 
serum lysins may he observed.38 The extensive investigations of 
Bail, Walker,39 and others, on the nature of changes in virulence in 
many invasive bacteria, and the knowledge more recently gained on 
the resistance to phagocytosis of virulent strains of streptococci and 
pneumococci are facts closely related in underlying principle to the 
inagglutinability of typhoid strains cultivated in immune sera. 
That no two strains of bacteria of the same species are exactly 
similar in their agglutinability in the same serum, moreover, is an 
observation which is made by every one who is in a position to carry 
out routine Widal tests in hospital practice. The spontaneous ag- 
glutination which occasionally occurs in the broth cultures of typhoid 
bacilli used for this test in many laboratories 40 may often be re- 
ferred, at least in the cases which have come to the writer’s notice, to 
an excessive acidity of the broth, a phenomenon which will be dis- 
cussed in a subsequent paragraph. As far as the phenomenon of 
variable agglutinability inherent in different strains is concerned, 
however, it is of great practical importance in carrying out routine 
Widal tests to bear this in mind and to control the strain of typhoid 
bacilli employed in the reactions from this point of view. A strain 
also which has been in use for a long time should be titrated with 
agglutinating animal sera from time to time to determine whether or 
not alterations in agglutinability have occurred. 
Group Agglutination.—That the reaction of bacterial agglutina- 
tion was specific was noted, we have seen, by Gruber and Durham 
from the very beginning. The closer study of the reaction in its 
application to bacterial identification has led to interesting data 
which have revealed certain definite limitations of this specificity. 
It has been found, for instance, that, while immunization with any 
given species of bacteria gives rise to a very marked increase of 
agglutinins for this species, there are formed at the same time, though 
38 Eisenberg. Centralbl. f. Bakt., Yol. 34, p. 739, 1903. 
39 Walker. Centralbl. f. Bakt., Yol. 33, 1903. 
40 See section on Agressins. 252 
INFECTION AND RESISTANCE 
to a lesser degree, agglutinins for bacteria of other species. This 
has been referred to as “group reaction,” and the agglutinins appear- 
ing in such sera are spoken of by German observers as “Haupt Agglu- 
tinine” and “Neben” or “Mit Agglutinine” In English texts they 
are usually referred to as “chief” or “major” agglutinins and “para” 
or “minor” agglutinins. Although, as a general rule, such group- 
agglutinin formation is evident most markedly in the cases of biologi- 
cally related micro-organisms like the typhoid, paratyphoid, and 
colon bacilli, this is not necessarily the case, and in some instances 
the immunization of an animal with a given bacterium may produce 
minor agglutinins for other bacteria that have no morphologically 
or culturally demonstrable biological relation to that which reacts 
with the major agglutinin. We may obtain the most graphic survey 
of these conditions by examining one of a number of experimental 
protocols in which such major and minor agglutinin formation is 
illustrated. Thus in the work of Hiss 41 on the dysentery bacilli the 
following relations were observed: 
A serum produced in rabbits by immunization with the Shiga 
bacillus agglutinated the Shiga bacillus in dilutions of 1 to 20,000, 
the “Baltimore,” “Harris,” “Gray,” and “Wollstein” bacillus 1 to 
1.200, the Y bacillus and others 1 to 200. 
An Anti-Y bacillus serum, which agglutinated this bacillus 1 to 
6,400, agglutinated the Baltimore bacillus 1 to 1,600, and the Shiga 
bacillus 1 to 100. 
Anti-“Baltimore” bacillus serum agglutinated this bacillus 1 to 
3.200, and the Y bacillus 1 to 400, and the Shiga bacillus 1 to 100. 
In this series there is fair correspondence between cultural bio- 
logical relations and agglutination, and from many such investi- 
gations it would seem that “group” agglutination might be taken 
to represent a method of determining biological classifications similar 
to the zoological relations revealed by the precipitin reaction. While, 
in a general way, this is undoubtedly true, nevertheless great caution 
must be exercised in relying upon such evidence for classification, 
since notable exceptions have been observed. Park,42 for instance, 
cites a case in which a horse, immunized with a paradysentery bacil- 
lus, agglutinated a colon strain in dilutions up to 1 to 10,000. Simi- 
larly Durham 43 found that two members of the colon group—one 
saccharose fermenting—reacted almost identically with the same 
agglutinating serum, while the agglutinations of two culturally 
identical bacilli of the hog cholera group were entirely at variance. 
The cause for the phenomenon of group agglutination must un- 
questionably be sought in the nature of the bacterial agglutinogens, 
and it is but reasonable to assume that living cells so little differen- 
41 Hiss. Jour. Med. Res., 13, N. S., Yol. 8, 1904. 
42 Park. “Pathogenic Micro-organisms,” 1910, p. 166. 
43 Durham. Jour. Med. Res., Yol. 5. THE PHENOMENON OF AGGLUTINATION 
253 
tiated biologically and morphologically should have much in common 
chemically as well. The bacterial cell, moreover, may contain several 
antigenic complexes and, beside its specifically peculiar constituents, 
therefore, we may suppose that every bacterium contains additional 
antigenic substances which it has in common with other species. It 
is the specific antigen in response to which the “chief” agglutinin 
is formed, while the others, present in smaller quantity, lead to 
the formation of the minor or para-agglutinins with an intensity pro- 
portionate to the amounts present in the bacterial cell. Thus, as 
Durham expresses it, if we assume one micro-organism to contain 
antigenic substances a, b, c, and d, and another d, e, f, and g, the 
antibodies produced by injections of the former would react with 
the common element d in the latter. 
The diagnostic value of the specificity, however, is plainly not 
affected by the phenomenon of group agglutination, since the action 
of minor agglutinins can be always 
easily eliminated by sufficient dilu- 
tion. Thus if we possess a typhoid- 
immune serum which agglutinates 
the typhoid bacillus in dilutions of 
1 to 10,000, the paratyphoid bacillus 
1 to 1,000 and the colon bacillus 1 to 
100 (as in the figure), we may still 
utilize this serum for the identifica- 
tion of suspected typhoid cultures, as, 
let us say, in the isolation of un- 
known bacteria from stools or urine, 
by using potent sera in dilutions as 
high or higher than 1 to 1,000, be- 
yond which point the action of minor 
agglutinins is eliminated. The dia- 
gram illustrates our meaning in the 
hypothetical case of a typhoid-im- 
mune serum which agglutinates 
typhoid in dilutions of 1 to 10,000, 
paratyphoid bacilli 1 to 800, and 
colon bacilli 1 to 100. The relation 
of agglutination to biologic relationship is not a simple problem in 
that individual strains even of the same species may vary consider- 
ably in agglutination by the same serum. Smith and Reagh 44 have 
studied particularly these conditions as they prevail in the colon, hog 
cholera and allied groups. I hey found that biologic relationship 
usually may be concluded from close agglutination affinities, and that 
minor biologic differences such as colony appearance, etc., do not 
Diagrammatic Representation 
op Group Agglutination. 
44 Smith and Reagh. “Studies from the Rockefeller Institute,” Vol 1. 
1904, p. 270. ’ 254 
INFECTION AND RESISTANCE 
exclude such affinities. On the other hand, closely related bacteria 
vegetating on mucous surfaces (different strains of diphtheria, dys- 
entery, and colon bacilli) may vary considerably in their agglutina- 
tive characteristics, while invasive species show a greater homo- 
geneity among their varieties or races. This brings in another impor- 
tant feature—that is, the modification in agglutinative characteristics 
induced in bacteria when they become parasitic upon different hosts, 
and Smith and Reagh conclude that such changes of parasitic habitat 
may modify the agglutinative properties (probably by adaptation to 
the peculiar reactions of each host), some of them being weakened 
and others strengthened. 
The animal species used for immunization indeed influences the 
quantity and nature of the produced agglutinin considerably. For 
instance, in Pfeiffer’s 45 experiments, a dog, a chicken, and a rabbit 
were immunized with the same strain of cholera spirilla. The sera 
obtained from these animals agglutinated this and other strains of 
cholera spirilla in an entirely irregular manner—showing that the 
constitution of the agglutinins in each case was an absolutely different 
one in regard to the relative concentrations of “major” and “minor” 
constituents. 
Agglutinin Absorption.—Castellani 46 found that the immuniza- 
tion of an animal with two or more different species of bacteria re- 
sults in the formation of agglutinins against all of these. Supposing, 
for instance, that species A and B are used for treatment, agglutinins 
against both A and B are formed in quantity, depending upon the 
intensity of the treatment in each case. JSTow, if to the serum so pro- 
duced an emulsion of A is added, agglutinin A only will be removed, 
while agglutinin B will remain in the serum almost undiminished. 
An example of this is seen in the following protocol taken from Cas- 
tellani’s paper: 
Titre of the serum 
Titre after 
absorption with 
B. typhi 
Titre after 
absorption with 
B. coli “31” 
Titre after 
absorption with 
B. coli and B. typhi 
B. typhi 4,000 
B. coli “31” 1,000 
B. typhi 0 
B. coli1131” 
1,000 > 300 
B. typhi 4,000 
B. coli “31” 0 
B■ typhi 0 
B.coli “31” 0 
In the preceding paragraphs, however, we have seen that im- 
munization with a single organism, say B. typhosus, will induce the 
formation of agglutinins, not only for this species, but also of para or 
minor agglutinins for biologically similar strains as well. In such 
45 Pfeiffer. Quoted from Paltauf in “Kolle u. Wassermann Handbuch,” 
Yol. 4. 
46 Castellani. Zeitschr. f. Hyg., Yol. 40, 1902. THE PHENOMENON OF AGGLUTINATION 
255 
cases, as Castellani showed, absorption of the serum with the or- 
ganism used for immunization takes out, not only the major agglu- 
tinins, but rather all of the agglutinins, major and minor. Con- 
versely, however, absorption of such a serum with the species agglu- 
tinated by the minor agglutinins takes out these antibodies only, 
leaving the major substances intact. These relations are well illus- 
trated by the two following protocols, also taken from Castellani’s 
paper: 
Serum of rabbit No. 1 immunized with B. typhi only 
Agglutination titre of serum 
Titre after absorption with B. typhi 
B. typhi 5,000 
B. typhi 0 
B. coli 600 
B. coli 0 
Agglutination titre 
After absorption with 
After absorption with 
of serum 
B. typhi 
B. coli 
B. typhi 10,000 
B. typhi 0 
B. typhi 10,000 
(heavy clumps) 
(small clumps) 
B. coli 800 
B. coli 0 
B. coli 0 
Serum of rabbit No. 7 immunized with B. typhi only 
Note: All of these protocols are taken from Castellani’s communication, loc. cit. 
These facts, variously confirmed, tend to corroborate the concep- 
tion of the production of major and minor agglutinins outlined 
above. 
It is of practical and theoretical importance to mention that 
complete absorption of specific agglutinin by a single exposure to 
homologous bacteria, however thickly emulsified, is not possible. It 
is always necessary to absorb repeatedly, and even then a minute 
trace of agglutinin may eventually remain. Eisenherg and Volk,47 
who have studied these conditions particularly, attribute this to the 
nature of the union of agglutinogen with agglutinin, wdiich they con- 
ceive as following the laws of mass action—this accounting for the 
persistence of a small “rest” of free agglutinin, even after repeated 
absorption by partial dissociation. The principle involved here is 
identical with that discussed in connection with antigen-antibody 
union in general. 
Normal Agglutinins.—It is not only in the sera of immunized 
animals, however, that agglutinins are found. Just as the other 
47 Eisenberg and Yolk. Zeitschr. f. Hyg.t Yol. 40, 1902. Eisenberg. 
Centralbl. f. Bakt., Yol. 34, 1903. 256 
INFECTION AND RESISTANCE 
antibodies, antitoxins, and bactericidal sensitizers may be found in 
the blood of normal animals, so agglutinins for various bacteria may 
be normally present. These normal agglutinins do not in any respect, 
further than that of quantity, differ from the immune agglutinins 
and follow the same laws of specificity which have been described 
for the latter. It has been shown a number of times that such normal 
agglutinins are not present in the new-born animal, but are acquired 
later in life, possibly because of the absorption of bacterial products 
from the intestinal canal. It has been variously shown 48 too, that 
living bacteria themselves may enter the lymphatics and the portal 
circulation from the intestine during apparently perfect health of 
the individual. 
This subject is of interest, not only in connection with the ag- 
glutinins, but has bearing upon the existence of normal antibodies 
in general. Ruffer,49 who has studied particularly the penetration 
of leukocytes and bacteria through the intestinal mucosa, demon- 
strated micro-organisms in the sub-mucous lymph nodes of normal 
rabbits, and Ribbert 50 and Bizzozero 51 have shown the presence of 
bacteria in apparently normal mesenteric lymph nodes. Adami and 
Nichols even claim that during health a certain number of living 
bacteria enter the portal circulation from the intestine, and from 
here may get into the systemic circulation, and are ordinarily de- 
stroyed by either leukocytes, liver lymphatic organs, or the kidneys. 
It is thus not surprising that normal agglutinins should occur, 
and that they should be qualitatively identical with the so-called 
“immune” agglutinins, since they probably arise by a sort of spon- 
taneous immunization through the intestinal canal. Researches of 
Ford 52 point in the same direction. However, Landsteiner and 
Reich 52“ in 1907 pointed out differences between normal and im- 
mune agglutinins in that the normal were more resistant and did 
not possess the specificity of the immune. For the details of these 
experiments, we refer to their publication. Their work prevents the 
unqualified acceptance of the identity of normal and immune anti- 
bodies. 
Ehrlich’s Conception of Agglutination.—An interpretation of 
the process of agglutination, according to the theory of Ehrlich, con- 
ceives it as a chemical union of agglutinin and bacteria (agglutino- 
gen). The agglutinin is regarded as consisting of two atom com- 
plexes, one the “haptophore,” having affinity for the bacterial protein, 
and concerned with the union, the other the “ergophore” or “zymo- 
48 Adami. Jour. Am. Med. Assoc., Dec., 1899. 
49 Ruffer. Brit. Med. Journal, 2, 1890. 
60 Ribbert. Deutsche med. Woch., 1885. 
51 Bizzozero. Centralbl. f. d. Med. Wiss., Yol. 23, 1885, p. 49. Quoted 
from Adami. 
52 Ford. Zeitschr. f. Hyg., Vol. 40, 1902. 
52a Landsteiner and Reich. Zeitschr. f. Hyg., Vol. 58, 1907, p. 213. THE PHENOMENON OF AGGLUTINATION 
257 
phore,” by means of which the actual agglutination is brought about 
after the union has taken place. Unlike the antibody concerned in 
the processes of hemolysis or bacteriolysis, the agglutinins are not 
dependent in their action upon the cooperation of alexin, and the 
agglutination power of a serum is therefore not destroyed by inactiva- 
tion or heating to 56° C., as is the case with the former. Although 
the accurate point of thermal destruction varies with different ag- 
glutinins (the agglutinins for the Bacillus pestis and a few other 
bacilli are said to be destroyed at 56° C.), as a rule agglutinins will 
not disappear from serum until the temperature is raised to between 
70° and 80° C. Once destroyed, however, no reactivation takes place 
upon the addition of fresh normal serum. Ehrlich, for this reason, 
has conceived the structure of agglutinins as “Haptines of the Second 
Order,” in which he supposes that the zymophore and the haptophore 
groups are inseparably connected, and in which we could assume an 
alteration of the less stable zymo- 
phore group without interference 
with the functional integrity of the 
haptophore group. Such an altered 
agglutinin could be spoken of as 
“agglutinoid,” and this could be- 
come united with a bacterial cell 
without inducing agglutination, but, 
by its union, prevent subsequent 
combination of the cell with un- 
altered agglutinin. This process of 
“agglutinoid Verstopfung” has been 
held responsible for the failure of 
agglutination when bacteria that 
have been in contact with heated serum are subsequently exposed to 
the action of actively agglutinating serum. It is assumed that the 
agglutinoids which were present in the heated serum have occupied 
the bacterial receptors and have thereby prevented the union of these 
with the agglutinins later added. 
The So-called Agglutinoid Phenomenon.—The work of Eisen- 
berg and Volk 53 has gone very thoroughly into these conditions and 
forms the strongest bulwark of this point of view. These workers 
showed that bacteria thus exposed have not only become less sensitive 
to agglutinins, but have, at the same time, lost much of their power 
to absorb agglutinins when compared with normal bacteria. The 
same loss of agglutinating power which is observable in heated ag- 
glutinating serum is evident to a lesser extent in serum which has 
been preserved at room temperature. Eisenberg and Volk have shown 
that such serum, in addition to a quantitative loss of agglutinin con- 
tent, loses the power to agglutinate in high concentrations. Thus 
Diagrammatic Representation of 
Ehrlich’s Conception of the 
Structure of Agglutinin. 
53 Eisenberg and Yolk. Zeitschr. f. Hyg., Yol. 40, 1902. 258 
INFECTION AND RESISTANCE 
a scrum which has been preserved in this way will no longer ag- 
glutinate bacteria in concentrations of 1 to 20, 1 to 40, or even 1 to 
100, but will agglutinate as before in higher dilutions. This is taken 
to mean that the agglutinoids formed during the period of standing 
possess a higher affinity for the bacterial antigen than do the true 
unaltered agglutinins. Since these so-called “proagglutinoids” are 
relatively small in amount, their action is masked when considerable 
dilution has sufficiently diminished their quantity, in proportion to 
the more plentiful unaltered agglutinins. In support of this assump- 
tion it has been further shown that sera which have been rendered 
inhibitory by either of the methods named can be deprived of their 
inhibiting characteristics by absorption with homologous bacteria. 
Together these observations might appear to constitute a strong argu- 
ment in favor of the agglutinoid theory. 
In practical experience the existence of such an inhibition zone 
is of great importance, since freshly taken sera will occasionally 
show this failure of agglutination in concentration, while strong 
agglutination follows when the dilution is increased. In clinical 
tests, as in the Widal reaction for the diagnosis of typhoid fever, we 
not infrequently encounter sera which will give no agglutination in 
dilutions of 1 to 20 and even 1 to 40, and the reaction would there- 
fore be regarded as negative unless the possibility of the proag- 
glutinoid zone were recognized and further dilutions carried out. 
While there is no question of the accuracy of the experimental 
data cited in the preceding paragraphs, the interpretation of the phe- 
nomena on the basis of Ehrlich’s haptine conception has not been 
universally accepted. 
The gist of the explanation of the so-called “agglutinoid” phe- 
nomena by the Ehrlich school, therefore, can be summarized as fol- 
lows: The agglutinin is conceived as consisting of a “haptophore” 
and a “zymophore” group; the “haptophore” group brings about the 
specific union of the agglutinin with the bacteria, the “zymophore” 
group causes the agglutination. Heating destroys the “zymophore” 
group, but not the “haptophore” group. In consequence, such altered 
agglutinin or “agglutinoid” can unite with bacteria, but can no 
longer cause agglutination. By uniting with the bacteria, it occludes 
the bacterial receptors for agglutinin in such a way that normal 
agglutinin, added subsequently, is unable to unite with the bacteria 
at all. In the transformation of agglutinin to “agglutinoid,” the 
affinity is changed so that the “agglutinoid” now unites more easily 
with the bacteria than does normal agglutinin, hence the term “pro- 
agglutinoid.” There seems to us almost nothing in this theory which 
can justly be retained. In the first place, Bordet’s salt experiment 
has pretty well excluded the assumption of “haptophore-zymo- 
phore” structure on the part of the agglutinin. Moreover, as we 
have seen in discussing complement fixation, and as we shall see in THE PHENOMENON OF AGGLUTINATION 
259 
discussing precipitins, the so-called “agglutinoid” zones are in entire 
analogy with the zone phenomena observed with other antibody re- 
actions, and dependent upon principles of quantitative union be- 
tween antigen and antibody that have nothing whatever to do with 
deterioration of the antibody by heat or otherwise. The only point 
of argument on the part of the Ehrlich school which is not easy to 
answer in this connection is the claim made by Eisenberg and Volk, 
as well as by Kraus and Joachim, that the agglutinin inhibiting prop- 
erties of a heated serum can be specifically absorbed out of such a 
serum with the appropriate homologous bacteria. If the matter of 
heating were only active by producing a protective colloidal action 
in the serum, then one would have to assume that such heated serum 
would prevent agglutination of any bacteria, or that normal heated 
serum could do the same thing. If, however, as Kraus and Joachim 
and others have claimed, it is possible to produce agglutinin inhibit- 
ing properties in typhoid immune serum by heating and then specif- 
ically absorbing out this inhibiting substance with typhoid bacilli, in 
such a way that the serum is deprived of its “agglutinoid” action, by 
the bacilli, and the bacilli are rendered inagglutinable by virtue of 
the material they have absorbed out of the serum, it would appear 
as though the inhibiting substance were really some inherent altera- 
tion property of the specific agglutinin, itself. Our explanation of 
this is as follows: We know from experiments of Porges, as well 
as from our own, that in various colloidal precipitations in which 
serum is involved, moderate heating of the serum will strongly re- 
duce its ability to precipitate a suspension. Thus, while minute 
amounts of fresh dog serum precipitate colloidal arsenic-sulphid, a 
very minute amount of heated dog serum will inhibit such precipita- 
tion by fresh dog serum. Kow, when we heat any normal serum, 
it is likely that we are changing it in its colloidal state, and produc- 
ing a certain amount of colloidal protective property in the serum. 
In all reactions by the bacteria and antiserum, it is likely that the 
antibody carries into the union a not inconsiderable amount of inac- 
tive serum constituents. Thus, in precipitin reactions, as we shall 
see, the bulk of the precipitate formed comes from the serum. It is 
our belief that the so-called specific action of “agglutinoids” is due 
to the fact that the antibody carries into the union with the bacteria 
inactive protein which has become colloidally protective by virtue 
of the heating. 
Hemagglutination.—In describing the investigations which led 
to the discovery of the mechanism of the lytic phenomenon, in the 
chapter on Cytolysis, we mentioned that Bordet and others had no- 
ticed the frequent agglutination of red blood cells in the sera of 
animals treated with such cells after the hemolytic property had 
been destroyed by heating to 56° C. Such hemagglutination is a 
phenomenon entirely analogous to the agglutination of bacteria by 260 
INFECTION AND RESISTANCE 
serum, and hemagglutinins regularly result when an animal is sys- 
tematically treated with the red blood cells of another species. Like 
the bacterial agglutinins, the hemagglutinins are relatively ther- 
mostable and are best observed, therefore, after the sera are inac- 
tivated. Otherwise rapid hemolysis will often obscure the agglutina- 
tion. Like other agglutinins the hemagglutinins thus produced are 
specific, acting only upon that variety of cells which are used in their 
production. Moreover, certain sera may normally contain hemagglu- 
tinins for the blood cells of animals of another species. An illustra- 
tion of this is the hemolytic and hemagglutinating property of normal 
goat serum for rabbit cells—but there are many other examples which 
might be cited. Such normal hemolytic and hemagglutinating prop- 
erties for the cells of other animals usually render the sera toxic for 
these animals, and some observers have attributed the toxicity to this 
agglutinating action, believing that the clumped erythrocytes form 
emboli around which clotting is initiated. The writer’s own investi- 
gations, however, seem to show that this is not the case, since the 
toxicity of such sera is completely removed after they have been 
heated, in spite of the fact that the hemagglutinative property re- 
mains unchanged. 
Bordet’s Analysis and the Importance of Electrolytes.—The 
fundamental principle underlying all the Ehrlich hypotheses is the 
conception that these reactions take place as do purely chemical 
reactions, following the law of multiple proportions. Such reason- 
ing has often necessitated the assumption of differences of affinity 
which, critically examined, are really ex post facto explanations, 
forcing the phenomena to conform with the theory. As a matter of 
fact, the bodies which participate in the antibody-antigen reactions 
are probably of the nature of the substances which are spoken of as 
colloids, and it is therefore more than likely that the quantitative 
and other relations governing the union of these reagents should be 
analogous to those governing colloidal reactions in general. The 
reaction of agglutination, like that of precipitation, has lent itself 
particularly to the study of the principles of the union from this 
point of view, and the first and fundamental progress made in this 
direction is found in the work of Bordet. 
Bordet 54 compared the formation of precipitates in bacterial 
emulsions to the precipitation of such inorganic emulsions as clay 
in distilled water, and noted that the precipitation of homogeneous 
emulsions of such substances is “often controlled by such insignifi- 
cant causes as the presence of salts.” Applying this analogy to the 
agglutination of bacteria, he performed the following experiment: 
Cholera spirilla, emulsified in salt solution, were treated with homol- 
ogous immune serum and, after agglutination had taken place, the 
bacteria were thrown to the bottom by centrifugation and divided 
54 Bordet- 4wn. ide VJnst. Pasteur, 1896, 1899, THE PHENOMENON OF AGGLUTINATION 
261 
into two parts. One part was again suspended in salt solution, and 
the other was washed, and then suspended in distilled water. The 
bacteria in the tube of salt solution rapidly agglutinated, while those 
in the distilled water, after thorough shaking, remained indefinitely 
suspended in an even emulsion. If, however, to these unagglutinated 
bacteria a small amount of pure sodium chlorid was added agglutina- 
tion occurred. 
The conclusions that can justly be drawn from this experiment 
are, first, that the bacteria could not agglutinate, even though they 
had been bound to agglutinin, when salt was removed from the en- 
vironment, and, second, that the addition of salt to such emulsions 
brought about immediate agglutination. The same principle can be 
demonstrated in other ways. If, for instance, a bacterial emulsion is 
rendered free of salt by dialysis, and this is added to an agglutinat- 
ing serum similarly dialyzed, no agglutination occurs. The suspen- 
sion may remain evenly clouded indefinitely unless salt is added. 
As soon as a little salt is added, however, perfect agglutination 
occurs. To this technique the very obvious criticism may be applied 
that perhaps the absence of salt has precipitated the agglutinins, 
which, as we know, are precipitated with globulin, which is insoluble 
in the absence of salt. However, this source of error is excluded by 
the first experiment cited, and, moreover, it can be shown by the last 
experiment that, even though the bacteria are not agglutinated in 
the salt-free serum, they have nevertheless absorbed agglutinin. For, 
if such a salt-free mixture is centrifugalized, the bacteria washed 
and suspended in distilled water, and salt is then added, agglutina- 
tion occurs. The supernatant fluid of the original suspension, 
furthermore, can be shown to have been deprived of agglutinins by 
suitable experiment. 
These facts, first observed by Bordet, and further elaborated by 
the studies of Joos,5° Friedberger,56 and others, have been inter- 
preted in different ways. Joos claims that there is a chemical union 
between the bacteria and the salt, and bases this upon the observation 
that the salt added to a salt-free mixture cannot be demonstrated in 
the supernatant fluid after agglutination has taken place. His ob- 
servations in this respect have not found confirmation at the hands 
of Friedberger and other workers, and it is generally agreed to-day 
that the role of the salts is, as Bordet first assumed it to be, a purely 
physical one. Bordet’s opinion is often spoken of as the “two phase” 
theory, in that he conceives the process of agglutination to consist of 
two distinct occurrences, first, an absorption of the agglutinin by 
the bacteria, and, second, an agglutination of the new complex by the 
salt. It is not the agglutinin which causes agglutination, but by 
union with the agglutinogen forms a complex which is altered in 
55 Joos. Centralbl. f. Bakt., 1, Yol. 33, 1903. 
56 Friedberger. Berl. klin, Woch,, 1902; Centralbl. f. Bakt., 1, Yol. 30, 262 
INFECTION AND RESISTANCE 
“colloidal stability,” and therefore is flocculable by the action of the 
electrolyte. 
The opinion of Bordet becomes clearer as we consider the con- 
ditions governing the flocculation of colloids in general. Without 
wishing to enter in this place into detail regarding the nature of 
colloidal suspensions, it nevertheless seems necessary in order to do 
justice to this phase of the question to recall briefly the conditions 
governing such flocculation. The so-called colloidal solutions are 
not true solutions as the term is applied to dissociable substances, but 
are looked upon as consisting of small particles in suspension. The 
particles are similarly charged, as can be demonstrated by their 
wandering when subjected to an electric current, and it is supposed 
that it is this fact of similarity of charge which, in the “sol” state, 
permits them to remain in suspension. For the similarity of the 
charges of the individual particles prevents their mutual approxima- 
tion.57 
The state of suspension of these substances, then, represents a 
delicately balanced equilibrium between the two forces of electrical 
repulsion and of surface tension, an equilibrium which may be dis- 
turbed by the action of a number of factors. Thus, studies on inor- 
ganic colloids have shown, long before these considerations were 
applied to the explanation of serum reactions, that the stability of 
these suspensions could be disturbed both by electrolytes and by the 
addition of other colloids. Thus they may be precipitated by various 
salts, acids, and bases and, as Schultze58 has shown, they react 
with that ion of the electrolyte which carries an opposite charge to 
that of the colloidal particles. For, although the colloidal units are 
similarly charged, this may be either negative or positive, according to 
the nature of the particular substance. In the case of the so-called 
amphoteric colloids reaction may take place, according to Pauli,59 
with both ions of the electrolyte. 
The probable mechanism of the process is postulated by Pauli 
in describing the precipitation of a colloidal metal by salts, acids, or 
bases in the following way: 
“The introduction of such electrolytes into a colloidal suspension 
is of course accompanied by a certain amount of dissociation. In 
consequence the weakly charged particles of the colloid collect about 
the ions of opposite charge until a sufficient accumulation of the 
particles leads to an electrical neutralization of the ion, and the 
57 That these relatively simple conceptions of the conditions in suspensions 
containing proteins and other colloids call for considerable revision in view 
of Loeb’s recent work will be discussed in a later section. They are crudely 
stated here because much of the work on agglutination cited below was based 
upon them. 
58 Schultze. Jour. f. prakt. Cltiem., 25, 1882, and 27, 1883. 
59 Pauli. “Hofmeister’s Beitrage,” 1905, and “Physical Chemistry in 
Medicine,” Wiley & Son, N, Y., 1907, THE PHENOMENON OF AGGLUTINATION 
263 
aggregation, if of sufficient size, will sink to the bottom, forming a 
precipitate.” 
In regard to the mutual influences exerted upon each other when 
two colloids are mixed, it has been shown by Biltz, Hardy, and many 
other observers that oppositely charged colloids precipitate each 
other, though this is not an absolute rule, as experiments by Pro- 
fessor Stewart Young, of Stanford, have shown. Thus colloidal gold 
and platinum will be precipitated by such colloids as ferric oxid or 
aluminum oxid. When such a precipitating colloid is added tc 
another oppositely charged suspension in quantities too small to 
bring about flocculation, moreover, the addition of a quantity of salt, 
likewise too small to precipitate alone, will in many cases bring about 
the flocculation. 
These and other phenomena of colloidal reaction have found 
close analogy in antibody-antigen studies, and have given support to 
the interpretation of the latter in the sense of Bordet. 
To return to the consideration of bacterial agglutination, we 
have spoken of the dependence of the reaction upon the presence of 
salts, and have seen that the researches of Friedberger and others 
have refuted the assumption that the action of the salt in bringing 
about agglutination depends upon chemical union of the salt with 
the bacteria. It is probable, therefore, that here, as in other colloidal 
precipitations, the function of the salt is to be regarded purely as 
an electrophysical phenomenon. 
The analogy becomes still closer when we consider the researches 
of Bechold,60 Neisser and Friedemann,61 Sears and Jameson,62 and 
others, which have shown that bacteria in suspension are to be com- 
pared very closely with true colloidal suspensions in that the bac- 
terial cells carry a definite and uniform electrical charge and wander 
in the electric field. 
Bacteria in salt solution emulsion, for instance, wander to the 
anode, thus giving evidence of their carrying a negative charge. 
This charge may be altered by adding to the emulsions definite con- 
centrations of acids or bases, a reversal of the charge taking place 
under the influence of NaOH or other hydroxids. Just how this is 
brought about is by no means clear, but it is not impossible that 
there is a selective absorption of OH ions by the bacteria, which 
therefore take on the charge of the ion. 
However this may be, and we must admit that explanations of 
these phenomena are as yet largely speculative, a fact which interests 
us particularly in connection with the phenomena under discussion 
at present is the influence exerted upon the charge of bacteria by 
60 Bechold. Zeitschr. f. physik. Chem., 48, 1904. 
“Neisser and Friedemann. Munch, med. Woch., Vol. 51, pp. 465, 827, 
1904. 
03 Sears and Jameson. “Thesis for M. A., Stanford University,1” 1912. 264 
INFECTION AND RESISTANCE 
exposure to the influence of serum. Bechold,63 as well as Neisser and 
Friedemann,64 assert that bacteria which have absorbed agglutinin 
no longer wander to the anode, but act as though they had been 
deprived of electrical charge, and precipitate or agglutinate between 
the electrodes. 
Bechold has suggested, for this reason, that it may be possible 
that bacteria in the normal condition are protected from the action 
of the electrolyte by a membrane or coating of protoplasm which 
acts as a protective colloid. The absorption of agglutinin may alter 
this in such a way that they become amenable to the flocculating 
effects of the salt ions. In keeping with such an opinion is the well- 
known observation of the inagglutinability of capsulated organisms, 
which, as Porges65 has shown, become agglutinable as soon as the 
capsules have been destroyed by heating in a weak acid. 
That the absorption of agglutinin indeed alters the electric sta- 
bility of the emulsified bacteria further appears from the fact that 
“agglutinin” bacteria 66 are agglutinated by concentrations of salts 
which are too slight to affect the normal micro-organisms. In this 
respect there is close similarity between the flocculation of agglutinin- 
bacteria and such emulsions as kaolin and mastic, whereas bacteria 
without agglutinin require much higher concentrations of the salts 
to produce like effects. The absorption of agglutinin may have re- 
moved a factor which protected the bacteria against the influence of 
the salt. On the other hand, it is equally just to assume—and this 
is more likely and corresponds with Bordet’s views—that the ab- 
sorption of agglutinin has “sensitized” the bacteria to the action of 
the electrolyte. The experimental facts upon which the above state- 
ments are formulated are largely found in the important papers of 
Neisser and Friedemann—whose work brought out, likewise, inter- 
esting analogies of the colloidal precipitations with the phenomenon 
which we have described above as the proagglutinoid zone.67 
In regard to the greater amenability of agglutinin bacteria to 
flocculation by electrolytes, the following protocol, adapted from the 
work of these authors, will explain itself. They were tabulated from 
experiments in which different quantities of normal " solution 
of various salts were added, on the one hand to emulsions of unal- 
tered bacteria, and, on the other, to bacteria which had absorbed 
agglutinin. It is seen that, with some salts, agglutination of the 
unaltered bacteria did not occur at all, whereas agglutination was 
03 Bechold. “Die Kolloide in der Biologie u. Medizin,” Steinkopf, Dres- 
den, 1912. 
64 Neisser and Friedeniann. Munch, med. Woch., Vol. 51, 1904, pp. 465 
and 827. 
65 Porges. Zeitschr. f. exp. Path. u. Therapie, 1905. 
fie “Agglutinin” bacteria—bacteria which have absorbed specific agglutinin. 
67 See work of Northrop and that of Coulter in its bearing on this point 
in subsequent sections on pages 267 and 313. THE PHENOMENON OF AGGLUTINATION 
265 
brought about in the treated bacteria with comparatively small 
amounts; in other cases the difference is a quantitative one only: 
Protocol constructed from the tables of Neisser and Friedemann, loc. cit. 
y solution of salt 
Quantity of salt sol. 
which brought about 
agglutination of 
1 c. c. of normal bacteria 
in emulsion. 
0 = no agglutination 
by the salt 
solution 
Quantity of salt sol. 
which agglutinated 
1 e. c. of 
agglutinin bacteria 
in 
emulsion 
NaCl 
0 
.025 
NaNO, 
0 
.025 
Na2S04 
0 
.025 
Rbl 
0 
.025 
MgS04 
0 
.0025 
ZnS04 
.01 
.001 
CaCl2 
0 
.005 
BaCl2 
0 
.005 
Cd(N03)2 
.01 
.001 
CuS04. 
.0025 
.0001 
CuCl2 
.0025 
.0005 
Pb(N03)2 
.0025 
.0001 
HgCl2 
.0025 
.0005 
The analogy between the experiment tabulated in the preceding 
protocol and the following from the work of the same writers is self- 
evident. Just as the absorption of agglutinin by bacteria rendered 
these more amenable to precipitation by salts, so the addition of 
minute quantities of gelatin to mastic emulsions had a similar sensi- 
tizing effect upon these. 
NaCl 
10% solution 
1 c. c. mastic 
(1-10 original emulsion) 
diluted to 3 c. c. 
1 o. c. mastic + .0001 c. o. 
of a 2% gel. solution, 
the whole diluted to 3 c. c. 
1.0 
+ + + 
+ + + 
0.5 
0 
+ + + 
0.25 
0 
+ + + 
0.125 
0 
+ + + 
0.05 
0 
0 
0.025 
0 
0 
Finally, one of the most important analogies yielded by the work 
of the above investigators is illustrated in the following protocol as 
follows: 266 
INFECTION AND RESISTANCE 
Colloidal iron 
hydroxid 
Precipitation of mastic 
emulsion 1 c. c. 
1. 
0 
0.5 
0 
0.25 
0 
0.1 
+ + 
0.05 
+++ 
0.025 
+++ 
0.01 
+++ 
0.005 
+++ 
0.0025 
++ 
0.001 
0 
Here we have an inhibition zone in the tubes containing the 
highest concentrations, accurately analogous to the previously dis- 
cussed proagglutinoid zone. It is a phenomenon similar also to the 
inhibition zones noticed in precipitin reactions and observed, though 
by a different technique, in bacteriolytic phenomena discussed in an- 
other place in connection with the Neisser-Wechsberg notion of com- 
plement-deviation or “Komplement Ablenkung.” It seems to be a 
universal fact governing the union of colloidal substances, that defi- 
nite quantitative proportions must be maintained in order to lead 
to reaction, this being, possibly, explicable on the basis that actual 
union can take place only after disturbance of the electrical balance 
which keeps the particles apart. These reactions will be found more 
accurately discussed in another place. Whatever the mechanisms 
may be, however, these and similar experiments have seemed to 
render unnecessary and unlikely the assumption of proagglutinoids, 
proprecipitoids, etc., to explain the inhibition zones so frequently 
observed in all reactions of this kind. 
A peculiar observation, which does not coincide with the above 
interpretation of these phenomena, and the significance of which is 
indeed doubtful, is one which Friedberger 68 made in researches in 
which he confirmed the work of Bordet on the absence of agglutina- 
tion in a salt-free environment. He found that not only the addition 
of various salts would bring about agglutination under such condi- 
tions, but that organic substances such as dextrose and asparagin 
could be substituted for salts and had similar agglutinating effect— 
although higher concentrations of these than of the salts were re- 
quired. Were these substances at all dissociable it might be pos- 
sible that they acted by a mechanism identical with that of the salts— 
but since such substances as dextrose either do not dissociate at all or 
do so to an infinitesimal degree only there does not seem any pos- 
sibility of reconciling these results with Bordet’s theory. 
68 Friedberger. Centralbl. f. Baht., 30, 1901. THE PHENOMENON OF AGGLUTINATION 
267 
It is difficult to explain Friedberger’s results. Possible impurity 
of his preparations and the presence of traces of electrolyte seem 
to be excluded by the fact that he was quite conscious of this possi- 
bility of error and used only substances which yielded no ash on 
combustion. 
It may be that the results of Friedberger in which glucose and 
asparagin were used may have brought about agglutination by an 
entirely different mechanism from that which we are discussing and 
form no analogy to this. 
The most important recent piece of work that has been done on 
agglutination in our opinion is that of Northrop and deKruif.69 
As we have seen above, it is well known that bacteria, as well as 
other particles in suspension, carry an electric charge with reference 
to the surrounding medium, and we have already referred to the idea 
that repellent forces due to the like charges carried by individual 
particles, represented the force by which agglomeration is prevented. 
As we have mentioned above, and as will be seen further discussed 
in the chapter on colloids, the two forces which are opposed to each 
other in suspensions of this nature are the electric charge above 
alluded to, on the one hand, and the surface tension which has been 
supposed to exert a force of attraction between particles. It is the 
balance between these two which until very recently was regarded as 
determining suspension or precipitation. The question of the charge 
on suspended particles has been investigated by a number of workers, 
a particularly interesting paper being that of Powis,70 who studied 
oil emulsions and found that coagulation occurred when the potential 
between the drops of oil and the surrounding liquid was reduced 
below a critical value of about 30 millivolts. Thus, the stability of 
the oil emulsion was definitely shown to depend upon the potential. 
Northrop mentions, however, that in the case of bacteria, this cannot 
be taken to have a direct bearing because agglutination by immune 
serum can occur without any apparent change of potential. For 
the first time in the investigation of agglutination, at least as far as 
we know, Northrop and deKruif have taken into account, in addition 
to the charge on the bacteria, the cohesive forces, and have attempted 
to measure both of these factors in relation to each other. 
They worked particularly with thoroughly washed suspensions of 
B. typhosus and the Bacillus of rabbit septicaemia. The potentials 
were calculated from the rate of migration of the organisms in an 
electric field. 
The cohesive force they measured by the following interesting 
technique which we quote verbatim. 
“Measurement of the Cohesive Force.—A piece of thick glass 
slide was covered with a thin film of very heavy suspension of washed 
68 Northrop and deKruif. Jour. Gen. Phys., 4, 1922, 639. 
70 Powis. Zeit. f. Physik. Chem., 89, 1914, 186. 268 
INFECTION AND RESISTANCE 
organisms (B. typhosus), the film allowed to dry and then heated 
to 60° for a few minutes. This causes the bacteria to adhere firmly to 
the glass. A heavy (No. 3) cover-slip was similarly prepared. The 
cover-slip was suspended by means of a fine platinum wire from the 
lever of the de Noiiy 71 surface tension apparatus. The glass slide 
was immersed in a dish containing the solution to be studied and 
the cover-slip allowed to rest on it with its own weight for 1 minute. 
The force required to pull the cover-slip from the slide was then 
determined, it was found that if the measurement was made imme- 
diately after the two surfaces came in contact, the value obtained 
depended on the force with which the two had been pressed together. 
If the slip had been pressed down firmly a much greater force was 
required than if it had simply been allowed to rest on the slide. 
After a short time interval, however, this difference became less, and 
eventually the same reading was obtained in both cases. This is 
due presumably to the fact that the distance apart of the two surfaces 
is regulated by capillary forces and comes to the same point from 
either side. The same smear was used as long as the same value 
was obtained on replacing the preparation in distilled water. The 
value obtained becomes less after ten or fifteen measurements, due 
to the gradual removal of the film. Control experiments with clean 
glass surfaces showed no significant variation under the conditions 
of the experiment. The values obtained in this way were surprisingly 
reproducible. They have been expressed as milligrams required to 
separate two surfaces each 2 cm. square. The results are not exactly 
comparable to the measurements of the potential since the organisms 
have been subjected to dry heat. It will be noted, in fact, that the 
results do not conform exactly to those expected from the potential 
measurements. In the case of NaCl, for instance, the concentration 
required to affect the cohesive force noticeably, is slightly higher 
than would be expected from the potential curve. 
“It has usually been considered that this force is a surface tension 
effect, but there does not appear to be any conclusive evidence as to 
its nature. It is better, perhaps, to refer to it simply as ‘cohesive’ 
without an exact definition of its nature.” 
The conclusions they reached in their work are extremely inter- 
esting and considerably at variance with many of the ideas previously 
held. In examining the effects of salts on the agglutination of 
washed bacteria, it was found that their experiments could be divided 
into two groups: first, those in which agglutination was caused by 
very low concentrations, less than 0.01 normal, and those in which 
high concentrations were needed. In the first group, it was found 
that agglutination occurred whenever the potential between the 
organisms and the solution was reduced below 13 millivolts, either 
positive or negative. This indicated that the change of potential 
71 deNoiiy. Jour. Gen. Physiol., 1, 1918, 521. THE PHENOMENON OF AGGLUTINATION 
269 
was the essential factor. In the second group, however, agglutination 
with many salts did not occur, even though the potential was reduced 
to an immeasurably small value, and in such cases they assume that 
the salt affected the cohesive force as well as the potential. This 
supposition was confirmed by experiment. In other words, they 
found that high concentrations of salts decreased the cohesive force. 
Therefore, although the potential required to keep them apart was 
also decreased and in concentrated salt was reduced to zero, the 
cohesive force was rendered so insignificant that agglutination did 
not occur. 
In regard to serum, they showed that if dialyzed serum is added 
to a suspension of organisms in distilled water, the potential of the 
organisms gradually decreases, but not enough to cause agglutination. 
If a salt solution now is added to this mixture until the potential 
is decreased to the critical value of 13 millivolts, agglutination occurs. 
The serum seems to prevent the salt from decreasing the cohesive 
force between the organisms, and the potential, therefore, determines 
the agglutination when the salt becomes sufficiently concentrated to 
reduce this to less than 13 millivolts. 
Conversely, if the experiment is reversed, that is, if serum is 
added to organisms in the presence of salt, sufficient in amount to 
show this low potential, the serum increases the cohesive force until 
it is greater than the repulsion due to the potential (which latter is 
not affected by the addition of serum), and agglutination occurs. 
Of course, their work does not touch upon the possible reasons 
why immune sera alter the bacteria more definitely than normal 
sera, or upon what the specific absorption of immune bodies from 
serum depends, but their work brings us closer to an understanding 
of the physical conditions which underlie the agglutination of bac- 
terial suspensions. 
Acid Agglutination.—In one of the preceding paragraphs we 
have mentioned the phenomenon spoken of as “acid agglutination.” 
By this is meant the spontaneous clumping, not only of bacteria, hut 
of small particles of any kind, in suspension, in the presence of 
certain concentrations of acid. Michaelis,72 Beniasch,73 and others 
who have studied this phenomenon in detail have come to the con- 
clusion that it is the concentration of the hydrogen ions which is 
responsible for the agglutination. This explanation is also applicable 
to the agglutination often observed about the anode when bacteria 
are subjected in suspension to the action of a direct current. In 
such experiments the organisms after concentrating at this electrode 
often flocculate, and it is here, of course, that hydrogen ions are 
present in the greatest concentration. How this takes place is prob- 
lematical, but the reasoning of Pauli, if applied to this, would favor 
72 Michaelis. Folia Serol., 7, p. 1010, and Deutsche med. JVoch., 37, 969. 
73 Beniasch. Zeitschr. f. 1mm., Yol. 12, 1912. 270 
INFECTION AND RESISTANCE 
the assumption that the weakly charged bacteria group themselves 
about the ions and, when a sufficiently large aggregation has formed, 
fall to the bottom as precipitate. This phenomenon of acid aggluti- 
nation is of course entirely different in nature from the specific serum 
agglutination which we are discussing. Nevertheless, Schidorsky 
and Iieim,74 Jaffe,75 and others have attempted to apply acid aggluti- 
nation to the isolation and differentiation of bacteria, on the concep- 
tion that different species are agglutinated by varying concentrations 
of hydrogen ions. The former investigators, even, claim to have 
been successful in isolating typhoid bacilli from the stools by this 
method in that the typhoid bacillus was agglutinated by concentra- 
tions of acid which had no effect upon the Bacillus coli. Sears 76 
has gone over this work carefully, and, while he has obtained results 
which bear out the contention that the agglutination is probably due to 
rhe concentration of the H ions, his experiments have revealed an 
irregularity in the behavior of bacteria of the same species in acid 
solutions and an overlapping of those of one species with those of 
another. Therefore the use of acid agglutination for differential 
purposes seems to us entirely hopeless. And indeed it would be sur- 
prising if any such distinctive and regular reaction differences be- 
tween simple bacterial cells, after all chemically and physically so 
essentially alike, could be found. 
The Functional Importance of Agglutination in Infectious 
Disease.—Since the property of specifically agglutinating bacteria is 
developed in the bodies of animals and man in the course of contact 
with bacteria in the course of infection, it is natural to speculate 
concerning the possible significance of this phenomenon. Of course, 
it is quite possible that it is merely an incidental, inconsequential 
effect of reactions between antigen and antibody, and it has been 
so looked upon by a number of investigators, such, for instance, as 
Metchnikoff,77 who says the part played by agglutination in acquired 
immunity is “merely accidental and subordinate.” Salimbeni 78 goes 
so far as to state that bacterial agglutination does not take place 
within the animal body. His work was done with cholera spirilla. 
On the other hand, many observers have actually determined aggluti- 
nation in vivo, the first observation of this kind emanating from 
Sawtschenko and Melkich,79 who found clumps of the spirochsetes 
of relapsing fever in the blood of infected patients. Also, in natural 
immunity, such as that, for instance, of pigeons against pneumococci, 
Keyes and others have seen clumps of bacteria in the capillaries of 
74 Schidorsky and Reim. Deutsche med. Woch., Vol. 38, p. 1125. 
75 Jaffe. Arch. f. Hyg., Yol. 76. 
76 Sears. Proc. Soc. of Exp. Biol, and Med., 1913. 
77 Metchnikoff. “Immunity in Infectious Disease,” Cambridge, 1905, p. 
263. 
78 Salimbeni. Ann. de VInst. Past., 11, 1897, 277. 
79 Sawtschenko and Melkich. Ann. de VInst. Past. 15, 1901, 497. THE PHENOMENON OF AGGLUTINATION 
271 
the liver and spleen. Recently, Bull 80 has studied the matter in 
more detail. Bull finds that with typhoid bacilli, in the circulating 
blood of normal rabbits, they are rapidly clumped, removed from 
the blood stream and accumulated in the organs where they are taken 
up by polymorphonuclear leukocytes, in the liver and spleen, espe- 
cially. In actively immunized animals, this is still more noticeable, 
both for the organisms mentioned, as well as for pneumococci, 
staphylococci, Shiga dysentery bacilli, and some other organisms. 
Bull’s observation in regard to the action of pneumococcus immune 
serum, in pneumococcus infected rabbits, is particularly interesting 
in that he found that the injection of sufficient amounts of anti- 
pneumococcus serum into septicsemic rabbits resulted in a rapid 
clumping of the pneumococci in the blood stream. Incidental to 
other experiments, we have had occasion often to confirm this ob- 
servation, and have been astonished by the speed with which the 
phenomenon occurs. Thus, if one injects a rabbit with pneumococci, 
allows him to go until the organisms are plentiful in the circulation, 
and then injects intravenously a sufficient dose of anti-pneumococcus 
serum, an immediate heart puncture will reveal clumps of the bacteria 
which were evenly distributed in the smears taken just before the 
injection. If one waits for 3 or 5 minutes after the serum injection, 
it may not be possible to find any organisms in the circulating blood 
by ordinary smear method. 
It will appear from this work that the clumping of bacteria in 
the body of the infected animal may have a very definite protective 
function, consisting in a preliminary concentration of the bacteria 
in the capillaries of various organs where phagocytosis is, in conse- 
quence, facilitated. 
80 Bull. Jour. Exper. Med.. 22, 1915, 475 and 484; also, 24, 1916, 25. CHAPTER X 
THE PHENOMENON OF PRECIPITATION 
(Precipitins) 
The establishment of the agglutinin reaction as a constant and 
specific serum-phenomenon by the work of Gruber and Durham led 
immediately to assiduous investigation of the many problems sug- 
gested by it, and among them, as we have seen, the question of the 
nature of the agglutinogen. It was found that agglutinins could be 
produced, not only by the injection of whole bacteria, but equally 
as well by treatment with dissolved bacterial extracts or with filtrates 
from old broth cultures. This naturally led to the thought that there 
might be a definite reaction if such extracts (instead of the bacteria 
themselves) were added to agglutinating sera in vitro. Rudolf 
Kraus 1 was the first to perform this very logical experiment. He 
was working with broth filtrates of Bacillus pestis and of the cholera 
spirillum, and found that when he mixed the perfectly clear filtrates 
of such cultures with their respective antisera the mixtures would at 
first become turbid and finally show a light flocculent precipitate. He 
named the reaction the “precipitin reaction” and, in analogy to 
agglutinins, spoke of the bodies in the serum which caused the pre- 
cipitation as “precipitins.” The reaction was found, like that of 
agglutination, to be specific; the cholera serum gave no precipitate 
with the plague extract and vice versa, and Kraus, after extending 
his observations to other bacteria, pointed out the practical diagnostic 
possibilities of his discovery. 
Though Kraus’ first observations were made entirely with bac- 
terial culture filtrates and antibacterial sera, it was soon discovered 
that his results were merely isolated instances of a broad biological 
law, and that specific precipitins were produced whenever animals 
were treated with injections of any kind of foreign protein. Thus 
Tschistovitch,2 in 1899, found that the blood serum of rabbits im- 
munized with eel-serum gave specific precipitates when mixed with 
eel-serum, and Bordet 3 obtained analogous results by treating rab- 
bits with defibrinated chicken blood and with milk. Thus rapidly 
the discovery of Kraus was developed into the generalization that 
1 R. Kraus. Wien. klin. Woch., No. 32, 1897. 
2 Tschistovitch. Ann. de Vlnst. Past., 13, 1899. 
3 Bordet, Ann, de Vlnst. Past., Yol. 13, 1899, pp. 225-273. 
272 THE PHENOMENON OF PRECIPITATION 
273 
the sera of animals that have been treated with foreign proteins of 
any kind—bacterial, animal, or vegetable—will develop the property 
of causing precipitates when mixed with clear solutions of the re- 
spective antigens. 
The substances which, after injection into the animal body, lead 
to the formation of precipitating antibodies are spoken of in the 
language of immunology as “precipitinogen.” In the case of bac- 
teria it has been shown that, while the injection of the whole bac- 
terial cell—dead or alive—will lead to precipitin formation, bacterial 
extracts produced in a variety of ways will lead to the same result. 
Such precipitinogen extracts can be obtained by allowing the bacteria 
to grow in flasks of slightly alkaline bouillon, keeping them in the 
incubator for from three weeks to three months, and then filtering 
them through Berkefeldt candles. Again, useful extracts can be 
more rapidly produced by growing large quantities of bacilli on 
agar, emulsifying in salt solution, and shaking in any one of the 
ordinary types of shaking machine for 48 hours or longer. On filter- 
ing an extract is obtained which will form precipitates with homol- 
ogous immune serum, or will incite precipitins when injected into 
animals. In fact, any one of the customary vigorous methods of 
extracting bacterial or other cells will yield precipitinogen. A rela- 
tively purified precipitinogen in the form of a dry, water-soluble 
powder has been obtained by Pick by the precipitation of culture 
filtrates with alcohol. 
Precipitinogens.—Regarding the chemical nature of the pre- 
cipitin-inducing substances, or precipitinogens, the same problems 
have arisen which have been discussed in connection with antigens 
in general. We may say that all soluble native proteins possess 
precipitin-inducing properties. Yet this does not sufficiently define 
the term, since many observations have been published which show 
that physically and chemically altered proteins may still induce 
specific precipitins; a few investigators, furthermore, have claimed 
that they have produced non-protein precipitinogen by various 
methods of breaking up the molecule of the original antigen. In 
the section on agglutination we have seen that moderate heating 
(56 to 65° C.) rather increases than decreases the agglutinogen char- 
acteristics of bacteria, and it is equally true that such heated 
bacteria or bacterial extracts may induce precipitins. The effects of 
various degrees of heat upon the specificity of precipitinogens have 
been extensively investigated and will be discussed, in considerable 
detail, below. 
Of more immediate, indeed of fundamental, importance is the 
problem of a non-protein antigen. The most important claims in 
this regard have been made by Pick,4 Obermeyer and Pick,5 and by 
4 Pick. “Hofmeister’s Beitrage,” Vol. 1, 1901. 
5 Obermeyer and Pick. Wien, klin. Woch., 1904, p. 265, 274 
INFECTION AND RESISTANCE 
Jacoby.6 Jacoby, working with a vegetable antigen, ricin, found 
that by trypsin digestion he could obtain a substance which still 
retained antigenic properties, but no longer gave any of the pro- 
tein reactions. Obermeyer and Pick, by the same method, claim 
that they have produced a non-protein precipitinogen from egg al- 
bumen. On the other hand, others have had negative results, and 
Kraus 7 himself, after reviewing the evidence on both sides, comes 
to the conclusion that available data do not justify us in separating 
the antigenic properties from the protein molecule. In unpublished 
experiments which the writer carried on in the laboratory of Profes- 
sor Friedemann in Berlin also attempts to produce a non-protein 
precipitinogen from horse serum by bacterial putrefaction were en- 
tirely negative. The putrefaction of the serum, though carried out 
in dialyzing bags for the removal of diffusible products, was ex- 
tremely slow, and when finally the Biuret reaction disappeared the 
serum was no longer precipitable by potent antisera. However, the 
flaw in these experiments is that the true test of the presence of 
precipitinogen is not the precipitable character of the solution in 
question, since actual precipitation is dependent, as we shall see, 
upon many modifying secondary factors, but rather the ability of 
the substance to induce precipitins in treated animals. 
The fact that Nicolle,8 and later Pick,9 were unable to obtain 
alcohol-soluble substances from bacteria and bacterial extracts which 
were still precipitable might also be taken to point toward the non- 
protein character of the precipitinogens, suggesting that these sub- 
stances may be of a lipoidal nature. However, as Landsteiner 10 
points out, mere solubility in organic solvents can no longer be taken 
as a proof of lipoidal character, since it is more than probable that 
non-lipoidal substances may go into alcoholic and other organic solu- 
tion when lipoids, such as lecithin, are present. Thus Muller 11 
found that the antigen of typhoid bacilli was soluble in chloroform 
in the presence of old preparations of lecithin. Pick and Schwartz,12 
who had previously studied similar antigen solubilities in the pres- 
ence both of lecithin and of other organ lipoids, suggest that possibly 
such solutions represent lipoid-protein combinations—colloidal “so- 
lutions”—which permit the presence of protein mechanically or 
chemically united to the lipoid in the organic solvents—alcohol, 
chloroform, etc. Here, too, then there is no evidence for the ex- 
istence of non-protein precipitinogen. 
6 Jacoby. “Hofmeister’s Beitrage,” Yol. 1, 1901. Oppenheimer. “Hof- 
meister’s Beitrage,” Yol. 4, 1904, p. 259. 
7 Kraus in “Kolle u. Wassermann Handbuch,” Yol. 4, p. 605. 
8 Nicolle. Ann. de I’Inst. Past., 12, 1898. 
9 Pick. “Hofmeister’s Beitrage,” Yol. 1, 1901. 
10 Landsteiner. “Weichhardt’s Jahresbericht,” Vol. 6, 1910, p. 214, 
11 Muller. Zeitschr. f. 1mm., Vol. 5, 1910. 
12 Pick and Schwartz. Biochem. Zeitschr., Yol. 15, 1909. THE PHENOMENON OF PRECIPITATION 
275 
In regard to the actual precipitable materials in the bacterial 
bodies, our own opinion is that they are not represented in the 
ordinary coagulable proteins alone. We do not wish to repeat too 
extensively observations reported in another part of this book. 
We, therefore, refer the reader to the section on alexin fixation, 
where we have set forth the manner of producing certain alcohol 
precipitable bacterial extracts from which all the ordinary proteins 
had been removed by boiling with acid. The substances there 
described, the chemical analysis of which is now occupying our own 
attention, are specifically and very sharply precipitable by homol- 
ogous antiserum. Indeed, in the case of such organism as pneu- 
mococci, meningococci and influenza bacilli, the reactions are 
sharper and more powerful than with whole bacterial extracts in 
which adventitious substances seem to exert some inhibiting influence. 
Of importance in connection with the problem of the nature of 
precipitinogen, also, is the claim of Myers,13 that specific precipitins 
may be produced in rabbits by treatment with Witte peptone, a sub- 
stance complex in constitution, but consisting largely of albumoses. 
This observation has failed of confirmation in the hands of Ober- 
meyer and Pick, Michaelis,14 ISTorris,15 and others, and cannot, there- 
fore, be accepted as an established fact. 
Whichever method of precipitinogen production is used bacterial 
precipitins appear in the serum of the immunized animal only after 
careful and continued immunization, usually later than the demon- 
strable appearance of the bactericidal or agglutinating properties of 
the serum. The most convenient material for such immunization 
consists of salt solution emulsions of agar cultures, killed at 60° to 
70° C. These may be injected subcutaneously, intraperitoneally, or 
intravenously, the last method leading to the most satisfactory and 
rapid results and, therefore, best employed unless great inherent 
toxicity of the particular bacteria contraindicates. When rabbits 
are nsed it is generally necessary to inject 3, 4, or 5 times at 5- or 6- 
day intervals, and to bleed the animals on the 8th or 9th day after 
the last injection. 
Specificity.—The bacterial precipitins so produced are, as we have 
said above, specific—but, again, specificity, as in the case of agglu- 
tinins, is limited by the so-called “group reactions.” In the chapter 
dealing with agglutination we have seen that the serum of a typhoid- 
immune animal which agglutinates typhoid bacilli strongly will also 
agglutinate, though far less powerfully, paratyphoid bacilli and, in 
some cases, even colon bacilli, this appearance of “minor” agglutinins 
being probably due to a close group relationship of these baoteria to 
the typhoid bacillus. In the case of bacterial precipitins the same 
13 Myers. Centralbl. f. BaJct., Yol. 28, 1900. 
14 Michaelis. Deutsche med. Woch., 1902. 
15 Norris. Jour. Inf. Dis., Yol. 1, 1904. 276 
INFECTION AND RESISTANCE 
thing is true, and has been made the subject of special studies by 
Zupnik,16 Kraus,17 Norris,18 and others. As in the case of ag- 
glutination, however, this fact does not in any way interfere with 
the practical value of the specificity of the reaction because elimina- 
tion of the secondary group reactions, which in agglutination is 
obtained by dilution of the antiserum, can here be obtained, as Kraus 
points out, by diminishing the quantity of the undiluted precipitat- 
ing serum added to the bacterial filtrates. Thus, while one volume 
of serum added to one, two, or three volumes of culture filtrate may 
still give error due to non-specific group reactions, a proportion of 
one part of serum to 8 or 10 parts of the filtrate will usually elimi- 
nate all secondary reactions and prove strictly specific. 
An illustration of such an elimination of “partial” or “minor” 
precipitins by diminution of the amount of the homologous anti- 
serum is given in the following table taken from the work of Nor- 
ris 19: 
ANTICOLI RABBIT SERUM 
TABLE III 
The precipitating action of the anticoli rabbit serum upon its corresponding 
filtrates and upon the filtrates of B. N° 1 (hog cholera) and B. typhosus. 
Coli filtrate 
0.5 c. c. 
0.5 c. c. 
0.5 c. c. 
0.5 c. c. 
Anticoli serum 
0.05 
0.10 
0.15 
0.25 
Cloudiness in all tubes in 1 hour at 37.5° C. which 
increases rapidly. Six hours well-marked precipita- 
tion—most copious in tube containing 0.25 serum. 
Fluid in all tubes becomes clear. 
B. N°1 filtrate 
0.5 c. c. 
0.5 c. c. 
Anticoli serum 
0.10 
0.25 
At 6 hours a slight precipitate in the form of fine 
granules appears on the sides of the tubes. After 
24 hours the precipitate in the tube containing 
0.25 c. c. serum compares in amount to that 
formed in the homologous filtrate with 0.05 c. c. 
of serum. 
B. typh. (Coll) 
filtrate 
0.5 c. c. 
0.5 c. c. 
B. typh. (Pfeif- 
fer) filtrate 
0.5 c. c. 
0.5 c. c. 
Anticoli serum 
0.10 
0.25 
0.10 
0.25 
Similar reaction obtained to that with B. N° 1 filtrate. 
Similar delay in reaction as obtained with B. typh. 
Coll. 
And, indeed, though the great practical value of the precipitin 
reaction has not been in the special field of bacteriology, it has been 
16 Zupnik. Zeitschr. f. Hyg., 49, 1905. 
17 Kraus. Wien. klin. Woch., 1901, No. 29. 
18 and 18 Norris. Jour. Inf. Dis., Yol. 1, 1904, p. 472. THE PHENOMENON OF PRECIPITATION 
277 
successfully utilized in the diagnosis of glanders by Wladimiroff,20 
and constitutes a valuable adjuvant to our methods of determining 
the biological relationship between micro-organisms. 
The production of precipitins against unformed proteins, egg 
albumen, animal sera, etc., is much more easily accomplished than 
the production of bacterial precipitins, and three intravenous injec- 
tions of from 2 to 5 c. c. of the protein at 5- or 6-day intervals usually 
give rise to a formation of potent precipitins. When a small quan- 
tity of the serum of such an animal, taken 9 or 10 days after the 
third injection, is mixed in a test tube with an equal quantity of a 
dilution of the protein which was injected, turbidity and rapid floccu- 
lation will result. In tests of this kind, unlike the bacterial precipitin 
tests in which the delicacy of the reaction is ordinarily determined 
by diminution of the amounts of antiserum, the same object may be 
more conveniently attained by dilution of the antigen. Thus, in test- 
ing the precipitating potency of, let us say, the serum of a rabbit 
immunized with sheep serum, we would proceed by setting up a 
series of small tubes, each of which contains a constant amount of 
antiserum (precipitin), but a progressively diminishing amount of 
antigen in the same volume—i. e., in dilution with isotonic salt solu- 
tion. The following example will make this clear: 
Antisheep serum 
from rabbit 
Sheep serum 0.5 c. c. 
of following dilutions: 
Precipitation 
0.5 c. c. 
+ 
1:10 
d= 
0.5 c. c. 
+ 
1:100 
+ + + 
0.5 c. c. 
+ 
1 500 
+ + + 
0.5 c. c. 
+ 
1:1,000 
+ + 
0.5 c. c. 
+ 
1:5,000 
1:10,000 
4- 
0.5 c. c. 
+ 
— 
In this test it will be noticed that the strongest concentration of 
the antigen (1:10) gave a relatively slight precipitation only. This 
phenomenon is analogous to the inhibition zones noticed in the case 
of agglutination and other antibody reactions and will be more espe- 
cially discussed in a succeeding paragraph. 
The delicacy of the above example, moreover, is by no means 
unusual, and sera have been obtained by careful immunization with 
which the specific antigen could be detected in dilutions as high as 1 
to 100,000 (Uhlenhuth). A serum which will react with antigen 
dilutions of 1 to 10,000 and 1 to 20,000 is not at all uncommon nor 
difficult to obtain. Apart from the additional advantage of the 
20 Wladimiroff. “Kolle u. Wassermann Handbuch,” article on “Glanders,” 
Yol. 5, 2d Ed. 278 
INFECTION AND RESISTANCE 
specificity of the reaction, therefore, this biological method of de- 
tecting proteins is more delicate than that of any of the known 
chemical methods; neither the Biuret nor Millon’s reaction will far 
exceed a delicacy of 1 to 1,000. By a modification known as the 
method of Complement or Alexin-fixation, furthermore, the delicacy 
of the biological reactions can be still further enhanced. This is 
discussed in detail in another place. 
The practical value of the precipitin reaction, however, lies, not 
in the mere detection of protein, but, by virtue of its specificity,21 in 
the determination of the variety of protein under examination. And 
here again the specificity, like that of bacterial precipitation, ag- 
glutination, and other serum tests, is relative rather than absolute. 
Thus a serum which has been obtained by the immunization of an 
animal with human serum may react, not only with human serum, 
but also with relatively higher concentrations of the sera of some of 
the higher apes. However, such non-specific partial reactions can be 
eliminated entirely by employing higher dilutions of antigen. Thus 
Uhlenhuth,22 on the basis of a large experience, has established 
a standard of antigen dilution at 1 to 1,000, beyond which no “para” 
or “minor” precipitation will occur. Since potency far exceeding 
this is easily procured, absolute specificity can be ensured by the 
very simple precaution of a sufficient dilution. 
Species Specificity of Precipitin Reactions and Its Uses.—The 
most important practical use for the reaction has been found in 
forensic medicine, where it is possible in this way to determine 
the species of animal from which have emanated the blood, sperm, 
etc., found in spots on wearing apparel, weapons, or other articles. 
The extensive investigations of fSTuttall 23 upon this subject have inci- 
dentally been of much value in furnishing a further method for the 
determination of zoological species relationships. ISTuttall carried 
out 16,000 precipitin tests, with precipitating sera, upon 900 speci- 
mens of blood which he obtained from various sources. He not only 
confirmed many of the accepted zoological classifications, but shed 
much light upon a number of disputed points. In working out the 
tests upon monkeys he found that the reactions carried out with anti- 
human serum become weaker as the species examined is farther re- 
moved from man zoologically. Thus as we read down the column 
21 Wassermann and Schiitze. Deutsche med. Woch., 1900, Yereinsbeilage, 
p. 178; Berl. klin. Woch., 1901; Deutsche med. Woch., 1902; Bordet, Ann. 
Fast., Yol. 13, 1899; Nolf, ibid, Yol. 14, 1900; Fish, Medical Courier, St. 
Louis, 1900, cited from Wassermann. 
22 Uhlenhuth. Deutsche med. Woch., 1900, 1901; Rob. Koch Festchrift, 
1903. Uhlenhuth and Weidanz. “Kraus u. Levaditi Handbuch,” etc., Yol. 
2, 1909. Uhlenhuth and Weidanz. Loc. cit., where other publications are 
summarized. 
23 Nuttall. “Blood Immunity and Blood Relationship,” Cambridge Uni- 
versity Press, 1904. THE PHENOMENON OF PRECIPITATION 
279 
from man to the hapalidse the precipitate becomes less and less in 
amount. 
Nuttall’s Tests with Antihuman Serum. (Nuttall, loc. cit., p. 165.) 
Tested against Precipitate 
34 Specimens human blood 100%* 
8 Simiidae, 3 species 100% 
36 Cercopithecidae 92% 
13 Cebidae 78% 
4 Hapalidae 50% 
2 Lemuridae 0 
ANTIHUMAN PRECIPITATING SERUM 
* The percentages refer to the volume of precipitate formed on standing for a given 
time, the amount formed by the antiserum with its specific antigen being taken as 100 
per cent. Antigen dilutions correspond throughout. 
In another series he finds: 
ANTIHUMAN PRECIPITATING SERUM 
Tested against Precipitate 
Man 100% 
Chimpanzee (loose precip.) 130% 
Gorilla 64% 
Ourang 42% 
Cynocephalus mormon 42% 
Cynocephalus sphinx 29% 
Ateles 29% 
Among the primates the highest figures with antihuman serum are 
given by the chimpanzee. Other bloods than those of the primates 
gave slight reactions or none whatever with the antihuman serum. 
In addition to these results the relationships within the dog 
family, the horse family, and many other kinships similar to these 
were confirmed. In every case the precipitin reaction was con- 
sistent with the results of other methods of classification, and Nut- 
tail’s work is an extremely valuable aid to zoologists in disputed 
questions of animal relationships. 
These facts are the more surprising in that they demonstrate 
species differences between the proteins of various animals which 
are not determinable by known chemical methods. How funda- 
mental these differences are and how delicate the reaction, is further 
shown by experiments of Uhlenhuth, in which he obtained a specific 
antihare serum by treating rabbits’ with hares’ blood, an astonishing 
result in view of the close zoological relations between these animals. 
Isoprecipitins, that is, precipitins resulting from the treatment 
of animals with blood of another individual of the same species, have 
also been described by Schiitze and others. They are not, however, 
regular in their appearance, nor are they very potent when obtained. 280 
INFECTION AND RESISTANCE 
Since the reaction is equally applicable to vegetable proteins, 
similar investigations on the interrelationship of different varieties 
of wheat have been carried out by Magnus.24 
The methods of performing precipitin tests for forensic or other 
purposes is extremely simple. Nevertheless, there are a number of 
theoretical considerations which we must take up in order to make 
clear the limitations of accuracy and conditions of control which are 
involved in these reactions. From our discussion of the nature of 
precipitinogen it follows that blood stains, etc., on linen or articles 
of any kind will be suitable for precipitin tests even after they have 
been exposed for considerable periods to unfavorable conditions, that 
is, an environment in which they are subjected to exposure to light, 
moderate heat, or drying. Thus blood spots, etc., if kept dry and in 
the dark, may give positive reactions even after years, as experi- 
ments by Uhlenhuth have shown. Meyer25 claims even to have 
obtained a precipitation with extracts of the material of mummies. 
One of his specimens was a mummy dating back to the first Egyptian 
Empire (5,000 years), the other about 2,000 years old. Pieces of 
the leg and neck muscles of these specimens were chopped up finely, 
extracted for 24 hours with salt solution, then filtered until clear. 
With antihuman serum they gave turbidity after one hour at 
37.5° O. 
Under conditions of putrefaction, of course, the precipitinogen 
is more rapidly destroyed, though blood putrefies with surprising 
slowness, even if, as in our own experiments, the conditions of mois- 
ture, temperature, and reinoculation with putrefactive bacteria are 
constantly observed. Under such conditions a weak reaction may be 
obtained after as long as a month or six weeks. 
In carrying out the tests with any material it is first necessary 
to get it into clear solution, a result which is best accomplished by 
soaking it in a small quantity of isotonic salt solution. Preliminary 
to this it is always necessary to scrape off a bit of the specimen and 
examine it microscopically to discover, if possible, whether blood cells, 
sperm, or other cellular constituents can be detected. The infusion in 
salt solution should be continued for several hours—if necessary for 
12 to 24 hours. After the first few hours in the incubator the material 
should be placed at room or refrigerator temperature so that the 
yield in unchanged protein may not be diminished by the action of 
bacterial growth. After extraction the solution may be filtered in 
order to clear it, but often mere centrifugation suffices for this pur- 
pose. The concentration of antigen in such an extract is always an 
uncertainty, but may be determined with sufficient accuracy for 
practical purposes by shaking and observing the formation of a 
lasting foam. Protein solutions will show foam on shaking in dilu- 
24 Magnus. Cited from Uhlenkuth, loc. cit. 
25 Meyer. Munch, med. Woch., Yol. 51, No. 15, 1904. THE PHENOMENON OF PRECIPITATION 
281 
tions as high as 1 to 1,000, and if the original amount of salt solution 
used in washing out the material is properly gauged to the amount 
of blood available in the stain, and the solution shaken and observed 
for the formation of foam, it is usually a simple matter to obtain a 
final concentration approximating one to one thousand.26 
The antiserum which is used should be of such a potency that 
preliminary titration with the specific antigen, diluted 1 to 1,000, 
should give an almost immediate cloudiness at room temperature. 
By testing this serum against a number of other varieties of 
protein—dog serum, beef serum, etc.—it must be determined that 
the precipitin in this case is strictly specific. 
The reaction can be observed with greater delicacy if it is first 
set up by the method recommended by Fornet and Muller,27 which 
we may speak of as the “ring test.” The antiserum is put into the 
tubes and the solution to be tested is allowed to flow slowly over this 
—as in Heller’s nitric acid albumin test. At the line of contact be- 
tween the two a fine white ring will rapidly appear, thickening and 
growing heavier as the preparation is allowed to stand. After taking 
the final readings from such a test, let us say after an hour or so, it 
is well to shake up the tubes, set them away in the ice-chest, and again 
read the amount of precipitates formed in the various tubes the next 
morning. Since every test of this kind necessitates a number of 
controls, the following example will serve as a basis for discussion: 
Forensic Blood Examination 
Material: Blood spot on trouser pocket, washed up in salt solution. Clear 
after paper filtration. 
Antiserum: Rabbit treated with three intravenous injections, 2, 5, and 5 c. c. 
of human serum at six-day intervals; bled on tenth day after last injec- 
tion. This serum has been titrated against human serum and gives 
precipitation in dilutions up to one to ten thousand. With one to one 
thousand there is clouding which begins in three minutes and is very dis- 
tinct in eight minutes, at room temperature.28 
Test 
Tubel. Known human serum 1 to 1,000... 1.0 c. c. + Antiserum... .0.2 c c. 
Tube 2. Unknown solution to be tested 1.0 c. c. + Antiserum... .0.2 c. c. 
Tube 3. Unknown solution to be tested 1.0 c. e. + Normal rabbit 
serum.... 0.2 c. c. 
Tube 4. Salt solution 1.0 e. c. + Antiserum... .0.2 c. c. 
Tube 5. Unknown solution 1.0 c. e. + Salt solution. .0.2 c. c. 
26 If there is enough material, the amount of dissolved protein can be 
also approximately gauged by adding to a little of it a drop of acid, boiling 
and observing the heaviness of the cloud which forms. A control test of a 
known dilution of the suspected variety of blood can be made at the same 
time and the heaviness of this cloud compared with that in the test solution. 
27 Fornet and Muller. Zeitschr. f . Hyg., Yol. 66, 1910. 
28 A mixture of too specific antisera should never be used, since such 
sera may often precipitate each other for reasons that are discussed below. 282 
INFECTION AND RESISTANCE 
In this test, if the original material was human blood, tubes 1 
and 2 should show ring formation within 5 minutes—while the 
other tubes remain clear. In addition to these controls it is well to 
be sure that the test extract is neither strongly acid nor alkaline, and 
that, as Uhlenhuth suggests, the material from which it is extracted 
does not contain other substances which can give precipitates by 
themselves when added to serum. This is especially necessary in the 
case of cloth fabrics, and a recent instance in our own experience 
has suggested to us the possibility that such materials may also 
contain colloidal dyestuffs or other extractable substances which can 
cause inhibition of the precipitation. In an apparently positive 
case the reactions with a blood extract from trouser cloth were suffi- 
ciently heavy, but regularly delayed, as in the flocculation of such 
colloidal suspensions as arsenic trisulphide in the presence of a 
protective colloid. 
In the ordinary criminal or civil case which would come under 
consideration for precipitin tests the spots or stains are made by 
blood as it flows from the wound and unchanged by chemical or 
physical agencies except as these are encountered. afterward, by 
exposure. In the case of meat inspection, in which the precipitin 
test is useful in detecting admixtures of horse flesh, dog flesh, or 
other less desirable varieties of meat, in sausages, chopped meat, 
etc., it often happens that such procedures as heating or smoking 
may vitiate the results of precipitin reactions. It is of practical im- 
portance, therefore, that we should know exactly what the effects of 
heating (boiling) may be upon precipitinogen. Moreover, this ques- 
tion possesses considerable theoretical interest since the coagulation 
of proteins by heat seems to involve chiefly a physical rather than a 
chemical change. 
Cohnheim 29 says in discussing this question: “It is still unclear 
what the changes are that take place in coagulation. It may be that 
there is merely an intramolecular ‘Umlagerung’—or there may be 
cleavage; or the process may be comparable to the flocculation of col- 
loidal clay emulsions by salts. . . . With coagulation all proteins 
have lost the differences which they possess in the native state in 
respect to solubility or precipitability by salts. Physically all coagu- 
lated proteins are alike; they are no longer native proteins, and with- 
out further decomposition are insoluble. Chemical differences, how- 
ever, variations of composition, and the cleavage products which 
they yield still distinguish them.” 
The question has been experimentally approached by Obermeyer 
and Pick 30 in connection with their general investigations upon the 
influence of chemical and physical alterations upon precipitinogen. 
29 Otto Cohnheim. “Chemie der Eiweiss Korper Vieweg Braunschweig,” 
1900, p. 8. 
30 Obermeyer and Pick. Wien. klin. Woch., 12, 1906. THE PHENOMENON OF PRECIPITATION 
283 
They found that precipitin produced with unchanged (native) beef 
serum does not react with heated beef serum, even if immunization 
was prolonged and a very potent serum was produced. On the other 
hand, when animals were immunized with beef serum which had 
been boiled for a short time (“Kurz aufgekocht” 31) the precipitin 
so produced reacted, not only with native beef serum, but also pre- 
cipitated the boiled serum and a whole row of split products which 
give no reaction to normal precipitin. The “coctoprecipitin” so 
produced, furthermore, was found by them to be specific, acting only 
upon beef protein or its derivatives. 
It is immediately evident that these investigations are closely 
analogous to those of Joos and others on the agglutinins. The anti- 
serum produced with the heated antigen here again reacts both with 
the native and with the heated antigen, whereas the antiserum 
produced with the native unheated antigen reacts only with the 
unheated. The “heat-precipitins” therefore may be also called 
“umfanglicher”—the term applied by Paltauf to the agglutinins 
produced with heated bacteria. 
Schmidt,32 who has studied the problem extensively, finds that 
heating serum protein to 70° C. for as long as 30 to 60 minutes 
alters its precipitability by “native precipitin” (precipitin produced 
by immunization with native unheated serum) only in so far as it 
diminishes the delicacy of the reaction by 10 to 30 per cent., and 
that heating to 90° C. for as long as an hour does not render it en- 
tirely non-precipitable, so that protein so treated may yet be detect- 
able by ordinary specific precipitins produced by injections of un- 
heated serum, though the delicacy of the reaction is lessened. Boil- 
ing, according to Schmidt, renders the antigen no longer precipitable 
by such “native precipitin,” but, on the other hand, it does not seem 
to destroy its antigenic property of inciting precipitins on injection 
into animals. Fornet and Muller, on the other hand, claim that even 
boiled protein can be detected by “native precipitins,” though the 
reaction is only about one-tenth as delicate as it is with unheated 
protein. 
Schmidt studied these relations especially as they affect the per- 
formance of specific precipitin reactions in the identification of 
boiled meat. He found that when he immunized rabbits with serum 
protein that had been heated at 70° C. for 30 minutes the antiserum 
so obtained gave strong and practically useful reactions with its 
specific antigen even if this had been boiled. Since “native pre- 
cipitin” gives weak reactions only with such a boiled protein, 
31 Sera or other proteins used in such tests are boiled in dilutions of 1 
to 10 or more, in order to avoid the formation of heavy flakes which cannot 
be injected. Boiled in sufficient dilution, an opalescent suspension is formed 
which easily passes through the syringe. 
32 Schmidt. Biochem. Zeitschr., 14, 1908; also Zeitschr. /. 1mm., Yol. 
13, 1912. 284 
INFECTION AND RESISTANCE 
Schmidt recommends the use of the “70° precipitin” (produced by 
injections of heated serum) for tests in which a heated antigen is to 
be identified. 
He states, however, that very prolonged heating may so com- 
pletely coagulate the antigen that none of it can be gotten into “solu- 
tion” (suspension), and in such cases results can be obtained neither 
with the “native” nor with the “70° precipitin.” He has at- 
tempted, therefore, to find a method whereby even such entirely 
insoluble proteins may be identified, and claims to have succeeded 
by preparing what he calls his “heat-alkali-precipitin.” 33 He di- 
lutes serum with equal parts of isotonic salt solution and heats it to 
70° C. for 30 minutes in a water hath. To 60 c. c. of such a solution 
he now adds 10 c. c. of " NaOH, and continues heating for 15 to 
20 minutes. At the end of this time he neutralizes with IIC1, cools, 
and injects 20 c. c. intraperitoneally into rabbits. (The neutraliza- 
tion is not absolutely necessary.) Five or more injections yield a 
serum sufficiently potent for use. 
A precipitin so produced will, according to Schmidt, react spe- 
cifically with heated proteins, and also with protein which has been 
solidly coagulated and brought into solution by means of NaOH and 
heat. It will not, however, react with normal unheated antigen. 
He tested this by coagulating horse serum by boiling for 3 hours. 
The coagulum was washed with salt solution, dried, and powdered. 
Tests were then made to prove that this powder was entirely in- 
soluble in NaCl solution. A little of it was then treated with 10 
c. c. of salt solution containing enough NaOH to correspond to an 
x solution. The exposure was continued for 20 minutes in a water 
bath at 60° to 70° C. Before the entire mass was dissolved the solu- 
tion was filtered and neutralized with Tn0- HC1. 
The rather complicated relations described by Schmidt are easily 
surveyed in the following protocol taken from his work (see table I, 
p. 285) : 
Schmidt speaks of the “heat-alkali-precipitin” also as “alkali- 
albuminate-precipitin.” It can be produced only if the NaOH treat- 
ment of the serum is cautiously performed. If the sodium hydroxid 
is allowed to act too vigorously in strong concentrations or for too 
long a time the antigen is completely destroyed, is no longer pre- 
cipitable, and no longer produces precipitin when injected into 
animals. 
The striking feature of these experiments is that they show a 
gradual alteration of the protein first by heat, then by alkali and 
heat, in such a way that the antigenic properties are changed but 
33 “Native precipitin” — precipitin produced by injections of normal un- 
heated serum. 
“70° precipitin” = precipitin produced by injections of serum heated 
to 70° C. for 30 minutes. THE PHENOMENON OF PRECIPITATION 
285 
TABLE I 
(W. A. Schmidt, Zeitschr. f. Imm., Vol. 13, 1912, p. 173) 
Solution 
of 
Natiye 
precipitin 
Heat (70°) 
precipitin 
Heat-alkali- 
precipitin 
Native serum 
Strong reaction 
Good reaction 
0 (very slight 
turbidity) 
70° serum (heated 30 min.) 
100° serum (heated 30 
Good reaction 
Strong reaction 
Strong reaction 
min.) 
70° serum treated with 
NaOH (used to produce 
0 
Good reaction 
Strong reaction 
heat-alkali-precipitin) . 
Boiled insoluble serum, 
brought into solution 
0 
0 
Strong reaction 
with NaOH 
Native serum treated with 
0 
0 
Good reaction 
NaOH in the cold 
0 
0 
Good reaction 
not destroyed. Each precipitin, moreover, seems to react most 
strongly with the particular antigen-alteration which produced it, 
and, according to Schmidt, retains its species specificity. This is 
not the case with the iodized proteins and nitroproteins and diazo- 
proteins produced by Obermeyer and Pick.34 Here iodized beef 
protein injected into animals produced a precipitin which reacted 
with the iodized protein, not only of the beef, but also similarly 
altered proteins of other animals—and the same was true of the 
nitro and diazo modifications. 
Although the experiments of Schmidt have great theoretical 
value, their practical utilization must depend upon the degree of 
specificity possessed by the heat-precipitins or the heat-alkali-pre- 
cipitins. In Obermeyer and Pick’s original investigations we have 
seen that they found the precipitin produced with heated serum as 
strictly specific as that induced by native serum. This has also been 
the experience of Schmidt. Eornet and Muller,35 on the other hand, 
report that the precipitins produced by them with heated muscle- 
protein were not as strictly specific as those produced with the un- 
heated—in that the former gave precipitates, not only with homol- 
ogous protein solutions, but with foreign proteins in moderate con- 
centration as well. In experiments carried out by the writer with 
Ostenberg 36 it was attempted to determine whether or not precipi- 
tins could be produced by injecting animals with protein that had 
been boiled, and if so what the action of these substances would be 
upon boiled proteins. Contrary to the results of Fornet and Muller, 
34 Obermeyer and Pick. Wien. klin. Woch., No. 12, 1906. 
35 Fomet and Muller. Zeitschr. f. Hyg., Vol. 66, 1910. 
36 Zinsser and Ostenberg. Proc. N. Y. Pathol. Soc., 1914. 286 
INFECTION AND RESISTANCE 
Cross titrations—dilutions of sera in salt solution boiled 5 minutes, pre- 
cipitated with antisera produced by injections with similarly boiled material. 
The readings here indicated were taken by “ring” test at the end of 30 
minutes. 
Experiments on Cocto-precipitin. Table II (March 23, 1913). 
Dilution 
Beef 
serum 
vs. 
anti- 
beef 
precipi- 
tin 
Beef 
serum 
vs. 
anti- 
dog 
precipi- 
tin 
Beef 
serum 
vs. 
anti- 
sheep 
precipi- 
tin 
Dog 
serum 
vs. 
anti- 
dog 
precipi- 
tin 
Dog 
serum 
vs. 
anti- 
beef 
precipi- 
tin 
Dog 
serum 
vs. 
anti- 
sheep 
precipi- 
tin 
Sheep 
serum 
vs. 
anti- 
sheep 
precipi- 
tin 
Sheep 
serum 
V8. 
anti- 
dog 
precipi- 
tin 
Sheep 
serum 
vs. 
anti- 
beef 
precipi- 
tin 
1:20 
+ 
+ 
+ 
+ + 
+ 
+ + 
1:50 
+ + + 
+ 
+ + + 
+ + 
— 
+ + 
+ + + 
+ 
+ + + 
1:100 
+ + + 
+ 
+ + 
+ 
— 
+ 
+ + 
+ 
1:500 
+ + 
— 
+ 
+ 
— 
+ 
+ 
1:1,000 
± 
— 
± 
— 
± 
± 
— 
— 
Controls of 
boiled serum 
alone* 
1:20 
1:50 
Serum 
control 
— 
— 
* These controls were necessitated by the fact that the boiled serum suspensions were them- 
selves turbid and occasionally showed slight settling on standing. 
it was actually found that sera boiled for 3 to 5 minutes injected into 
rabbits induced precipitins which acted upon boiled proteins, but at 
the same time it was determined that the antibodies so produced 
were no longer strictly specific. The protocol given at the top of the 
page will illustrate these experiments. 
Summarizing these results together with those of Fornet and 
Muller and of Schmidt it would seem that the injection of boiled 
proteins induces precipitins which no longer act on native antigen, 
which act powerfully on boiled antigen, but are no longer strictly 
specific. This seems to us of great theoretical interest as showing 
an alteration by heating in the species adherence of the antigen. 
Practically, therefore, precipitins produced with boiled protein are 
of little value, and forensic determinations of boiled proteins should 
be done, as advised by Schmidt, by the “70° precipitins,” or with 
native precipitin as practiced by Fornet and Muller. 
Organ Specificity.—The specificity which is the basis of the prac- 
tical value of the reactions that we have discussed is spoken of as 
“species” specificity since it has been found that the blood serum of 
rabbits or other animals into which the serum of another animal has 
been injected reacts, not only with the homologous blood serum, but 
also with extracts of the various organs of the particular species of 
animal which furnished the serum. Thus if we immunize rabbit, 
let us say, with sheep serum the resulting precipitin will react, not 
only with sheep serum, but also with extracts of sheep spleen, sheep THE PHENOMENON OF PRECIPITATION 
287 
liver, etc. It seems that every species of animal possesses throughout 
its tissues a particular variety of protein, fundamental to its general 
metabolism and peculiar to its species. On the other hand, we have 
seen in the preceding discussions how chemically slight the changes 
in a protein may be which can alter materially its antigenic nature, 
and it is a logical deduction that different organs of the same animal 
might contain antigenic constituents qualitatively different from the 
general serum protein. There are undoubtedly in many organs pro- 
tein complexes which are peculiar to them and not present in other 
organs, and it would be reasonable to expect therefore that im- 
munization with separate organ substances would lead to the pro- 
duction of sera of specific precipitating power for the protein of that 
particular kind of organ. This is not ordinarily obtainable, how- 
ever, because it has been impossible to isolate from organs their pe- 
culiar, characteristic proteins, and immunization of animals with 
organ extracts or solutions has necessarily implied the injection of 
much blood protein and other albuminous material of a character 
general to many organs of the animal, i. e., to the species. These 
quantitatively overshadow the organ-specific substances which may be 
present, and give rise, therefore, to a “species” precipitin. That 
“organ specificity,” however, is a fact has been shown by the experi- 
ments of Uhlenhuth with the protein of the crystalline lens of the eye. 
Immunization with this substance induces a precipitin which does 
not react with the serum of the animal from which the lens was taken, 
hut does react, not only with the crystalline lens proteins of this spe- 
cies of animal, but also with crystalline lens proteins in general, 
though taken from another animal species. Analogous to this are the 
experiments of von Dungern and others upon the protein derived 
from the testicle. 
In both of these cases, as well as in other less sharply defined 
examples, the specificity is attached, not to the species of animal, 
but rather to the nature of the organ from which the particular 
protein is derived. These facts—first ascertained by means of the 
precipitin reaction—have been recently confirmed by means of the 
reaction of anaphylaxis by Uhlenhuth and Haendel, and by Kraus, 
Doerr, and Sohma. (See chapter on Anaphylaxis.) They have 
been discussed, moreover, in connection with the problem of spec- 
ificity in general. 
Biologically they probably signify that, although there are fun- 
damental species differences between the general body proteins of 
various animals, there are still, in certain highly specialized organs, 
varieties of protein which, possibly because of functional exigencies, 
have developed similar chemical characteristics. These have been 
determinable by our present methods, however, only for organs like 
the lens, the testicle, and the placenta from which the organ-specific 
protein can be gotten in a relatively pure state. The pathological 288 
INFECTION AND RESISTANCE 
importance of these phenomena lies in the fact that, although guinea 
pig serum injected into a guinea pig will not give rise to antibodies, 
lens protein apparently will do so—an observation which opens the 
possibility of autocytotoxins. The significance of this is indicated in 
such investigations as those of Homer,37 who, using the complement- 
fixation technique to determine antibody, found that the serum of 
adult human beings possessed antibodies for their own lens protein, 
but that such antibodies were absent in the sera of children. 
The study of agglutination and that of precipitation reveal,’ 
throughout, a close similarity between the two reactions, and indeed 
in physical principles they are probably the same, although the one 
(agglutination) consists in the flocculation of large particles in sus- 
pension—the bacteria—while in the other the precipitation is one 
of smaller units—the precipitable colloidal particles of the protein 
solutions. This phase of the subject will be more thoroughly dis- 
cussed directly. 
Schematic Representation op Ehrlich’s Views on the Structure op Pre- 
cipitins. 
Ehrlich’s Conception of Precipitins.—Meanwhile, it is noticeable 
also that, even without drawing the physical parallel between the two 
reactions, there is much in the behavior of the antibodies—the 
agglutinins and the precipitins as conceived by Ehrlich, which led 
him and his school to attribute to them a similar receptor structure. 
Like the agglutinins, the precipitins are not inactivated by 56° C., 
but when once rendered ineffectual by higher temperatures (70° C. 
or over) they can no longer be reactivated by the addition of fresh 
normal serum. For this reason chiefly Ehrlich attributed a similar 
structure to both agglutinins and precipitins, calling them “haptines 
of the second order.” 
Ehrlich assumed that when dissolved protein substances—ordi- 
narily suitable for body nutrition—are injected into animals, they 
become anchored to the cells by such receptors of the second order. 
When overproduction occurs in response to repeated stimulation of 
the cells by consecutive injections (see Side-Chain Theory), these 
haptines of the second order circulate as agglutinins or precipitins. 
37 Romer. Klin. Monatsbl. f. Augenheilkunde, Sept., 1906. Ref, from 
“Weichhardt’s Jahresbericht,” Yol. 2, 1906, p. 348. THE PHENOMENON OF PRECIPITATION 
289 
Since they act without the apparent cooperation of alexin, he sup- 
poses that they carry within themselves the “zymophore,” or ferment 
groups, by means of which the agglutination or coagulation is ac- 
complished. It is this zymophore group which, it is assumed, accom- 
plishes the digestion of the foreign protein before its assimilation, 
when these receptors are still parts of the living cell. 
Thus the conception of precipitins is identical with that formu- 
lated by the same school concerning the agglutinins, and the deduc- 
tions from these premises have been essentially similar. Thus, anal- 
ogous to the conditions prevailing in agglutination, Pick,38 and 
Kraus and v. Pirquet 39 have shown that when precipitating serum 
is inactivated by heat, and then is added to bacterial filtrates, it will 
prevent their subsequent precipitation by active precipitin. An 
illustration of this is found in the following protocol taken from the 
paper by Kraus and v. Pirquet (loc. cit., p. 69). 
(a) 5 c. c. cholera filtrate + 0.5 c. c. inactiv. (60°) cholera serum = no pre- 
cipitate after 10 hours at 37° C. 
After 10 hours add 0.5 c. c. active cholera serum — no precipitate. 
(b) Omitted. 
(c) Omitted. 
(d) 5 c. c. cholera filtrate + 0.5 c. c. active cholera serum = after 10 hrs. 
typical precipitate. 
From this it was concluded that heat may destroy the zymophore 
or coagulating group of precipitins, leading to the formation of 
“precipitinoids” which, like agglutinoids, may have a higher affinity 
for the antigen than is possessed by the uninjured antibody. 
Subsequently there were opposed to these views the physical in- 
terpretations which have been outlined sufficiently in the section on 
Agglutination (see p. 267). In the case of precipitation the anal- 
ogy between colloidal reactions and the serum phenomena is fully as 
striking as in the former, an analogy in the delineation of which the 
first credit belongs to Landsteiner,40 and important further contri- 
butions have been made by Keisser and Friedemann, Porges, Gen- 
gou, and a number of others. As in agglutination and colloidal floc- 
culation, the presence of salts (electrolytes) fundamentally influ- 
ences the occurrence of precipitin reactions; and in both colloidal 
and precipitin reactions the relative concentration of the reacting 
bodies is paramount in determining whether or not precipitation 
takes place. In this connection the most frequently observed inhibi- 
tion occurring in serum precipitations is that which is caused by an 
excess of antigen. An example of this is as follows: 
38 Pick. “Hofmeister’s Beitrage,” Yol. 1, 1902. 
39 Kraus and v. Pirquet. Centralbl. f. Bakt., Vol. 32, 1902. 
40 Landsteiner and Jagic. Munch, med. Woch., No. 18, 1903; No. 27, 
1904; Wien. klin. Woch., No. 3,1904. Landsteiner and Stankovic. Centralbl. 
f. Bakt., Yols. 41 and 42, 1906. 290 
INFECTION AND RESISTANCE 
Sheep serum 0.5 c. c. 
Antisheep rabbit serum 
Precipitate 
1:10 
+ 
0.5 c. c. 
— 
1:50 
+ 
0.5 c. c. 
=fc 
1:100 
+ 
0.5 c. c. 
+ + 
1:500 
+ 
0.5 c. c. 
+ + + 
1:1,000 
+ 
0.5 c. c. 
+ + 
1:5,000 
+ 
0.5 c. c. 
+ 
This is entirely analogous to the inhibition which may occur 
when, let us say, a weak gelatin solution is added to a colloidal sus- 
pension of arsenic trisulphid; or blood serum is added to mastic or 
arsenic suspensions. In both cases inhibition zones appear which 
show that the relative quantities of the two reacting bodies are quite 
as significant as their chemical or physical constitution in determin- 
ing the occurrence of flocculation. This, according to Bechold, Bil- 
litzer,41 and others depends upon the fact that the reason for floc- 
culation is one of electrical charge. One hydrosol—say arsenic 
trisulphid—can be flocculated by the oppositely charged colloidal 
aluminium hydroxid, but this will occur only when the quantitative 
relations are properly adjusted. If one or the other is in excess, no 
flocculation may occur, and, if subjected to a direct current, both 
colloids, though ordinarily wandering in opposite directions, will 
now wander in that of the one which is now present in the largest 
amount. We will not elaborate here upon the causes for this, since 
they have been indicated in the section on Agglutinins and in the 
section on Alexin-fixation where Dean’s accurate analysis of the 
importance of the relative amounts of antigen and antibody for 
precipitin reactions are discussed. 
This effect of quantitative proportions would explain not only 
the absence of precipitation in the presence of too much antigen, but 
also the converse phenomenon, already mentioned, that precipitation 
may be inhibited when the precipitin is in excess. 
The fact that heated precipitating serum when added to its an- 
tigen not only does not cause flocculation, but may even prevent sub- 
sequent precipitation by active precipitin, also finds its analogy in 
colloidal reactions in the so-called protective colloids. Thus arsenic 
trisulphid may be protected from precipitation by gelatin, if a small 
amount of gum arabic is added, and the analogy has been brought 
even closer by Porges,42 who showed that heated serum will protect 
mastic suspension from precipitation by normal serum. This obser- 
vation of Porges is so closely similar to the results obtained by Kraus 
and v. Pirquet and others on the inhibition of precipitation by heated 
41 Billitzer. Cited from Bechold, “Die Kolloide, etc./’ p. 79. 
42 Porges. Chapter on “Colloids and Lipoids” in “Kraus u. Levaditi 
Handbuch ” Yol. 1. THE PHENOMENON OF PRECIPITATION 
291 
precipitating serum that it would seem, on first consideration, effec- 
tually to refute the conception of “precipitoids.” 
However, it does not explain the specificity of such inhibition on 
the part of heated precipitating serum, as reported by Kraus and 
v. Pirquet, an observation which is one of the strongest arguments in 
favor of the derivation of the inhibiting factor from the specific 
precipitin (a precipitoid).43 
This apparent specificity of the precipitoids we think can be 
explained on the same basis on which we have explained the so- 
called specificity of “agglutinoids,” where we have attributed it 
to the colloidal protective action of the inactive protein mate- 
rials carried into combination with the antigen by the specific 
precipitin. 
In spite of the strong evidence in favor of the colloidal inter- 
pretations, such contrary evidence, brought forward by careful and 
experienced workers, must be borne in mind and positive acceptance 
of the colloidal explanations, however attractive, must be withheld 
until much further investigation has been done. 
Another important and interesting phase of the study of precipi- 
tins is that associated with the occasional presence in the same serum 
of remnants of antigen and of precipitins which, though present 
side by side, do not unite to form precipitates. This condition 
is frequently seen in such sera as those produced by Fornet and 
Midler44 for rapid precipitin production for forensic work, a 
method in which the foreign serum is injected into rabbits in large 
amounts (2 to 10 c. c.), on consecutive days, and the animals are 
bled 6 to 8 days after the last injection. That such sera contain 
both antigen and antibody is shown by the fact that, though clear 
when taken, they will show precipitation not only when mixed with 
dilutions of the antigen, but also when added to homologous precipi- 
tating sera.45 
This phenomenon has been noticed by Linossier and Lemoine,46 
Eisenberg,47 Ascoli,48 and others, and has been extensively studied 
43 Although normal sera may gradually precipitate on standing, this takes 
place much more rapidly in precipitin-sera. The spontaneous precipitation 
of normal sera as well as of those under consideration is analogous to what 
Bechold and others call the “ageing” (altern) of colloidal suspensions, which, 
though originally stable, will eventually settle out, even in the presence of 
protective colloids. 
44Fornet and Muller. Zeitschr. f. biol. Technik u. Methodik, Yol. 1, 
1908. 
45 For instance, a rabbit was injected on three consecutive days with 
sheep serum. It was bled on the fifth day after the last injection. The 
serum was clear when taken, but a precipitate was formed when it was 
added to sheep serum and also when it was added to serum from another 
rabbit similarly treated and containing sheep serum precipitin. 
46 Linossier and Lemoine. C. R. de la Soc. de Biol., 54, 1902. 
47 Eisenberg. Centralbl. f. Bakt., 34, 1903. 
48 Ascoli. Munch. med. Woch., Yol. 49, No. 34, 1902. 292 
INFECTION AND RESISTANCE 
by von Dungern.40 Gay and Husk 50 have recently observed it in 
connection with the rapid method of precipitin production of Fornet 
and Muller, and have noted that such sera, although containing both 
antigen and precipitin, do not possess complement-fixing properties. 
According to Uhlenhuth and Weidanz,51 the antigen may persist in 
the sera of protein-immunized animals, in demonstrable amounts, 
as long as fifteen days after the last injection, and it is constantly 
present during this period, but in progressively diminishing amounts, 
eventually disappearing. 
We are thus confronted by the apparently paradoxical phenom- 
enon of the presence in these sera, side by side, of an antigen and its 
homologous precipitin, incapable of reacting with each other, al- 
though each of them readily reacts with precipitin or antigen, re- 
spectively, when these are added from another source. 
Many attempts have been made to account for this. A number 
of observers, notably Eisenberg, have concluded from extensive anal- 
yses of quantitative relationships, both of agglutinin and precipitin 
reactions, that these take place according to the laws of mass action. 
In consequence, in addition to the combined precipitin-antigen com- 
plex present in all mixtures of the two, there should also be present 
free dissociated fractions of each, in amounts dependent upon rela- 
tive concentrations. This might explain conditions such as those 
described above. 
Von Dungern, whose paper forms one of the most extensive studies 
of the phenomenon with which we are concerned, does not believe 
that precipitin reactions can follow the laws of mass action, and 
explains the simultaneous presence of precipitin and antigen in the 
same serum by assuming a multiplicity of precipitins. He believes 
that every proteid antigen contains a number of related partial an- 
tigens which give rise in the immunized animal each to a partial 
precipitin. In sera in which both antigen and precipitin are found 
side by side and free, he believes that the antigen is of a nature that 
has no affinity for the particular partial precipitin present with it. 
He says: “Auch hier handelt es sich nicht um zwei reaktionsfahige 
Korper, deren Verbindung aus irgend welchen Griinden unterbleibt, 
sondern um Substanzen, welche keine Affinitat zu einander besitzen. 
Die betreffenden Kaninchen haben zu dieser Zeit noch nicht alle 
mdglichen Teilprazipitine gebildet, sondern nur einzelne derselben. 
Diese zunachst produzierten, nur auf bestimmte Gruppen der prazi- 
pitablen Eiweisskorper passenden Partialprazipitine sind es, welche 
nach der Absattigung aller zur Verfiigung stehenden zugehorigen 
Gruppen der priizipitablen Substanz in Serum nachweisbar werden. 
49 Yon Dungern. Centralbl. f. Bakt., 34, 1903, 
50 Gay and Rusk. “Univ. of Cal. Public, in Pathology,” Vol. 2, 1912. 
51 Uhlenhuth and Weidanz. “Praktische Anleitung zur Ansfiibrung, etc.,” 
Jena, 1909. THE PHENOMENON OF PRECIPITATION 
293 
Daneben bleibt aber ein anderer Teil der prazipitablen Substanz, 
der keine Affinitat zu dem gebildeten Prazipitin bestizt, bestehen, 
solange bis ein anderes Partialprazipitin von den Kaninclienzellen 
geliefert wird, welches sich mit Gruppen der in Losung geliebenen 
Eiweisskorper vereinigen kann.” 
Zinsser and Young52 have also studied these phenomena and 
have attempted to explain them on the basis of protective colloidal 
action. In considering the theories that have been advanced to ex- 
plain these occurrences, the conception of mass action as accounting 
for the simultaneous presence of the two reacting bodies in the same 
serum seemed entirely incompatible with our own observations and 
with those of Gay and Rusk, that these sera do not of themselves 
fix alexin. Were the conception of the manner of union of these 
two reagents, according to the laws of mass action, representative 
of the true state of affairs, it would be necessary to assume the pres- 
ence, in such sera, not only of the two reacting bodies free and disso- 
ciated, but also of a definite quantity of the united complex of the 
two, a state of equilibrium being established. If this were the case 
the sera should, in agreement with all experience on the phenomenon 
of complement fixation, exert definite complement-binding power. 
Moreover, it has not been experimentally shown that colloidal sub- 
stances react in accordance with the laws of mass action as observed 
for simpler chemical substances. 
As regards the opinion of von Dungern, this seemed incom- 
patible with another occurrence, observed by many writers, namely, 
that such sera, although clear at first, eventually, after prolonged 
standing, do actually precipitate spontaneously; that is, the union 
of the precipitin and the precipitinogen does actually take place, but 
goes on with extreme slowness. 
How a notable and strange feature of this phenomenon is the 
fact that two such sera, both containing antigen and precipitin, but 
neither of them precipitating by itself, will precipitate each other 
when mixed. For this reason Uhlenhuth has advised against the 
use of mixtures of precipitin sera for forensic tests. For it is not 
unusual that precipitin sera, even when produced by the slow method, 
may contain traces of antigen, and this may lead to precipitate 
formation if such a serum is mixed with another homologous pre- 
cipitin and thereby simulate a positive forensic test. 
In seeking analogy for this serum phenomenon with the various 
colloidal suspensions, the problem consisted in protecting two 
mutually precipitating colloids by a third, and this in such propor- 
tions that the mixing of two such protected suspensions, each con- 
taining all three of the elements, would be followed by precipitation. 
This was obtained by the use of gum arabic, gelatin, and arsenic tri- 
sulphid. Thin emulsions of gelatin will precipitate arsenic tri- 
52 Zinsser and Young. Jour. Exp. Med., 1913, Yol. 17. 294 
INFECTION AND RESISTANCE 
sulphid suspensions. Small amounts of gum arabic will act as a 
protective agent, preventing the precipitations. 
The amount of the protecting substance necessary to prevent 
precipitation in any one mixture varies apparently with every 
change in the relative proportions of the two. Thus a considerable 
number of mixtures of the three can be made which will remain 
stable for days, the actual and relative quantities of the three 
ingredients differing in each of the mixtures. When two such mix- 
tures are poured together, in many cases precipitation will result, 
varying in speed and completeness, according to the particular quan- 
titative relationship arrived at in the mixture. 
An example of such an experiment follows: 
Two solutions of colloidal arsenic sulphid were prepared, one containing 
1 gm. per liter, the other containing 5 gra. per liter. With Kahlbaum’s “Gold- 
ruck” gelatin a solution containing 1 gm. per liter was prepared. A solution 
of gum arabic was prepared which contained 10 gm. per liter, this being made 
stronger than the gelatin solution to avoid too great dilution in the final mix- 
tures. The gelatin solution was prepared twenty-four hours before being used, 
as freshly prepared gelatin has but slight precipitating power for arsenic 
sulphid, this power appearing to increase greatly with the ageing of the 
solution. 
For the purpose of demonstrating this analogy two protected 
solutions were prepared as follows: 
Solution 1.—This consisted of 2 drops of gum arabic, 2 c. c. of gelatin, 
and 5 c. c. of the weaker arsenic solution. 
Solution 2.—This consisted of 10 drops of gum arabic, 1 c. c. of gelatin, 
and about 4 c. c. of the stronger arsenic solution. 
In each case the arsenic sulphid was added until there were signs 
of increasing opalescence or turbidity, this being done in order that 
the two solutions should each be as little overprotected as possible. 
Portions of the two solutions were then mixed in equal propor- 
tions. In the course of a few minutes the mixture was noticeably 
more turbid than either of the original solutions. This turbidity 
continued to increase quite rapidly, and on the following morning 
after about sixteen hours of standing, the mixture was found to be 
completely flocculated out, while the original protected mixtures re- 
mained unprecipitated and showed about the same degree of opales- 
cence as on the preceding night. The same condition of affairs was 
found to have persisted after five days. On the fifth day the less 
concentrated of the clear protected suspension began to settle out, 
and was completely precipitated within twenty-four hours. The 
other remained clear for four days more, but on the ninth day it 
began to precipitate slightly, the precipitation remaining incom- 
plete. 
In these cases it appears, therefore, that a complete analogy to THE PHENOMENON OF PRECIPITATION 
295 
the observed conditions of the serum reactions has been found, and 
that all data observed in connection with sera in which antigen and 
precipitin are found side by side without reacting can be most simply 
explained on the conception of protective colloid action. Moreover, 
the chemical nature of the substances involved seems to add weight 
to this point of view. 
These relations have been gone into here at some length, since 
they seem to us to possess considerable theoretical and practical sig- 
nificance. Tor it may be that the presence of a protective colloid 
may, by inhibiting the union of antigen and precipitin within the 
body, protect the animal from intoxication during the early stages 
of immunization when antigen and antibody are present simulta- 
neously for longer or shorter periods. Were union between the two 
possible at such times in the circulation, an assumption necessitated 
both by the hypotheses of mass action and of multiplicity of precip- 
itins, there would probably be an absorption of complement by these 
complexes, with, as shown by Friedberger, a consequent formation 
powerful toxic products. (See chapter on Anaphylaxis.) It is not 
impossible by any means, therefore, that the injection of antigen 
in an animal in which such a balance has been established may 
lead to a sudden elimination of the colloidal protective action, union 
of the antigen and antibody, and, by the mechanism just outlined, 
anaphylactic shock. 
The fact, moreover, that mere heating will change the precipi- 
tating action, which certain sera have on inorganic colloids, to a 
protective one seems to show that this latter function may justly 
be associated with delicate physical or chemical alterations of animal 
sera. 
Furthermore, this point of view is strengthened by the fact that 
the mutual precipitation of sera, such as those described, takes place 
slowly, as does the mutual precipitation of two protected colloidal 
mixtures, in contradistinction to the more rapid precipitation which 
takes place when any of these sera is added to an antigen dilution, 
where the element of protection may be assumed to be practically 
eliminated by more extensively changed quantitative relations. 
This point of view has lately been disputed by Weil, who has gone 
back to the older view of von Dungern, largely on the basis of pre- 
cipitation experiments he carried out with crystallized egg albumen. 
Weil claims that if a pure protein, like crystallized egg albumen, is 
used for immunization, antigen and antibody are never found simul- 
taneously in the blood stream. If this were true it would indeed con- 
stitute a very important contradiction of our point of view. In con- 
sequence Bayne-Jones carried out similar experiments in our labora- 
tory, and found that even with the purest obtainable recrystallized 
egg albumen both antigen and antibody are at times demonstrable 
in the blood. He showed this both hy precipitation and by cornple- 296 
INFECTION AND RESISTANCE 
ment fixation. We do not consider the question closed, however, be- 
cause it is indeed true that when working with a purified protein it 
is more difficult to demonstrate the two substances in any quantity, 
than it is when the crude serum antigen is used. This may be, of 
course, due to the fact that pure egg albumen may be more rapidly 
assimilated and remains in the circulation for a less extensive period 
than does the crude antigen. However, further experiments in this 
direction will unquestionably clear the matter up because it is a 
simple question of fact amenable to experiment. CHAPTER XI 
ISO-ANTIBODIES 
Early in their researches, Ehrlich and Morgenroth were led to 
speculate upon the possibility of the formation of lytic antibodies 
within the animal against its own tissue cells. It would be of 
the greatest importance to pathology, they pointed out, if it could 
be shown that an animal could produce hemolysins, for instance, 
against its own blood cells. Thus, if an extensive internal hemor- 
rhage occurred from trauma or other cause, in the course of which 
considerable quantities of erythrocytes are subjected to disintegra- 
tion and absorption, it is at least conceivable that specific “auto- 
hemolysins” might appear which would lead to a chronic destruction 
of the red cells, with consequent anemia. This form of reasoning, 
as we shall see, has been extensively applied in the case of the cyto- 
toxins for the explanation of a variety of pathological conditions. 
Ehrlich and Morgenroth approached the question experimentally in 
their further work on the hemolysins in goat blood. They found that 
it was comparatively easy to produce hemolysins in one goat by 
treatment with the erythrocytes of other goats, isohemolysins, as 
they called them. 
Although, however, the blood serum of such an immunized goat 
was strongly hemolytic, not only for the blood cells of the goats 
whose blood had been injected, but also for the erythrocytes of cer- 
tain other goats (though not, as we shall see, for goats in general), 
it was never in any case active against this goat’s own cells. More- 
over, while the other sensitive erythrocytes could absorb the hemo- 
lytic antibody out of the inactivated serum, the insensitive corpuscles 
of the goat himself seemed to possess no affinity whatever for the 
lysin of his own serum; mixed with the serum they failed to absorb 
out the hemolysin. This was in no sense, therefore, an autolysin. 
These experiments show a remarkable individual variation be- 
tween the similar tissues of animals of the same species, since Ehr- 
lich and Morgenroth were indeed able to show that the insensibility 
of the goat’s own corpuscles depended upon a complete absence of 
receptors for the isolysin. For, to explain the lack of “autolytic” 
action of such a serum, two possibilities could be assumed. One, as 
above, that the corpuscles of the goat possessed no receptors by means 
of which the isolysin could be “anchored” or, second, that, although 
297 298 
INFECTION AND RESISTANCE 
such receptors were present, they were already satisfied, or saturated 
with the lysin in the blood stream. In the latter case it would be 
hard to understand why hemolysis had not taken place. 
In order to completely disprove the latter possibility, Ehrlich 
and Morgenroth did not allow the matter to rest upon conjecture, but 
resorted to an ingenious method of experimentation which yielded a 
further important result, namely, the discovery that the injection of 
antibodies into animals may give rise to “anti-antibodies.” They 
injected inactivated hemolytic serum into goats whose corpuscles 
were sensitive to its action, and found that an “anti-isolysin” was 
formed, which, mixed with hemolysin and sensitive corpuscles, pre- 
vented hemolysis. Injection of such an isolysin into the goat from 
which it had been obtained, however, did not yield anti-isolysin, and 
it was therefore reasonable to suppose that its tissue cells possessed 
no suitable receptors. This failure of the production of antibodies 
by an animal against its own tissue cell has been spoken of by Ehr- 
lich as “Horror Autotoxicus.” 
These rather involved experimental data will be shown to have a 
more than purely academic value when we come to speak of the 
problems of cytotoxin formation, and although they seem to show 
that auto-antibodies do not form as a rule, exceptions to this gener- 
alization have been observed. The most notable of these is the ob- 
servation of Landsteiner and Donath 1 made in connection with the 
condition of paroxysmal hemoglobinuria. It was found that in such 
cases, in which hemoglobinuria follows exposure to cold, the blood 
serum of the patient contains an “autohemolysin.” If the blood of 
such a case is taken into oxalate or citrate solution, and allowed to 
stand at ordinary or incubator temperature, nothing occurs. If, 
however, such blood is cooled to 0° to 10° C. and then warmed grad- 
ually to the temperature of the body, rapid hemolysis occurs. In 
this case the “amboceptor” of the serum is apparently fixed or an- 
chored by the blood cells only at a low temperature, the complement 
becoming active as the blood is warmed. Although Landsteiner’s 
observations are undoubtedly accurate, it is likely that this mechan- 
ism does not explain all such cases. The writer has had occasion to 
examine carefully a number of clinically diagnosed cases of this 
sort with a partially successful “Landsteiner” phenomenon in one 
of them only. Other observers have, however, confirmed Land- 
steiner’s observation in well-established cases of the condition. 
Taking its departure from these early observations, the problem 
of iso-antibodies, in general, aroused a considerable amount of at- 
tention among serologists and pathologists because it revealed an 
unexpectedly wide range of possible antigenic differences in the 
proteins of similar cells in animals even of the same species. The 
practical bearing on pathology was obvious. 
1 Landsteiner and Donath. Munch, med. Woch., 1904, p. 1590. ISO-ANTIBODIES 
299 
The peculiar facts unearthed by Ehrlich and Morgenroth 2 indi- 
cated specific differences between red blood cells of individuals in 
the same species (goats), which could only be recognized by the 
development of immune isolysins. Work on other species of ani- 
mals has indicated that this fact has a broad significance and that 
similar differences between individuals of the same species occur in 
many, if not all, species of animals. Isolysins similar in principle 
to those of Ehrlich and Morgenroth were produced by Ascoli 3 in 
rabbits; by Todd and White,4 in oxen; by Ottenberg, Kaliski, and 
Friedmann 5 in dogs; by Ottenberg and Thalhimer 6 in cats, and by 
Hada and Rosenthal 7 in chickens. In all these instances the iso- 
lysins developed showed the same peculiarities, namely, that they 
attacked the cells of certain individuals and left the cells of other 
individuals of the same species unharmed. Recent work on the 
isolysins occurring naturally in the human blood has thrown con- 
siderable light on the nature of immune isolysins. 
The occurrence of iso-antibodies in human blood was first noted 
by Maragliano 8 in 1892, and a large amount of work was done be- 
fore it became perfectly clear that the occurrence of the iso-hemo- 
lysins noticed by him was not a characteristic of disease. Analogous 
to isolysins, iso-agglutinins against blood were first described in 1901, 
independently, by Landsteiner9 and by Schattock. Landsteiner 
studied 22 individuals whom he could divide into three groups with 
respect to iso-agglutinins. The three groups were designated by 
A, B, and C, as follows: In Group A, the sera agglutinated the 
corpuscles of Group B, but not of Group C; the corpuscles of Group 
A, however, were agglutinated by the sera of both B and C; Group 
B, the sera agglutinated the corpuscles of Group A, but not those of 
C; the corpuscles of Group B were agglutinated by the sera of Groups 
A and C. In Group C, the sera agglutinated the corpuscles of both 
the other groups, but its corpuscles were not agglutinated by the 
sera of either of the other two. This preliminary classification was 
slightly changed in 1902 by Decastello and Stiirli who added a fourth 
group, the corpuscles of which were agglutinated by both the other 
groups, but the serum of which agglutinated none of the others. A 
2 Ehrlich and Morgenroth. “Uber Hamolysine,” Berl. klin. Woch., 1900, 
No. 21. 
3 Ascoli. Munch, med. Woch., 1901. 
4 Todd and White. Nature, June 23, 1910. 
5 Ottenberg, Kaliski, and Friedmann. Jour. Med. Res., Yol. 28, 1913. 
6 Unpublished personal communication. 
7 Hada and Rosenthal. Zeitschr. f. Imm., 1913, 16, p. 524. 
8 Maragliano. IX Kongr. f. Innere Med., 1892. 
9 Landsteiner. Wien. klin. Woch., 14, 1901, 1132. Landsteiner and 
Richter. Zeits. f. Med., 3, 1902. See also Ascoli, Munch, med. Woch., 1901. 300 
INFECTION AND RESISTANCE 
concise description of these groups was finally given by Jansky 10 
in 1907, whose classification of the four definite groups is shown in 
the following diagram. 
J~am,shy I - 40 % more or le.53 
" H - 35 % " " 
« nr = 15 % " " 
« IE = 5 % " " 
Tanoky's i" and Moaa' “4-" ia spoken of some- 
times as the '"universal" donor—for reasons explained 
in the text. 
In 1910, Moss 11 who studied particularly the parallelism of iso- 
lysins and iso-agglutinins in human sera, at that time unfamiliar 
with Jansky’s classification, reported observations which are essen- 
tially identical with Jansky’s, but, unfortunately, for uniformity of 
10 Jansky. Originally published in the Klincky Sbornik, 2, 1907. In- 
formation obtained from abstract in Report of Committee of Soc. of Amer. 
Pathol. Article has been abstracted in the “Jahresbericht iiber Leistungen 
und Fortschritte auf dem Gebiete der Neurologie und Psyehiatrie,” 11, 1907, 
1092. 
11 Moss. Johns Hopkins Hosp. Med. Bull., 11, 1910, 63. ISO-ANTIBODIES 
301 
classification happened to designate his groups in a manner directly 
opposite to that of Jansky, in that Jansky’s Group I becomes Moss’s 
Group IV, and vice versa. Thus, the two classifications are alike, 
but Groups I and IV are reversed. We have indicated this difference 
in the table, and wish to call particular attention to the fact that 
all clinicians who utilize these facts for transfusion purposes must be 
perfectly clear in their minds concerning the classification which 
they are using. Failure to do this may lead to accident, evident from 
examining this grouping that the phenomena can be explained (as 
Landsteiner has suggested) if it is assumed that there are two agglu- 
tinins (a and (3) and two corresponding agglutinogens present in 
the red cells (A and B). The blood of the first group possesses both 
agglutinins, but no agglutinogens, the blood of the second group pos- 
sesses agglutinin a, agglutinogen B, the blood of the third group 
possesses agglutinin (3, agglutinogen A, the blood of the fourth group 
possesses no agglutinin but both agglutinogens. 
The correctness of this conception can be proved by experiments 
in which various agglutinins are absorbed out of serum by the cells 
of the different groups, a line of study which has been followed par- 
ticularly by Hooker and Anderson,12 and recently, again, by 
Gichner.13 
These agglutinins are present in weak dilution only, being gen- 
erally active in dilutions only of 1-15 to 1-30. They are separately 
absorbed from the serum by the suitable red cells (Hektoen).14 Ot- 
tenberg noticed that they were inherited, and this was also shown 
in 1908 and in 1910 by von Dungern and Hirschfeld,15 who further 
found that this inheritance followed the Mendelian law strictly. The 
agglutinogens are the unit characters. The agglutinogens apparently 
are present at an earlier embryonic stage than the agglutinins. The 
agglutinogen, or agglutinahility of the red cells, thus, is usually pres- 
ent at birth, while the specific agglutinative power of the blood serum, 
or agglutinin, may be absent at birth, and may not appear until 
several months later. 
The laws of inheritance in regard to the transmission of iso- 
antibodies in human beings have been studied extensively since von 
Dungern and Hirschfeld’s first observations. Ottenberg,16 particu- 
larly, has followed a considerable series of families, and recently 
made the suggestion that in a very limited way, medico-legal use 
might be made of the reaction. This would, of course, as he admits, 
be necessarily a limited use for reasons that will become obvious when 
12 Hooker and Anderson. Jour. Immunol., 6, 1921, 419. 
13 Gichner. Jour. A. M. A., 79, 1922, 2143. 
14 Hektoen. Jour. Inf. Dis., 1907, p. 297. 
15 Von Dungern and Hirschfeld. Zeitschr. f. Imm.} 4, 1910, p. 53 and 
5, P- 284. 
16 Ottenberg. Jour. A. M. A., 77, 1921, 682. 302 
INFECTION AND RESISTANCE 
we discuss its possibilities. Buchanan 17 discussing this suggestion, 
disputed Ottenberg’s assertion on genetic grounds. But since that 
time the question has been submitted to Thomas H. Morgan 18 whose 
discussion in the recent lecture before the New York Pathological 
Society is so interesting an example of the application of genetic 
reasoning to problems of immunology, that we think it worth quoting 
at length from his paper. 
“Since Ottenberg’s statement has recently been disputed by 
Buchanan, from an entirely wrong interpretation of Mendel’s prin- 
ciples, I should like to point out that on the Mendelian assumption 
of two pairs of factors, all the known results are fully accounted for. 
If we represent one pair of genes by A and a and the other pair by B 
Mating of blood, group AaBb to same AaBb 
Fig. 8. Representing the kinds of individuals expected when an indi- 
vidual of the blood group type AaBb marries an individual of the same blood 
type, namely AaBb. Sixteen kinds of individuals are possible in the ratio 
of 9: 3: 3:1. These belong to four blood types, namely, class IV that con- 
tains at least one A and one B; class II that contains at least one A but 
no B; class III that contains at least one B but no A; and class I that 
contains neither A nor B. 
and b, and if we represent an individual with the genetic consti- 
tution AaBb mating with another individual of like constitution 
(AaBb), then each will contain four kinds of germ cells, viz., AB, 
Ab, Ba, and ab. Thus sixteen possible combinations may be formed 
if any sperm may fertilize any egg. 
17 Buchanan. Jour. A. M. A., 79, 1922, 180. 
18 Morgan, T. H. Goldsmith Lecture, Rep. from Transac. N. Y. Pathol. 
Soc., 1922. ISO-ANTIBODIES 
303 
These sixteen individuals fall into four groups according to 
whether they have both A and B, or only A, or only B, or neither 
A nor B (i.e., ab) in the proportion of 9AB: 3A: 3B: lab. These 
four genetic classes correspond to the four recognized blood types 
IV, II, III, I, as indicated in the diagram. How these sixteen 
kinds of individuals are found in all populations, so far studied, 
although in somewhat different proportions in different “races.” 
It is very simple to tell what the kinds of genetic offspring 
will be where any one of these sixteen individuals marries any 
other one. These possibilities are summarized in the following 
statement taken from Ottenberg: 
Unions of I and I give I 
I nl 
II Ilj 
I, II 
I III] 
III III] 
I, III 
Unions of II and III give I, II, III, IV. 
iy i i, n, III, IY. 
IV II I, II, III, IV. 
IV III I, II, III, IV. 
IV IV I, II, III, IV. 
From a knowledge of the blood group to which the child be- 
longs it is possible to predict to what groups its parents may have 
belonged, and in certain cases it is possible to state that an indi- 
vidual of a certain group could not have been the parent of a par- 
ticular child. This gives these considerations a practical medico- 
legal value which, it is true, is limited, but definite as far as it goes. 
From the practical medical point of view, the importance of this 
subject is based on the increasing use of transfusion of human blood 
from one individual to another, especially in surgical practice. In 
selecting donors, it is, of course, desirable to select an individual of 
the same group as the recipient. Infusion from some of the other 
groups may bring about both agglutination, and as Hopkins first 
showed, phagocytosis of red blood cells in the circulation, with con- 
sequent injury to the recipient. In judging of the possibilities of 
harmless infusion, it is not necessary to memorize the combinations, 
hut merely to remember that the chief danger occurs when the pa- 
tient’s serum is agglutinative for the donor’s red cells, and a number 
of such transfusions have had fatal results. When the opposite 
occurs, that is, when the patient’s corpuscles are agglutinated by the 
serum of the donor, the danger is far less, owing to the dilution of 304 
INFECTION AND RESISTANCE 
the injected serum in the circulation of the recipient. Thus, in 
Jansky’s classification where the serum of group I agglutinates the 
corpuscles of all the other groups, but the corpuscles of which are 
agglutinated by none of the other groups, group I could be regarded 
as a universal donor and used when a donor of the same type as the 
patient is not available. Conversely, this would apply to Moss’s 
group IV. 
For practical purposes, it is seen that tests can easily be carried 
out if one merely possesses stock sera of group II and III. The 
5 era 
Ce\U 
Jansky I 
Moss 44- 
Jansky H 
Moss 2. 
Jansky III 
Moss 3 
Jansky EZL 
Moss 1 
following diagram illustrates this. Formerly these reactions 
carried out with washed blood corpuscles in test tubes. This method 
has been greatly simplified of recent years by carrying out the entire 
test in drops on slides, a drop of a salt solution dilution of the blood 
of one individual being mixed with a drop of II or III serum, and 
observed for agglutination, either under the low power microscope, 
or with the naked eye. The reactions are prompt, sharply defined, 
and offer no difficulties, but precautions must be taken to prevent 
the drying up of the drops so that the increasing concentration of the 
salt does not bring about false agglutination. ISO-ANTIBODIES 
305 
Because of the confusion that the Jansky and Moss classification 
brought to the nomenclature;, and the possible dangers accruing 
therefrom, a joint committee of the American Immunological and 
Bacteriological Societies took up the matter and after some study on 
the basis of priority, recommended a general adoption in the United 
States of the Jansky classification.19 
The importance of iso-antibodies in such procedures as tissue 
transplantation, skin grafting, etc., has been variously studied with 
results which one might have expected, namely, that heterologous 
types did not lend themselves easily to such procedures. In general, 
this confirms our own observations on the toxicity of normal serum 
of one species to the red cells of another, in that it shows that aj 
serum which has hemo-agglutinating or hemolytic action, also has a 
certain amount of action on other tissues of the body, being a gen- 
eral expression of incompatibility, not only of one type of cell. 
Of the greatest interest in connection with iso-agglutination of 
red blood cells are studies recently made by L. and H. Hirschfeld.20 
Hirschfeld, working in Serbia, carried out thousands of iso-agglutina- 
tion tests on different races. He found the agglutinable substances, 
A and B, present in all races, but a great preponderance of A in 
Europeans, and of B in Asiatics and Africans. Arabians, Turks, 
Jews and Russians were intermediate. The anthropological pos- 
sibility of such an observation in pointing out the different racial 
intermixtures and the possibility of a double origin in Europe, is 
obvious. 
Similar iso-agglutinins have been observed in the blood of lower 
animals, in horses (Klein,21 1902) ; rabbits (Boycott and Douglas,22 
1910) ; cats (Ingebrigtsen) ; dogs, rats, and steers (Ottenberg).23 
The iso-agglutinins have been developed in dogs (von Dungern and 
Hirschfeld).24 In most of the lower animals they have occurred 
with peculiar irregularity, indicating probably the presence of, not 
two, but of a larger number of agglutinins and agglutinogens. In 
steers, however, they fall into simple groups, indicating the presence 
of only one agglutinin and agglutinogen. In many animals the 
agglutinins are entirely latent, and the biochemical differences rep- 
resented by the agglutinogens are present in the red cells, and the 
correct agglutinin is developed by the animal only when it is im- 
munized with blood whose cells contain agglutinogen not present in 
the animal’s own blood cells. 
19 Report of the Committees of the Amer. Assoc. Immunologists and Bac- 
teriologists. Jour. A. M. A., 72, 1921, 130. 
20 Hirschfeld, L. and H. Lancet, October, 1918, Yol. 2, p. 675, abstracted 
in the Jour. A. M. A., 1919, 1641. 
21 Klein. Wien. klin. Woch., 1902, p. 413. 
22 Boycott and Douglas. Jour, of Path, and Bad., Jan., 1910. 
23 Epstein and Ottenberg. Tr. N. Y. Path. Soc., 1908. 
24 Yon Dungern and Hirschfeld. Zeitschr. f. Imm,, 1909, 1910, p. 53L CHAPTER XII 
A FURTHER CONSIDERATION OF THE NATURE OF 
ANTIBODIES 
In the immediately preceding chapters we have considered the 
course of the occurrences which follow upon contact of the animal 
body with various infectious agents and their products. We have 
seen that one of the expressions of reaction in such cases is the 
appearance of the so-called specific antibodies which have the prop- 
erty of reacting with the particular bacteria or bacterial products 
which have incited them, in various ways, both within the animal 
body and in the test tube, and we have attempted to analyze such 
reactions. We hope, also, that wre have made it plain that the 
phenomena with which we are dealing are by no means limited to the 
reactions of the body to bacteria, but represent general biological 
laws which govern the response of the living body to a large class 
of substances spoken of as antigens. 
In the case of the antibodies formed in response to antigens of 
bacterial origin, the alterations in the physiology of the body of 
which antibody production is an expression, leads, under certain cir- 
cumstances, to an ability of the body to defend itself against injury 
by bacterial poisons, and to remove the invading bacteria. In this 
process, circulating antibodies may, as in the case of the antitoxins, 
agglutinins and opsonins, be directly responsible for an important 
part of the protective process. Again, as we shall see in the chapters 
dealing with hypersusceptibility, the production of antibodies and 
their presence both in the cells which form them, or to winch they 
become secondarily attached, and in the circulation, may lead to con- 
ditions of hypersusceptibility quite the reverse of protection in its 
practical consequences. The criteria which govern this are complex, 
and are considered in detail in the individual chapters dealing with 
the several antibodies and with anaphylaxis. Immunological analy- 
ses have, of necessity, dealt almost entirely with the consequences 
of antibody formation and the mechanism by which these conse- 
quences are governed, but have, up to the present time, left us almost 
entirely in the dark concerning the fundamental cellular physiology 
upon which antibody formation is based. 
The fundamental source of all changes induced by antigen injec- 
tions is, of course, the tissue cell. 
306 CONSIDERATION OF THE NATURE OF ANTIBODIES 
307 
It is reasonable to assume, as Ehrlich did, that the tissue cells 
of higher animals are normally prepared only for the metabolic 
processes concerned in the nutritional and excretory functions essen- 
tial to the maintenance of life, the liberation of energy and the per- 
formance of any special secretory functions peculiar to their respec- 
tive provinces of activity in the general body economy. Compared 
with the cells of the simpler forms of life, the normal functions of 
the mammalian tissue cells are considerably less wide in range. A 
protozoan cell, for instance, may take up boiled starch granules or 
bacteria and deal with them by complex processes, killing the bacteria 
in an acid vacuole, digesting them subsequently in an alkaline me- 
dium and extruding waste products. In only a few cells of the 
body, such as the phagocyting ones, either free or fixed, does such 
an atavistic ability to deal with substances unprepared by digestion 
survive. 
We are not equipped to enter upon matters of cell nutrition in 
an authoritative manner, and, indeed, this is not necessary in the 
present connection. The essential consideration is that in the 
bodies of the higher animals the substances which reach the cells for 
nutritional purposes through the blood or lymph stream, or perhaps 
through the intercellular fluids beyond the lymphatics, come into 
actual cellular contact only after elaborate preparation by preliminary 
digestive processes. Thus, proteins probably reach the ultimate; 
cells which they nourish only in the form of aminoacids; the fats 
are in the form of glycerol, fatty acids and perhaps soaps, and the 
carbohydrates in that of simple sugars. The cell is thus normally 
attuned only to dealing with substances in the chemical and physical 
conditions into which such preliminary digestion has put them. 
It is only such substances, apparently, which can come into repeated 
or continuous contact with cells without in some way altering the 
quality or degree of the cellular reactions aroused by them. It seems 
to be a fairly general biological principle that most materials which 
are not in this chemical and physical class of predigested nutritive 
matter give rise sooner or later to a specifically altered state of 
reaction capacity on the part of the cells. Thus, even when dealing 
with substances as relatively low in molecular structure as alcohol, 
some of the narcotics, quinine, morphine and other alkaloids, etc., 
the mere facts of habituation and occasional idiosyncracy indicate 
that such altered conditions have been produced. And, indeed, such 
drug tolerance or habituation cannot always be explained entirely by 
increased powers of excretion or destruction of the tolerated drug. 
Eor, although such increased eliminating properties are involved in 
the process, still it has been shown, at least in the cases of alcohol 
and morphine tolerance, that equal concentrations of the drugs in the 
blood stream of normal and of tolerant individuals may cause a 
deeper state of intoxication in the abstainers than in the habituated. 308 
INFECTION AND RESISTANCE 
(Rubsamen, Van Dongen, cited from Wells.) Thus, it is necessary 
to assume, as Wells puts it, “a certain refractoriness or cellular im- 
munity in addition.” So, too, as we proceed upward toward the 
more complex substances with which the bacteriologist and im- 
munologist more particularly deal, toxins and the proteins proper, 
contact of cells with these substances, incident to processes of in- 
fection or artificial parenteral administration, invariably induces an 
altered reaction capacity which is recognized either as some form of 
hypersusceptibility or of tolerance: 
When we consider the varied chemical and physical constitution 1 
of the many substances which may thus come into contact with 
the body cells, from the drugs at the bottom of the list to unchanged 
proteins, it need not astonish us if the manifestations accompanying 
the development of such susceptibilities or tolerances are subject to 
a wide range of variations. Common to all of them, whether it is 
the fluctuating tolerance to morphine or the quasi-permanent ac- 
quired immunity to plague, cholera or typhoid, is a fact that the 
tissue cell is basically changed in its reaction to the particular mate- 
rial concerned with a degree of specificity which seems to be more 
striking as chemical complexity of the foreign substance increases. 
Important differences, however, exist in the laws governing the ac- 
quisition of the altered cell reaction capacity and in the manifesta- 
tions by which they may be recognized. These differences have been 
analyzed by a number of recent reviewers, and it is not necessary 
for our purposes to enter into the detailed controversial points in- 
volved. 
Basic to such discussions has been the classification of all these 
phenomena into two main subdivisions: Those in which definite 
antigen-antibody reactions are involved and in which passive transfer 
to normal animals is, therefore, feasible; and those in which no 
antibodies seem to be concerned and passive transfer is consequently 
unsuccessful. The problem has received particularly careful analysis 
in relation to reactions of hypersusceptibility. Here the various 
manifestations of increased susceptibility to proteins in which anti- 
body formation is unquestionably and easily demonstrable constitute 
a sharply defined class, the so-called condition of anaphylaxis, about 
the basic mechanism of which there is very little essential difference 
of opinion among workers. Apart from this well-defined group, 
however, stand many other phenomena of hypersusceptibility, in 
which the antibody mechanism so clearly concerned in typical protein 
anaphylaxis is either uncertain or can be definitely excluded. 
The ability of the cell thus to express its response in the produc- 
tion of specific antibodies wdiich, in the course of the reaction, may 
become free in the blood stream is the chief point which separates 
1 See Zinsser. Newbold Lecture, Transact. Phil. College of Physicians. 
1922. CONSIDERATION OF THE NATURE OF ANTIBODIES 
309 
the mechanism of one type of reactions (the protein ones) from all 
the others. But even in reactions in which antibodies occur (im- 
munity to plague, typhoid, etc., or anaphylaxis in the very early or 
very late stages after active sensitization) the changed reactions 
can exist on a purely cellular basis without the intervention of de- 
tached antibodies. Indeed, the cellular properties by virtue of which 
the typhoid convalescent remains resistant for years after recovery 
are just as little understood today as is the mechanism of the 
tolerance to morphine. Regarded in this light, the formation of free 
antibodies, practically important as it is, may still be looked upon 
as an incident rather than as a fundamental difference, dependent 
upon the manner in which the chemical or physical properties of the 
inciting substance permits it to react with the cell. 
Ehrlich, whose lines of reasoning we have followed in the main, 
with others of his time conceived the cell as a giant molecule, a 
homogeneous chemical system with an enormous molecular weight, 
provided with a large number of “side chains;” these “side chains” 
were conceived as normally adapted for union with the various 
types of foodstuff that might be brought to the cell in the course 
of nutrition, but also capable of reacting with other substances 
reaching the cell under unusual conditions if they happened to 
have chemical affinities which fitted with one or another of the 
normally provided “side chains.” It is not necessary for us to 
follow this well-known, purely chemical reasoning in his explana- 
tion of antibody formation. Since his time the conception of the 
cell entity has changed considerably. The material is perhaps most 
thoroughly brought together in Bayliss’s book, and a clear discus- 
sion of some of the problems involved may be found in recent 
articles by Alsberg on chemical structure and physiological action. 
The cell can no longer be regarded as a large molecule of living 
matter, the life of which is maintained by a constant interchange 
of chemical unions and dissociations, but must be conceived, as 
Alsberg puts it, “as a heterogeneous series of phases separated from 
each other by semipermeable surface layers ” 2 resulting from the 
concentration at the surfaces, particularly, of lipoidal substances. 
But the semipermeable membranes which separate the internal 
subdivisions of the cell and delimit it from its environment prob- 
ably do not consist of lipoids alone, but are composed of complex 
“colloidal intermixtures of lipoids and proteins.” 3 It is clear that 
the reaction of an extraneous substance with a cell must depend 
not only upon its chemical but also upon its physical properties, 
according to which it may enter the cell and react with the sub- 
stances in its interior or may come into direct contact only with 
2 See also work of Clowes. 
8 Bayliss. “Principles of General Physiology,” Longmans, Green & Com- 
pany, 1920. 310 
INFECTION AND RESISTANCE 
the outer-surface phases of the protoplasm. It is also quite prob- 
able that the permeability of the cell membrane may change under 
various functional conditions, and that in the problems of hyper- 
susceptibility a part of the process may consist of an increased 
ability of the injurious substances to get into the cell. In fact, 
though we have not yet made the matter one of special study, we 
have noticed on a number of occasions that the isolated uterus of 
a sensitized guinea-pig was more irritable to nonspecific extraneous 
substances than were the uteri of normal guinea-pigs. This matter, 
however, needs more definite investigation. At any rate, the ques- 
tion of cell permeability to extraneous substances is quite as im- 
portant in guiding our reasoning about immunological problems as 
is the matter of chemical structure. 
Alsberg points out three general ways in which substances may 
affect cells: They may attack the surfaces of the cells either by 
precipitating, coagulating or dissolving parts of the constituents 
of the surface, or by coming into chemical or purely physical union 
with only these surface layers. Again, they may enter the cell and 
cause chemical and physical alterations within the protoplasm. 
Other substances, again, may react with the cell in an indirect way 
only by altering the concentrations within the cell. 
Thus, in reconstructing our ideas of the changes which may take 
place in cells under conditions of immunization or sensitization, 
we must depart from the purely chemical conception of Ehrlich 
to the extent of including in our considerations the physical prop- 
erties of the foreign substances concerned, especially as regards 
their ability to enter the cells or to react with the surfaces only. 
When we look upon the conditions governing altered cell reac- 
tions from this point of view, we are struck by the fact that anti- 
body formation, a property by virtue of which the inciting substance 
is classified as an “antigen,” is exclusively an attribute of materials 
which are practically nondiffusible, proteins or substances that have 
not been chemically separable from proteins hitherto. Considering 
this in connection with the fact that in test-tube reactions between 
such antigens and their antibodies the phenomena which are ob- 
served follow closely the analogy of colloidal reactions and are 
intimately dependent upon physical conditions of the environment, 
the thought suggests itself strongly that nondiffusibility, the property 
of reacting only with cell surfaces, and antibody formation are in 
some way connected. Although this thought is at the present time 
one that cannot be approached by direct experimentation, it is still 
of sufficient importance for the shaping of experimental thinking to 
be expressed. Thus, tolerance and hypersusceptibilities to drugs and 
other substances that pass easily through semipermeable membranes 
would be conceived as based on intracellular processes in which, per- 
haps, substances functionally analogous to antibodies may play a part, CONSIDERATION OF THE NATURE OF ANTIBODIES 
311 
but in which the ease with which the foreign substance can enter the 
cells renders unnecessary a mechanism for the production of circulat- 
ing antibodies. In the case of proteins, on the other hand, where 
diffusion does not normally occur, the process takes place on the 
surface of the cell, and here the reaction products, as we know by 
observation, jire eventually discharged into the circulation, and rep- 
resent those factors in the circulating blood by virtue of which the 
tolerance or the hypersusceptibility can be transferred to normal 
animals. 
The suggestion, thus, that antibody formation is in some way 
a consequence, not only of its chemical structure but also perhaps 
of the molecular size of the antigen, is one that must be borne in 
mind. Landsteiner came to this conclusion in a recent study of the 
chemistry of antigens, as we have from our own studies on ana- 
phylaxis. 
The Union of Antigen with Antibody.—In considering the vari- 
ous individual antibodies in the chapters on the toxin-antitoxin 
reaction, as well as in those dealing with agglutination and precipita- 
tion, we have seen that many different points of view have been 
expressed concerning the nature of the specific union. Ehrlich fol- 
lowed a purely chemical analogy in which atom groups of the tissue 
cell were assumed to have definite chemical affinity for atom groups 
of the injected antigen; that these cellular atom groups, when over- 
produced by the cell and discharged into the circulation, represented 
the antibody, and that, subsequently, therefore, the union of this anti- 
body with the antigen was again based upon chemical affinity between 
definite parts of the molecules of each. According to Ehrlich’s 
analysis, the reaction was supposed to take place by processes anal- 
ogous to those by which a strong acid and a strong base unite. 
Quantitative irregularities in the curve of gradual saturation of 
one by the other Ehrlich explained on the basis of alteration products, 
such as toxoids, agglutinoids, etc., incident to which affinities of one 
for the other were altered. This conception of Ehrlich was modified 
by the school of Arrhenius, Madsen and their coworkers. While 
they admitted the chemical nature of the union, they explained the 
irregularities in the quantitative relations between the two, as frac- 
tional amounts of one were added to a unit amount of the other, by 
assuming that the antigen-antibody combination was dissociable or re- 
versible, and followed laws of mass action, coming finally, in the case 
of each mixture, to a definite equilibrium. 
Bordet and his followers, on the other hand, for reasons which 
have been set forth at some length in previous sections, denied that 
antigen-antibody unions under any circumstances took place by 
chemical laws of equivalents, asserting that the process was one of 
adsorption, comparable to that which occurs when any anilin dye 
is brought into contact with, let us say, filter paper. Bordet held that 312 
INFECTION AND RESISTANCE 
the antigen-antibody union was in every way analogous to colloidal 
reactions in which the laws of adsorption, and, in this case, of 
course, specific adsorption, dominated the phenomenon. 
It is not unlikely that it will become necessary in the future to 
alter all of these views, particularly on the basis of the recent work 
of Loeb.4 Loeb’s investigations upon the so-called colloids have 
led to a change in our conception of the reactions which take place 
between proteins and other substances which immunologists cannot 
afford to neglect. The adsorption theories applied to immunological 
processes, such as particularly those of Bordet, were largely based 
on analogy, and assumed that when protein reacted with other sub- 
stances, colloidal or otherwise, the union was one in which the entire 
molecules of the reacting substances united. Loeb’s work seems to 
indicate definitely that proteins may be regarded as amphoteric elec- 
trolytes which can exist in three states, according to the hydrogen 
ion concentration. Colloidal particles, as is well known, carry a 
definite charge which is dependent upon the hydrogen ion concen- 
tration. The particles carry a negative charge on the alkalin side of 
the so-called iso-electric point, and a positive charge on the acid side. 
For every particular colloidal suspension, there is a definite hydrogen 
ion concentration at which there is no electric potential difference 
between the particles and the medium in which they are suspended, 
so that they will not move in the electric field and are uncharged in 
their relationship to the media. Loeb has shown that at the iso- 
electric point, protein is in its purest condition, exists in a prac- 
tically non-ionized condition, and is able to form neither metal 
proteinate nor protein acid salt. On either side of the iso-electric 
point, however, the protein combines like any other chemical com- 
pound, with acids, salts and perhaps other substances, in a character- 
istic way which is determined by the hydrogen ion concentration. 
Thus, in the case of gelatin, for instance, at which the Ph for the 
iso-electric point is 4.7, the hydrogen ion concentrations on the acid 
side of this combination can take place only with the anion of an 
electrolyte, forming, let us say, gelatin chloride. When the Ph is 
greater than 4.7, or on the alkalin side, it can unite with the cation 
only, forming sodium gelatinate, etc. Loeb conceives the protein 
molecule as presenting various atom groups for combination with 
materials with which it comes in contact, which, on the alkalin side 
of the iso-electric point, act something like fatty acids which are 
able to form salts with metals, etc., perhaps by combination with the 
COOII groups of the molecule. On the acid side of the iso-electric 
point the reverse is true, the proteins reacting like ammonia and 
being capable of uniting with only the anion of electrolytes, perhaps 
through their NH2 groupings. 
4 Loeb, J. “Proteins and the Theory of Colloidal Behavior,” McGraw- 
Hill, New York, 1922. CONSIDERATION OF THE NATURE OF ANTIBODIES 
313 
It is impossible for us in this space to go into the various proofs 
of such a conception of the reactions of proteins and the tremendous 
importance for the understanding of colloidal phenomena in general, 
which Loeb’s conception carries with it. It is plain that it must alter 
and simplify considerably our points of view. It is, of course, as Loeb 
admits, quite possible that this is not the whole story, and that the 
difference in the behavior of proteins on the opposite sides of the iso- 
electric point may be accompanied by further intra-molecular changes 
in the protein molecule. However, this, until we know more about 
it, must remain undecided. There can be no doubt about the fact, 
however, that many of the substances which in immunology we have 
classified as behaving by the laws of colloidal reactions, must be 
regarded as amphoteric electrolytes and great attention must be 
given to the control of hydrogen ion concentration and the influence 
of cells, etc., in the reactions which are carried out with them. 
When we are working with bacterial suspensions, or suspensions of 
blood cells and the various normal and immune sera which are the 
instruments of immunology, we are using materials all of which 
come under the classification of colloids. As a matter of fact, one 
of the workers in our laboratory discovered by chance in following up 
this line of thought, recently, that practically all the ordinary serum 
reactions done in laboratories, were done at a Ph of at or about 8. 
It will be impossible in the future to ignore these facts in the 
performance of serum reactions and, especially, in quantitative work. 
Moreover, the principles elucidated by Loeb have already in- 
directly led to what seems to us an important advance in our under- 
standing of the union of antigen and antibody in the hands of 
Coulter. 
Coulter 5 determined the movement of normal and sensitized red 
blood cells in an electric field and found, as others had before him 
with bacterial and other colloidal suspensions, that the movement 
of such cells in the field is the function of the hydrogen ion concen- 
tration. The iso-electric point for the cells he used was Ph 4.6. 
On the alkaline side they carried a negative charge and on the acid 
side a positive one. There is an old observation by Joos 6 that the 
salt which is necessary for the agglutination of bacteria did not act 
indirectly as supposed by some, but combined chemically with the 
bacteria. Joos using the minimum amount of salt necessary, found 
that it disappeared from the supernatant fluid; but until very re- 
cently this observation was neither confirmed nor further pursued. 
Coulter reinvestigated this in the light of Loeb’s demonstration of the 
chemical nature of the relations of protein with acids, bases and 
salts, at various hydrogen ion concentrations, and found that, like 
colloidal protein solutions, both normal and sensitized cells combined 
5 Coulter. Jour. Gen. Phys., 3, 1921, 309 and 513. 
6 Joos. Zeit. f. Hyg., 36, 1901, 422. 314 
INFECTION AND RESISTANCE 
chemically with inorganic ions. By adding decimolecular HaCI and 
HC1 solutions, he found that the cells combined with the chlorine ions 
on the acid side of the iso-electric point, taking up amounts much 
greater than any they would take up on the alkaline side. On adding 
NaOH to such mixtures, he found that chlorin is actually given off 
by the cells. Conversely, he found that when he suspended the cells 
In isotonic barium chloride solutions, the barium ion was absorbed 
on the alkalin side of the iso-electric point. This behavior corre- 
sponded to that found by Loeb for his gelatin and other protein 
suspensions which combined with the cations on the alkalin side of the 
iso-electric point, and the anion on the acid side. He found, further, 
that the optimum agglutination for normal cells is at Ph 4.75, a 
point at which the cells supposedly carried no charge in relation to 
the surrounding medium, and are, if we follow out Loeb’s views, in 
their most chemically pure form, and uncombined with inorganic 
ions. The optimum for the agglutination of sensitized cells was at 
Ph 5.3, a point probably related to the optimum for the flocculation 
of the immune bodies in the serum. 
Of still greater significance are the further experiments of 
Coulter on the equilibrium between sensitizer and red cells in relation 
to hydrogen ion concentration. Working in a salt-free medium, 
namely, isotonic saccharose solution (9.2 per cent.), he measured the 
proportionate amounts of sensitizer which combined with the cells 
at various hydrogen ion concentrations, and, conversely, the amount 
of sensitizer which dissociated from the saturated cells when they 
were brought into a similar range of hydrogen ion concentrations. 
Both measurements corresponded with considerable accuracy, showing 
that in the combination a definite equilibrium is maintained between 
the two for each Ph. He found that the maximum combination be- 
tween sensitizer and cells took place at or about Ph 5.3, the approxi- 
mate iso-electric point of the serum globulin in which the immune 
bodies are probably carried, according to Bona and Michaelis.7 At 
this point the combination between the two was almost 100 per cent. 
On both sides it diminished rapidly, on the alkalin side being re- 
duced to not much more than 5 per cent, at a Ph of about 10. On 
the acid side it also diminished, but could not he measured to the 
same degree because of the hemolytic action of the increasing 
acidity. 
The addition of sodium chloride greatly increased the proportion 
of sensitizer combining with the cells at all reactions except those 
near the iso-electric point where the union between the two seemed 
independent of the salt. 
Apparently, if we follow Coulter in analyzing these facts in ac- 
cordance with Loeb’s observations, wre can assume that the maximum 
union is at a point where almost all the immune body is present 
7 Rona and Michaelis. Biochem. Zeit., 28, 1910, 193. CONSIDERATION OF THE NATURE OF ANTIBODIES 
315 
as an undissociated molecule not in combination with inorganic ions. 
The dissociated ions of the sensitizer formed either by its acid or 
basic dissociations, do not appear to unite with the cells, so that, 
as we pass from the iso-electric point in both directions, there would 
be a gradually smaller and smaller amount of undissociated, that 
is, unionized sensitizer present, and the union -with cells would de- 
crease. The salt solution, again in conformity with Loeb’s observa- 
tion, seems to depress the ionization, and, therefore, render the com- 
bination relatively independent of the Ph. 
It need hardly be pointed out wrhy a conception of this kind would 
render it imperative to take actively into account the Ph at which 
every antigen-antibody union or titration is carried out. It enforces 
a completely changed interpretation of the significance of the Bordet 
salt experiment, and definitely introduces the factor of calculable 
dissociation and reversibility into antigen-antibody reactions, making 
it possible to think of the reaction at every given Ph, as one at which 
an equilibrium is established. 
The Dissociation of Antigen and Antibody.—In the preceding 
section we have gone into Coulter’s recent work at some length be- 
cause it seemed to us to follow out purposefully in immunological 
experiments, the newer conception of colloids and colloidal reac- 
tions which recent investigations have made necessary. The re- 
versibility of the antigen-antibody reaction, however, has been the 
subject of many previous investigations. Landsteiner,8 and Land- 
steiner and Jagic 9 in 1902 found that when red blood cells were 
agglutinated by abrin, and the agglutinated cells then rapidly washed 
in cold salt solution and finally emersed in a salt solution at 42° C., a 
certain amount of the abrin was split off from the combination and 
could be recovered. Experiments with normal agglutinins showed 
the same thing, the most successful dissociation of agglutinins from 
agglutinated typhoid bacilli being obtained at 55° C. At about the 
same time, Morgenroth 10 showed that when red blood cells were 
sensitized with hemolysin or amboceptor, and were then brought 
into contact with unsensitized cells, some of the hemolytic antibody 
was dissociated from the sensitized cells and became attached to the 
freshly added unsensitized ones. Similar observation of dissocia- 
tion of antigen and antibody after union was subsequently made by 
a considerable number of observers. Bail and Tsuda,11 Spaet,12 
Hahn and Trommsdorf,13 Von Liebermann and Fenyvessy 14 digested 
8 Landsteiner. Munch, med. Woch., 49, 1902, p. 1905. 
9 Landsteiner and Jagic. Munch, med. Woch., 50, 1903, 764. 
10 Morgenroth. Munch, med. Woch., 50, 1903, 61. 
11 Bail and Tsuda. Zeit. f. Immunitdts., 1, 1909, 546. 
12 Spaet. Zeit. f. Immunitdts., 7, 1910, 712. 
18 Hahn and Trommsdorf. Munch, med. Woch., 47, 1900, 413. 
14 Yon Liebermann and Fenyvessy. Centralbl. f. Baht., 47, 1908, 274. 316 
INFECTION AND RESISTANCE 
sensitized pig corpuscles in N/100 hydrochloric acid in salt solution, 
precipitated the extracts with alkali, purified the precipitate with 
ether, and found that the final solution contained antibodies with too 
little protein to be determinable by qualitative tests. This may be 
cited as one of the earliest attempts to produce protein-free antibodies. 
In 1918, Kosakai 15 washed sensitized sheep cells in saccharose 
solutions, and succeeded in recovering about five-sixths of the anti- 
body combined with the antigen. 
The most important recent contribution to this problem has been 
made by Huntoon.16 Huntoon worked with various materials, 
but chiefly with the pneumococci. He has developed a method since 
then, by which he can dissociate a large proportion of the antibody 
from sensitized pneumococci by washing the agglutinated pneu- 
mococci with 0.5 per cent, sodium carbonate solution in salt solution. 
Such solutions could be repeatedly filtered without loss of antibody, 
and, therefore, obtained in a sterile manner. His antibody solutions 
did not give any of the ordinary qualitative protein reactions, were 
not affected by trypsin, and were not precipitated in solutions con- 
taining little or no electrolytes. They were not soluble in ether, and 
not dissolved by dilute alkalis or acids. They were still destroyed 
by temperatures over 60°. 
With his so-called pure antibody solutions, Huntoon has been do- 
ing a considerable amount of clinical work which is not yet sufficiently 
analyzed for appraisal. 
The results that he and clinical workers using his antibody solu- 
tions have obtained, indicate to us that, in addition to the anti- 
bodies, a considerable amount of bacterial material, possibly in the 
form of the residue antigens studied by us, is present in his solutions. 
From a theoretical point of viewT, his researches are of considerable 
importance in that they form a beginning of the possibility of our 
being able to study antibodies free from admixtures of inactive serum 
constituents. 
Many of these investigations w7ere made before we had a definite 
idea of the influence of hydrogen ion concentration on dissociation. 
With this knowledge available, it ought to he possible in the future 
to remove almost any desirable part of the united antibody by prop- 
erly planned experiments. 
Investigations such as these justify a certain amount of hope, 
therefore, that we may eventually be able to define chemically what 
the antibody consists in, but for the present none of this work has 
led to sufficiently accurate and detailed investigation to make i£ 
possible for us to come to a definite conclusion. 
15 Kosakai. Jour. Immunol., 3, 1918, 109. 
16 Huntoon. Jour. Immunol., 6, 1921, 117 and 123 and 185. CONSIDERATION OF THE NATURE OF ANTIBODIES 
317 
On the Essential Identity of the Antibodies.17—Since Ehr- 
lich’s 18 first classical analysis of antibodies, it has been a generally 
accepted conception of immunity that agglutinins, precipitins, sensi- 
tizers, bacteriolysins, hemolysins, or the so-called amboceptors, opso- 
nins and the anaphylactic antibodies are separate substances formed 
in the animal body, often in response to treatment with a single anti- 
gen. Kraus,19 in the first edition of Kolle and Wasserman’s Hand- 
book, summarizes this point of view unambiguously in the following 
words: 
“Just as the bacterial body contains a variety of different anti- 
gens, so we may assume that animal protein is made up of a large 
number of different antigenic elements. If the animal body is treated 
with such substances and finds corresponding receptors, there results 
the formation of a variety of qualitatively different antibodies. ...” 
When Gengou,20 in 1902, noted that alexin or complement was 
fixed when a precipitating antiserum was added to its homologous 
antigen, he interpreted this as meaning that, in addition to precipitins, 
the antiserum contained other antibodies, the “albuminolysins.” To 
be sure, Gay and Moreschi 21 showed that the fixation of alexin was 
chiefly a property of the precipitate which was formed, but this was 
regarded as signifying that the protein sensitizers were mechanically 
carried down during the precipitation. Ehrlich, commenting upon 
this in 1910, says: 
“It seems reasonable to assume, in accordance with Gengou’s first 
explanation, that the property of binding the complement is exercised 
by the albuminous bodies sensitized with a specific amboceptor and, 
j ust as when immunizing with cells, agglutinins and amboceptors are 
formed, so also when immunizing with dissolved albuminous bodies 
two kinds of antibodies are formed, precipitins and amboceptors.” 
To be sure the idea that agglutinins and precipitins might rep- 
resent one and the same antibody has been more or less prevalent 
since their first observation. It originated, we believe, with Pal- 
tauf 22 and was expressed as a widely accepted view by v. Eisler 23 
in a review of precipitin reactions published in 1909. But, in regard 
to identification of other antibodies with these two, there has been 
either a complete disregard of such a possibility or, when suggested, 
it has been thrown out of consideration because of the frequent and 
17 Zinsser. Jour. Immunol., 6, 1921, 289. 
18 Ehrlich. “Collected Studies on Immunity,” translated by Bolduan, 2nd 
Edit., p. 585. 
19 Kraus, Kolle and Wassermann. Handb., 1st Edit., 4, 1904, 617 and 618. 
20 Gengou. Ann. de I’Inst. Past., 16, 1902, 734. 
21 Moreschi. Berl. klin. Woch., 37, 1905, 1181, and 1906, 100. 
22 Paltauf. Quoted from V. Eisler, loc. cit. 
23 V. Eisler, Kraus and Levaditi. Handb., 2, 1909, 835. 318 
INFECTION AND BESISTANCE 
complete lack of quantitative parallelism between the curves of tlie 
various antibody functions in one and the same serum. 
The idea of such a possible identity, however, has cropped up 
again and again and, in our own work, has gradually forced itself 
upon us so insistently that we have thought it important to bring 
it forward again. 
The fundamental idea was expressed quite definitely in 1908 
by Bail and Hoke 24 who made an extensive study of the bacteriolytic, 
precipitating and agglutinating actions of normal beef serum and 
immune rabbit serum upon cholera spirilla. They clearly expressed 
the opinion that there were no separate bacteriolytic, precipitating or 
agglutinating antibodies in these sera; that the essential fact was the 
existence, in the sera exerting these effects, of a single antibody 
which united with the bacterial substances and that the various 
reactions which followed were the results of the conditions under 
which the different observations were made. Although a similar idea, 
on somewhat less valid evidence, had been expressed by Biirgi,25 
and although Bail and Tsuda 26 followed out the thought in subse- 
quent publications, the view has made little actual headway, in spite 
of much corroborative, though scattered, evidence. 
In order that there may be no ambiguity as to just what is 
meant by what we call the “unitarian” view of antibodies, let us 
begin by formulating it clearly. 
By such a conception of antibodies, we do not, of course, imply 
a complex cell like, for instance, the typhoid bacillus can give rise 
to one variety of antibody only. There may be formed a specific 
sensitizing antibody against the major chemical constituent, and other 
sensitizers against other antigenic substances enclosed in the same 
cell body or contained in the same antigenic solution. But we do 
mean that, were we working with a single antigen, in a pure state, 
one variety of antibody only would be produced. This would be 
present in the form of a serum constituent specifically capable of 
uniting with the antigen. As a result of the union, the antigen is 
altered in its physical, and perhaps, to some extent, in its chemical 
behavior. The resultant reactions which may be observed with this 
sensitized antigen (agglutination, precipitation, complement fixation, 
bactericidal phenomena, bacteriolysis, opsonization or sensitizing 
effects in the anaphylactic sense) would be determined not by 
differences in the nature of the antibodies with which the antigen had 
united, but rather by the physical state of the antigen itself, the 
nature of the cooperative substances (alexin, leucocytes, tissue cells), 
and by the environmental conditions under which the observations are 
made. 
24 Bail and Hoke. Arch. f. Hyg., 64, 1908, 313. 
25 Biirgi. Arch. f. Hyg., 62, 1907, 239. 
26 Bail and Tsuda. Zeit. f. Immunitats., Orig., 1, 1909, 546. CONSIDERATION OF THE NATURE OF ANTIBODIES 
319 
Thus, if the antibody comes in contact with a very finely divided 
antigen, as in a bacterial extract or in, let us say, horse serum, if elec- 
trolytes are present and perhaps other necessary physical factors fur- 
nished by the presence of serum, etc., precipitation occurs. 
When we are dealing with whole bacteria of relatively large mass 
and correspondingly small surface exposure, agglutination is the re- 
sult, and quantitative parallelism with the precipitin reaction is not 
to be expected because of the much greater dispersion of the antigen 
in the latter test. 
When alexin is present, complement fixation or hemolysis or bac- 
tericidal effects result, since the changes produced by the sensitiza- 
tion have not permitted union with the complement. 
When there are leucocytes present the union makes possible the 
phagocytosis of the antigen, and when the antibody is absorbed by 
the cells of an animal, anaphylactic “sensitization” occurs. 
As we have stated, the earlier opposition to such a view was 
largely based upon lack of quantitative parallelism between agglu- 
tination and precipitation curves on the one hand, and bactericidal or 
protective antibody curves on the other, and this in spite of the rela- 
tive inaccuracies which biological measurements of this nature can- 
not fail to involve. 
In appraising such objections, however, we must not forget that 
agglutination and precipitation are actually only secondary phe- 
nomena, after the union of antigen and antibody has taken place, and 
are dependent upon a great many environmental factors which may 
not, to the same degree, influence phenomena in which alexin, the 
leucocyte or the body cells of animals are involved. We need only 
to point out the frequently observed alterations and diminutions of 
the agglutination and precipitating powers of sera by heat. Heating 
antibacterial sera even to 56° to 60° C. will often materially diminish 
their precipitating effects for bacterial extracts, an observation which 
is entirely analogous to the influence of heating (to 60° to 70° C.) 
on the flocculating effects of serum for various colloidal suspensions, 
like arsenic trisulphide, etc. In addition to this, the flocculation re- 
actions depend upon the presence and the concentration of electrolytes, 
upon reaction, upon mutual relations of concentration, and perhaps 
upon viscosity. Moreover, the suspension equilibrium of the sensitized 
antigen must to some extent depend upon the varying factor of the 
inactive serum constituents carried into the union with the antibody. 
For we know that in precipitation reactions the hulk of the precipi- 
tate comes from the serum, and yet relatively protein-free antibodies 
can be split off from such a complex (Gay and Chickering,27 Chick- 
ering,28 Huntoon,29 conditions which prove that, in the union, much 
27 Gay and Chiekering. Jour. Exper. Med., 21, 1915, 115. 
28 Chiekering. Jour. Exper. Med., 22, 1915, 248. 
29 Huntoon. Jour. Immunol., 6, 1921, 117. 320 
INFECTION AND RESISTANCE 
inactive protein substance is carried along, which inevitably must 
influence reactions of flocculation. Indeed, Tulloch has suggested 
that these “presumably inactive constituents” may even protect the 
united antigen-antibody complex from flocculation, a conception which 
we believe explains the “agglutinoid” phenomena. 
It is not to be wondered at, therefore, that agglutination and 
precipitation curves should not run parallel with the curves of other 
antibody functions. And it is worth noting that, in regard to such 
lack of parallelism, while it has frequently been noted that agglu- 
tinating and precipitating functions were often weaker than other 
antibody effects or even absent entirely in such sera, it has rarely been 
observed that they were powerfully and specifically present when 
other effects were lacking.* 
In addition, however, to these purely quantitative objections to the 
“unitarian” conception, other arguments have been advanced, such 
for instance as those of Gay and Stone 30 who observed, with cholera 
extracts and cholera sera, that the formation of precipitates removed 
little or no lytic substance from the supernatant fluid. As against 
this, however, we have the experiments of Bail and Tsuda who found 
that from specific cholera precipitates, obtained as above and injected 
into the peritoneal cavities of guinea pigs, lytic and bactericidal anti- 
bodies were liberated and effective in producing a Pfeiffer reaction. 
We must confess that such failure of quantitative removal of 
antibodies from the supernatant fluid, after agglutination and pre- 
cipitation reactions had been carried out with these sera, was our 
own experience with typhoid serum, and there are certain points here 
that require further investigation. But when one considers, as we 
have stated above, that the amount of sensitizer necessary to bring 
about an agglutination or precipitation may be very slight, even in 
the presence of an excess of antigen, and that from such agglutinated 
or precipitated bacteria or bacterial extracts dissociation can take 
place; and if we then further consider the great inaccuracies coin- 
cident to the quantitative determination of bactericidal and bacteri- 
olytic effects, such experiments lose much of their weight. One need 
only recall such experiments as those of Topfer and Jaffe 31 who 
found absolute lack of uniformity between plating tests and Pfeiffer 
phenomenon in determining bactericidal substances in one and the 
same sera. There are too many and uncertain secondary factors in all 
these purely biological methods to give one much confidence in any 
reasoning based purely on quantitative results. 
More recently the growing interest in the purification of antibodies 
by dissociating them from antigen-antibody unions has again led 
30 Gay and Stone. Jour. Immunol., 1, 1916, S3. 
31 Topfer and Jaffe. Zeit. f. Hyg., 52, 1906, 393. 
* See also work of Northrop and deKruif on Agglutination, in a preceding 
chapter. CONSIDERATION OF THE NATURE OF ANTIBODIES 
321 
to attempts to separate antibodies in the pure state. Landsteiner 32 
in his early work on dissociation makes no particular point of this 
phase of the problem. Gay and Chickering found that protective 
antibodies could be extracted from specific precipitates with dilute 
sodium carbonate, but interpreted this as a mechanical absorption 
of the protective bodies during precipitation. Huntoon succeeded in 
sensitizing pneumococci, heavily, and then, by treating them with 
slightly alkaline salt solution at 55° C. dissociating from them pro- 
tective antibodies in considerable concentration, and in solutions 
which are practically protein-free. With these antibody solutions 
he has been able to protect mice. It is not our function to go into 
the experiments of Huntoon more extensively, since these are all 
published or in the process of publication. The important point for 
us is the fact that these highly protective antibody solutions fail 
to agglutinate pneumococci in a large number of tests made by him, 
although they possessed distinct protective powers. 
Having considered some of the more important objections, let 
us see what may be said in favor of the “unitarian” view apart 
from its simplicity. 
The identification of agglutinins with precipitins should hardly 
require much argument since both are specific flocculation reactions 
between a serum and the same antigenic substance, depending upon 
analogous environmental conditions. That they should differ from 
each other quantitatively is to be expected from the fact that in 
one case the antigen is present in relatively large masses and in the 
other case it is finely dispersed. 
That there is at least a strong likelihood that precipitins and 
bactericidal and protective antibodies may be identical is indicated 
by the experiments of Bail and Tsuda who dissociated bacteriolytic 
cholera antibodies from precipitates obtained with cholera extracts 
and normal beef serum. 
Bail and Hoke obtained similar results not only with normal 
beef serum, but with immune rabbit serum, though in the latter case 
somewhat less sharply. The latter point is important because the 
conglutination question is involved in their experiments with normal 
serum for they obtained much better precipitates with the normal 
beef serum in its active state. However, since sensitization with an 
antibody is necessary for the conglutinin effect, the principle of 
Bail and Hoke’s experiment in regard to the liberation of bacteriolytic 
sensitizer from precipitated cholera extracts remains the same. 
Also the identity of precipitins and protective antibodies is very 
strongly suggested by the experiments of Gay and Chickering, inas- 
much as they succeeded in extracting protective pneumococcus anti- 
bodies from specific precipitates, and the identity of the two is at 
least likely as mere mechanical carrying down with the precipitate. 
32 Landsteiner. Miinch. med. Woch., 49, 1902, 1905. 322 
INFECTION AND RESISTANCE 
The probable identity of precipitins with complement fixing anti- 
bodies is indicated by the work of Dean 33 and by many observa- 
tions of our own. 
In 1912, Dean, analyzing the relationship between alexin fixation 
and precipitation, concluded that the proportions of antigen and 
antibody which favor rapid and complete precipitation do not favor 
complete alexin fixation. He did not believe that the two reactions 
followed a parallel course, but said that he thought they represented 
twro phases of the same reaction, a “flocculation representing the 
first stage of a change that can be recognized in its early stage by 
complement fixation.” In 1912 and 1913 we were engaged in a 
similar analysis from which we came to the conclusion that there 
was no need for assuming that the antibodies which were involved 
in the fixation of the alexin were essentially different from those that 
brought about the precipitation. Indeed, in his Croonian lecture of 
1917, Dean elaborated this idea, and came to the same conclusion 
that we did, namely, that all the various so-called antibody reactions 
wrere attributable to one and the same specific substance in the serum. 
This lecture of Dean, published during our absence, was, unfor- 
tunately not in our hands when our own paper on the same subject 
was written, but though he approached the problem from a slightly 
different point of view, his conclusions were essentially the same. 
Studying the relationship between precipitin reaction and com- 
plement fixation in the reactions between sheep serum and anti-sheep 
rabbit serum, like so many other observers, we found that amounts 
of antibody which did not precipitate would, nevertheless, produce 
complement fixation reactions in such mixtures and this, of course, 
is the basic principle of the forensic complement fixation, introduced 
and described practically by Neisser and Sachs.34 In order to 
eliminate the possibility of non-specific mechanical carrying down 
of the complement fixing antibody, we carried out experiments with 
complement fractionation, and found that the fixation by the pre- 
cipitates and by the supernatant fluids in such mixtures took place 
in the same way that fixation by sensitized cells occurs; namely, that 
the precipitates first fixed end-piece and mid-piece, secondarily. 
From this and other experiences, we expressed the conclusion at 
that time that “there is but one variety of specific sensitizer,” and 
that “the visible precipitation is merely secondary, occurring be- 
cause of the colloidal nature of the reacting bodies under quantitative 
and environmental conditions which favor flocculation.” Further- 
more, the experiments of the last ten years have shown again and 
again that the physical state, the formation of precipitates and the 
changes in sizes and perhaps in electrical conditions of the substances 
in serum reactions are intimately related to the fixation of alexin. 
33 Dean. Zeit. f. Immunitats., Orig., 13, 1912, 84. 
34 Neisser and Sachs. Berl. klin. Woch., 1905, 1388. CONSIDERATION OF THE NATURE OF ANTIBODIES 
323 
The recently developed observations on the Wassermann reaction, 
the Yernes test and the Sachs-Georgi reaction now successfully em- 
ployed in our laboratory are all evidence in this direction. Recently 
the writer with J. T. Parker observed a pneumococcus horse serum 
which gave powerful precipitation with the residue antigens pre- 
pared from bacteria by us and described in another place. This 
serum, however, gave no complement fixation with the same antigen, 
and since this was consistently so, it seemed to contradict our ideq. 
of the identity of the two antibodies involved. We subjected this 
to analysis, and found that the discordant results were due to a special 
interfering substance in the horse serum which prevented the fixation 
of complement even when an anti-pneumococcus serum obtained 
from rabbits, was employed, with which perfect complement fixation 
was obtained unless the horse serum was added. It was thus found, 
that the failure to fix complement on the part of the horse serum 
could not be interpreted as indicating two separate antibodies.35 
That the anaphylactic passive sensitizing effect of sera was found 
to be quantitatively proportionate to the precipitin contents of these 
sera, was shown by Doerr and Russ 36 in experiments which seem 
completely to justify the identity of precipitating and anaphylactic 
antibodies suggested by Friedberger in 1908. 
In regard to the dissociation experiments of Huntoon, some of 
whose solutions were sent to us by Dr. Huntoon, himself, we have, 
indeed, found that these solutions exert considerable protective effect 
upon mice and fail either entirely, or almost entirely to produce 
agglutination “in vitro” of the pneumococcus against which they 
protect. 
This would seem a potent argument against the “unitarian” idea. 
However, it must be considered that the “protein-free” antibody of 
Huntoon is physically, and, therefore, in its agglutinating and pre- 
cipitating functions quite a different substance from the original anti- 
body in the unchanged serum, which in its native state is associated 
with the inactive pseudo-globulin substances which it carries down 
with it into precipitates. For, as we have stated before, it is of 
course known that most of the substance which comes down in pre- 
cipitin reactions is derived from the immune serum. Moreover, by, 
to some extent, restoring the Huntoon antibodies to their original 
environment in the circulation of an animal, we were able to observe 
powerful agglutination. We repeated, with these substances, the ex- 
periment on the mechanism of serum protection against pneumococci 
in rabbits, carried out by Bull 37 in 1915. A rabbit of approximately 
1200 grams was intravenously injected at 5 P. M. with about 8 c.c. 
of a broth culture of pneumococcus I of relatively low virulence. 
35 Zinsser and Parker. Jour. Immunol., 1922. 
36 Poerr and Russ. Zeit. f. Immunitdts., 3, 1909, 181. 
37 Bull. Jour. Exper. Med., 22, 1915, 466, 484, and 24, 1916, 25. 324 
INFECTION AND RESISTANCE 
At 10 A. M. the next morning the rabbit was very sick; a heart’s 
blood puncture was done which showed numerous pneumococci 
evenly distributed through the blood stream. Ten cubic centimeters 
of Huntoon’s material was injected intracardially in this animal, the 
needle was withdrawn and about two minutes later blood was taken 
from the heart and smeared. Similar punctures were done five 
minutes later and fifteen minutes later. As will be seen from the 
attached illustration, smears from the heart’s blood taken two minutes 
later showed the pneumococci in the blood in clumps of varying 
sizes, with very few small clumps of two and three, and a very few 
individual organisms. But the large majority of the organisms 
were now in clumps of 10 or more members. Smears after five min- 
utes showed a few clumps and a few individual organisms, and after 
fifteen minutes there were very few single organisms and no clumps, 
but it was then hard to find organisms, whereas before injection every 
field showed ten or more. 
We did not succeed in thus restoring the agglutinative functions 
of Huntoon’s materials in all cases. But the reason for this was re- 
vealed, we believe, by experiments carried out for us by J. T. Parker. 
She found that the Huntoon substances, produced by dissociating 
sensitized pneumococci in dilute sodium bicarbonate solutions at 
55° C., had a reaction of Ph 8.8 to Ph 9.4. When powerfully 
agglutinating pneumococcus sera were subjected to similar treatment 
at similar reactions, their agglutinating powers were largely lost 
and but partially restorable by neutralization. It would seem that 
these various considerations should serve materially to weaken the 
validity of objections to the “unitarian” theory based upon the 
separation of antibodies by dissociation experiments. 
Furthermore, observations recently made by Coulter 38 seem to 
us to have considerable bearing upon this question in indicating that 
agglutination and the sensitization to hemolysis of red cells, are both 
due to the action of one substance. Coulter found that “the optimum 
hydrogen ion concentration for the agglutination of sensitized cells 
(rabbit-antisheep sensitizer), in a salt-free medium, occurs at Ph 5.3 
which corresponds with the optimal point for the precipitation of the 
serum-globulin itself.” The optimum for precipitation of serum- 
globulin, in which the immune bodies are carried, is stated by Rona 
and Michaelis 39 as Ph 5.2 and the iso-electric point for typhoid 
immune bodies as Ph 5.4. This reaction is more alkalin than that 
which is optimal for the agglutination of normal sheep cells in sac- 
charose solution, so that Coulter concludes, as far as agglutination is 
concerned, the behavior of sensitized cells is closely related to the 
properties of the immune serum. We quote directly from a personal 
communication from Coulter as follows: 
38 Coulter. Jour. Gen. Physiol., 3, 1921, 309. 
80 Rona and Michaelis. Biochem. Zeit., 28, 1910, 193. CONSIDERATION OF THE NATURE OF ANTIBODIES 
325 
“Furthermore, an equilibrium has been found to exist between 
the amount of hemolytic sensitizer free and that combined with cells, 
the amount which combines at a Ph of 5.3 being at the maximum, 
and approximating 100 per cent.” 
These observations indicate that the hydrogen ion concentration 
at which the agglutination of sensitized cells is most perfect corre- 
sponds with the iso-electric point of that part of the serum which 
makes the antibodies and also corresponds with the point at which 
the largest amount of hemolytic sensitizer is absorbed by the red cells. 
This, to our mind, would be strong evidence in favor of identity of 
the hemolytic sensitizer and the agglutinin. 
We do not wbsh by any means to convey the impression that we 
consider the “unitarian” view as absolutely and rigidly proven. We 
do believe, however, that the denial of such a view necessitates the 
assumption that the injection of a pure antigen calls forth five or 
six fundamentally different reactions on the part of the tissue cells, a 
theory which would be justified only on the basis of incontrovertible 
proof. 
If there is nothing further to be said in favor of the “unitarian” 
view, one might at least wait for further evidence before one tried 
to deny an existing view, however complicated. But there is much 
to be said in favor of it and evidence in this direction is accumulating. 
We have believed in the probable truth of the “unitarian” view 
for a considerable number of years, with sufficient conviction to 
teach it as the most likely state of affairs. And while we cannot 
prove it in all the ramifications of the difficult experimental problems 
involved, we believe it has gone far enough certainly to throw the 
burden of proof upon those who still cling to the separation and the 
conception of separate structure for agglutinins, precipitins, bac- 
teriolysins, etc. 
FURTHER ATTEMPTS TO APPLY PHYSICAL MEASUREMENTS TO ANTI- 
BODY REACTIONS 
In a number of discussions in preceding chapters we have em- 
phasized the fact that serum reactions are gradually being recognized 
as having more in common with so-called colloidal phenomena than 
with those taking place during the union of crystalloids. The funda- 
mental physical principles underlying colloidal reactions are as yet 
pretty vague and the serologist especially (ourselves included), when 
he speaks of colloidal reactions is often groping in the dark. Never- 
theless, even without such fundamental understanding we can recog- 
nize the close analogies which exist between the various serum re- 
actions and those taking place between substances recognized by 
their attributes as colloids. We have sufficiently discussed this in 326 
INFECTION AND RESISTANCE 
the chapters on agglutination and precipitation, and need not en- 
large upon it here except in indicating how the gradual tendency to 
pay attention to the physical factors involved has led to the actual 
working out of serum experiments in which the reasoning from the 
beginning and the technique finally developed were physical rather 
than chemical. Perhaps the earliest phases of such work were those 
in which complement fixation was shown to occur when complement 
was brought together with non-specific inorganic and organic colloids. 
It is due chiefly to the work of Landsteiner and his associates that 
attention was called to such phenomena. Landsteiner with Stan- 
kovic showed that complement may he fixed by silicic acid, and 
Seligmann soon afterwards found that in the precipitation of mastic 
suspensions complement may he fixed. Much work has been done 
that shows that a variety of colloidal suspensions of known chemical 
constitution fix complement, probably by adsorption. Perhaps the 
most striking instance of such complement adsorption is that occur- 
ring in the Wassermann reaction. Here, as we know, complement is 
fixed by a combination of syphilitic serum and various lipoidal 
suspensions, which may he entirely non-specific in origin. When 
the reaction occurs, as Jacobstahl and, later, Bruck have shown, a 
precipitation occurs which can he seen in the ultramicroscope. 
Furthermore, this precipitation takes place more rapidly in the ice- 
chest than in the incubator, a principle upon which is based the so- 
called refrigerator method of performing the test, and which strik- 
ingly suggests that the reaction is an adsorption rather than a true 
chemical union. It is this precipitate which fixes the complement; 
whether or not this is due to quantitatively increased globulin or to 
purely physical change in the syphilitic serum, is a matter which we 
cannot discuss at present. Whatever it is, it is unquestionable that 
the availability of the antigen for the Wassermann reaction depends 
not only upon its lipoidal nature but also on its state of dispersion. 
Since it is not possible, as we have found, to make available antigens 
for this reaction with non-lipoidal substances, like mastic, gelatin, 
gum arabic, salicic acid, albumins, and a number of other substances, 
even when in dispersion more or less similar to that of the Wasser- 
mann antigen, it seems that the secret of the Wassermann antigen 
must lie in the fact that substances of the chemical and physical con- 
stitution of lipoids when brought into a definite state of dispersion 
offer surface tension conditions not easily obtained with colloids of 
another nature. It is, therefore, at least in our opinion at present, the 
physical condition of the Wassermann antigen which makes it avail- 
able for the test, a physical condition which is secondarily dependent 
upon the chemical nature of the dispersed substance. The importance 
of the state of dispersion and therefore the surface tension properties 
is quite apparent from the fact noted by many Wassermann workers 
that a considerable difference in the fixing power of the antigen may CONSIDERATION OP THE NATURE OF ANTIBODIES 
327 
be obtained by in one case adding tbe salt solution to the alcoholic ex- 
tract quickly, and in another case adding it very slowly, the two 
separate preparations showing, one a very transparent, the other a 
very turbid, condition. 
The Bordet-Danysz phenomenon has already been discussed, but 
is another case in point. It consists of the fact that if to a definite 
amount of antitoxin a definite quantity of toxin is added, the re- 
sult is one of greater toxicity in the mixture if the poison is added 
to the antitoxin fractionally than when the entire amount is added 
at once. An interesting difficulty of this phenomenon is the fact that 
such a reaction seems to force upon us the assumption that the toxin- 
antitoxin union is not reversible, whereas later work on immuniza- 
tion with neutralized mixtures of these substances seems to necessitate 
the assumption that within the body they are reversible. 
The “Epiphanin” Reaction.—it is a consideration of these and 
many other apparently physical factors of immune reactions which 
has led experimenters to seek to utilize physical changes for practical 
serological purposes. One of the reactions resulting from such 
studies is that known as Weichhardt’s Epiphanin reaction. Weich- 
hardt noticed that when two solutions of toxin and antitoxin are 
brought in contact with each other on exactly horizontal glass plates, 
diffusion takes place between them much more rapidly than in con- 
trols in which one or the other, antigen or antibody, was lacking. 
He tested this at first on the horizontal glass plates by adding coloring 
matter to the two solutions and observing the diffusion cur- 
rents. Later he developed a method in which he tried to determine 
such increased diffusion by means of a delicate chemical balance. 
His method, briefly described in his own words, was as follows: “To 
the two arms of scales, to the right and left, a little bell-shaped jar 
is attached into which dilutions of serum are placed, on the one 
side specific immune serum, on the other normal serum. These 
little bell-jars are closed at the bottom with ‘Schleicher-Schiill’ filter 
paper and are immersed into solutions containing antigen in salt 
solution. Through the filter-paper membrane diffusion takes place, 
and in observations carried on for from several hours to several days 
it can be determined that the scales sink on that side in which specific 
immune serum had been placed into the little bell-jar. In other 
words, the antigen solution which is contained in the salt solution 
diffuses more rapidly on the side on which the specific immune 
serum was present. In consequence, the weight of the little bell in- 
creases and a definite reading can be made.” Similar observations 
have been made by other observers, Kraus and Amiradzibi doing 
th6 experiment as follows: they employed a U-tube in the connecting 
horizontal part of which there was a glass stopcock. Intq. one side 
they put diphtheria toxin with a little aqueous methylene blue, and 
into the other side antitoxin. For each experiment two controls were 328 
INFECTION AND RESISTANCE 
set up, one with salt solution and the other with normal horse serum, 
and they noticed that after a definite period of time after the stopcock 
connecting the two had been opened, methylene blue could be ob- 
served to diffuse across in the tube in which antigen and antibody 
had been present and not in the controls. Many modifications of 
technique to demonstrate this principle have been made by Weich- 
hardt himself, and since the reaction is not likely to find much im- 
mediate diagnostic application because of its great delicacy, we need 
not describe them at length but refer the reader to Friedemann’s 
excellent description in the Kolle and Wassermann, second edition, 
Vol. Ill, 
Similar in principle but by no means identical is the so-called 
Meiostagmin reaction, chiefly developed by Ascoli. 
The Meiostagmin Reaction.—Ascoli and Izar 40 have attempted 
to work out a diagnostic reaction which depends upon an alteration 
of surface tension of a fluid when an antigen unites with its specific 
antibody. Ascoli in his first experiments worked with typhoid 
bacillus extracts and the sera of typhoid patients, and found that 
when the two suspensions were mixed a reduction of surface tension 
resulted after time for union between the two had been allowed. 
They determined the reduction of surface tension by Traube’s 41 
method by the use of apparatus spoken of as the “stalagmometer.” 
The principle of this method depends upon the fact that as surface 
tension is reduced the number of drops to a given quantity of fluid 
is increased. 
Diluted serum of patients was mixed with diluted antigen, and 
the number of drops contained in one cubic centimeter of the mix- 
ture was immediately determined and again measured after the mix- 
ture had remained for two hours in the incubator at 37° C. An 
example of one of Ascoli early measurements is given in the protocol 
cited below. 
Reduction of surface tension results when various antigens are 
brought together with normal sera, but this can be easily controlled 
by suitable dilution, and must be carefully taken into consideration 
in each individual case. Ascoli and Izar have applied this method to 
the diagnosis of tuberculosis, typhoid, and various other diseases, 
and have reported what seemed to them reliable results. So far ex- 
perience with the meiostagmin reaction has not been very extensive; 
not all observers have been able to obtain results as apparently reliable 
as those of Ascoli and his collaborators. It is not possible therefore 
to express a final opinion regarding this method of investigation; 
it contains, however, an interesting principle which with more exact 
methods of measurement may well become very important in serum 
diagnosis. 
40 Ascoli and Izar. Munch, med. Woch., Nos. 2, 7, 18, 22, 41, 1910. 
41 Traube. Pfliiger’s Archiv, Yol. 123, 419. CONSIDERATION OF THE NATURE OF ANTIBODIES 
329 
1 c. c. of serum of typhoid patient diluted to 1-10. 
Number of drops 
Immedi- 
ately 
57.8 
58.1 
57.5 
57.5 
57.0 
57.0 
58.1 
57.7 
57.4 
57.6 
57.0 
56.9 
56.5 
56.5 
56.5 
56.6 
56.7 
56.5 
56.6 
56.7 
After 
2 hours in 
incubator 
59.7 
59.9 
59.4 
59.6 
59.3 
59.2 
59.7 
59.6 
59.4 
59.2 
59.2 
59.4 
58.0 
57.8 
57.5 
57.4 
57.4 
57.5 
57.5 
57.6 
1 0/00 
1 0/000 
1 0/000 
1 c. c. alcoholic typhoid extract diluted to.... 
1 0/00 
1 0/000 
1 0/000 
1 c. c. alcoholic typhoid extract diluted to 
1 0/00 
1 0/000 
1 0/0000 
1 c. c. alcoholic precipitate taken up in distilled 
water - 
1 c. c. in 1 0/00 alcohol in 1 c. c. 0.85 per cent. 
NaCl solution 
Colloidal Gold Reaction.—An important reaction, depending 
entirely upon the principle of colloidal precipitation, is the so-called 
colloidal gold reaction extensively used for the diagnosis of syphilis of 
the central nervous system. In 1901, Zsigmondi, knowing that pro- 
tein substances will protect various colloids from precipitation by 
electrolytes, attempted to work out a method by means of which he 
could quantitatively estimate protein in solution by its protective 
effect. He worked with colloidal metals, using especially colloidal 
gold. He determined what he called the “Goldzahl” for various pro- 
teins, by which he meant the amount in weight of a protein which 
was just enough to protect from precipitation 10 c.c. of colloidal gold, 
of a percentage of 0.0053, in the presence of 1 c.c. of 10 per cent. 
NaCl. It, of course, had long been known that in various pathological 
conditions of the central nervous system the protein contents of the 
fluid varied. Thus, the reactions of Honne, of Hoguchi, etc., were 330 
INFECTION AND RESISTANCE 
all aimed at revealing an abnormal globulin content of the spinal 
fluid. Lange, attempting to apply Zsigmondi’s method to the quan- 
titative detection of proteins in the spinal fluid, observed a curious 
reaction, quite the opposite of what he had set out to find. He 
observed that spinal fluid, especially of syphilitic cases, in which 
there was protein material beyond the normal, precipitated rather 
than protected colloidal gold. He also observed that the dilution at 
which such precipitation took place was more or less constant in 
syphilitic cases and might therefore be utilized for the diagnosis of 
such diseases as paresis, locomotor ataxia, etc., etc. The reaction was 
taken up and carefully studied by a large number of observers, 
among whom were Zaloziecki, Jaeger and Goldstein, Flesch, Kafka, 
Lee and Hinton, Miller and Levy, and many others. One of the 
most practical discussions and clear descriptions available for Ameri- 
can and English readers is that of Miller, Brush, Hammers, and 
Felton. The chief difficulty of the test consists in the preparation of 
a proper colloidal gold solution. 
Lange adopted the method used by Zsigmondi with slight modi- 
fications. He added 10 c.c. of a 1 per cent, solution of gold colloid 
and 10 c.c. of a 2 per cent, solution of potassium carbonate to a liter 
of carefully and freshly distilled water. The solution was rapidly 
warmed almost to a boil and just before boiling, while actively stir- 
ring 10 c.c. of 1 per cent, formalin solution is added. The solution 
has remained colorless up to this point but upon this addition should 
almost instantaneously become a deep, transparent red. There should 
be no iridescent or smoky “Schimmer.” The utmost care in getting 
pure distilled water must be exercised and Jena glass must be used 
throughout. 
Other methods are those of Eicke, who adds dextrose, and many 
workers, such as Flesch, Kaffge, Eskuchen, and others, have found 
great difficulty in making up the solution by Lange’s technique or by 
any of the other methods described. Miller and Levy state that 
with Lange’s method good solutions can always be obtained if the 
water is absolutely pure and the glassware is satisfactorily cleaned; 
later, Miller, with the collaborators mentioned above, found that this 
was not universally true. Also, these writers, as well as Eskuchen, 
found that occasionally solutions which appeared all right did not 
react, and special studies followed dealing wfith the technique of the 
preparation of the gold sol. Miller and his collaborators find that 
the reduction method of producing this solution is the best one, and 
proceed as follows. They use the following reagents: 42 
1. The Gold Solution: 
AuCl,—Merck’s yellow crystals hermetically sealed in brown 
glass ampules 1 gram 
Water triply distilled, up to 100 c.c. 
42 The descriptions are taken direct from the thorough papers of Miller 
and Levy. CONSIDERATION OF THE NATURE OF ANTIBODIES 
331 
This stock solution is kept well stoppered in dark glass bottles 
away from any bright light. 
2. The Alkaline Solution: 
Merck’s Blue Label Potassium Carbonate (desiccated) 2 grams 
Water triply distilled, up to 100 c. c. 
3. The Reducing Agents: 
a. Formaldehyde, Merck’s 40 per cent, stock solution, highest 
purity 1 c. c. 
Water triply distilled, up to 40 c. c. 
b. Oxalic acid, Merck’s Blue Label, Crystals 1 gram 
Water triply distilled, up to 100 c. c. 
Solutions dSTo. 2 and No. 3 must he made up immediately prior 
to use. 
4. Bichromate cleaner for glassware: 
Potassium Bichromate, powdered 200 grams 
Water, distilled, up to 1500 c. c. 
Sulphuric Acid cone 500 c. c. 
The potassium bichromate should be well dissolved before the 
sulphuric acid is added. If this solution is reserved for cleaning 
glassware only, it can be used repeatedly. 
Great attention must be paid to the cleaning of glassware. They 
boil their Jena beakers in Ivory soap solution, then brush under hot 
tap water. After being rinsed for five minutes, hot bichromate and 
sulphuric acid is added for half an hour. The beaker is then emptied 
and washed in running water for five minutes, rinsed with distilled 
water and then in triply distilled water. Similar methods are used 
with other glassware. The chief errors are insufficient brushing, 
failure to get all bichromate out, and allowing beakers to dry in air 
before using. 
In obtaining the distilled water, they first take ordinary distilled 
water which they then re-distil from Jena flasks. They wash out 
the collecting flasks with the first 200 c. c. of the second distillate, 
then collect and re-distil in the same way, again rinsing with the 
first 100 c. c. of the second distillate. After all these steps have been 
performed the rest is more or less simple. A beaker rinsed out 
with triply distilled water is filled to the liter mark and the tem- 
perature raised to 50° C. gradually, then the gas is turned on full 
and when the temperature has reached 60° C. 10 c.c. of the 1 per cent, 
gold solution and 7 c.c. of the 2 per cent, potassium carbonate solu- 
tion are added. The solution should remain perfectly clear. At 
80° C. while stirring with a clean thermometer, 10 drops of oxalic 
acid are slowly added. The solution may now turn a delicate bluish- 
pink, often due to an excess of alkali. Otherwise the solution re- 
mains colorless until 90° C. has been reached. When 90° C. has been 
reached, the gas is turned out and, while stirring, 5 c. c. of 1 per cent, 
formaldehyde is added, drop by drop. If a pink color makes its 332 
INFECTION AND RESISTANCE 
appearance before all the reducing agent has been added, Miller 
advises to stop at once. He also states that the best solutions are 
those in which the color change is slow. For further technical points, 
the reader is referred to the paper of Miller and his collaborator in 
the Johns Hopkins Hospital Bulletin, 1915, xxvi, 391, from which 
this description is almost bodily taken. 
The test is done as follows: into a clean test tube 1.8 c. c. of 
fresh sterile 0.4 per cent. NaCl solution is placed. Into ten further 
tubes 1.8 c. c. of the same salt solution is placed. To the first tube is 
added 0.2 c. c. of spinal fluid to be tested and well mixed. 1 c. c. of 
this 1: 10 solution is added to the second tube, mixed, and of this 1 
c. c. is added to the third tube, in consequence of which dilutions are 
obtained running from 1: 10, 1: 20, 1: 50. How to each of these is 
added 1 c. c. of the colloidal gold solution. Changes begin to take 
place in five minutes, which usually reach their completion in about 
a half hour. The spinal fluids must be free from blood and clear 
from bacterial contamination. 
Normal spinal fluids produce usually no reaction. The so-called 
luetic curve is usually one in which the greatest precipitation occurs 
in the tubes ending in 1: 40 to 1: 160. In suppurative lesions, etc., 
the strongest precipitation occurs in higher dilutions, say, from 320 
to 280, which Lange has called “verschiebung nach oben.” A so- 
called paretic curve means complete precipitation in, say, from 110 
to 160, with a gradual fading out toward the higher dilutions. A 
complete precipitation means a colorless tube; partial precipitation 
shades from pale blue to the complete red of the unaffected colloidal 
gold. Fluids from early stages of syphilis without nervous system 
involvement react usually like normal fluids. 
That the reaction has unquestionably found a permanent useful- 
ness cannot be doubted. It seems to be dependent entirely upon 
the technique of making the colloidal gold solution. CHAPTER XIII 
PHAGOCYTOSIS 
Early investigations into the fate of bacteria within the infected 
animal body were largely carried out by pathological anatomists, and 
the observation of the presence of micro-organisms within the cells 
of the animal and human tissues was definitely made as early as 
1870. Hayem,1 Klebs,2 Waldeyer,3 and others, saw leukocytes con- 
taining bacteria but failed to interpret this in the sense of possible 
protection. The process was regarded rather as a means of trans- 
portation of the bacteria through the infected body, or it was as- 
sumed that possibly the micro-organisms had entered these cells be- 
cause of the favorable nutritive environment thus furnished. 
The first to suggest that such cell ingestion might represent a 
method of defence was Panum,4 who referred to it as a vague possi- 
bility. A similar hut more convinced expression of this opinion 
was made in 1881, according to Metchnikoff,5 by Roser in explaining 
the resistance of certain lower animals and plants against bacteria. 
But Roser brought no experimental support for his contention, and 
little attention was paid to his assertion. 
The significance of cell ingestion as a mode of protection against 
bacterial invasion, therefore, was hardly more than a vague sugges- 
tion when Metchnikoff, who, though a zoologist, had become intensely 
interested in the problem of inflammation, began to experiment upon 
the cell reaction which followed the introduction of foreign material, 
living or dead, into the larvae of certain starfishes (Bipinnaria). 
Pathologists, at this time, held complicated views of inflamma- 
tion which involved complex coordinated reactions of vascular and 
nervous systems, and Metchnikoff’s primary purpose was to observe 
reactions to irritation in simple forms devoid of specialized vascular 
or nervous apparatus. He noted in these transparent, simple forms of 
life that the foreign particles were rapidly surrounded by masses 
of ameboid cells and reached a conclusion which, in his own words, 
is expressed as follows: 
1 Hayem. C. R. de la Soc. Biol., 1870. 
2 Klebs. Pathol. Anat. der Schusswinden, 1872. 
3 Waldeyer. Arch. f. Gynekol., Yol. 3, 1872. 
4 Panum. Virch. Arch., Yol. 60, 1874. 
5 Metchnikoff. “L’lmmunite dans les Maladies Infectieuses.” 
333 334 
INFECTION AND RESISTANCE 
“L’exsudat inflammatoire doit etre considere comme une reac- 
tion contre toutes sortes de lesions et l’exsudation est un phenomene 
plus primitif et plus ancien que le role du systeme nerveux et des 
vaisseaux dans Pinflammation.” 6 
He compared the process of cell ingestion or phagocytosis of for- 
eign particles, as here observed, to that taking place in the most 
simple intracellular digestion which occurs in unicellular forms, a 
hereditary cell function now specialized in certain mesodermal cells, 
and passed on in the evolution of higher forms to other specialized 
cells. And indeed in animals of the most complex structure the 
leukocytes which carry on this phagocytic process may he considered 
as, in a way, representing a primitive form of cell, since they are 
only nucleated elements of the body which wander from place to 
place, and are anatomically independent of nervous control. In 
1883, at the Naturalists’ Congress in Odessa, Metchnikoff 7 first 
expressed his views and communicated the first of the splendid re- 
searches upon which our modern conception of phagocytosis is largely 
based. 
His earlier studies were carried out with a small crustacean, the 
daphnia, in which he studied the reaction which followed the intro- 
duction of yeast cells. He observed the struggle which ensued be- 
tween the ameboid leukocytes of the crustacean and the infecting 
agents and determined that complete enclosure of the yeast within 
the leukocytes assured protection to the daphnia, while a failure of 
this process, either from fortuitous causes or because of too large a 
quantity, or too high a virulence of the infecting agents, resulted in 
disease and rapid death. 
This early work of Metchnikoff forms the beginning of a long 
train of investigations to which we owe most of the basic facts we 
possess concerning the role of the phagocytic cells in the protection 
of the body against infection. Just as the various serum phenomena, 
of which we have spoken, have a general biological significance apart 
from their importance in relation to bacterial invasion, so the process 
of phagocytosis must be looked upon as an attribute of the animal 
and vegetable cell which has important physiological bearing entirely 
apart from infection. 
In fact, the ingestion of bacteria and other foreign particles by 
the leukocytes and other phagocytic cells of higher planets and ani- 
mals is entirely analogous to the intracellular digestive processes 
which take place, as the ordinary manner of nutrition, among the 
unicellular forms. Among the rhyzopods, in general, food is taken 
6 Inflammatory exudation should be considered as a reaction against all 
sorts of injuries, and exudation is a phenomenon more primitive and ancient 
than are the parts played by nervous system and blood vessels in the process 
of inflammation. 
7 Metchnikoff. Arb. a. d. zool. Inst., Wien, Yol. 5, 1883. PHAGOCYTOSIS 
335 
in by means of the ingestion of other smaller forms of life, bacteria, 
algae, etc. (or particles of dead organic matter), into the cell body 
of the protozoon. 
These materials are gradually engulfed by the body of the ameba, 
which flows about them with its pseudopods, and within the cyto- 
plasm undergo gradual digestion. The process has been carefully 
studied by Mouton.8 In symbiotic cultures of amebse with colon 
bacilli on agar plates, the bacteria are taken up in large numbers 
and about them are formed small vacuoles. That the digestion takes 
place in a slightly acid medium with the vacuoles can be proved by 
adding a drop of neutral red to the hand-drop preparation of amebse 
and observing the brownish-red color taken by the materials in the 
vacuoles. Mouton was able to obtain a digestive ferment from the 
amebse, by glycerin extraction, which exerted strong proteolytic action 
upon various albuminous substances, liquefied gelatin, and digested 
dead colon bacilli in vitro, acting best in slightly alkaline, but also in 
slightly acid, reactions. It is plain, therefore, that the most prim- 
itive form of digestion is an intracellular one carried on by ferments 
comparable in every way to the secreted digestive enzymes which 
accomplish the same purpose outside of the cells in higher animals. 
In essence, however, there is no fundamental difference physiolog- 
ically between intra- and extracellular digestions, and the intracellu- 
lar manner of assimilating solid nutritive particles may be retained 
in forms much higher in the scale of evolution than the rhizopods. 
It has been studied by Metchnikoff and others in certain of the flat 
worms (Dendrocelum lacieum) in which typical phagocytosis is car- 
ried on by the cells of the intestinal mucosa. Many of these pla- 
naria obtain their nourishment by sucking the blood of higher ani- 
mals. Placed under a microscope after feeding, it may be seen that 
the foreign blood cells are rapidly taken up by the intestinal epithe- 
lial cells, which engulf them by means of pseudopodia not unlike 
those of the ameba. After ingestion, here, too, the blood cells are 
surrounded by vacuoles within which their gradual disintegration or 
digestion is accomplished. Similar intracellular digestion seems to 
be general among the coelenterates, and has been thoroughly studied 
by Metchnikoff in the actinia. Here the food particles are carried 
by the tentacles into the esophagus, and are taken up by the endo- 
dermal cells of the so-called “mesenteric filaments,” where they are 
digested hy a trypsin-like enzyme. In these animals digestion is 
entirely intracellular, though the ingesting cells are the parts of a 
specialized tissue. In other forms, still higher in the scale, although 
there is persistence of intracellular digestion, the extracellular 
process begins to be developed. Thus in certain mollusca the solid 
food is taken into the intestinal canal, where it first undergoes a 
preliminary digestion by secreted intestinal juices. After it has 
8 Mouton. C, R. de VAcad. des Sciences, Yol. 133, 1901. 336 
INFECTION AND RESISTANCE 
been reduced to small amorphous particles in this way, these are 
seized by the ameboid cells, and intracellular digestion completes the 
process which has been begun extracellularly. 
As we study the process among higher animals, it appears that, 
among vertebrates, the intracellular methods of digestion have been, 
at least for normal metabolism, entirely displaced by the extracellu- 
lar as it occurs in the intestine, where solid particles are rendered 
completely amorphous, dissolved, and reduced to a diffusible condi- 
tion by the digestive juices before they are offered to the cells for 
utilization. However, the capacity for intracellular digestion is not 
entirely lost, and is retained of necessity in certain body cells. For 
were there not such an emergency arrangement the body would lack 
an available mechanism with which to meet such accidents as ex- 
travasations of blood, or the entrance of bacteria and other foreign 
solid particles into the tissues. It seems reasonable to classify both 
the phagocytic action of body cells and the formation of anti-bodies 
in the blood plasma, primarily as emergency devices for the diges- 
tion of foreign materials both formed and unformed which, under 
abnormal conditions of injury or disease, penetrate into the physio- 
logical interior of the body (blood stream or tissue spaces), and 
must be disposed of. 
In the lowest animals the single cell is called upon to perform 
all necessary functions. In the course of evolution, however, as the 
body becomes more and more a community of many cells, a division 
of labor takes place which is expressed morphologically in the differ- 
entiation of tissues and organs, and physiologically in the adaptation 
of individual tissue cells to the performance of specialized functions. 
Nevertheless, it is necessary, both for certain normal processes, as 
well as for provision against such complex emergencies as those 
mentioned, that certain cells of the complex community 
the primitive abilities of the more independent cells of the lower 
forms. Thus, among many animals, the phagocytic action of cells 
performs definite services in the course of normal development. 
This is seen most markedly in some insects (diptera) in which the 
destruction of larval organs, useless to the adult animal, may be en- 
tirely accomplished by the action of phagocytic cells, and a similar 
process may accompany the transformation of the tadpole to the adult 
in many amphibia.9 In higher animals the removal of extravasations 
of blood is accompanied by a train of occurrences which is readily 
subjected to study.10 In such cases the leukocytes rapidly enter the 
area of extravasation and an engulfment of the blood cells occurs, 
followed by a process of digestion entirely analogous to the digestion 
of similar blood elements by the various forms of intestinal hem- 
amebse. In the latter case it is a process of normal digestion, in the 
9 See Henneguy. “Les Insectes,” Paris, 1904, p. 677. 
10 Langhans. Virchow’s Archiv, Yol. 49, 1870. PHAGOCYTOSIS 
337 
former an emergency procedure carried out by virtue of the retained 
ancestral characteristics of the special phagocytic cells. 
The leukocytes, whose chief functions seem to be associated with 
such processes of intracellular digestion, may, therefore, be looked 
upon as cells retaining primitive characteristics for definite physio- 
logical purposes. We shall see, however, that, to meet exceptional 
conditions, the process of phagocytosis may be carried out also by 
many other cells which are associated ordinarily with functions en- 
tirely apart from this phenomenon. 
During normal life in higher animals, too, constant destruction 
of red blood cells by phagocytosis takes place in the spleen and liver, 
and is described by Dickson 11 as occurring in the bone marrow as 
well; and similar phagocytosis of red cells is seen in the hemolymph 
nodes. It is claimed by Metchnikoff, furthermore, that many of the 
degenerative and retrogressive processes which take place in the 
human body are carried on by the mechanism of phagocytosis. The 
rapid return of the puerperal uterus to the normal state is explained 
in this way, and work by Helme 12 seems to show that there is an 
actual phagocytosis of the hyperplastic uterine musculature during 
this period. The atrophic changes of senility, too, are attributed by 
Metchnikoff13 to the same processes. The involution of the 
ovaries is accompanied by active phagocytosis of portions of this 
organ, and Metchnikoff claims further to have shown that the de- 
generation of the nervous system during old age is accomplished by 
the phagocytosis of nerve cells by phagocytic elements derived either 
from the leukocytes or the neuroglia, or from both.14 The whitening 
of the hair, both in human beings and in old animals (dogs), is simi- 
larly due, he claims, to phagocytosis of the pigment by cells which 
wander in from the root sheaths. It is, up to the present time, im- 
possible to determine the stimulus by which this phagocytosis is 
initiated. 
Since the subject is a very important one, many studies have been 
made to determine which cells of the body of higher animals can 
take in and digest foreign particles and to classify them according 
to this power. Metchnikoff has distinguished between the “motile” 
and “fixed” phagocytes, the former the leukocytes of the circulating 
blood, the latter certain connective tissue cells, endothelial cells, 
splenic pulp cells, and certain cellular elements of the lymph nodes, 
11 Dickson. “The Bone Marrow,” Longmans, Green, London, 1908. 
12 Helme. Transact. Roy. Soc. of Edinburgh, Vol. 35, 1889. Cited from 
Metchnikoff. 
13 Matschinsky. Ann. de VInst. Past., Vol. 14, 1900; Vol. 15, 1901. 
14 That the leukocytes are concerned in the destruction and resorption 
of dead tissues has been shown by Leber especially (Leber, “Die Entstehung 
der Entziindung,” Leipzig, Engelmann, 1891). An accumulation of leukocytes 
about a bacterial focus or from any other stimulus is followed by tissue lysis 
due to leukocytic enzymes. 338 
INFECTION AND RESISTANCE 
the neuroglia tissue, and, in fact, all phagocytic cells which are 
ordinarily confined to some definite localization in the body. Among 
phagocytic cells Metchnikoff further distinguishes between “micro- 
phages,” by which he designates the polymorphonuclear leukocytes 
of the circulating blood and 
“mgcrophages.” The ma- 
crophages include the fixed 
cells mentioned above, to- 
gether with the large mono- 
nuclear elements of the 
blood, in short, all phago- 
cytic cells except the micro- 
phages. 
Although no absolute 
functional differentiation is 
possible between the two, it 
is true, in a general way. 
that the microphages are 
concerned primarily with 
the phagocytosis of bacteria 
and especially of those 
which invade acutely, while 
the macrophages are con- 
cerned especially with the resorption of cellular detritus, foreign 
bodies, and such bacteria as are more chronic in their activities, or 
are peculiarly insoluble 
On the other hand, micro 
phages may take up foreigr 
particles and bacteria of all 
kinds under suitable condi- 
tions, and no sharp line can 
be drawn between the twc 
varieties in this respect 
Metchnikoff further be- 
lieves that the two classes 
of phagocytic cells differ in 
the nature of the protective 
substances they secrete and 
furnish in the blood 
plasma. This, however, is 
a problem concerning which 
there is much difference of 
opinion and which calls for 
separate discussion in an- 
other place. 
The property of phairo- 
Polynuclear Leukocytes Taking up Sta- 
phylococci. 
Kupper Cells Containing Malarial Pig- 
ment . Diagrammatically Drawn 
from a Section of Malarial Liver 
Kindly Furnished by Dr. R. Lambert. PHAGOCYTOSIS 
339 
cytosis is therefore 
an attribute of a 
considerable num- 
ber of different va- 
rieties of cells. In 
the circulating 
blood the polynu- 
clear leukocytes are 
the most actively 
motile and phago 
cytic elements. The 
eosinopbile cells 
may also take up 
foreign particles 
and bacteria, as 
may also the large 
lymphocytes. The 
small lymphocytes 
and mast cells are 
either entirely inac- 
tive in this respect, or, at least, possess phagocytic powers under ex- 
ceptional circumstances only. This does not mean, however, that 
these last-named cells may 
not accumulate at the point 
of invasion nor that they 
may not play an important 
part in the defence of the 
body. It is well-known, of 
course, that, in tuberculosis 
and a number of other con- 
ditions, the lymphocytes 
may form the majority of 
the cellular elements which 
accumulate at the site of 
the lesion. Among the 
fixed cells of the body it is 
probable that phagocytosis 
may be carried on by cells 
of many different origins, 
though the identification of 
cells in tissues is often a 
purely morphological prob- 
lem, and therefore fraught 
with many possibilities of 
error. Probably the most active fixed tissue cells are the endothelial 
cells of the blood vessels and those which line the serous cavities, the 
Rat Leprosy Bacilli Grouped in the Remains of 
Dead Spleen Cells Growing in Plasma. 
Drawn after illustration in Zinsser and Carey, 
Journal of the A. M. A., Vol. 58, 1912. 
Phagocytosis of Sensitized Pigeon Cor- 
puscles by Alevolar Cells of Lung. 
Drawing’ made after photomicrograph pub- 
lished by Briscoe, Journal of Path, and 
Bad., Yol. 12, 1908. 340 
INFECTION AND RESISTANCE 
sinuses of the lymphnodes, and of the spleen. However, there are 
many other cells in addition to these which may be phagocytic. The 
writer, with Carey,15 has observed the active phagocytosis of leprosy 
bacilli by cells, probably of connective tissue origin, growing from 
plants of rat spleen in plasma. Phagocytosis by the cells lining the 
alveoli of the lungs has been observed by Briscoe.16 This author made 
the interesting observation that in cases of mild infection such cells 
can free the lungs of micro-organisms entirely without aid from the 
leukocytes of the circulating blood. It is these cells, too, which, in 
the ordinary conditions of life, take up the inhaled particles of dust 
and are, therefore, often spoken of as dust cells. The origin of the 
dust cells has often been the subject of controversy. In the embryo 
the alveoli of the lung, like the bronchi, are lined with columnar cells 
which are transformed into flattened epithelium as the alveoli ex- 
pand at the first inspirations after birth. These flattened cells, 
which constitute the alveolar or dust cells, are probably of epithelial 
origin, and as such are probably the only epithelial cells which act 
as phagocytes under ordinary conditions. Although no positive 
general statement is justified, we can yet say with reasonable accuracy 
that among the phagocytic fixed tissue cells the most important are 
the connective tissue and endothelial cells. 
The type of phagocy- 
tosis and the variety of cell 
which participates in it 
seem to depend to a great 
extent upon the nature of 
the substance which incites 
the process, or rather at 
which the process is aimed. 
Thus the large cells which, 
in tissues, take up the lep- 
rosy bacillus, those which 
are characteristic of tuber- 
culous foci, or those caused 
by blastomycetes, or by for- 
eign bodies, all have special 
appearances which are suf- 
ficiently characteristic to 
have diagnostic value. 
However, it is difficult to determine with certainty the origin 
of the cells which participate. The chemical nature of the substances 
taken up, moreover, often complicates the phagocytic process in such 
a way that different cellular elements are enlisted in succession in 
order that the ingested substances may be disposed of. Thus tubercle 
Giant Cell in Tuberculosis. 
15 Zinsser and Carey. Jour. A. M. A., March, 1912, Yol. 58. 
16 Briscoe. Jour. Path, and Bacter., Yol. 12, 1907. PHAGOCYTOSIS 
341 
or leprosy bacilli which are injected into an animal may be at first 
taken up by polynuclear leucocytes or microphages, by which they may 
even be carried into the lymph channels and distributed, perhaps to 
the detriment of the host. But these cells, probably because they lack 
a lipolytic ferment by means of which the waxes of the acid-fast or- 
ganisms can be digested, cannot destroy the bacteria, which are then 
attacked by other cellular elements at the site of their final deposit. 
That fixed tissue cells as a matter of fact play a very important 
role in the disposal of invading bacteria is becoming more and more 
clear. Kyes 17 showed a few years ago that the immunity of the 
pigeon to pneumococcus infection is largely due to active removal of 
the bacteria by the Kupfer cells in the liver. Bull’s 18 observations 
on the intravascular agglutination of typhoid bacilli which are then 
phagocyted by cells in the liver and spleen point in the same direc- 
tion, and recently Hopkins and Parker 19 in our laboratory have 
shown that streptococci injected into rabbits and cats are rapidly re- 
moved from the circulation, the removal being to a great extent due 
to phagocytosis carried on by the endothelial cells in the lungs and 
by similar cells in the liver and spleen. 
In many such cases the further resolution of the foreign sub- 
stance is accomplished by an important type of phagocytosis which is 
characterized by the formation of the so-called giant cells. These 
cells are of varying appearance in different conditions and locations. 
Thus the giant cells which form about foreign bodies, such as the 
small cotton fibers occasionally left in wounds, or injected particles of 
paraffin or iron splinters, etc., are quite characteristic and distinct 
from the giant cells of tuberculous foci, or of rhinoscleroma, glanders, 
or leprosy. They are all large cells, containing often numerous 
nuclei which form either by the fusion of several cells, as claimed by 
Borrell,20 Hektoen,21 and others, or by the cleavage of the nuclei 
alone, without coincident divisions of the cytoplasm. 
Although it is, of course, impossible to decide definitely upon 
purely morphological grounds, the researches of Hektoen especially 
would lead one strongly to favor the former view. It is equally 
difficult to decide the origin of giant cells, and endothelial, connec- 
tive tissue, and even leucocytic origin has been claimed for them. 
Yet in no case has it thus far been possible to actually observe their 
formation by a method which could positively decide this point. 
In order to gain a clear conception of the participation of phago- 
cytes in the response of the body to injury or invasion, it will be 
useful to follow out the process of inflammation as it occurs in the 
17 Kyes. Jour. Inf. Dis., Yol. 18, 1916, p. 272. 
18 Bull. Jour. Exp. Med., Yol. XX, p. 237. 
19 Hopkins and Parker. Paper in Manuscript. 
20 Borrell. A nn. de VInst. Past. 7, 1893. 
21 Hektoen. Jour. Exp. Med., 3, 1898, p. 21. 342 
INFECTION AND RESISTANCE 
higher animals. Inflammation may be incited by a large number of 
agencies—chemical irritants, mechanical injury, or even by the in- 
troduction of inactive and isotonic substances such as broth or salt 
solution.22 Yet in these cases the response, though essentially 
similar in principle to that following invasion by bacteria, lacks 
certain features especially interesting in the present connection, and 
it will be most profitable for our purpose to consider in detail the 
result of infection with pathogenic micro-organisms. 
If an emulsion of pyogenic staphylococci is injected into an ani- 
mal subcutaneously the site of injection will soon become reddened 
and swollen and microscopic examination will show, within a few 
hours, a swelling and engorgement of the blood vessels. 
The injected cocci will be found to lie partly scattered in the 
tissue spaces, in part within polynuclear leukocytes and connective 
tissue cells which have begun to ingest them. The tissue space 
will be swollen and 
stretched by the 
exudation of blood 
serum from the ves- 
sels. This condi- 
tion will begin in 
from 4 to 6 hours 
after injection and 
increase during the 
next 24 hours in ex- 
tent and severity, 
according to the 
quantity and viru- 
lence of the cocci 
injected. The con- 
ditions which pre- 
cede the wandering 
of the p o 1 y m o r- 
phonuclear leuko- 
cytes out of the ves- 
sels have been care- 
fully studied in such thin tissues as the mesentery of a frog after 
injury by trauma or acid. Within the vessels of the affected area 
there is at first an acceleration of the blood stream, then a dilatation 
of the capillaries and a slowing of the current. Leukocytes may now 
be observed moving more slowly than the main stream, and keeping 
close to the periphery along the walls of the vessels. Here and there 
they seem interrupted in their movements and adhere to the vascular 
wall. A little later these cells appear to pass through the wall of the 
vessel by sending out pseudopodia which slowly penetrate it. Adami 
Diagrammatic Representation of Leukocytes 
Wandering Through Capillary Walls. 
Adapted from Ribbert, “Lehrbuch der Allgemeinen 
Pathologic,” p. 337. 
22 Adami. “Inflammation,” Macmillan, London, 1909. PHAGOCYTOSIS 
343 
states that if, at this stage, the tissues be excised, fixed in osmic 
acid, and stained, leukocytes may be seen crowding the inner sur- 
face of the vessel in all stages of transition from its anterior to 
the lymph spaces on the outside. 
In the staphylococcus infection, after from 12 to 48 hours, there 
will be seen the results of an active and destructive struggle between 
the invading bacteria and the defending cells. In the center of the 
area of invasion tissue has been destroyed and disintegrated. Amid 
the necrotic detritus, closely packed, lie leukocytes and cocci and 
active phagocytosis has taken place. In some cases the intracellular 
bacteria appear swollen and disintegrating, in others the leukocyte 
itself, overcome by the larger number of bacteria it has taken in, 
becomes vacuolated, indefinite in outline, and apparently is being 
itself destroyed. The presence of blood serum, which is aiding in 
the destruction of bacteria both by its bactericidal powers and its 
reenforcement of the phagocytic process, renders this mass fluid or 
semi-fluid, and the whole mixture constitutes what is known as pus. 
Around the periphery cocci and leukocytes become more scattered 
and sparse, and bacteria, together with leukocytes, loaded with cocci, 
may be seen lying within large mononuclear cells (macrophages). 
Whether the process goes on to further extension or is eventually 
walled off into a distinct abscess by the formation of granulation 
tissue and new connective tissue depends upon the balance of forces 
between attacking agent and defensive factors. 
If we inject a similar emulsion of cocci into the pleural or peri- 
toneal cavity of an animal a process similar in principle may be 
observed. 
formally the peritoneum contains a small amount of this serous 
fluid and a moderate number of white blood cells, chiefly lympho- 
cytes. When any substance, broth or salt solution, an aleuronat or 
a bacterial emulsion, is injected into the peritoneal cavity, there 
follows a brief period during which there is a diminution of the 
free cellular elements in the peritoneal fluid. At this time there is 
a clumping of cells in the folds of the omentum and mesentery, a 
transient stage of flight away from the point of injury. This, how- 
ever, is soon over. Within one to two hours an active immigration 
of leukocytes into the serous cavity occurs and if, during the next 
12 to 24 hours small quantities of fluid are, from time to time, with- 
drawn with a capillary pipette, a rapid and constant increase of 
leukocytic elements, chiefly of the microphage or polynuclear type, 
is observed. If the injected substance has been a sterile, harmless 
fluid, a gradual return to normal within 48 hours then ensues. If, 
however, we have injected bacteria, a struggle similar to the one 
described above takes place within the peritoneum, and active 
phagocytosis of the micro-organisms takes place. 
Let us suppose that the injected bacteria have been small in 344 
INFECTION AND RESISTANCE 
quantity and moderate in virulence. In such a case a rapid phago- 
cytosis gradually rids the fluid of micro-organisms and within 24 
hours after injection few, if any, free bacteria are visible. 
A little exudate taken at this time shows large numbers of inicro- 
pliages varyingly crowded with well-preserved and disintegrating 
bacteria. Some of the phagocytes, having literally taken up more 
than they can digest, are vacuolated and disintegrating, but, in gen- 
eral, the victory lies with the cells. A little later large mononuclear 
elements appear, and here and there will be seen to take up dead 
leukocytes together with ingested cocci. In this way gradually a 
cleaning out of the peritoneum takes place, the animal recovers, and 
the peritoneum returns to normal. 
Let us suppose, on the other hand, that the bacteria injected are 
in larger doses and of greater virulence. In such a case, after a 
period of active phagocytosis, there may be a gradual increase of 
bacteria over leukocytes. The phagocytic cells are found to be under- 
going degeneration in larger numbers, the free bacteria increase, 
and the impending death of the animal can often be foretold by the 
appearance of the exudate. Finally, the peritoneal fluid may con- 
sist chiefly of free and rapidly multiplying bacteria with a practical 
absence of phagocytic cells. 
In all of the processes so far as described the burden of the 
defence has fallen upon the microphages or polynuclear leukocytes, 
while the macrophages—endothelial and connective tissue cells— 
have taken a purely secondary part in the reaction, forming, to some 
extent, a second line of defence, or, more probably, taking part only 
in the final removal of degenerated and disintegrating combatants 
and tissue detritus. In order to obtain a complete conception of 
phagocytosis in its entire significance it will be necessary to consider 
a further example, namely, the process which takes place within tis- 
sues in the course of the efforts of macrophages to remove bacteria 
and other substances which, either because of their insolubility or for 
other unknown reasons, are refractory to the attacks of the mi- 
crophages. Since we are interested in this subject chiefly from the 
point of view of the defence against bacteria, we may illustrate this 
process best by the description of the reaction which takes place when 
tubercle bacilli become localized anywhere within the animal body. 
When tubercle bacilli are injected into the peritoneum they are 
actively taken up by the polynuclear leukocytes just as are other 
bacteria and many entirely inactive solid particles. A similar inges- 
tion by microphages may take place in the folds of the intestinal 
mucosa if tubercle bacilli are fed to guinea pigs. However, this 
preliminary phagocytosis is probably of but secondary significance 
in the combat of the body against tuberculosis, since it has still to 
be shown that polynuclear leukocytes are capable of digesting and 
destroying acid-fast bacilli. Indeed, much evidence tends to show PHAGOCYTOSIS 
345 
that the ingestion of tubercle bacilli by microphages may be a detri- 
ment to the host, since the bacilli by this means are carried through 
the lymphatics and variously distributed throughout the body. Poly- 
nuclear leukocyte extracts, though containing, as we shall see, pro- 
teolytic enzymes, do not, according to Tschernorutzky, contain any 
lipase, and it may well be that for this reason they are unable to 
attack the waxy substances which form an integral part of these or- 
ganisms. This is in keeping with the observations made by Terry 
in our laboratory, that rat leprosy bacilli may be kept within leu- 
kocytes for weeks without losing their acid-fast properties, whereas 
the same bacilli, as the writer and Cary found, were rapidly disin- 
tegrated in spleen cells growing in plasma. Moreover, it is well 
known that the estimation of tuberculo-opsonin contents of the sera 
of tuberculous patients has been peculiarly unsatisfactory in throw- 
ing light on the progress of the disease. It would seem, therefore, 
that in this disease, as well as in others caused by acid-fast organ- 
isms, the microphages play only an unimportant part in the defence 
of the body. 
On the other hand, when tubercle bacilli are deposited either in 
a lymphnode (through the vehicle of leukocytes) or in a capillary 
anywhere by the blood stream, a train of cellular changes is initiated 
in which the predominant part is played by the macrophages. The 
tubercle bacilli so deposited are rapidly surrounded by large mono- 
nuclear cells, probably endothelial in origin. Some of the micro- 
organisms may even be phagocyted and taken into these cells. These 
cells, spoken of as “epithelioid cells,” surround the clump of bacteria 
in more or less concentric rings, and around these there is an accumu- 
lation of leukocytes, largely of the lymphocyte variety, with an ad- 
mixture of a very few microphages. Then by the fusion of endothe- 
lial cells, or possibly by division of the nuclei of some of these cells 
within the individual cell bodies, giant cells are formed which take 
up the bacilli. The further progress of the tubercle now greatly 
depends upon the balance of power. Often such a tubercle may 
heal, possibly because of complete intracellular digestion of the ba- 
cilli. On the other hand growth and multiplication may lead to a 
slow and dry necrosis of the center of such a mass of cells, leading 
to the condition spoken of as caseation. Epithelioid cells lose their 
outlines and staining properties, and go to pieces. The center of the 
lesion is a grumous mass, the periphery shows a few giant cells and 
connective tissue proliferation. 
It is always surprising to those who study these lesions for the 
first time how rarely they succeed in finding tubercle bacilli in 
microscopic sections prepared from such tubercles by the ordinary 
Ziehl-Eeelsen method of staining. Repeated and careful examina- 
tion of such material may fail to reveal any acid-fast organisms, 
though inoculation into guinea pigs is nevertheless successful, pro- 346 
INFECTION AND RESISTANCE 
ducing typical tuberculosis. Much 23 has studied this peculiar state 
of affairs particularly and has shown that, although such lesions may 
show no tubercle bacilli by the ZiehKNeelsen carbol-fuchsin method, 
staining by a modified Gram technique will reveal numerous Gram- 
positive rods and granules which have lost their acid-fast properties. 
This, too, if true, and the evidence is very much in its favor, would 
point to an ability of the macrophages to digest the waxy substance 
of the tubercle and other acid-fast bacilli, a property not possessed 
by the microphages. It may, of course, mean on the other hand that 
the tubercle bacilli in the lesion have not developed the waxy condi- 
tion. 
Chemotaxis and Leukocytosis 
The part played by the phagocytic cell in the defence of the body 
against the entrance of bacteria and other foreign substances consists, 
then, of two functionally different phases. The first is an active 
motion of the cells toward the point attacked, and their accumulation 
about the noxious agent, the second consists in the act of ingestion 
itself. 
The motion of the leukocytes toward the invading substances 
indicates a sensibility on the part of the cell to changes in its environ- 
ment incited by the foreign agent, and since the stimuli most likely 
to reach the leukocytes and bring about this alteration in the direc- 
tion of their movements are chemical in nature, the phenomenon is 
spoken of as “chemotaxis.” This term was borrowed from Pfeffer,24 
who studied similar phenomena in connection with many freely 
motile plant cells, spermatozoa, and bacteria. Since the change of 
direction brought about in a moving cell by such influences may be 
such as either to attract or to repel, the term “positive chemotaxis” 
is used to designate the former and that of “negative chemotaxis” 
the latter. 
The property of chemotaxis is of vital interest in the present 
connection, since, whatever may be our opinion regarding the relative 
values of phagocytosis and serum protection in immunity, the great 
importance of the phagocytic process cannot be questioned, and any 
agency which repels the approach of the phagocytes must be a detri- 
ment, while any factor which attracts them is, of necessity, a power- 
ful means of defence. In the investigations upon the nature of infec- 
tious diseases attention has been concentrated upon the phenomenon 
of phagocytosis, and the relations governing the act of ingestion 
have been very thoroughly studied. The details of the chemotactic 
phenomenon, however, though of equal importance, are much more 
23 Much. “Beitrage zur Klinik der Tuberk.,” Yol. 8, 1907, Hft. 1 and 4. 
24 Pfeffer. “Untersuch. a. d. Botan. Inst. Tubingen,” Yols. 1 and 2, 1884 
and 1888. PHAGOCYTOSIS 
347 
obscure. A large part of our sparse knowledge in this connection, 
moreover, has been gained by studies not related to infection. 
The stimuli which determine the motion of cells are, of course, 
not necessarily chemical, and extensive studies have been made upon 
the effect of light waves in this connection. Although these inves- 
tigations are of great biological importance, they have little direct 
bearing upon the problems of tropism as related to bacteria and leu- 
kocytes and cannot therefore be considered here. 
Some of the earlier researches upon chemotaxis were those made 
by Stahl 25 upon the slime-molds or myxomycetes. These organisms 
possess the power of ameboid motion, and were observed by Stahl to 
move toward or away from any given region, according to the nature 
of the substances with which they came in contact. Pfeifer sub- 
jected this phenomenon to closer analysis. Working with the sperma- 
tozoa of ferns, swarm spores, bacteria and infusoria, he elaborated 
an ingenious technique by means of which he was enabled to de- 
termine directly the negative or positive chemotactic properties of 
various substances in solution upon these motile forms. His tech- 
nique was exceedingly simple. Capillary glass tubes, about 8 to 10 
mm. long and 0.1 mm. in diameter, were sealed at one end in the 
flame, and then dropped into a watch-glass. The solution which was 
to be tested was poured over the tubes and the watch-glass then 
placed under the bell of an air-pump. When the air was evacuated 
and pressure reduced the tubes became partly filled up with the 
liquid. They were then removed, washed in water, and placed under 
a cover slip under which a preparation of the motile cells was swim- 
ming. Positive chemotaxis was indicated by entrance of the cells 
into the tubes, negative, by their refusal to enter. Failure of the 
solution to exert any chemotactic influence resulted in their moving 
into and out of the tubes indiscriminately.26 
By this technique a large number of interesting observations were 
made which threw much light upon the causes underlying the move- 
ments of plant cells. For instance, in investigating the spermatozoa 
of the ferns it was found that they were attracted strongly by malic 
acid and its salts, while no other substance investigated approached 
these compounds in the intensity of positively chemotactic stimula- 
tion. From this Pfeffer concludes that the bursting of the fern 
archegonia is accompanied by the liberation of malic acid, this at- 
tracting the male to the female cell. 
Similar experiments have been carried out since then by numer- 
ous naturalists, among them Buffer,27 Lidforss,28 and Jennings,29 
25 Stahl. Botanische Zeit., 1884. 
26 Buller. Annals of Botany, Yol. 16, No. 56, 1900. 
27 Buller. Loc. cit. 
28 Lidforss. “Jahrbiicher f. wissensch. Botanik,” 41, 1904. 
29 Jennings. “Behavior of Lower Organisms,” Columbia Univ. Press, 
Macmillan, 1906. 348 
INFECTION AND RESISTANCE 
and it lias been found that in addition to malic acid compounds many 
other substances, organic and inorganic, occurring in plant cells and 
cell-sap exert positive chemotactic power. Lidforss has shown, for 
instance, that calcium elilorid in 0.1 per cent, solution may strongly 
attract plant spermatozoids (equisetum—horsetail). When the solu- 
tion is concentrated to 1 per cent., attraction is still exerted, but the 
spermatozoids immediately lose their motility upon entrance into the 
fluid. 
The same worker has shown that a substance which is positively 
chemotactic for one variety of plant cell may be negatively chemo- 
tactic for another, showing a certain selective variation which should 
be of great biological importance. Thus capillaries with a 1 per cent, 
solution of potassium malate actively attracted the spermatozoids of 
marchantia (a liverwort), while not a single spermatozoid of equi- 
setum would enter these tubes. Low 30 has applied these methods of 
study to the investigation of the chemotaxis of mammalian sperma- 
tozoa and found that these cells were actively attracted by weakly 
alkaline solutions. 
Studies upon the factors determining the movement of bacteria 
and amebse toward some substances and away from others have been 
numerous, and are valuable for the understanding of leukocytic 
chemotaxis, because they have led to the formulation of a number of 
important general theories. The fact that the motions of bacteria 
in suspensions are, to a certain extent, determined by the negative 
electrical charge which they all carry in neutral media, has been 
touched upon in the section on agglutination. Attempts on the part 
of Young and the writer to determine whether the attraction of 
leukocytes toward bacteria might he due to the carrying of an elec- 
tropositive charge by the white cells have met with no result, owing 
so far to the failure to elaborate a reliable technique. However, this 
thought is not an impossible one and should be borne in mind. 
That certain bacteria will Avander actively toward a source of 
oxygen was shown by Engelmann’s 31 classical experiment in which a 
diatom, half in the shade and half in the light, was surrounded by an 
emulsion of bacteria, and these were seen to collect about the lighted 
half only, where oxygen was being liberated by virtue of the chloro- 
phyll. The extreme delicacy of chemotactic reactions is illustrated 
in these experiments in that Engelmann calculated that the bacteria 
reacted to one one-liundred billionth of a milligram of oxygen. The 
selective reaction of bacteria to various chemical substances, further- 
more, has been shown by allowing different solutions to diffuse into 
bacterial emulsions from capillary tubes, and by observing attraction 
or repulsion from the point of contact. 
The chemotaxis of leukocytes has opposed more difficulties to 
30 Low. Sitzungs Berichte kais. Akad. d. TFiss., Wien, Yol. 3, Abf. 3. 
31 Engelmann. Arch. f. d. ges. Physiol., Yol. 57, p. 375. PHAGOCYTOSIS 
349 
direct study, since the conditions within the living body are subject 
to a large number of modifying factors, and experiments upon the 
isolated cells, in vitro, even under conditions of the most careful' 
technique, are fraught with much unavoidable injury to the cells. 
However, enough has been learned to indicate that these cells are 
subject to the phenomena of chemotaxis or tropism just as are inde- 
pendent unicellular forms, and that they may be attracted or re- 
pelled by a variety of organic and inorganic substances. Leber 32 
was one of the first to study this in his work upon inflammation. 
He found that leukocytes were actively attracted by powdered cop- 
per and mercury compounds, but not by powdered gold or iron. He 
also observed that dead bacteria exerted a similar positive chemo- 
tactic influence, and Buchner 33 later succeeded in extracting sub- 
stances from various bacteria which possessed similar properties. It 
appears, from these and other investigations, that the power of stim- 
ulating positive chemotaxis is a general property of bacterial pro- 
teins, equally evident in bacterial extracts, dead bacteria, or the 
living organisms. It is likely, therefore, that the attraction of leu- 
kocytes toward the point of bacterial invasion is, in part at least, 
due to the properties of the bacterial proteins themselves. That this, 
however, is not the whole story is evident from the work of Massart 
and Bordet,34 who showed that the products of cell destruction and 
disintegration possess similar positively chemotactic properties. This 
is true not only of the products of disintegrated tissue cells, but of 
those of the destroyed leukocytes themselves. Thus it appears that 
when any injury of tissue takes place, a stimulus which attracts 
leukocytes results, even when the injury is not accompanied by bac- 
terial invasion. This would explain the participation of leukocytes 
in reactions to injury, and in inflammations not of bacterial origin, 
and their local accumulation following the injection of insoluble 
inorganic substances. 
When bacteria are actually present, however, the added stimulus 
due to the diffusion of bacterial proteins probably increases the 
process to a degree often sufficient to meet the added requirements 
for protection. Following this, both the destruction of tissues, of 
bacteria, and of leukocytes may together exert a cumulative chemo- 
tactic power which continues the process proportionately with the 
extent of the lesion. 
It is of the utmost importance, therefore, to ascertain whether 
or not any substances derived from bacteria may, under any circum- 
stances, exert a repellent or negatively chemotactic power. If we 
infect an animal intraperitoneally with virulent bacteria, in doses 
32 Leber. Fortschr. der Med., 1888; also “Die Entstehung der Ent- 
ziindung,” Engelmann, Leipzig, 1891. 
33 Buchner. Berl. klin. Woch., Yol. 27, No. 30, 1890. 
34 Massart and Bordet. Ann. de VInst. Past., Yol. 5, 1891. i 350 
INFECTION AND RESISTANCE 
sufficient to lead to death, and examine the peritoneal exudate just 
before the lethal outcome, we may observe that leukocytes are gradu- 
ally disappearing, and that finally but a few will be present and the 
fluid will be swimming with free micro-organisms. In the same way 
it is well known that the diminution of leukocytes in the circulating 
blood—or even the failure of these cells to increase in the circulation 
in the course of such diseases as pneumonia, or general infections 
with staphylococci or streptococci—is seriously prognostic of fatal 
outcome. The conditions here observed point strongly to the ex- 
istence of substances of negative chemotactic influence which protect 
the bacteria, not from phagocytosis itself, but from that necessary 
forerunner of phagocytosis, the approach of the leukocyte. It is 
necessary to draw this distinction since these phenomena are not 
merely, as often believed, “antiopsonic,” but in truth largely “anti- 
chemotactic.” It is true that Kanthack,35 and more especially 
Werigo,36 have denied the existence of negatively chemotactic bac- 
terial products, the latter basing his assertion upon the observation 
that active phagocytosis occurs in the lungs, liver, and spleen of 
animals dying of infection with virulent germs. However, the argu- 
ments of these authors are not conclusive and a mass of experi- 
mental and clinical evidence which points to a direct failure of 
leukocyte accumulation in the presence of virulent bacteria in the 
animal body would alone suffice to render such conclusions unlikely. 
Moreover, strong evidence in favor of the existence of negatively 
chemotactic influences is brought by the extensive experiments of 
Bail upon the so-called aggressins, discussed in another place, and 
such observations as those of Vaillard and Vincent 37 and Vaillard 
and Kouget,38 which showed that the injection of a little tetanus 
toxin together with tetanus spores would prevent the ingestion of the 
spores by leukocytes, and thereby furnish an opportunity for germi- 
nation and consequent fatal toxemia. 
Similar observations have been made by Besson 39 in the case of 
the bacillus of malignant edema by the use of the original technique 
of Pfeiffer. Capillary tubes containing the toxin remained free of 
leukocytes after subcutaneous introduction into guinea pigs, while 
similar tubes containing the culture medium alone, or the bacilli and 
their spores, attracted leukocytes in considerable numbers. 
It is possible, of course, to interpret such phenomena as due to a 
failure of positive chemotaxis rather than to an active negative 
chemotaxis. 
Although the phenomena of chemotaxis are most easily studied 
35 Kanthack. Quoted from Adami, loc. cit. 
36 Wei'igo. Ann. de I’Inst. Past., Yol. 8, 1894. 
37 Yaillard and Vincent. Ann. de VInst. Past., Yol. 5, 1891. 
38 Vaillard and Rouget. Ann. de VInst. Past., Yol. 6, 1892. 
39 Besson. Ann. de VInst. Past., Yol. 9, 1895. PHAGOCYTOSIS 
351 
in extravascular inflammatory changes, there is none the less a 
regular and apparently purposeful attraction or repulsion of leuko- 
cytes evident in the circulating blood during infectious diseases. 
That infection of the body with many micro-organisms results in the 
increase of leukocytes, and that in others there is either no increase 
or even a decrease, is too well known and too generally applied in 
diagnosis and prognosis to warrant our giving up much space to a 
review of the facts. Nevertheless, the causes which lead to a leuko- 
cytosis in the one case, a leukopenia in the other, are still very 
obscure and deserve discussion. 
In the first place it is by no means certain whether a leuko- 
cytosis signifies an active discharge of new leukocytes from the bone 
marrow or whether it means simply an altered distribution in that 
the phagocytes accumulated in the lymphatic and other organs are 
attracted by chemotaxis into the peripheral circulation. Studies of 
the bone marrow during infection as well as the occasional appear- 
ance of myelocytes and other cells ordinarily found only in the bone 
marrow during health would point toward a participation of active 
bone-marrow hyperplasia in the increase of peripheral leukocytes. 
There is no good reason to doubt, moreover, that a chemotactic stimu- 
lus exercised in the circulation should withdraw leukocytes from 
any place of accumulation to the circulation. Probably both proc- 
esses take part. When bacteria are injected into the circulation of 
an animal there is, at first, a moderate diminution of the leukocytes 
just as there is after injection of bacteria or other substances into 
the peritoneum. This is soon followed in most cases by a rapid and 
progressive increase, in which, whenever the leukocytosis is one of 
considerable degree, the polynuclear leukocytes preponderate. The 
extensive clinical study of the white cells in infectious disease of the 
human being give us more material for reasoning in this respect 
than we have available from animal experiment. Infection with 
invasive bacteria such as the pneumococcus (and Neufeld and others 
have shown that most lobar pneumonias are accompanied by pneu- 
mococcus bacteriemia), streptococci, staphylococci, and others is 
always accompanied by an increase of the leukocytes, while, in 
typhoid fever, influenzal infection, tuberculosis, and a number of 
other infections, the leukocytes do not increase and may even de- 
crease. How are we going to account for this ? That all these bac- 
teria contain a substance which is positive in its chemotactic effects 
is easily demonstrated by injecting them into the peritoneum and 
observing an accumulation of leukocytes and a consequent phago- 
cytosis, even in the cases of those organisms which do not call forth 
a leukocytosis in the blood of the diseased human being. Thus it 
has been our experience as well as that of others invariably to ob- 
serve the rapid and complete polynuclear phagocytosis of both 
leprosy bacilli and tubercle bacilli after injection of these micro- 352 
INFECTION AND RESISTANCE 
organisms into the peritoneal cavities of guinea pigs. Yet a chronic 
tuberculous peritonitis or pleurisy is characterized usually by an 
exudate which contains but few polynuclears and relatively many 
lymphocytes. A final explanation of these conditions is not pos- 
sible at present. No adequate explanation for the selective accu- 
mulation of lymphocytes and the absence of polynuclear cells about 
tuberculous foci has yet been advanced. The absence of polynuclear 
leukocytosis may possibly be due to the great insolubility of these 
bacilli, in consequence of which little or no leukocytosis-stimulating 
substances are liberated. 
Pearce 40 has suggested a similar reason for the absence of poly- 
nuclear accumulations about chronic localized lesions of any kind 
in which tissue encapsulation may prevent the contact of the inciting 
agents with the body fluids and there is a consequently slow or slight 
production of such chemotactic stimulating materials. 
In typhoid fever, where the slight primary leukocytosis is rap- 
idly succeeded by a leukopenia with a relative lymphocytosis, the 
conditions are somewhat different. Here, as in some other infec- 
tions, as Friedberger and others have shown, we are dealing with a 
generalized infection by an organism which is easily subject to the 
action of alexin with consequent production of anapliylatoxin. (See 
chapter on Anaphylaxis.) This poison, it seems, exerts a nega- 
tive chemotaxis, and probably during the height of the disease, 
therefore, leads to the low leukocyte count observed. That this is 
at least likely seems to follow from the studies which have been 
made upon the nature of the typhoid poisons, and also from the 
observation of Gay and Claypole, that typhoid-immune rabbits react 
to the infection of typhoid bacilli with a rapid and powerful increase 
in the polynuclear leukocytes, whereas similar injections into the 
normal animal lead to leukopenia. 
If the supposition regarding tuberculosis, made above, is correct, 
it would follow that a sudden and considerable increase in the poly- 
nuclear leukocytes in a case of tuberculosis would indicate a dis- 
charge of organisms into the circulation and a tendency toward gen- 
eralization of the infection in this manner. (See Weigert’s view of 
the manner in which tuberculosis may spread by the destruction of 
the wall of a vein by a localized lesion.) However, although specu- 
lation in the absence of experimental proof is justified, it must not 
be forgotten that the problems of selective chemotaxis are too ob- 
scure to permit of our laying much weight on any of these views. 
Gabritchewsky,41 who investigated this subject extensively, has 
classified various substances according to their positively, negatively, 
or neutral chemotactic activities. It is not necessary to recapitulate 
these, but it is interesting to note that he found that some substances 
40 Pearce. Jour. A. M. A., Yol. 61, 1913. 
41 Gabritchewsky. Ann. de VInst. Past., Yol. 4, 1890 PHAGOCYTOSIS 
353 
which were positively chemotactic in certain concentrations became 
neutral or even negative when the concentration was altered. 
We have seen that the action of the leukocytes in moving toward 
some substances and away from others is entirely analogous to simi- 
lar phenomena occurring among lower, unicellular forms of life, and 
the explanations applied to the apparently conscious acts of the 
ameba, such as the motion toward and the engulfment of food, have 
been applied to the activities of the leukocytes as well. Many of 
the theories developed concerning the free living forms, however, 
have been easily excluded in the case of the leukocytes, because of the 
environment in which their activities are developed. Thus the many 
interesting reactions of paramecia and other organisms to light 
(heliotropism) have little bearing upon this subject, and the views 
based on the theories of orientation may be excluded on the ground 
of the symmetry of the normal leukocyte. The observations of 
Garrey,42 that indicate that it is the dissociated ions of various acids 
and bases which are responsible for the directive stimuli exerted 
upon certain flagellates, may yet result in throwing some light upon 
leukocytic movements, especially if we can come to accept the con- 
ceptions of ion-proteins upheld by Loeb 43 and his pupils. However, 
the facts concerning these phenomena, as well as the possibility, 
previously mentioned, of the opposite electrical charges carried by 
the leukocytes and the substances attracting them cannot be regarded 
at present as more than interesting thoughts. Of more than merely 
speculative interest, however, are the views of chemotaxis which are 
based upon the study of conditions of surface tension. In order to 
consider these properly it will be useful to review briefly the funda- 
mental principles governing these conditions. 
The molecules of any fluid are held together by mutual attraction 
due to the force generally spoken of as cohesion. This force is ex- 
erted by like molecules upon each other in solids more strongly than 
in liquids, and in gases less strongly. Since we are dealing in this 
connection with occurrences taking place in liquids, we will restrict 
our consideration to these. The force of cohesion is influenced in 
a number of ways. Thus, for instance, heat reduces it, and this 
is the cause that solids are converted into liquids and liquids into 
gases, provided of course that the heat brings about no chemical 
change. In large masses of fluids the force of gravitation over- 
comes that of cohesion and larger masses of liquid assume the shape 
of the containing vessel. In smaller masses the force of cohesion 
tends to bring about the spherical shape. This comes about in 
the following way: Within the interior of a drop of liquid all the 
molecules attract each other, and since the force of attraction is 
equal in all directions it neutralizes itself, and the molecules are 
42 Garrey. Am. Jour. Phys., 3, 1900. 
43 Loeb. Am. Jour. Phys., 3, 1900. 354 
INFECTION AND RESISTANCE 
uninfluenced by it, mobile and free. The molecules on the surface 
are in a different condition, however. They are subjected to the 
force of cohesion from -within, but not from without, and are there- 
fore drawn strongly toward the center. The result is the same as 
though the drop were subjected to pressure from without and the 
surface layers were in a state of compression. There is in conse- 
quence a constant tendency of all the surface molecules to be drawn 
toward the center and a resulting tendency to a diminution of the 
surface area. It is as though the surface of such a drop were a thin, 
elastic membrane which tended to contract and diminish in size and 
surface. The force with which this takes place is spoken of as sur- 
face tension,44 and the energy underlying it is called, by Ostwald, 
surface energy. Since a drop of one fluid suspended in another with 
which it cannot mix is relieved of the disturbing factor of gravitation, 
its surface tension has the effect of contracting the small mass into a 
form which, for the given volume, will expose the smallest possible 
surface, and this is, of course, the sphere. It is for this reason that, if 
-we shake up such systems as water and chloroform, or oil and water, 
the chloroform or the oil will be distributed through the water as 
small droplets. The degree of surface tension of any fluid is meas- 
urable by a number of reasonably accurate methods which may be 
found in any text-book of physics and which we need not consider 
here. It is of course dependent in each case upon the nature of the 
surrounding medium. We have taken into consideration above only 
the force which is exerted within the drop by the cohesion, that is, 
the attraction toward the center. This would be uninfluenced from 
without only in a vacuum. In nature the surface molecules, though 
forcibly drawn toward the center, are also affected from without by 
the attraction exerted by the molecules of the substances surrounding 
the drop. There is a constant balance, therefore, at any part of the 
surface of a drop of fluid between the cohesion tension from within 
and attractions from without. The resultant of the two forces de- 
termines the surface tension, which will be greater or less in inverse 
ratio to the attraction from without for any given drop, and a varia- 
tion of the external attraction at different points on the periphery of 
the drop will naturally influence the shape of the drop. For a relief 
of attraction at one point would tend to permit that part of the sur- 
face to retract, and an increase in this attraction would tend to 
allow it to bulge, with the formation of a sort of pseudopod. 
In studying the importance of surface tension 45 in determining 
the motions of unicellular organisms a number of important attempts 
have been made to imitate cell motion by means of the suspension of 
various substances of strong cohesive properties in liquid media. The 
44 Michaelis. “Dynamik der Oberflachen,” Steinkopf, Dresden, 1909. 
45 For a thorough discussion of this phenomenon see also Gideon 
“Chemical Pathology/’ Saunders, 1911. PHAGOCYTOSIS 
355 
idea was suggested by Quincke,48 and later by Biitschli,47 but has 
been most extensively studied by Rhumbler.48 The result has been 
the production of a number of “artificial amebse” which in almost 
all respects behave like the living organisms. Thus if a small mass 
of mercury is placed into a dish filled with water acidified with 
nitric acid, and a small crystal of dichromate of potassium is dropped 
near the mercury, the dichromate will dissolve and a yellow cloud 
will gradually diffuse from it toward the mercury. As soon as the 
yellow cloud touches this it will begin to show change of form and 
to elongate in the direction of the dichromate, often moving to- 
ward it. The motion of the quicksilver will resemble with con- 
siderable accuracy that of an ameba moving toward a particle of 
food or sending out pseudopodia. A more striking and complete 
imitation is that obtained by Rhumbler when he placed a drop of 
clove oil into a mixture of alcohol and glycerin. The changes of sur- 
face tension produced upon the surface of the clove oil by the alcohol 
give rise to movements in the oil entirely analogous to those of mo- 
tile cells in favorable media. The similarity has been extended 
even to the processes of engulfment of the food as observed among 
amebse. Thus a drop of chloroform in water will flow about a 
particle of shellac and dissolve it. If a piece of glass coated with 
shellac is placed in contact with the drop it will engulf it, 
but will cast out the glass after the shellac coating has been dissolved 
away. 
The similarity between phenomena purely referable to surface 
tension and those taking place in the living cells is therefore very 
striking and has been clearly analyzed in regard to its bearing 
upon leukocytic chemotaxis by Gideon Wells in his “Chemical Path- 
ology.” The chemotactic substances, diffusing to the leukocyte, will 
lower its surface tension on the side at which they come in contact. 
Pseudopodia will be thrown out on this side in consequence, and the 
leukocyte will move in this direction. The motion will be continued 
in this direction as long as the concentration of the chemotactic sub- 
stance, and therefore the diminution of surface tension is greater on 
this side than on other parts of the periphery, until a point is reached 
at which the chemotactic substance is equally diffused on all sides, and 
motion will cease. The actual engulfment may then occur or the 
nature and concentration of the chemotactic substance may be so 
great that injury is done to the leukocyte. Whether or not the purely 
physical explanation of chemotaxis tells the whole story it is of course 
not possible to decide. At any rate, it furnishes a rational basis for 
46 Quincke. Quoted from H. G. Wells, “Chemical Pathology,” Saunders, 
1907. 
47 Biitschli. “Untersuch. iiber mikroskopische Schaume und das Proto- 
plasma.” Leipzig, 1892. See also H. G. Wells, loc. cit., pp. 220 et seq. 
48 Rhumbler. Arch. f. Entwickelungs Mechanik, 1898, 356 
INFECTION AND RESISTANCE 
the study of the phenomenon more promising than any of the others 
so far offered. 
It is true, on the other hand, that such a theory in no way ac- 
counts for the apparently selective positive chemotaxis which is 
exerted by different substances. Thus the preponderance of poly- 
nuclear leukocytes in foci and serous cavities containing organisms 
like staphylococci, meningococci, streptococci, and others is in con- 
trast to the lymphocytic accumulation in the pleural, subarachnoid, 
and peritoneal spaces infected with tubercle bacilli. Some writers 
have spoken, therefore, of active and passive leukocytosis according 
to whether or not the cells attracted seemed to possess ameboid 
motility. That surface tension phenomena alone do not account for 
this is clear. But it must be remembered that even tubercle bacilli, 
though eventually attracting few polynuclears and many lympho- 
cytes, will cause an active polynuclear accumulation in the perito- 
neum and pleura when first injected, and are actively phagocyted. 
Later when the lesion is established and the bacilli are lodged in the 
tissues the polynuclears give way to the lymphocytes, which even 
then never accumulate in such proportion as do the microphages in 
acute suppurative lesions. It may well be that the chemotaxis origi- 
nally attracting the polymorphonuclear leukocytes is the same in 
every case, but that a continued irritant, especially one well sur- 
rounded by tissue elements as are the organisms within the tubercles, 
may cease to exact any chemotactic influence, the accumulation of in- 
active lymphocytes possibly being due to a progressive death of these 
cells carried into the neighborhood of the lesion by the normal circu- 
lation of the lymph. CHAPTER XIV 
THE DELATION OF THE LEUKOCYTES AND OF 
PHAGOCYTOSIS TO IMMUNITY 
In Metchnikoff’s earliest work upon the daphnia or water flea 
he observed clearly that there was a direct relation between the de- 
gree of phagocytosis and the outcome of the infection. When 
phagocytosis of the invading yeasts was energetic and complete the 
daphnia recovered. When the yeast cells penetrated the intestinal 
wall of the daphnia in large numbers, and were enabled to multiply 
before the phagocytic cells could accumulate in suflicient numbers to 
engulf them, then the body of the daphnia was soon swamped with 
the parasites and death ensued. 
This simple observation fostered the thought that the basic prin- 
ciple underlying all processes of immunity was represented in this 
struggle between the invading bacteria and the phagocytic cells. 
To the activity of the latter, entirely, he attributed the power of 
resistance. 
In support of this contention Metchnikoff and his pupils have 
marshaled many facts, most of which are set forth in his classical 
work “L’lmmunite dans les Maladies Infectieuses.” Tt will he 
manifestly impossible here to do more than outline the plan of 
study which these investigations have followed and the conclusions 
to which they gave just foundation. 
The original study upon the infectious disease of daphnia led to 
analogous experiments upon higher animals and, by the prolonged 
and patient investigations of Metchnikoff and his pupils, it was 
shown that, throughout the field of infectious disease, there was a 
striking parallelism between the resistance of the infected subject 
and the degree of phagocytosis which occurred. 
Earlier studies concern themselves chiefly with the natural im- 
munity possessed by many animals against certain infection. The 
infectious disease which at this time had been most thoroughly 
studied was anthrax, and Koch had shown that frogs and other cold- 
blooded animals were markedly resistant against this micro-organism. 
Taking advantage of this observation, Metchnikoff studied the phago- 
cytosis of anthrax bacilli in frogs and found that it took place rap- 
idly and effectively, all of the injected bacilli being soon engulfed 
by the accumulating cells. Similarly, active phagocytosis of anthrax 
bacilli was demonstrated in such naturally resistant animals as dogs 
357 358 
INFECTION AND RESISTANCE 
and chickens, while almost no cell ingestion occurred in delicately 
susceptible animals like guinea pigs and rabbits. Eats, on the other 
hand, more resistant to anthrax than guinea pigs, less so than dogs, 
showed a degree of phagocytosis intermediate between that observed 
in the cases of the other animals mentioned above. And yet, in these 
more susceptible animals, the normal bactericidal action of the blood 
upon anthrax bacilli, though never extreme, was often more marked 
than that of the naturally immune animals mentioned above. 
It is well known, for instance, that the serum of dogs possesses 
almost no bactericidal properties for anthrax bacilli,1 although the 
animals are highly resistant to this infection, while the serum of 
rabbits is probably more strongly bactericidal for these bacilli than 
the serum of most other animals, and yet rabbits are extremely 
susceptible. That the lack of bactericidal powers of the serum is 
not always a sign of susceptibility on the part of the animal was 
shown as early as 1889 by Lubarsch. (We must remember, how- 
ever, that lack of bactericidal power does not necessarily mean lack 
of sensitizer. For bacteria may be sensitized without being killed 
extracellularly as can be shown by the alexin-fixation reaction.) 
The study of anthrax infections was a peculiarly fortunate 
choice of subject, since in this bacillus resistance to serum lysis is 
especially well marked and phagocytosis seems indeed to be the chief 
mode of bacterial destruction. Studies analogous to those originally 
made with anthrax, however, were subsequently carried on with 
streptococci, pneumococci, and staphylococci chiefly by Bordet,3 
Marchand,3 and others, and results coinciding with those of Metchni- 
koff were obtained. In every case naturally resistant animals 
showed marked phagocytosis, and susceptible ones failed to show it 
to the same degree. It is a strong support of the same opinions, too, 
that Marchand’s studies, later extensively confirmed, showed that 
the more virulent and invading strains of streptococci, the less active 
is the phagocytosis—a converse, but equally conclusive, observation. 
Further support for this point of view is manifold and cannot be 
considered with anything like completeness. We may refer briefly, 
however, to the experiments of Vaillard, Vincent, and Rouget 4 with 
tetanus, and those of Leclainche and Vallee5 with symptomatic 
anthrax, because they are especially valuable in illustrating the 
importance of phagocytosis in another class of infection. The poi- 
sons of these micro-organisms are extremely toxic for rabbits, and if 
a small amount of culture material, together with agar, broth, or 
any foreign substance which may inhibit or divert phagocytosis from 
1 Petterson. Centralbl. f. Bakt., 1, 39, 1905. 
2 Bordet. Ann. de Vlnst. Past., Vol. 11, 1897. 
3 Marchand. Archiv. de med. Exp., Vol. 10, 1898. 
4 Vaillard, Vincent, and Rouget. Ann. de Vlnst. Past., Vols. 5, 6, 1891- 
1892. 
5 Leclainche and Vallee. Ann. de Vlnst. Past., Vol. 14, 1900. RELATION OF LEUKOCYTES TO IMMUNITY 
359 
the spores, is injected into these animals rapid proliferation and 
death with toxemia result. If, on the other hand, the spores are 
carefully washed of foreign material and toxin rapid phagocytosis 
results and the animals recover. 
The parallelism which was followed out so extensively between 
natural immunity and phagocytosis was even more closely marked 
in the case of artificial acquired immunity. The first observations 
of this kind made by Metchnikoff, again on the subject of anthrax 
infection, were carried out by the active immunization of rabbits. 
The subcutaneous injection of virulent anthrax bacilli into normal 
rabbits is usually followed by a rapid growth of the bacteria, with 
much serous exudation but hardly any leukocytic accumulation. In 
immunized animals, on the other hand, the bacilli are taken up by 
hosts of phagocytes, just as this occurs in naturally resistant dogs or 
other animals. Similarly Bordet6 has shown that cholera spirilla 
injected into the blood stream of cholera-immune animals are taken 
up by leukocytes even before they can be subjected to lysis by the 
circulating lytic antibodies. 
It would add little to clearness were we to multiply the examples 
in which it has been demonstrated that the acquisition of increased 
resistance is accompanied by enhancement of the phagocytic process. 
This statement may be regarded as an axiom, and indeed our later 
discussions of the opsonins and bacteriotropins will show clearly why 
such a state of affairs is to be expected. Taken by itself, however, 
it does not necessarily prove that the destruction of the invading 
germs is entirely due to the leukocytes. It might still be possible 
that the bacteria are injured or even killed by the antibacterial 
serum constituents before they can be taken up and carried away 
by the cellular elements; the phagocytes then would act only as 
scavengers for the removal of the dead bodies. Indeed, this opinion 
was long held by a number of the adherents of the purely humoral 
school. However, such a point of view is no longer tenable—espe- 
cially in the light of the later opsonin studies just referred to. 
Moreover, long before these more recent studies it was clear that 
bacteria may often grow within the leukocytes—finally destroying 
these—and that they may even remain fully virulent after ingestion. 
For, as Metchnikoff showed, if guinea pigs were injected with a 
little of the exudate formed after the injection of anthrax bacilli 
into immunized rabbits (an exudate in which there were no longer 
any extracellular bacteria because of energetic phagocytosis) death 
often resulted. It was clear, therefore, not only that the ingested 
bacteria were still alive, but that they were, at least in part, still 
fully virulent. 
A further method of investigation employed by Metchnikoff in 
his endeavors to prove his point consisted in the attempt to demon- 
6 Bordet. Ann. de I’Inst. Past., Yol. 9, 1895. 360 
INFECTION AND RESISTANCE 
strate that virulent bacteria could be protected from destruction in 
the bodies of resistant animals if the leukocytes could be held at bay. 
This resulted in a number of ingenious experiments, the most con- 
vincing of which is the one carried out with anthrax bacilli and 
frogs by Trapetznikoff.7 Anthrax spores were inclosed in little sacks 
of biter paper and these were introduced subcutaneously into frogs. 
In consequence the spores, bathed in the tissue fluids, but protected 
from phagocytosis, developed into the vegetative forms, multiplied, 
and remained virulent for several days. Taken up by the frog’s 
phagocytes under ordinary conditions, they would rapidly have been 
taken up, digested, and destroyed. Here again it was shown that the 
body of fluids alone were unable to dispose of the bacteria and that 
the natural resistance of the frog was due entirely to phagocytic 
processes. 
Other experiments have been aimed at a general reduction of 
phagocytic activity by the use of narcotics. Thus, Cantaeuzene 8 
showed that animals treated with opium are very much more sus- 
ceptible to infection than are normal controls. And since opium 
markedly inhibits the activity of the white cells it may possibly be 
that these experiments furnish a further support for Metchnikoft’s 
opinion. At any rate, it is worth noting that, even though these 
experiments are not convincing in their assertion that the increased 
susceptibility was due entirely to the interference with the leuko- 
cytes, they indicate very definitely the inadvisability of using mor- 
phin and similar narcotics in infectious diseases. 
It is quite clear at any rate, then, that the process of phagocy- 
tosis increases in energy as immunity is acquired and, so far, Metcli- 
nikotf’s assertions are entirely upheld by later knowledge. In his 
contention that all properties upon which the resistance of the ani- 
mal against infection depends center directly or indirectly in the 
phagocyte, however, many subsequent amendments have been neces- 
sary, which will become self-evident in the following discussions of 
individual phases of the destruction of invading bacteria. 
We cannot at the present time attempt to correlate these extreme 
views of Metchnikoff with the equally extreme opinions of those in- 
vestigators who formerly attributed immunity entirely to the prop- 
erties of the body fluids, assigning to the cellular activities a merely 
secondary role. Many of the apparently opposed contentions have 
become reconciled, and we now realize that neither process alone tells 
the whole story, both being parts of the complicated correlated proc- 
esses which together constitute the mechanism of resistance. It was 
indeed to the eager controversy between the two schools that we owe 
much of the clearness of conception which recent years have given. 
After the bacteria are taken up by phagocytes they undergo a 
7 Trapetznikoff. Ann. de VInst. Past., Vol. 5, 1891. 
8 Cantacuzene. Ann. de VInst. Past., Yol. 12, 1898. RELATION OF LEUKOCYTES TO IMMUNITY 
361 
gradual disintegration or dissolution comparable to that by which 
a particle of food is digested within the cell body of a rhizopod. 
With the exception of such particularly insoluble micro-organisms as 
the tubercle bacillus, the leprosy bacillus, blastomyces, and a few 
others, there is in all cases an eventual complete resolution of the 
bacterial body. As in amebse the digestion takes place often after 
the formation of digestive vacuoles, and by staining at this time with 
neutral red it may be demonstrated that the process goes on in a 
weakly acid environment. 
Problem of the Leukocytic Origin of Alexin.—Metchnikoff natu- 
rally assumed, therefore, that the intracellular digestion of bacteria 
by microphages (polynuclear leukocytes), or of cellular elements, 
etc., by macrophages, was a process carried on most probably by 
enzymes, and that these enzymes were identical with the bactericidal 
bodies described as “alexin” and “sensitizer” or “amboceptor” in the 
blood stream. To follow without confusion the development of his 
ideas, however, it is necessary to bear in mind that much of his 
earlier work was done at a time when the discovery of the cooperation 
of two substances in bacteriolysis and hemolysis (the amboceptor and 
the complement) had not yet been made by Bordet, and when the 
bactericidal effect of normal serum was attributed entirely to a single 
substance—the alexin of Buchner. 
Buchner 9 himself had suggested that alexin was an enzyme-like 
body probably derived from the leukocytes. 
In his experiments Buchner had noticed that escudates, produced 
by intrapleural injections of aleuronat in rabbits and dogs, possessed 
a bactericidal value for Bacillus coli which exceeded the bactericidal 
power of the blood serum itself. The influence of active phagocytosis 
could be excluded by the fact that the leukocytes of the exudate had 
been killed by repeated freezing and thawing. Similar results were 
obtained by Hahn 10 with B. typhosus. 
Denys and Kaisin,11 working along similar lines, found that the 
pleural exudates of rabbits, obtained by the injection of dead staphy- 
lococci and freed of cells by centrifugalization, were more highly 
bactericidal for staphylococci than the blood serum of the same ani- 
mals, but found also that the inactivated exudate could not be reacti- 
vated by the addition of leukocytes. Denys offered as an explanation 
for these phenomena that the living leukocytes in the original exudate 
secreted alexin or complement which enhanced the bactericidal 
activity of the exudate, that the leukocytes, subsequently added to 
inactivated exudate, however, had lost vitality during the process of 
isolation and washing, and no longer possessed secretory power. 
9 Buchner. Munch, med. Woch., No. 24, 1894. 
10 Hahn. Archiv f. ILyg., Yol. 25, 1895. 
* 11 Denys and Kaisin, Denys and Havet. La Cellule, Yol. 9, 1893; Yol. 
10, 1894. 362 
INFECTION AND RESISTANCE 
Hankin,12 Kanthack and Hardy 13 had gone even farther than 
this, and had attributed the production of alexin to the eosinopliile 
leukocytes particularly. 
Metchnikoff,14 basing his opinion on his own studies, those of 
his pupils, and many other investigations similar to those mentioned 
above, came to the conclusion that, under ordinary conditions, the 
normal blood contains no free bactericidal substances. He assumes 
that these substances are entirely intracellular, being constituents of 
the various phagocytic elements, by means of which the cells digest 
the foreign elements they take up. He believes that there are essen- 
tially two varieties of such digestive enzymes or “cytases”—just as 
there are two varieties of phagocytes. The microphages, chiefly con- 
cerned in the digestion of bacteria, secrete the bactericidal alexin, or, 
as Metchnikotf calls it, “microcytase.” The macrophages, the large 
mononuclear lymph and endothelial cells, primarily concerned in the 
phagocytosis of cellular elements (red cells, etc.) contain another 
variety of digestive enzyme, the “macrocytase,” or cytolytic (hemo- 
lytic) alexin. The supposition that the hemolytic “cytase” is largely 
derived from the macrophages was based particularly upon the in- 
vestigations of Metchnikoff’s pupil, Tarassewitch,15 who found that 
the extracts obtained from lymph nodes, and other organs rich in 
macrophages, possessed hemolytic properties. Both this work and 
the preceding studies regarding the extraction of alexin from poly- 
nuclear leukocytes will be more fully discussed below. 
Maintaining that these cytases are purely intracellular under 
ordinary conditions, Metchnikotf believes that, in normal animals, 
the destruction of invading bacteria or of injected cellular substances 
(blood cells, etc.) is accomplished entirely by the phagocytic process, 
with subsequent intracellular digestion. In immunized animals, 
however, there is present in the circulating blood another substance, 
not identical with the cytases, but also derived from the leukocytes or 
from the blood-forming organs—the “fixateur” (Ehrlich’s “ambo- 
ceptor”—Bordet’s “sensitizer”). This specific “fixateur” sensitizes 
the bacteria or other antigens to the action of the cytases. For his 
assumption regarding the origin of this sensitizer he finds support 
largely in the researches of Pfeiffer and Marx, and others mentioned 
in our section on the origin of antibodies, as well as in the simi- 
lar investigations of Deutsch,16 carried on under Metchnikoff’s per- 
sonal supervision. 
Final digestion of the sensitized antigens (bacteria or blood 
cells), however, can take place only under the influence of the 
12 Hankin. Centralbl. f. Bakt., Yol. 12, 1892. 
13 Kanthack and Hardy. Proc. Roy. Soc., Yol. 52, 1892. 
14 Metchnikoff. Ann. de VInst. Past., Vol. 7, 1893; Vol. 8, 1894; Vol. 
9,1895. 
10 Tarassewitch. Ann. de VInst. Past., Yol. 16, 1902. 
16 Deutsch. Ann. de VInst. Past., Vol. 13, 1899. RELATION OF LEUKOCYTES TO IMMUNITY 
363 
cytases intracellularly, unless by previous leukocytic injury these 
enzymes have been liberated into the blood stream. 
While it is admitted, then, that bacteria may be killed and di- 
gested both intra- and extracellularly in the animal body, the cytases, 
which accomplish this, are regarded as the same in both cases, being 
derived from the phagocytic cells. In immunized animals “fixateur” 
may be produced under the stress of active immunization and fur- 
nished to the blood stream by the blood-forming organs. By this 
substance bacteria and cells may be sensitized. However, the enzyme 
by which digestion is actually accomplished, “cytase” or alexin, is 
not present free in the blood even in immune animals unless it has 
become free and extracellular by injury to the leukocytes. 
How, then, on this basis does Metchnikoff account for the Pfeiffer 
phenomenon, in which the extracellular destruction of bacteria takes 
place rapidly in the peritoneal exudate ? His explanation is the fol- 
lowing: When bacteria or other substances are injected into the 
peritoneum there is a preliminary injury of leukocytes (phagolysis), 
and by this alexin or cytase is liberated. When cholera spirilla, for 
instance, are injected into the peritoneal cavity of an immunized 
guinea pig there follows a short period during which few if any 
leukocytes are present in the exudate, but many may be found gath- 
ered in motionless clumps in the folds of the peritoneum and mesen- 
tery, incapable of phagocytosis and apparently injured. If such 
leukocytic injury can be avoided Metchnikoff claims that the extra- 
cellular lysis of cholera spirilla will fail to take place. Thus if 
sterile broth or salt solution is injected into the peritoneum of a 
guinea pig a preliminary phagolysis will be followed by an accumu- 
lation of leukocytes. If cholera spirilla are now introduced no 
extracellular digestion is seen, but, instead of this, rapid phagocytosis 
takes place. This he says is due to the fact that the freshly accumu- 
lated, healthy phagocytes, collected in response to the preliminary 
broth injection, are not easily injured and do not undergo phago- 
lysis ; no cytase is liberated and, in consequence, no serum bacterio- 
lysis can take place. In the same way he claims that the hemolysis 
of red blood cells (goose blood) in the peritoneum of specifically 
immunized guinea pigs may be prevented if, by a previous injection 
of broth, healthy leukocytes have been caused to accumulate. In 
such a case the goose blood corpuscles are rapidly ingested by the 
phagocytes and no hemolysis occurs. 
It is self-evident that the validity of this interpretation of the 
occurrences is strictly dependent upon the demonstration that the 
circulating blood normally contains no alexin or complement. This 
is rigidly maintained by Metchnikoff, and is indeed one of the most 
important uncertainties of serology. He asserts that alexin appears 
in the blood serum only because changes in the leukocytes take place 
during coagulation. It is not, by any means, settled that Metchni- 364 
INFECTION AND RESISTANCE 
koff is right in this—in fact, more recent investigations seem to show 
that he is wrong, and that we may assume definitely that alexin is 
present in considerable amounts in the circulating blood plasma of 
normal animals. 
Metchnikoff s denial of this is based chiefly on the experiments 
of Gengou. Gengou 17 took the blood from various animals into 
paraffined tubes and centrifugalized it at low temperature before it 
could clot. This freed the plasma from the cells before clotting, 
though coagulation of course took place as soon as this plasma was 
removed to tubes and kept at room temperature. The serum ex- 
pressed from this clotted plasma he compared for alexin contents 
(bactericidal properties) with that obtained from clotted whole 
blood. 
He found that, in all cases examined (dogs, rabbits, and rats), 
the plasma contained practically no bactericidal substances, or at any 
rate an incomparably smaller amount than was present in the serum 
obtained from the clotted blood. 
These experiments of Gengou would be conclusive in establishing 
Metchnikoff's theory if they were confirmed by other observers. 
This, however, has not been the case. Petterson 18 found no differ- 
ence between the bactericidal properties of serum and oxalate plasma, 
and Lambotte 19 arrived at similar results when he compared serum 
with the coagulable plasma obtained by tying off a section of a vein 
and centrifugalizing the blood without opening the vessel. Hew- 
lett,20 Falloise,21 Schneider,22 and more recently Dick 23 and Addis,24 
whose work has been done with particular attention to technical 
accuracy, fail to confirm Gengou, finding no appreciable difference 
between plasma and serum. 
In favor of Gengou’s results are the investigations of Herman 25 
and the more recent ones of Gurd.26 Further supporting Gengou’s 
conclusion is the observation recorded by a number of workers (Wal- 
ker,27 Longcope,28 and others) that the complement or alexin con- 
tents of serum will increase somewhat as the serum is allowed to 
stand on the clot. This observation, too, has been rendered incon- 
clusive by contrary reports from other investigators. Longcope,29 
17 Gengou. Ann. de VInst. Past., Yol. 15, 1901. 
18 Petterson. Arch. f. Hyg., Yol. 43, 1902. 
19 Lambotte. Centralbl. f. Bakt., Vol. 34, 1903. 
20 Hewlett. Zeitschr. f. Heilkunde, 24, 1903. 
21 Falloise. Bull, de VAcad. Boy. de Med., 1905. 
22 Schneider. Arch. f. Hyg., 65, 1908. 
23 Dick. Jour. Inf. Dis., Yol. 12, 1913. 
24 Addis. Jour. Inf. Dis., Vol. 10, 1912. 
25 Herman. Bull, de VAcad. Boy. de Med., 1904. 
26 Gurd. Jour. Inf. Dis., Vol. 11, 1912. 
27 Walker. Jour. Hyg., 3, 1903. 
28 Longcope. Med. Bull. Univ. of Pa., 1902, Vol. 15, p. 331. 
29 Longcope. Jour. Hyg., Vol. 3. RELATION OF LEUKOCYTES TO IMMUNITY 
365 
further, has found that alexin was more plentiful in the blood of 
individuals suffering from leukemia—in which of course a larger 
percentage of leukocytes is present in the circulation. This, too, 
has been contradicted by other workers, but even if upheld would not 
influence the possibility of there being alexin in the normal circula- 
tion. On the whole Gengou’s contentions with their consequent 
bearing upon MetchnikofFs theory cannot be accepted as final. In 
fact, the greater part of available experimental evidence seems to 
point to the actual presence of alexin in the normal circulating blood. 
This seems also indicated by the unquestionable fact that active 
phagocytosis may take place in the circulation of an animal and, as 
we shall see below, free alexin is probably necessary (as opsonin) in 
this process. Further evidence in this direction also is furnished by 
the immediate anaphylactic shock which follows the injection of 
antigen into the blood stream of a sensitized animal, a process in 
which we have much reason to believe that alexin takes an active 
part. However, the problem is a difficult one, and, while we 
favor the opinion that free alexin is present in the intravascular 
blood, we must admit that a crucial experiment has not yet been 
formulated. 
Leukocytic Bactericidal Substances.—How, as regards the ap- 
parent extraction of alexin from leukocytes and lymphatic organs, 
we have already called attention to the fact that most of the re- 
searches associating these cells with the bactericidal substances were 
carried out before the dual mechanism of sensitizer and alexin in 
bacteriolysis had been fully worked out. In consequence conclusions 
were formulated from the mere facts of the presence of bactericidal 
or hemolytic properties in cell-extracts without the further determina- 
tion of heat stability or the possibility of reactivation. Most of the 
earlier work also was done without sufficient attention to the separa- 
tion of the cells and the serum of leukocytic exudates. The first one 
to do this carefully was Hahn,30 who, like his predecessors, con- 
cluded that the bactericidal leukocytic substances, undoubtedly en- 
countered by him, were identical with alexin. Doubt was first cast 
upon this by Schattenfroh,31 who worked with leukocytes suspended 
and extracted in physiological salt solution. He found that bac- 
tericidal substances were, indeed, obtained in this way, but that, 
unlike alexin, these substances were relatively thermostable, with- 
standing exposure to a temperature of 56° C. and destroyed only 
by exposure to temperatures as high as 75° C. to 80° C. continued 
for thirty minutes. 
Moxter,32 a little later, working with cholera spirilla, also came 
to the conclusion that the leukocytic bactericidal substances were not 
30 Hahn. Arch. f. Hyg., Yol. 25. 
31 Schattenfroh. Arch. f. Hyg., Yols. 31 and 35, 1897. 
32 Moxter. Deutsche med. Woch., 1899, p. 687. 366 
INFECTION AND RESISTANCE 
identical with those found in the blood serum. Petterson,33 too, made 
thorough investigations into the nature of the bactericidal substances 
extracted from the leukocytes, and, working chiefly with B. proteus 
and B. anthracis, found such substances in the leukocytes of dogs, 
rabbits, and guinea pigs active against the bacteria mentioned above, 
but failed to tind them active, at least in guinea pig and cat leuko- 
cytes, against B. typhosus or the cholera spirillum. He expresses the 
opinion that bactericidal leukocytic substances are normally given 
up to the blood in minute quantity only or not at all, and that these 
substances hold no definite relationship to the bactericidal sub- 
stances found in blood serum. In a later investigation he showed 
that the “endolysins,” as he now calls the leukocytic bactericidal 
substances, may, like many enzymes and serum bacteriolysins, be 
precipitated out of solution with alcohol and ether; but he separates 
them absolutely from serum lysins and complement. The latter, 
while they may be, in part at least, secreted by the leukocytes, are, 
according to Petterson, induced easily to come out of the cells during 
life by slight injury or other stimulation, while the endolysins them- 
selves are abstracted from the cells only after extensive extraction or 
maceration. 
Schneider 34 has come to similar conclusions and speaks of the 
endocellular bactericidal substances as “Leukine.” In a recent in- 
vestigation of the same subject the writer 33 has in all essentials con- 
firmed Schattenfroh’s original conclusions regarding the heat sta- 
bility of the extracted leukocytic bactericidal substances, and has 
shown that after inactivation by heat these substances are not reacti- 
vate by the addition of fresh leukocytic extracts, and that the yield 
obtained from the leukocytes of immunized animals is not greater 
than that obtained from normal leukocytes. 
It appears, therefore, that, contrary to Metchnikoff’s first sup- 
position, the enzymes which bring about the digestion of phagocyted 
bacteria within the cell are not identical with those which bring 
about a similar extracellular digestion. In addition to the demon- 
stration of a definitely different structure possessed by the bacteri- 
cidal leukocytic extracts, as evidenced by their heat stability, we have 
the negative evidence that neither true alexin nor sensitizers have 
ever been successfully extracted from such cells. 
It is still possible that this may eventually be done—but, al- 
though indirect evidence like that of Denys, Longcope’s observations 
in leukemia, and the occasional increase of the alexic powers of serum 
after standing on the clot points to a probability of this, no direct 
evidence has so far been satisfactorily produced. In the hope that 
the leukocytes would give up alexin—possibly as a secretion as sug- 
33 Petterson. Centralbl. f. Bakt., i, 39, 1905; 46, 1908. 
34 Schneider. Archiv f. Hyg., Vol. 70, 1909. 
35 Zinsser. Jour. Med. lies., Yol. 22, 1910. RELATION OF LEUKOCYTES TO IMMUNITY 
367 
gested by Denys—the writer, with Cary, some years ago kept 
washed leukocytes at 37.5° C. in Locke’s solution, but was unable 
to find any evidence of alexin production within 48 hours. 
The apparent extraction of hemolysin from macrophages by 
Tarassewith, moreover, has met wuth a similar refutation. Korschun 
and Morgenroth 36 have shown that these hemolytic extracts are ex- 
tremely heat resistant, are alcohol and ether soluble, and do not act 
as antigens. They are quite different from the serum hemolysins, 
therefore, and probably closely related to the hemolytic lipoidal 
substances described by Noguchi and others. 
Further strong arguments against the assumption of the pres- 
ence of hemolytic alexin in the body of the macrophages have been 
advanced by Gruber 37 and by Neufeld.38 Gruber showed that no 
extracellular hemolysis takes place when leukocytes are brought 
together with sensitized red blood cells, and Neufeld showed that 
even after the phagocytosis of such sensitized cells the hemolysis is 
very much slower, and of a different character from extracellular 
serum hemolysis. In the intracellular digestion there are no mere 
solution of the hemoglobin and formation of a shadow form 
(stroma), but there occur a gradual degeneration with the forma- 
tion of a granular detritus of hemoglobin. 
It is probable, then, that the digestion of bacteria and cells 
within the phagocytes is carried out by substances not identical with 
those taking part in serum lysis. It is not unlikely that the intra- 
cellular process is a quite complicated one, not depending on a single 
enzyme. 
Leukoprotease.—In addition to the bactericidal substances ex- 
tracted from leukocytes a number of true enzymes have indeed been 
obtained by various investigators. We have mentioned in another 
place that one of the earliest observations in this respect was that of 
Leber,39 who noticed that sterile pus could liquefy gelatin. It may 
be commonly observed in paraffin or celloidin sections of staphy- 
lococcus abscesses that a ring of apparently digested or degenerating 
tissue is formed about an accumulation of leukocytes—in foci in 
which the bacteria may be too sparse to be held accountable for the 
changes. These leukoproteases have subsequently been carefully 
studied by Opie,40 Jochmann and Muller,41 and a number of others. 
Opie found that two distinct proteolytic enzymes could be ex- 
tracted from the cells of exudates obtained by turpentine injections. 
One—peculiar to the polynuclear leukocyte, and similar to one pre- 
36 Korschun and Morgenroth. Berl. klin. Woeh., 1902. 
37 Gruber. Quoted from Sachs, in “Kraus u. Levaditi Handbuch,” Vol. 2, 
p. 991. 
38 Neufeld. Arb. a. d. kais. Gesundheitsamt, Yol. 28, 1908. 
39 Leber. “Die Entstehung der Entziindung,” Leipzig, 1891. 
40 Opie. Jour. Exp. Med., Vol. 7, 1905; Yol. 8, 1906; Vol. 9, 1907. 
41 Muller and Jochmann. Munch, med. Woch., Nos. 29 and 31, 1906. 368 
INFECTION AND RESISTANCE 
viously described by Miiller 42—acts in a weakly alkaline medium. 
The other, present particularly in exudates containing a predominat- 
ing number of mononuclear cells, acts in a weakly acid reaction. 
Tschernorutski also found proteolytic ferments in both micro- and 
macrophages, but found no lipase in the polynuclear extracts. This 
seems particularly interesting in view of the great resistance to 
intracellular digestion noticed in acid-fast bacteria, a point of some 
importance in connection with the destruction in the body of such 
micro-organisms as the bacilli of tuberculosis and leprosy.43 Joch- 
mann 44 states that the action of the leukoprotease, which acts in an 
alkaline medium upon casein, results in the formation of tryptophan 
and ammonia, and believes it to be functionally very similar to 
trypsin. It is interesting to note that the most active protease is 
obtained from pus as it forms about acute infections or other stimuli 
which lead to the accumulation of polynuclear leukocytes, whereas 
it is apparently completely absent from tuberculous pus. 
The question immediately arises, are these leukoproteases identi- 
cal with the bactericidal substances extracted from leukocytes as de- 
scribed above ? For it might well be that bacterial death resulted 
merely from the digestive action of the enzyme upon the bacterial 
protein. Jochmann,45 who has approached this problem experi- 
mentally, has answered it in the negative. By repeated alcohol 
precipitation of glycerin extracts of leukocytes he obtained an en- 
zyme preparation which possessed absolutely no bactericidal prop- 
erties, though it was still actively proteolytic. 
ISTot only did this relatively pure ferment possess no bactericidal 
action, but bacteria actively proliferated when suspended in it. Joch- 
mann believes that living bacteria are not amenable to the enzyme 
possibly because of their possession of an “antiferment,” at least 
this would follow in some cases from the experiments of Kantoro- 
wicz.46 
The leukoproteases, therefore, appear to possess no direct sig- 
nificance in bacterial immunity. Their function seems rather to 
lie in the resorption of dead tissues, fibrin, blood clots, etc. Friedrich 
Muller 47 has pointed out their possible importance in the rapid de- 
struction and liquefaction of the massive fibrinous exudates remain- 
ing after the crisis in lobar pneumonia. 
Effects of Leukocytic Substances upon Infections.—From the 
discovery of antibacterial properties in the extracts of leukocytes 
it is but a logical step to the attempt to utilize these substances thera- 
42 Muller. Kongr. f. inn. Med., Wiesbaden, 1902. 
43 Zinsser and Cary. Jour. A. M. A., 1911. 
44Jochmann. Leueozyten Fermente, etc., “Kolle u. Wassermann Hand- 
buch2d Ed., Yol. 49, 2. 
45 Jochmann. Zeitschr. f. Hyg., 61, 1908. 
48 Kantorowicz. Munch, med. Woch., No. 28, 1909. 
47 Friedrich Muller. Verhgnd. d, Kongr. f. inn. Med.f 1902, RELATION OF LEUKOCYTES TO IMMUNITY 
369 
peutically. Petterson 48 was probably the first to study this phase of 
the problem systematically in connection with anthrax infection in 
dogs and rabbits. In preliminary studies he claimed to have de- 
termined that when leukocytes are left in contact with serum for four 
hours or longer there develops in the mixture a bactericidal power 
far superior to that which is possessed by these elements when sepa- 
rately kept in salt solution and mixed only just before the bactericidal 
tests. He attributes this to the fact that in dogs, at least, the leuko- 
cytes furnish bactericidal substances to the serum—an assumption 
which is entirely in accord with the earlier opinion of Denys and 
Kaisin,49 which we have mentioned in another place. In direct con- 
tinuance of these experiments he injected leukocytes into dogs at 
the same time at which he infected them with anthrax and observed 
a moderately protective influence, which, however, he admits was 
not very great. He followed this work in 1906 with similar observa- 
tions on the protective influence of leukocytes in intraperitoneal in- 
fections of guinea pigs with typhoid bacilli. In these experiments 50 
he made the curious observation that, although such protective influ- 
ence was unquestionable, the guinea pig leukocytes contained no 
bactericidal substances active against typhoid bacilli. In conse- 
quence he concluded that the destruction of these bacteria in the 
guinea pig was due entirely to the serum-antibodies absorbed by the 
micro-organisms before phagocytosis, even when the actual destruc- 
tive process took place intracellularly. The protective effect follow- 
ing on the injection of the leukocytes he attributed to an indirect 
influence of the leukocytic substances in stimulating the more rapid 
accumulation of alexin or complement in the peritoneum, with 
consequently more powerful phagocytosis. Following this, in 1908, 
Opie 51 carried out experiments in which he observed that leukocytes 
injected intrapleurally into dogs, together with tubercle bacilli, 
exerted a distinct protection. 
In the same year extensive observations on the protective prop- 
erties of leukocyte extracts were published by Hiss. 
Hiss 52 worked at first with extracts of dog, rabbit, and guinea 
pig leukocytes; later he confined himself entirely to rabbit leuko- 
cytes. He extracted the leukocytes at first by repeated freezing and 
thawing in physiological salt solution, but the technique of his sub- 
sequent work was uniformly as follows: Intrapleural injections of 
aleuronat emulsions were made in rabbits and, after about 24 hours, 
the resulting exudates were taken away with sterile pipettes and 
centrifugalized before clotting could take place; the serum was de- 
48 Petterson. Centralbl. f. Baht., Yol. 36, 1904. 
49 Denys and Kaisin. “La Cellule,” Yol. 9, 1893. 
50 Petterson. Centralbl. f. Baht., Yols. 40 and 42, 1906 
51 Opie. Jour. Exp. Med., 1908. 
£2 JIjss, Jour. Med. Res., new series, Yol. 14, 1908, 370 
INFECTION AND RESISTANCE 
canted and the leukocytes then emulsified in distilled water, in quan- 
tity about equal to the amount of serum poured off. In this the leu- 
kocytes were allowed to stand for a few hours at incubator tempera- 
ture, and then in the ice-box until used. For his experimental work 
in both animals and man, in most instances, not only the clear super- 
natant fluid was injected, but the cell residue as well. 
With leukocytic extracts so prepared Hiss treated staphylococ- 
cus, typhoid bacillus, pneumococcus, streptococcus, and meningococ- 
cus infections in rabbits and obtained results which justified him in 
concluding that the leukocyte extract exerted strong protective action 
in all of these cases. Many of his animals survived infections fatal 
to controls even when the treatment was delayed as long as 24 hours 
after infection. Subsequently Hiss and Zinsser 53 treated series of 
patients, ill with pneumonia, meningitis, and staphylococcus infec- 
tions, with leukocyte extracts prepared by the method of Hiss, and 
felt that they were justified in concluding that in many cases, at 
least, the course of the disease was favorably influenced by the leuko- 
cyte extract. Favorable results have since then been obtained also by 
Lambert in erysipelas, and by Hiss and Dwyer in a variety of con- 
ditions. 
While there seems to be little question about the actually favor- 
able influence of the leukocyte extract, both in experiments with 
animals and in the treatment of human cases, there has been consid- 
erable difficulty in determining the reasons for this influence. In 
subsequent studies Hiss and Zinsser (loc. cit.) were able to show that 
the extracts did not favor phagocytosis and that the moderate bacteri- 
cidal properties possessed by the leukocytic substances could not ac- 
count for their effectiveness. There did seem to be a more rapid 
accumulation of phagocytes in the peritoneal cavities of guinea pigs 
infected with cholera spirilla when leukocyte extract was injected 
with the bacteria, and it is not impossible, in fact, it seems probable 
to the writer, from subsequent experience, that the protective prop- 
erties of the leukocyte extracts are attributable, in part at least, to 
their positively chemotactic effect. 
We are inclined to believe at present that the beneficial effects of 
leukocyte extracts are based on the same principles as those which de- 
termine the reactions following on the injection of bacterial and any 
other protein. 
The Problem of Leukocytosis.—In this connection a very inter- 
esting problem has arisen—namely, that spoken of as the phenomenon 
of Specific Hyperleukocytosis. Bordet, as early as 1896, made the 
following statement, “Active immunity has also other characteristics, 
in that there is an increase of the number of leukocytes, that is, an 
‘exaltation’ of the chemotactic sensibilities of the leukocytes.” He 
suggests herein that an immunized animal may respond with a more 
53 Hiss and Zinsser. Jour. Med. Res., new series, Yol. 14, 1908. RELATION OF LEUKOCYTES TO IMMUNITY 
371 
powerful leukocytic reaction to the injection of the infectious agent 
than would a normal animal similarly treated. This idea has re- 
cently found experimental elaboration in the work of Cay and Clay- 
pole, who found that the reinjection of immune animals with the 
homologous bacteria produced a specific hyperleukocytosis, that is, 
typhoid immune animals receiving typhoid bacilli would respond 
with counts ranging up to 150,000 leukocytes per cu. mm., whereas 
the normal animals rarely showed more than 40,000 to 50,000. 
This observation would tend to indicate a great advantage of the 
specific over the non-specific methods of treatment. 
Unfortunately the results of Gay and Clay pole 54 have not found 
confirmation. McWilliams 55 in similar experiments found no differ- 
ences in the degree of leukocytosis between normal and immune ani- 
mals in response to the injection of bacteria and reported that the 
same degree of response followed in typhoid immune animals when 
injected with Bacillus coli as when typhoid bacilli were adminis- 
tered. In part, this is also stated to be the experience of Jobling and 
Petersen. 
The question is such an important one that the writer, with Dr. 
Tsen,56 reinvestigated it in connection with work on the therapeutic 
effect of leukocytic extract. Our conclusions showed little agreement 
with those of Gay and Claypole. 
We found that when homologous Gram-negative bacilli are in- 
jected into immunized animals there seems to be a definitely higher 
leukocytosis in the immunized animals than in normal controls simi- 
larly treated. The contrasts in our experiments, however, were 
nothing like as striking as those reported by Gay and Claypole. 
Indeed the contrast in general is so slight and so irregular that in 
the case of the Gram-negative bacilli we were at first inclined to agree 
with McWilliams. There was, however, a sufficiently definite differ- 
ence in an average of many counts to convince us that this was more 
than coincidence. 
In the case of the Gram-positive cocci there was a more marked 
difference, in that the immunized animals reacted more promptly 
and very much more energetically than did the normal animals to 
similar injections. 
It seems reasonably clear, then, that an animal reacts more ener- 
getically as far as its mobilization of leukocytes is concerned when 
reinjected than does a normal animal treated with the same variety 
and quantity of bacteria. 
The reaction is dependent upon a number of factors, chief among 
which are: (1) the condition of the animal (loss of weight, etc.) ; 
(2) the amount of bacteria injected, and (3) the interval between 
54 Gay and Claypole. Arch, of Int. Med., V, XIV, 1914, p. 662. 
55 MeWilliams. Jour. Immunol., Vol. I, No. 2, 1916, p. 159. 
56 Zinsser and Tsen. Jour. Immunol., Vol. II, No. 3, 1917, p. 247. 372 
INFECTION AND RESISTANCE 
injections. These factors all very naturally signify the necessity of 
avoiding too profound an intoxication of the animal. 
When immune animals are treated with heterologous bacteria— 
that is, when prodigiosus bacilli or colon bacilli are injected into 
typhoid immune animals and vice versa—there seems to be no specific 
difference in response. That is, the injection of colon bacilli into 
typhoid animals has shown no marked difference in leukocytic re- 
sponse from that observed when typhoid bacilli were injected into a 
typhoid animal. In this our figures correspond with those of McWil- 
liams. They are also in keeping with the clinical experience of 
Kraus, Ichikawa, and others which have been mentioned above. 
The injection of leukocytic extract does not arouse as vigorous a 
leukocytic response as does the injection of bacillary protein. 
In reading these facts superficially they at first seem to be con- 
tradictory in significance, inasmuch as specificity seems to exist in 
the fact that typhoid immune rabbits or streptococcus immune rabbits 
respond somewhat more vigorously than do normal controls injected 
with the same substance. 
However, we think that these relations are explained by the fact 
that animals that have reacted to such organisms as the typhoid bacil- 
lus, etc., develop a certain amount of non-specific tolerance against 
the proteose-like substances which are probably responsible for a not 
unimportant part of the symptoms caused by the bacteria. Such 
tolerance has been shown in the experiments made by the writer with 
Dwyer. 
To summarize, therefore, we do not think that at present a specific 
leukocytosis in the sense of Gay and Clavpole has been demonstrated, 
but believe that an animal, immune to one micro-organism, will have 
a slight non-specific increase of resistance to other organisms. 
This does not mean that the immunity is non-specific. The de- 
struction of living bacteria is still a purely specific process, and this, 
of course, would determine the occurrence of outcome of an infection. CHAPTER XV 
FACTORS DETERMINING PHAGOCYTOSIS 
OPSONINS, TROPINS AND THE OPSONIC INDEX 
From the very beginnings of his researches upon phagocytosis 
Metchnikoff recognized that the process was profoundly influenced 
by the properties of the fluid constituents of the blood plasma in 
which the phenomenon occurred. Both he and his pupil Bordet,1 
at this time working in Metchnikoff’s laboratory, noticed that the 
phagocytic activity of leukocytes was greater in immune than in 
normal sera and associated this with the specific properties of the 
immune substances or antibodies in these sera; Metchnikoff himself 
interpreted the phagocytosis-enhancing power of the serum as a 
stimulation of the leukocytes and referred to the serum constituents 
by which this effect was produced as “stimulins.” A closer analysis 
of the factors involved in this interrelationship, however, was not 
attempted at this time by him or his pupils, although indirect ref- 
erence was made to it in a number of articles emanating from this 
school in the course of investigations on kindred problems of phago- 
cytosis. Thus Gabritschewsky,2 in 1894, published a paper on 
“Leukocytose dans la Diphtherie,” in which he concluded that the 
poison of diphtheria bacilli, among other harmful effects, diminished 
the phagocytic power of the leukocytes, and that one of the beneficial 
influences of the curative serum was to render these and other cells 
“less sensitive to the bacterial poisons.” This may be interpreted 
as indicating an assumption that the action of an immune serum iu 
increasing phagocytic activity rested rather upon its influence upon 
the bacterial products than upon any stimulation of the phagocytes 
themselves. However, in diphtheria the action of the leukocytes 
was, even at this time, recognized as a merely secondary one, and 
Gabritschewsky’s results did not materially influence the “stimulin” 
conception. 
The first extensive investigation which occupied itself directly 
with these problems was that of the Belgian bacteriologists Denys 
and Leclef.3 The publication of these workers deals primarily with 
the nature of streptococcus immunity in rabbits. It established, first 
1 Bordet. Ann. de VInst. Past., 1895. 
2 Gabritschewsky. Ann. de VInst. Past., 1894. 
3 Denys and Leclef. La Cellule, 11, 1895. 
373 374 
INFECTION AND RESISTANCE 
of all, the paramount importance of phagocytosis in the resistance 
of animals against these bacteria, and made clear that the destruction 
of bacteria was carried out equally as well by the leukocytes of nor- 
mal as by those of immune animals, but was powerfully enhanced 
when either normal or “immune” leukocytes were combined with 
immune serum. Their work, therefore, indicated again that the in- 
creased phagocytosis of virulent bacteria, taking place in immune 
animals, depended clearly upon alterations in the functions of the 
serum rather than in those of the cells, and they suggested that 
the influence of this serum was not necessarily one of leukocyte 
stimulation, but might rather consist in action upon the bacteria, 
rendering them less resistant to phagocytosis. They say in sub- 
stance : “A notre avis, on pourrait tout aussi bien admettre que la 
substance vaccinante ou antitoxique agit, non pas sur le leukocyte, 
mais sur un poison renferme dans le corps du microbe ou dissous 
dans le milieu, et qui preserve le micro-organisme contre les at- 
teintes du leukocyte.” 4 
In this statement we have, in brief, the distinct formulation of 
our present view of the “opsonins.” 5 
Observations with pneumococci and streptococci carried out 
after this by Marchand 6 and by Mennes,7 whose investigations wre 
cannot discuss in detail, beside confirming most of the observations 
of Denys and Leclef, brought out especially the relation of the 
virulence of micro-organisms to phagocytosis, showing that very 
virulent strains were taken up to a slight degree only in the presence 
of normal serum, but were subject to active phagocytosis when im- 
mune serum was employed. This, too, seemed to point primarily to 
the fact that the serum influenced rather the bacteria than the 
phagocytes, although no convincing proof is brought for this in their 
publications. Though much that had bearing indirectly on this 
problem was written during the following years, no definite progress 
was made beyond the results of Denys and his pupils until 1902, 
when Leischman 8 introduced a technique by means of which it be- 
4 In our opinion one can just as well believe that the vaccinating or anti- 
toxic substance acts not upon the leukocyte but upon a poison inclosed within 
the body of the bacteria or dissolved in the medium, which preserves the 
micro-organism against the attacks of the leukocyte. 
6 Denys formulated this view with still greater clearness and positiveness 
at the Congress of Hygiene held at Brussels in 1903. We take our citation 
from the discussion on opsonins by Gruber (3d meeting Freie Yereinigung f. 
Mikrobiol., Vienna, 1909, Centralbl. f. Bakt., I Ref., Vol. 44, Suppl. p. 3). 
Following is Denys’ statement: 1. The phagocytosis in immune sera is de- 
pendent upon substances which are precipitated with the euglobulins. 
2. These substances cause phagocytosis by inciting a physical alteration of 
the micro-organisms. 3. These substances are specific. 
6 Marchand. Arch, de Med. Exp., 1898. 
7 Mennes. Zeitschr. f. Hyg., Vol. 25. 
8 Leischman. Brit. Med. Jour., Vol. 2, 1901, and Vol. 1, 1902. FACTORS DETERMINING PHAGOCYTOSIS 
375 
came possible to observe the process of phagocytosis with fresh serum 
and leukocytes in vitro. 
By utilizing this technique and improving upon it Wright and 
Douglas in the following year (1903) evolved a method by means of 
which phagocytic activity could be quantitatively measured with 
reasonable accuracy. They worked at first with staphylococcus 
phagocytosis by human leukocytes in the presence of human citrate 
plasma, a research undertaken primarily because Wright,9 in collabo- 
ration with Windsor, had previously determined that human blood 
serum possessed practically no bactericidal power for this organism, 
and that phagocytosis was probably the chief mechanism of protection 
which the human body possessed against these bacteria. The re- 
searches of Wright and Douglas 10 were carried out chiefly by mixing 
equal volumes of bacteria, serum, and leukocytes (in citrate sus- 
pension),11 allowing these elements to remain together at 37.5° C. 
for varying periods, then staining on slides and determining the 
degree of phagocytosis by counting the numbers of bacteria taken up 
by each polynuclear leukocyte. Though many technical difficulties 
had to be overcome, and although the method at its best still permits 
of much personal error, careful work and untiring repetition made 
possible a considerable degree of accuracy, and definite facts regard- 
ing the mechanism of phagocytosis, heretofore merely suspected, 
could be recorded. The most important result of these investiga- 
tions was the unquestionable establishment of the function of serum 
in the process of phagocytosis, namely, that it in no way “stimu- 
lated” the leukocytes in the sense of Metchnikoff, but rather acted 
entirely upon the bacteria, preparing them for ingestion. For this 
reason Wright coined the word “opsonins” (6\f/oveoo = I prepare food) 
for the serum constituents which brought about this effect, believing 
them to be new antibodies, entirely distinct from the other serum 
antibodies heretofore recognized. 
Wright and his followers now concluded that the role of the 
leukocyte in taking up bacteria was entirely dependent upon the 
opsonin contents of the serum. In a menstruum containing no 
serum, or in a serum in which the opsonins had been destroyed by 
heat, they found practically no phagocytic action on the part of 
washed serum-free leukocytes, and they, therefore, doubted the oc- 
currence of spontaneous phagocytosis on the part of leukocytes 
themselves. 
In this point it is not unlikely that Wright is mistaken, since 
9 Wright and Windsor. Jour. Hyg., Yol. 2, 1902. 
10 Wright and Douglas. Proc. Roy. Soc., 72, 1903, 73 and 74, 1904. 
See also Wxight, “Studien iiber Immunisierung,” Fischer, Jena, 1909. 
11 At first bacteria were merely mixed in equal volumes with eitrated 
whole blood. 376 
INFECTION AND RESISTANCE 
other observers, notably Lohlein,12 have observed the phagocytosis of 
various bacteria by washed leukocytes in indifferent, opsonin-free 
media. Although we may take it as assured that such spontaneous 
phagocytosis may take place (Metchnikoff and a number of others 
having obtained results similar to those of Lohlein), this is prob- 
ably never very intense. 
In fact, Wright, in some of his later work, does not insist rigidly 
upon the non-occurrence of spontaneous phagocytosis, but attempts 
to associate such phenomena with the salt contents of the medium 
in which it occurs. Together with Reid,13 he determined that spon- 
taneous phagocytosis of tubercle bacilli unquestionably takes place, 
is most intense at a concentration of about 0.6 per cent. ISTaCl, and 
diminishes as the concentration is increased. This, as we shall see, 
has bearing on the possible physical explanations advanced to ac- 
count for opsonic action, and has its parallels in experiments on the 
influence of electrolytes on agglutination and precipitation. 
The fact remains that Wright demonstrated by his work that 
Metchnikoff’s original view, which interpreted the difference be- 
tween susceptibility and immunity as a difference between the in- 
herent phagocytic powers of the leukocytes, is incorrect, and that 
the essential regulating influence affecting phagocytosis rests upon 
the action of the serum upon the bacteria. 
The following experiment from the work of Hektoen and Rue- 
diger 14 illustrates this point with exceptional clearness. It shows 
that human leukocytes in the presence of normal defibrinated blood 
will take up bacteria energetically. When the leukocytes, however, 
are washed free of blood and added to untreated bacteria phago- 
cytosis is practically nil. If, however, such washed leukocytes are 
mixed with bacteria that have been previously in contact with serum 
active phagocytosis will take place. In other words, the bacteria 
have been altered by the serum in such a way that they are now 
amenable to phagocytosis by washed leukocytes. The serum then 
acts upon the bacteria and not upon the leukocytes. 
TABLE II 
Phagocytosis by Human Leukocytes of Sensitized Bacteria 
Average 
Phagocytosis 
Human leukocytes (defibrinated blood) + Staphylococcus aureus... 22. 
Human leukocytes (washed in NaCl solution) + Staphylococcus 
aureus 1.2 
Human leukocytes (washed in NaCl solution) + Staphylococcus 
aureus (treated with human serum) 10. 
Human leukocytes (defibrinated blood) + Streptococcus 300 22. 
12 Lohlein. Centralbl. f. Bakt., 38, 1906, Beiheft, p. 32; also Munch, med. 
Woch., 1907, p. 1473. 
13 Wright and Reid. Proc. of Roy. Soc. B., Yol. 77, 1906. 
14 Ilektoen and Ruediger. Jour. Inf. Dis., Yol. 2, 1905, p. 132. FACTORS DETERMINING PHAGOCYTOSIS 
377 
Average 
Phagocytosis 
Human leukocytes (washed in NaCl solution) + Streptococcus 300.. 1. 
Human leukocytes (washed in NaCl solution) + Streptococcus 
(treated with human serum) 14. 
Human leukocytes (washed in NaCl solution) + Streptococcus 
(treated with guinea pig serum) 12. 
Human leukocytes (washed in NaCl solution) + Streptococcus 
(treated with rabbit serum) 14. 
Wright and Douglas’ 15 work was done at first with normal 
serum or normal citrate plasma, and in this case they found that the 
opsonins were essentially unstable, being easily weakened by ex- 
posure to light, or heat, and even when preserved in sealed tubes in 
the dark they diminished noticeably on standing for 5 or 6 days. 
Other writers who have worked with the opsonic substances in nor- 
mal serum have confirmed this instability of the normal opsonin, 
although even Wright himself admits that heating to 60° C. does 
not entirely destroy the opsonic power, though it reduces it to a mini- 
mum. A protocol from Wright and Douglas’ first paper will best 
illustrate the degree of reduction of opsonic power resulting from 
the exposure of normal serum to 60° to 65° C. for 10 to 15 minutes. 
A. Unheated serum Wright—Staphylococcus suspension 1 vol.-—Blood cells 
Wright 3 vols. 
(1) Phagocytic average 20 cells 17.4 
(2) Phagocytic average 20 cells 19.8 
B. Heated serum as above. 
(1) Phagocytic average 52 cells 0.6 
(2) Phagocytic average 46 cells 3.4 
The experiments just cited refer only to the opsonic powers of 
normal serum. When an animal is immunized with any particular 
micro-organism or other cellular antigen, such as red blood cells, 
etc., a marked specific increase of opsonins occurs, but unlike the 
opsonins of normal serum these newly formed elements in the im- 
mune serum seem to possess a much greater resistance to heat. 
Neufeld and Kimpau,16 who have studied these constituents of 
immune serum with especial thoroughness, have shown that heating 
to 62° to 63° C. for as long as three-quarters of an hour does not 
destroy them,, and that such sera may he preserved for as long as 
several years without their complete disappearance.17 
We may accept as definitely determined, therefore, that there is 
a qualitative difference between the serum components which initiate 
15 Wright and Douglas. Cited in Wright, “Studien iiber Immun., etc.,” 
P- 9. 
16 Neufeld and Rimpau. Deutsche med. Woch., No. 40, 1904; Zeitschr. 
f. Hyg., Yol. 51, 1905. 
17 Leishman. Trans. London Path. Soc., Yol. 56, 1905. 378 
INFECTION AND RESISTANCE 
phagocytosis in normal serum (normal opsonins) and those which 
carry out the same function to a much more powerful degree in 
immune serum. This is the more surprising since, in the case of all 
other antibodies (lysins, agglutinins, etc.), it has been shown that in 
structure and mode of action the antibodies of immune serum are 
in every way qualitatively similar to the corresponding ones of nor- 
mal serum,18 representing merely a specific quantitative increase 
of substances originally present in small amount. 
This difference between the normal and immune opsonic sub- 
stances has added much difficulty to the investigation of the nature of 
these bodies, and we may approach the problem with greater clearness 
by considering them separately, at first, attempting to define the 
relations between them after we have set down the facts ascertained 
in connection with each. 
Relation of Opsonins to Alexin and Other Antibodies.—In their 
earliest investigations upon the normal opsonins Wright and 
Douglas19 regarded them as new antibodies, separate and dis- 
tinct from those already known. There is no convincing proof of 
this, and a number of other interpretations of the observed phe- 
nomena are possible. Indeed, the burden of proof is rather upon 
those who would establish the existence of a new antibody, for before 
this can be done it must be shown that the new function is not merely 
another property of the serum constituents already known. For, as 
Gruber has justly said, “One of the most important attributes of the 
natural scientist is economy of hypotheses.” And in the case of the 
normal opsonins there are many good reasons for regarding them as 
possibly identical with known serum constituents. The two possi- 
bilities suggested have been (1) Are the opsonic substances identical 
with the alexin or complement ? or (2) Do they represent the com- 
bined action of the normal sensitizer of the serum activated by the 
alexin ? 
The similarity of normal opsonin with alexin or complement has 
been brought out especially by Muir and Martin,20 by Baecher,21 
and by Levaditi and Inmann.22 The fact that both are thermolabile 
has been mentioned above. 
In addition to this, as Muir and Martin 23 have shown, all antigen- 
antibody complexes which absorb alexin out of serum at the same 
time remove the normal opsonin. Thus sensitized red corpuscles, 
18 Dean. Proc. Roy. Soc., 76, 1905. Neufeld and Hiine. Arb. a. d. kais. 
Gesundh. Amt., Vol. 25, 1907. 
19 Wright and Douglas. Loc. cit. 
20 Muir and Martin. Brit. Med. Jour., Vol. 2, 1906; Proc. Roy. Soc. B., 
Yol. 79, 1907. 
21 Baecher. Zeitschr. f. Hyg., Yol. 56, 1907. 
22 Levaditi and Inmann. C. R. de la Soc. Biol., 1907, pp. 683, 725, 817, 
869. 
23 Muir and Martin. Loc. cit. FACTORS DETERMINING PHAGOCYTOSIS 
379 
sensitized bacteria, and specific precipitates added to normal serum 
take out its opsonic substances. From this fact they also concluded 
that the normal opsonins like alexin were non-specific. For just as 
the alexin of a serum may serve to activate a considerable variety of 
sensitized antigens, so the opsonic action of a normal serum may 
functionate upon a large variety of bacteria. Muir and Martin were 
probably wrong in this and, as we shall see below, normal opsonins, 
like normal sensitizers, may be regarded as specific. 
Similar to the observations of Muir and Martin are those of 
Neufeld and Hiine,24 which showed that yeast cells will absorb both 
alexin and opsonin out of serum. 
A further similarity between the two serum constituents is the 
fact that both are absent from the normal fluid of the anterior cham- 
ber of the eye, but they together appear in it after injury (puncture 
for the first removal of fluid). A like parallelism between the ab- 
sence and presence of both has been shown for edema fluids. 
Furthermore, phosphorus poisoning which reduces alexin likewise 
reduces opsonin. 
Although this parallelism is very striking, it does not on this 
account mean that necessarily the two are identical. It may signify 
merely that the alexin is a necessary participant in normal opsonic 
action, essential in that it activates a thermostable opsonic constitu- 
ent just as it activates hemolytic or bactericidal sensitizer. 
This opinion has been expressed by Levaditi, Neufeld,25 Dean,26 
and others, and indeed it is a conception which seems most logical. 
For the procedures which remove both alexin and opsonin, as stated 
above, do not, as a matter of fact, remove all the opsonic action. 
(Although Neufeld maintains this.27) Studies of Hektoen and 
others have definitely proved that, though reduced to almost nil, 
nevertheless heated serum shows definite though slight opsonic action 
as compared with indifferent menstrua such as salt solution. A 
similar slight remnant of opsonic action after absorption of normal 
serum with sensitized cells, bacteria, and precipitates is evident in the 
protocols of Muir and Martin. The significance of this point be- 
comes immediately clear when we consider the properties of the bac- 
teriotropins or immune opsonins, which are heat stable and capable 
of initiating opsonic action in the entire absence of alexin or comple- 
ment. It is possible, therefore, that there may be present in normal 
serum a slight amount of specific thermostable opsonin, which, 
though capable of acting feebly by itself, is nevertheless powerfully 
24 Neufeld and Hiine. Arb. a. d. kais. Gesundh. Amt., Vol. 25, 1907. 
25 Neufeld. “Kolle u. Wassermann’s Handbuch,” Erganzungsband 2, 
p. 313. 
26 Dean. Brit. Med. Jour., 2, 1907, p. 1409. 
27 In fact he states that heated normal serum may be used as a control 
in opsonic experiments instead of salt solution. 380 
INFECTION AND RESISTANCE 
activated by alexin—just as bactericidal or hemolytic antibody is 
similarly activated. 
One of the most thorough studies upon this question is that of 
Cowie and Chapin.28 Dean 29 had previously shown that, although 
heated immune serum was capable of exerting opsonic action by 
itself, this action could nevertheless be enhanced by the addition of 
a little diluted fresh normal serum. The particular significance of 
Dean’s work will be discussed later. Cowie and Chapin, however, 
carried on similar experiments with normal serum in which they at- 
tempted to reactivate heated normal serum by the addition of small 
amounts of diluted fresh serum, by itself but slightly opsonic. One 
of their experiments may serve to illustrate this point, as follows: 
Experiment 10. June 13, 1907 
Phagocytic 
count * 
1. Unheated serum 15.44 
2. Salt solution 0.18 
3. Heated serum, 57° C 1.08 
4. Diluted serum (1:15) 1.56 
5. Heated serum 57° C. + diluted serum (1:15) 12.40 
6. Unheated serum + unheated serum 16.08 
* Phagocytic count = average number of bacteria in each leukocyte. 
This experiment and others like it seem to demonstrate clearly 
that the opsonic action of normal serum, though dependent largely 
upon alexin, is nevertheless also dependent upon a heat-stable body, 
comparable to the sensitizer or amboceptor, in that it is reactivable to 
almost the full power of the original condition (before heating) by 
slight amounts of alexin—in themselves almost inactive.30 
These findings were later confirmed by Eggers,31 and it is plain 
from this work that the apparent opsonic inactivation of normal 
serum by heat depends upon the destruction of the heat-sensitive 
constituent only—the heat-stable substance—surely involved in the 
process, remaining intact, and reactivable. 
Closely associated with this phase of the problem is that of the 
specificity of the normal opsonins. For if, as at first supposed, the 
normal opsonins are, like complement or alexin, non-specific, the 
above amboceptor-complement structure of this mechanism would 
be rendered unlikely. Earlier work upon this question was con- 
28 Cowie and Chapin. Jour. Med. Res., Yol. 17, 1907, pp. 57, 95 and 
213. 
29 Dean. Loc. cit. 
30 In earlier experiments Hektoen and Ruediger33 did not succeed in 
reactivating heated sera and concluded that normal opsonins had the hypo- 
thetical structure of toxins in that they possessed a haptophore and an 
opsonophore group. From this point of view Hektoen has subsequently 
receded largely because of work done under his own direction. 
31 Eggers. Jour. Inf. Dis., Yol. 5, 1908. FACTORS DETERMINING PHAGOCYTOSIS 
381 
tradictory. Bulloch and Western,32 working with staphylococci and 
tubercle bacilli, found that each of these organisms absorbed out 
separately specific opsonins from normal serum, leaving those for 
other bacteria but slightly reduced. Slight reduction of the opsonic 
action for other micro-organisms might easily be explained by a 
partial removal of complement which is bound to take place in such 
experiments. Simon, Lamar and Bispham,33 and some others failed 
to find any such specificity. Russell,34 Axamit and Tsuda,35 and a 
number of others obtained similar negative results—in that a num- 
ber of different bacteria seemed to absorb opsonins out of normal 
serum indiscriminately and without specificity. On the other hand, 
more recent careful work by Rosenow,36 by Macdonald,37 and by 
Hektoen 38 has upheld the original contention of Bulloch and West- 
ern. The work of Rosenow, in which pneumococci were shown to 
absorb out their specific opsonins from normal human serum, taking 
out in part only those for streptococci, staphylococci, and tubercle 
bacilli, is especially convincing, and the experiment of Hektoen with 
normal hemopsonins (opsonins which cause the phagocytosis of red 
blood cells) bear him out. 
It seems fair to conclude, therefore, that normal opsonins de- 
pend upon the cooperation of a heat-stable and a heat-sensitive body. 
The heat-stable body, analogous to normal sensitizer or amboceptor, 
is specific and reactivable by the heat-sensitive body which appears to 
be identical with alexin. This statement merely asserts the facts of 
the dual mechanism of the process without assuming necessarily the 
identity of the heat-stable body with sensitizer or that of the heat- 
sensitive one with alexin, though this seems extremely probable. 
This question we will discuss again more particularly in connec- 
tion with the bacteriotropins or immune opsonins. 
Further proof for such a complex constitution of the normal 
opsonins has been adduced by means of absorption experiments at 
0° C.—by Cowie and Chapin. In our discussion of the lytic anti- 
bodies we have seen that sensitizer or amboceptor may be absorbed 
from serum by its specific antigen at 0° C.—but that the attachment 
of alexin takes place only when the temperature is raised above this. 
Practically no alexin is bound at the low temperature. Cowie and 
Chapin, applying this method of investigation, showed: 
1. That normal human serum may have its opsonic power for 
staphylococci removed by absorption with staphylococci at 0° C. 
32 Bulloch and Western. Proc. Boy. Soc. B., 77, 1906. 
33 Simon, Lamar, and Bispham. Jour. Exp. Med., Yol. 8, 1906. 
34 Russell. Johns Hopkins Bull., Vol. 18, 1907. 
35 Axamit and Tsuda. Wien. klin. Woch., Yol. 20, No. 35, 1907. 
36 Rosenow. Jour. Inf. Dis., Vol. 4, 1907. 
37 Macdonald. Quoted from Hektoen, loc. cit.; Aberdeen TJniv. Studies, 
Vol. 21, 1906, p. 323. 
38 Hektoen. Jour. Inf. Dis., Yol. 5, 1908. 382 
INFECTION AND RESISTANCE 
2. Serum so treated retains the power of reactivating the op- 
sonin of heated normal serum. 
3. Staphylococci so treated are more easily subject to phago- 
cytosis in the presence of dilute normal serum, or normal serum 
which has been inactivated by contact with staphylococci in the cold, 
than are the same bacteria untreated. 
Kurt Meyer 39 has carried out similar experiments with paraty- 
phoid bacilli and normal serum, and, though his work is less exten- 
sive, he reaches the same conclusion as Cowie and Chapin. 
We may accept, therefore, as fairly well established that the 
opsonic power of normal serum depends upon a complex mechan- 
ism consisting of (a) a thermostable substance comparable to 
amboceptor or sensitizer, probably specific, but present in very 
small amount, and (b) a thermolabile substance probably identical 
with alexin or complement which powerfully, but non-speci- 
fically, enhances the slight opsonic power of the thermostable 
substance. 
In considering this conception, together with the subsequent dis- 
cussion of bacteriotropins or immune opsonins, it will be well to 
remember that in normal inactivated sera the thermostable opsonic 
constituent differs in its action from the bodies we speak of as ambo- 
ceptors or sensitizers in that it may functionate for phagocytosis by 
itself—entirely without alexin—while neither bactericidal nor he- 
molytic effects can be brought about by sensitizer alone. Does this 
definitely exclude the identity of this thermostable opsonic substance 
and sensitizer? It is indeed an argument against identification, but 
in opsonic action, we must remember, there is merely a sensitization 
to the action of the phagocyte. This phagocyte may in itself be 
capable of furnishing a small amount of substance comparable in 
action to alexin—in fact, we have seen that the origin of alexin from 
leukocytes is still suspected by a number of workers. At any rate 
the phagocyte is a living cell which may well be capable of supplying 
in itself to some degree the necessary activation, and therefore the 
difference cited above is not necessarily a proof that the normal 
thermostable opsonic constituent is different from normal sensitizer 
or amboceptor. 
The difference between the opsonic action of normal serum and 
that of immune serum, then, is the fact that heating to from 56° to 
60° C. almost completely destroys the former, whereas it has but 
slight if any diminishing effect upon the latter. The immune op- 
sonins, or, as Neufeld and Rimpau have called them, bacteriotropins, 
therefore are thermostable. This was determined as early as 1902 
by Sawtschenko,40 and was subsequently studied with great accuracy 
39 Kurt Meyer. Berl. klin. Woch., 1908, p. 951. 
40 Sawtschenko. Ann. de VInst. Past., Yol. 16, 1902, quoted from Levaditi. FACTORS DETERMINING FHAGOCYTOSIS 
383 
by Neufeld and Rimpau,41 Neufeld and Topfer,42 Dean,43 Hektoen,44 
and others. It was shown that when an animal is immunized with 
any given bacterium or other cellular antigen (blood corpuscles, 
etc.) opsonic substances specific for the particular antigen appear in 
considerable quantities, and these are but slightly, if at all, dimin- 
ished when the serum is heated; Neufeld and Hiine 45 found that 
heating for as long as three-quarters of an hour to 63° C. did not 
noticeably reduce the activity of the bacteriotropins of immune 
serum, and that, again, unlike the normal opsonins, prolonged pres- 
ervation, under sterile conditions, changes them but slowly. 
These facts alone indicate a close similarity between the bac- 
teriotropins and the other well-known thermostable constituents of 
immune sera, and the question here again immediately arises whether 
we are to regard them as identical with any of the other specific 
antibodies or as distinct substances independent of these. 
It was suggested early in these investigations by Muir and Mar- 
tin that bacteriotropins might be identified with agglutinins, inas- 
much as they possessed resistance to heat, were active without ap- 
parent dependence upon alexin, and could not, at least in the earlier 
studies, be reactivated by the addition of fresh normal serum when 
once inactivated. The supposition was that for this reason the bac- 
teriotropin might have a structure like the hypothetical “haptines 
of the second order” which Ehrlich attributes to the agglutinins. 
This supposition has found no experimental support in that ag- 
glutination and bacteriotropic effects did not run parallel. We our- 
selves are not ready to admit that such lack of parallelism is proof 
against their identity. However, since it is very probable that both 
agglutination and precipitation are merely phenomena of colloidal 
flocculation effects which follow certain quantitatively adjusted com- 
binations of antigen and specific antibody, and that it is not at all 
necessary to assume separate agglutinating or precipitating serum 
constituents, this problem becomes merely another version of the 
question of the identity of bacteriotropins and sensitizer or ambo- 
ceptor. 
Apart from thermostability, further similarity lies in the fact 
that bacteriotropins are strictly specific and may be specifically ab- 
sorbed out of immune sera by their respective bacteria. 
Like amboceptor or sensitizer they are specifically increased to a 
powerful degree by the treatment of animals with any given micro- 
organism and may be incited not only by the injection of bacteria 
but by that of blood cells as well. In spite of these points of likeness, 
41 Neufeld and Rimpau. Deutsche med. Woch., No. 40, 1904; Zeitschr. 
f. Hyg., 51, 1905. 
42 Neufeld and Topfer. Centralbl. f. Bakt., 1, 38, 1905. 
43 Dean. Proc. Boy. Soc. B., 76, 1905. 
44 Hektoen. Jour. Inf. Dis., 3, 1906, and loc. cit. 
45 Neufeld and Hiine. Arb. a. d. kais. Gesundh. Amt., Vol. 25, 1907. 384 
INFECTION AND RESISTANCE 
however, Neufeld 46 and his associates maintain rigidly that the two 
substances are not the same and that the bacteriotropins are distinct 
and independent antibodies. 
Among the reasons advanced in support of this opinion are the 
facts that certain immune sera, both antibacterial and hemolytic, 
may contain bacteriotropins without containing lysins and vice versa. 
That this is undoubtedly true has been shown not oidy by Eeufeld 
and his associates but by Hektoen 47 and others, and it is likewise a 
fact that in sera in which both functions are demonstrable they fre- 
quently do not run quantitatively parallel. These are unquestionably 
strong arguments, but their force is somewhat weakened, as Levaditi 
has pointed out, by the fact that there are many varieties of bacterial 
immune sera which undoubtedly sensitize the specific bacteria (as 
can be shown by alexin fixation), but which do not lead to bacteriol- 
ysis. Wassermann 48 also attaches little value to the lack of parallel- 
ism between the lytic and opsonic functions, expressing the belief 
that the solubility of the particular antigen may determine whether 
sensitization leads to phagocytosis or to lysis. With bacteria like 
the cholera spirillum rapid lysis takes place, but when, as in pneumo- 
cocci or streptococci, there is great resistance to lysis, sensitization 
may lead to delayed lysis anticipated by leukocytic accumulation, 
phagocytosis, and intracellular digestion. 
It by no means follows from mere lack of parallelism, therefore, 
that the two serum functions are dependent upon separate antibodies. 
Another important argument advanced against the identification 
of bacteriotropins with the bactericidal sensitizers or amboceptors is 
the fact that the former lead to phagocytosis without the participa- 
tion of alexin, whereas the latter become active for lysis only when 
alexin is present. 
This point also has constituted Neufeld’s strongest support for 
maintaining that the bacteriotropins or immune opsonins are entirely 
distinct from the normal opsonins. It is true, indeed, that immune 
serum, unlike normal serum, may opsonize powerfully even after 
heating to temperatures which destroy alexin. 
If we regard the heat-stable lytic antibody as an amboceptor in 
the strict sense of Ehrlich, as a specific “Zwischenkorper” with a 
complementophile group, this argument would have considerable 
weight. Even in this case, however, strong sensitization of the bac- 
teria may make them amenable to the living cells—the phagocytes— 
which in itself may furnish a slight amount of alexin or alexin-like 
substances. 
We may regard the action of the immune serum upon the antigen 
as rather a sensitization in the sense of Bordet, and it does not seem 
48 Neufeld and Topfer. Centralbl. f. Baht., 1, 38, 1905. 
47 Hektoen. Jour. Inf. Dis., 6, 1909. 
48 Wassermann. Deutsche med. Woch., Yol. 33, No. 47, 1907. FACTORS DETERMINING PHAGOCYTOSIS 
385 
logical to assume that the heat-stable bodies, similar in other respects, 
are different merely because they can sensitize bacteria both to the 
action of an alexin and to that of a living cell, which in itself surely 
contains a number of different enzymes, comparable functionally to 
alexin, though possibly not identical with it. 
Indeed, the experiments of Dean have given much positive evi- 
dence in favor of regarding the immune opsonins or bacteriotropins 
as true amboceptors or sensitizers. Dean49 found that, although 
heated immune serum may unquestionably opsonize by itself, its 
action may be still further enhanced by the addition of a little diluted 
normal serum (compare these results with those of Cowie and Chapin 
on normal opsonins). Hektoen’s 50 experiments with hemopsonic 
immune sera are analogous. We cite one of these as illustrating the 
point in question: 
Phagocytosis of Goat Corpuscles under the Influence of Goat-hlood-immune Rabbit 
Serum, and Normal Guinea Pig Complement (Table from Hektoen, loc. cit.) 
TABLE i 
Immune serum 
Normal guinea pig serum 
Phagocytosis 
.001 
4. 
.001 
+ 
.01 
20. 
.01 
0 
Here, therefore, the diluted immune serum, but slightly cyto- 
tropic in itself, was powerfully activated by a diluted unheated nor- 
mal serum, which in itself was entirely inactive. 
Indeed, an experiment by Heufeld himself, with Bickel,51 points 
in the same direction. They found that, when a heated specific hemo- 
lytic serum was added to the homologoiis cells in such small quanti- 
ties that it no longer exerted cytotropic (opsonic) action, the addi- 
tion of a small amount of alexin, too small to lead to hemolysis of 
the cells (and not by itself cytotropic or hemopsonic), caused active 
phagocytosis. Analogous experiments upon bacterial antisera were 
carried out by Levaditi and Inmann. It thus appears that, even in 
the case of the immune opsonins or bacteriotropins, we may think of 
a participation of two substances—a sensitizer-like one and one com- 
parable to alexin or complement. We may, at least, infer that the 
full opsonic action both of normal and immune sera is dependent 
upon the cooperation of two such bodies. It is likely, therefore, that 
the mechanism of normal and of immune opsonic action may, after 
all, differ only in quantitative relations between the two. 
For assuming this to be an antibody-alexin mechanism like hemol- 
49 Dean. Proc. Roy. Soc. B., 79, 1907. 
50 Hektoen. Jour. Inf. Dis., Yol. 6, 1909, p. 67. 
51 Neufeld and Bickel. Arb. a. d. kais. Gesundh. Amt., Yol, 27, 1907, 386 
INFECTION AND RESISTANCE 
ysis, we may recall the work of Morgenroth and Sachs on the rela- 
tions between amboceptor and complement in hemolysis. There we 
saw that a large amount of amboceptor would cause hemolysis in the 
presence of a small amount of complement and vice versa. There- 
fore, here, too, in normal serum the small quantity of amboceptor or 
specific thermostable opsonin (bacteriotropin) may act very power- 
fully in the presence of the alexin. When the latter is destroyed, 
however, the minute quantity of specific thermostable opsonin is 
hardly enough to do more than initiate slight phagocytosis of com- 
paratively non-resistant bacteria, whereas the large amount of spe- 
cific sensitizer left in immune sera after inactivation may still lead 
to strong bacteriotropic action. In outlining this explanation we 
have consistently drawn upon the analogy between thermostable op- 
sonin and amboceptor or sensitizer. Whether or not these two sub- 
stances are identical is by no means positively determined and must 
be considered an open question for the present. However, from the 
above, it seems to us that much testifies in favor of such an identi- 
fication.52 
The preceding discussions have ignored the possibility that apart 
from opsonic or bacteriotropic action on the bacteria there may be 
a difference in phagocytic energy which depends upon inherent prop- 
erties of the leukocyte itself. 
Indeed, the technique by which the researches of Wright and 
his followers were carried out does not in any way take into account 
the source of the leukocytes as a possible variable factor. There is, 
however, a considerable amount of evidence which points to differ- 
ences in phagocytic powers residing in the leukocytes themselves 
independent of the serum. Park and Biggs 53 have claimed such 
differences for the leukocytes of normal persons in the phagocytosis 
of staphylococci, and more extensive researches have been made with 
similar results, in the case both of staphylococci and tubercle bacilli 
by Glynn and Cox.54 
The last-named authors, moreover, recognized the necessity, in 
making such investigations, of experimenting with leukocyte emul- 
sions containing approximately the same number of cells, for, as 
Fleming55 had shown, if unequal leukocytic emulsions are used, 
less phagocytosis per cell occurs in the emulsion containing the 
greater number of leukocytes. This phase of the subject has been 
taken up most thoroughly by Hektoen 56 and his associates, and Bose- 
52 Pfeiffer (quoted from P. Th. Muller) regards opsonic action as due 
to a combined action of amboceptor and complement and speaks of it as an 
“Andauung” of the bacteria for the leukocyte—which we may translate best 
as a partial predigestion. 
53 Park and Biggs. Jour. Med. Res., Vol. 17, 1907. 
54 Glynn and Cox. Jour. Path, and Bad., 14, 1910. 
55 Fleming. Praditioner, London, Vol. 80, 1908. 
56 Hektoen. Jour. A. M. A., Vol. 57, No. 20, 1911. j FACTORS DETERMINING PHAGOCYTOSIS 
387 
now,>7 has made careful comparative studies on pneumococcus 
phagocytosis, in which he standardized the leukocytic suspensions by 
actual cell counts. His work as well as that of Tunnicliff,58 of the 
same school, has shown definitely that the inherent phagocytic power 
of leukocytes may vary not only in health and disease, but differences 
may exist between the cells of apparently normal people. Tunni- 
cliff showed, for instance, that at birth the leukocytes are less active 
than in adult life. 
For accurate experimental work, therefore, as well as in theoret- 
ical reasoning upon problems of phagocytosis, it is necessary to bear 
in mind the possible inherent variations in the leukocytes themselves. 
Of the three factors concerned in the process of phagocytosis, 
then, we have considered two, the serUm and the leukocytes. The 
former we have seen exerts a powerful determinative influence on 
the process, the latter a less marked influence, though still definite 
and measurable. We have still to discuss the bacteria themselves 
as variable factors in determining the degree to which phagocytosis 
may take place. 
This problem was first investigated by Denys and Marchand in 
connection with their work upon streptococcus immunity, and was 
further studied in detail by Marchand. Marchand 59 showed that 
leukocytes would readily take up non-virulent streptococci in the 
presence of normal serum, but that under similar conditions virulent 
streptococci were not phagocyted at all or to a very slight degree 
only. He determined further that this resistance to phagocytosis 
remained unchanged after the virulent organisms had been killed 
by heat, and washed clean of culture fluid. It seemed, therefore, 
that the resistance depended upon a condition of the bacterial body 
and not upon substances secreted and given off to the environment. 
These experiments, as well as similar work bv Mennes,60 Gruber and 
Futaki,61 and others make it clear that differences in virulence 
between different species of bacteria, as well as between different 
strains of the same micro-organism, depend, at least in part, upon 
the resistance which the bacterial bodies oppose to ingestion by the 
leukocytes. We must distinguish clearly here between these appar- 
ently purely “antiopsonic” bacterial properties and those supposedly 
“antichemotactic” substances which are conceived as a cause for 
virulence by Deutsch and Feistmantel 62 and by Bail 63 in his so- 
called “aggressins.” The latter are supposed to be secreted bacterial 
57 Rosenow. Jour. Inf. Dis., 7, 1910. 
58 Tunnieliff. Jour. Inf. Dis., 8, 1911. 
59 Marchand. Arch, de Med. Exp., No. 2, 1898. 
60 Mennes. Zeitschr. f. Hyg., Yol. 25, 1897. 
61 Gruber and Futaki. Munch, med. Woch., 1906. 
62 Deutsch and Feistmantel. Quoted from Sauerbeck, Eubgrsch und 
Ostertag, Yol. 2, 1906. 
33 Rail. Arch. f. Hyg., Yol. 52, 19Q5., 388 
INFECTION AND RESISTANCE 
substances by means of which the leukocytes are held at bay. The 
properties we are, at present, considering are probably in no wav 
antichemotactic, but oppose purely the actual ingestion by the 
leukocyte, nor do they seem to depend upon the secretion of sub- 
stances which injure the leukocytes. For, in the first place, a 
profuse accumulation of leukocytes may follow the injection of 
virulent micro-organisms, and Denys (quoting from Gruber) has seen 
active phagocytosis of virulent pneumococci, but none of virulent 
streptococci when antipneumococcus serum was injected with the 
mixture. 
Rosenow 64 has carried out a thorough investigation dealing wfith 
these relations in pneumococcus infection. Seventy-five strains of 
this organism were all found non-phagocytable when first isolated 
and the resistant condition was associated with virulence for rabbits 
and guinea pigs. It was found, moreover, that the resistance to 
phagocytosis was dependent upon the inability to absorb opsonin. 
For, while phagocytable non-virulent pneumococci absorbed specific 
opsonin from serum, the virulent ones failed to do this in proportion 
to the degree of their virulence. Furthermore, extraction of the 
bodies of the virulent organisms in RaCl solution yielded a substance 
which inhibited the action of pneumococcus opsonin—a true anti- 
opsonin—which he speaks of as “virulin.” This discovery, if con- 
firmed, would supply us with a very simple explanation for some 
phases of the problem of virulence. It is, indeed, likely that the 
antiopsonic property is closely bound up with chemical and struc- 
tural changes which take place in the bacterial cell as it adapts itself 
to the parasitic conditions. This is plain from the fact that pneumo- 
cocci and some other bacteria will rapidly lose their virulence when 
cultivated on artificial media devoid of animal serum, will retain it 
longer if grown on some serum media, and will rapidly regain it if 
passed through animals. The formation of a capsule is unquestion- 
ably a morphological evidence of such a change. Habitually capsu- 
lated bacteria, like the Friedlander bacillus, and Streptococcus muco- 
sus, are of fairly constant virulence, while in other micro-organisms 
like the pneumococci, anthrax bacillus, plague bacillus, and certain 
other streptococci, the formation of a capsule goes hand in hand with 
an increase of virulence. By the aid of this morphological earmark 
of virulence, moreover, Gruber and Futaki have obtained further 
proof that the resistance to phagocytosis in these cases is due to the 
nature of the bacterial cell body rather than to any secreted anti- 
opsonic substances. For, after the injection of anthrax bacilli into 
guinea pigs, they saw that leukocytes would take up uncapsulated 
bacilli, apparently picking them out of the midst of surrounding 
capsulated organisms which they were unable to ingest. 
64 ftosenow, Jour. Inf. Dis., Yol, 4, 1907, FACTORS DETERMINING PHAGOCYTOSIS 
389 
THE OPSONIC INDEX 
Wright’s65 investigations upon phagocytosis were, indirectly, 
the outcome of his earlier work upon antityphoid vaccination. His 
purpose in these studies had been a purely practical one, and he 
had attempted to obtain a guide for the dosage and the interval be- 
tween injections by measuring the bactericidal and agglutinating 
powers of the blood serum. In the case of typhoid immunization 
this was indeed a practicable method of control, since the bacteri- 
cidal power of the blood serum rose directly as the immunization of 
the patient was attained. In the cases of many other bacteria, how- 
ever, this method of study was not practicable, and Wright, as.others 
before him, did not find a regularly increased specific bactericidal 
power in the blood sera of immunized animals or of patients con- 
valescing from infections with such bacteria as the staphylococcus, 
streptococcus, Micrococcus melitensis, the Bacillus pestis, and a num- 
ber of others. In fact, together with Windsor,66 he showed that nor- 
mal human blood has practically no bactericidal power for pyogenic 
staphylococci and that antistaphylococcus inoculations or recovery 
from an infection do not result in the production of such proper- 
ties in the serum. These determinations are practically identical 
with Nutt all’s 67 earlier studies on the same bacteria and, indeed, cor- 
respond with the data obtained by Metchnikoff and his followers in 
their work on anthrax infection. For, in discussing these investi- 
gations, we saw that very often the serum of a comparatively resist- 
ant animal is less potently bactericidal than that of a more suscep- 
tible one. We need only recall the difference between rabbits and 
dogs in this respect. The serum of the former is more strongly bac- 
tericidal than that of the latter, and yet rabbits are the far more 
susceptible animals. These relations have been studied with great 
care, also, by Petterson.68 It was logical in such cases to look for 
the cause of resistance in the activity of the phagocytes, and this, 
we have seen, Metchnikoff did successfully in a large series of cases, 
both as regards natural and acquired immunity. 
Yet the controversy between the strictly humoral and the cellular 
schools was by no means regarded as closed, especially since, in such 
cases as typhoid infection, the parallelism between increased resist- 
ance and extracellular bactericidal power was so plainly evident, 
while in this disease particularly (for technical reasons which will 
become clear as we proceed) no such parallelism with phagocytosis 
could at first be shown. It was because of such apparent confusion 
65 Wright. Lancet, 1902; Practitioner, Yol. 72, 1904. 
66 Wright and Windsor. Jour. Hyg., Vol. 2, 1902; and Wright, Lancet, 
1900 and 1901. 
67 Nuttall. Zeitschr. f. Hyg., Yol. 4, 1888. 
68 Pettersou. Centralbl. f. BaJct., Vol, 39, 390 
INFECTION AND RESISTANCE 
that Leishmann 69 undertook to study again the relation of phago- 
cytosis to active immunity, chiefly upon staphylococcus cases that 
were being “vaccinated” therapeutically by Wright himself. 
In order to obtain a numerical measure of the degree of phago- 
cytosis, he developed a simple technique which, though crude, served 
to give him the information he sought. It consisted in taking small 
quantities of the blood of patients and mixing these in capillary 
pipettes with equal volumes of bacteria suspended in salt solution 
and thus incubating them. 
The mixtures were then placed on slides, covered with a cover- 
slip, and incubated at 37° C. for varying periods. At the end of 
incubation the preparations were smeared upon slides and stained 
by Leishmann’s modification of the Romanowski method, the num- 
ber of bacteria in a large series of leukocytes counted and an average 
taken. 
This method had many serious flaws, chief among them bein.<r 
the liability to coagulation of the preparations and the fact that, in 
each test, the fluid constituents as well as phagocytes, both of them 
variable factors, came from the same individual. While, therefore, 
it was possible to estimate an increase or decrease of general phago- 
cytic power, it was impossible to analyze this in reference to its de- 
pendence either upon the condition of the cells, on the one hand, or 
that of the plasma or serum, on the other. Moreover, the relation of 
the number of leukocytes to that of bacteria in individual tests neces- 
sarily differed, and this, we have seen, adds a variable factor which 
renders it impossible to compare any two experiments with accuracy. 
In spite of these difficulties, however, Leishmann succeeded in 
establishing, in a number of cases of staphylococcus infection, that 
an increased resistance was accompanied by an increased energy of 
phagocytosis. 
Leishmann, however, went no further than this, and interpreted 
his results on the basis of the “stimulin” theory of Metchnikoff. 
The subsequent studies of Wright, which began at the point at 
which Leishmann stopped, have been described in the preceding chap- 
ter and had, as their main result, we have seen, the discovery of the 
opsonins and the final confirmation of Denys’ conception of the true 
mechanism of cooperation between serum and leukocytes in phago- 
cytosis. In order to carry out these studies the technique of Leish- 
mann was quite inadequate, and Wright’s first task was to modify it 
in such a way that reasonably accurate comparative estimates of 
phagocytosis could be made. 
It is necessary to outline Wright’s method briefly in this place 
in order that we may consider possible sources of error and obtain 
69 Leishmann. Brit. Med. Jour., 1, 1902; Trans. London Path. Soc.f Yol, 
56, 1905. FACTORS DETERMINING PHAGOCYTOSIS 
391 
a clear understanding of the conclusions he based on his observa- 
tions. 
Wright recognized that the determination of the degree of phago- 
cytosis, induced by the opsonin of any given serum in a single test, is 
by itself of no value, since the actual number of bacteria taken up 
by each leukocyte, apart from the 
opsonic contents of the serum, de- 
pends also upon such purely technical 
factors as the concentration of the 
bacterial emulsion, the relative num- 
ber of leukocytes, and the length of 
time of incubation. Two individual 
tests, therefore, carried out with the 
serum of the same patient at the same 
or at different times, with different bacterial emulsions or leukocytes 
in each, would give variable results, even though the opsonin con- 
tents themselves were entirely alike. 
In order, therefore, to obtain a relative estimate of the opsonic 
contents of any serum it is necessary to compare the phagocytic ac- 
tivity induced by this serum with 
the similar power of another sup- 
posedly normal serum, both tests 
being carried out, under exactly 
similar conditions, with the same 
bacterial emulsion and the same 
leukocytes. The average number 
of bacteria found in each leuko- 
cyte in each one of the prepara- 
tions is then the “phagocytic in- 
dex.” The relation of the phago- 
cytic index of the unknown serum 
to that of the supposedly normal 
serum constitutes what Wright has 
called the “opsonic index.” 
Instead of using the whole 
blood of the patient Wright takes 
a small amount of blood in glass 
capsules, allows it to clot, and uses 
the expressed serum in his test. 
For comparison with this he em- 
ploys a “pool” of a number of 
specimens of serum from sup- 
posedly normal individuals. By the use of such a serum mixture 
any slight possible variations from the normal in any one of the sera 
are likely to be equalized, and a closer approach to a normal standard 
is attained. 
Wright Capsule for Taking 
Blood to Obtain Serum for 
Opsonic Tests. 
Method of Producing an Even 
Emulsion of Bacteria for Op- 
sonin Determination. 392 
INFECTION AND RESISTANCE 
The leukocytes used in both tests are the same and taken, as a 
rule, from the blood of the worker or from some other supposedly 
healthy person. They are obtained by taking 15 or 20 drops of 
blood from the finger or ear into 5 to 10 c. c. of sodium citrate solu- 
tion, in which the blood does not clot. Brief centrifugalization throws 
down the blood cells, with a thin, buffy coat of leukocytes on top, and 
these are gently taken off with a pipette. This constitutes the leu- 
kocytic cream of Wright’s experiments, and furnishes a uniform leu- 
kocyte factor for the two tests which are to be compared. The bac- 
teria are obtained by emulsifying carefully in salt solution. It is 
very important to obtain an emulsion free from clumps and neither 
Method of Taking Up Equal Volumes of Leukocytes, Blood Serum and 
Bacterial Emulsion in Wright’s Technique for Opsonic-index Deter- 
mination. 
too thick nor too thin, a result which can be secured only by experi- 
ence. 
Equal quantities of serum (unknown and normal “pool” respec- 
tively) are mixed with equal quantities of the bacterial emulsion and 
the leukocytes in capillary pipettes, and the mixtures are incubated 
for fifteen to thirty minutes under exactly similar conditions. At 
the end of this time smears are made upon slides, the preparations 
stained, and the numbers of bacteria in a hundred or more leuko- 
cytes counted in each of the two experiments. The average is taken, 
and from the phagocytic indices thus obtained the opsonic index is 
calculated. For instance, if 
Phagocytic index (normal pool) = 8 
Phagocytic index (patient’s serum) = 6 
then the opsonic index (patient’s serum) = 0.75. Or, if the phago- 
cytic index of the normal pool had been 10. and that of the patient’s 
serum 15., then the opsonic index of the patient’s serum, higher than 
normal, would be 1.5. 
For the insurance of accuracy in carrying out this method Wright 
calls especial attention to the caliber of the capillary pipettes that 
are used, the concentration of the sodium citrate solution, which 
should he 1.5 per cent., and the freshness of the leukocytes. But 
it is still necessary to remember that with the greatest care in tech- 
nique uncontrollable sources of error influence this method. Most 
important among them are the differences necessarily existing be- 
tween different normal sera used for comparison and differences FACTORS DETERMINING PHAGOCYTOSIS 
393 
in the agglutinative powers of the sera used in the two specimens. 
For it is plain that different degrees of agglutination may bring 
about great variations in the number of bacteria with which the 
individual leukocyte comes into contact. 
Wright’s method has also been particularly unsatisfactory in 
taking the opsonic index against such bacteria as the tvnhoid bacillus 
and the cholera spirillum, or- 
ganisms which are very rap- 
idly digested after being 
taken up by the leukocytes. 
In consequence, even after as 
short an incubation time as 
five or ten minutes, the in 
gested bacteria are partly dis- 
integrated, are stained in- 
distinctly, and cannot be 
counted with accuracy. Ii 
order to avoid this source oi 
error Klien 70 has devised £ 
modification which depend 
upon gradual dilution of the 
serum in a series of pha 
gocytic tests with the 
same leukocytic and bacteria' 
emulsions. In this wav he 
determines the degree of di 
lution of the serum to be 
tested at which phagocytosis no longer exceeds that taking place in 
salt solution alone. The degree of dilution at which this result was 
obtained has been called by Simon the “coefficient of extinction.” A 
comparison of sera with regard to this value, it is clear, furnished an 
estimate of their quantitative opsonic properties quite as instructive 
as the direct estimations by the Wright method, and in our opinion, 
at least, more reliable. Though also subject to some of the objections 
advanced against the Wright method, it has the definite advantages 
mentioned above, and is not so closely dependent upon irregularities 
in counting, agglutinin influences, and differences in relative pro- 
portions of bacteria and leukocytes employed. Jobling 71 has used 
this method with success for the standardization of antimeningitis 
serum. 
A further modification suggested by Simon, Lamar, and Bis- 
nham 72 denenrl ft 11 iron fi PHTTiTtinatirm nf tVm V»zvrl nnrl c% 
Leukocytes Containing Bacteria. 
Drawing op Field as Seen in 
Wright’s Method op Opsonic-index 
Estimation. 
70 Klien. Johns Hopkins Hosp. Bull., Yol. 18, 1907. 
71 Jobling. “Studies from the Rockefeller Inst.,” Yol. 10, 1910, p. 614. 
72 Simon and Lamar. Johns Hopkins Hosp. Bull., Vol. 17, 1906; Simon, 
Lamar, and Bispham, Jour. Exp. Med., Vol. 8, 1906; Simon, Jour. A. M A 
Vol. 48, 1907, p. 139. 394 
INFECTION AND RESISTANCE 
modification in the method of counting. They make comparative 
tests of the same serum, diluted from 1 to ten to 1 to one hundred in 
salt solution, and estimate the opsonic power, not by determining 
the average number of bacteria to the leukocyte, but by taking a per- 
centage of the total number of leukocytes which take part in the 
phagocytosis, that is, contain any leukocytes at all. The bacterial 
emulsion for this method should be so thin that, in normal serum, 
only about 50 per cent, of the leukocytes will contain bacteria. 
That Wright’s method, or any of the others, gives absolutely 
accurate results will hardly be claimed by any one who has worked 
upon opsonic-index estimations. There are certain uncontrollable 
variable factors, some of which have been pointed out above; and, 
apart from these, the delicacy of the technique is such that reliable 
results can ordinarily be obtained only by trained workers after con- 
siderable practice and experience. Even in such hands the per- 
centage of personal error is more likely to be above than below 10 
per cent. For ordinary clinical purposes, therefore, in the control 
of cases the estimation of the opsonic index is not often practicable. 
On the other hand, there can be little doubt about the fact that 
careful comparative estimation, by Wright’s method and by some 
of the modifications, carried out by workers with experimental train- 
ing and consequent attention to extensive controls, have yielded 
results of sufficient accuracy to permit the recognition of definite 
facts concerning opsonins. it is beyond question, therefore, that the 
Conclusion regarding the relation of opsonic fluctuations to clinical 
conditions and the general significance of opsonins emanating from 
laboratories like those of Wright, jSfeufeld, Hektoen, and some others 
may be accepted as fact—especially since in most essentials such 
workers have agreed. In consequence we are now in possession of 
knowledge regarding the opsonic constituents of the blood in health 
and disease, and in the course of active immunization with bacterial* 
vaccines, which is of the greatest practical importance. We may 
summarize the results of such investigations by saying that in many 
of the infections of man the resistance of the patient is roughly pro- 
portionate to the opsonic index—and that properly spaced inocula- 
tion with suitable quantities of dead bacteria (vaccines) will raise 
the opsonic index and lead to recovery in many of the localized 
subacute and chronic conditions. 
Relation of Opsonic Index to Clinical Condition.—As to the use- 
fulness of the treatment in various infections and the limitations 
within which we may hope for results opinions differ, and these will 
be discussed more fully below. Before we proceed to this, however, 
it will be useful to consider the studies upon which the parallelism 
between opsonic index and clinical condition was founded. 
Wright’s own earlier studies were made chiefly upon staphylococ- 
cus infections and tuberculosis. Since then the method has been FACTORS DETERMINING PHAGOCYTOSIS 
395 
applied to almost all known infections with varyingly successful 
results. 
One of the first steps in determining such a parallelism between 
the resistance of a patient and the opsonic index consisted, of course, 
in comparing the index of the sera of normal individuals with that 
of patients suffering from infection. Wright and Douglas did this 
in a large series of studies. In the case of staphylococcus infections 
the following experiment will illustrate their results: 
(Wright and Douglas, Proc. Royal Soc., Vol. 74, 1904.) 
Showing the ratio in which the phagocytic or opsonic power of the 
patient’s blood stood in each case to the phagocytic or opsonic power of the 
normal individual who furnished the control blood. (The phagocytic power 
of the control blood is taken in each case as unity.) 
TABLE i 
Initials of Patient 
Form of Staphylococcus Invasion 
Opsonic Index 
E. E. 
Furunculosis 
0.48 
F. F. 
Sycosis 
0.49 
J. E. 
Acne 
0.64 
J. H. 
Furunculosis 
0.87 
W. B. 
Acne 
0.55 
E. H. 
Acne 
0.82 
W. H. 
Furunculosis 
0.79 
R. G. 
Furunculosis 
0.7 
G. L. 
Acne and sycosis 
0.74 
S. C. 
Furunculosis 
0.87 
W. L. 
Furunculosis 
0.88 
W. P. 
Furunculosis 
0.39 
S. F. 
Very aggravated sycosis 
0.1 
E. F. D. 
Acne 
0.73 
D. C. 
Sycosis 
0.8 
J. M. 
Acne 
0.48 
W. M. 
Sycosis 
0.37 
E. P. 
Acne 
0.6 
M. S. 
Pustular affection of lips 
0.6 
F. V. 
Repeated staph, infection 
0.47 
In this series, as in others investigated by Wright and his col- 
laborators, staphylococcus infection was uniformly associated with 
a low index. He concludes that there is probably a causative rela- 
tion between the two facts, in that under conditions of depressed 
phagocytic powers staphylococci may gain a foothold, while under 
ordinary normal conditions they would fall prey to phagocytic de- 
struction soon after entering the body. 
The study of the opsonic index during the treatment of such 
cases with dead staphylococcus cultures (usually with the organisms 
cultivated from the patient’s own lesions—“autogenous vaccines”) 396 
INFECTION AND RESISTANCE 
revealed a striking coincidence between the rise of the opsonic index 
and improvement in the clinical conditions. A number of further 
interesting and practically important points were brought out by 
the systematic study of these relations which may be illustrated by 
reproducing a plan of the opsonic index curves constructed from 
cases. 
The curve shown above, and taken from a paper by Wright and 
Douglas, illustrates the course of the opsonic fluctuations in the 
case of a medical student who Tiad suffered for four years from boils. 
When first seen the opsonic index (1. being normal) was 0.6, 
Curve I.—Result Upon Opsonic Index of Vaccine Treatment in Two Cases of 
Chronic Staphylococcus Furunculosis. 
(After Wright and Douglas, Proc. Royal Soc., Vol. 74, 1904, p. 156; also from 
“Studies on Immunity,” p. 41.) 
and there were 2 boils oil the neck. For 3 days after this there was 
a spontaneous rise in the index accompanied by an improvement of 
the lesions. 
On the third day 2 billion staphylococci were injected. This 
was followed by an immediate drop of the phagocytic power—(the 
negative phase) ; together with this a new boil began to form. Soon, 
however, the opsonic power began again to rise, this time consid- 
erably above normal, reaching its highest point on the 8th day, when 
it again began to diminish. A second inoculation on the 12th day 
was followed by a similar preliminary negative phase, then a steady 
and rapid positive phase, which was accompanied by cure. 
The rise and fall of the opsonins after the injection of bacteria 
is entirely analogous to the similar fluctuations of other antibodies FACTORS DETERMINING PHAGOCYTOSIS 
397 
after antigen injections. Measurements of this kind are numerous 
in the literature. Thus Salomonsen and Madsen, measuring the 
antitoxin contents of the blood and milk of a mare which were being 
immunized by injections of diphtheria toxin, obtained the following 
curve, which is entirely similar in essential features to those con- 
structed for the opsonic index by Wright and Douglas: 
Curve Describing Quantitative Measurements op Antitoxin in a Mare in 
Response to Toxin Injections. 
(Taken from article by Salomonsen and Madsen, Ann. de I’Inst. Pasteur, Yol. 11, 
1897, p. 319.) 
Results having the same general significance are apparent in the 
measurements made upon a tetanus toxin goat by Ehrlich and 
Brieger,73 and in the observations upon the fluctuations of bacteri- 
cidal power of the sera of patients treated with typhoid vaccines 
made by Wright74 himself. Similar, again, are the various ag- 
glutinin curves constructed by Jorgensen and Madsen 75 and others. 
Apart from the purely theoretical value of such measurements, 
they demonstrate features which are therapeutically of the greatest 
importance. They show that in all processes of active immunization 
the injection of antigen is followed almost immediately by a rapid 
decline of specific antibodies in the blood serum. This “negative” 
phase, as it is called, is possibly due to a neutralization of existing 
antibodies and lasts for varying periods, which must, of course, de- 
pend upon complex relations between the degree of resistance (or 
amount of antibody constituents of the serum), the quantity of anti- 
gen injected, and the general recuperative powers of the subject. 
73 Ehrlich and Brieger. Zeitschr. f. Hyg., Yol. 13. 
74 Wright. Practitioner, Yol. 72, 1904, p. 118. 
75 Jorgensen and Madsen. Festschrift. Serum Institut. Kopenhagen, 
1902. 398 
INFECTION AND RESISTANCE 
Therefore, without some control like that furnished by the measure- 
ment of opsonins or other antibodies it is impossible to determine 
whether the negative phase has ended or is still in progress unless the 
clinical condition is of such a nature or location that degrees of im- 
provement or exacerbation are well marked and easily observed. 
Even then clinical observation alone is at best not an absolutely 
reliable guide. 
The practical importance of the question lies in the harm which 
may accrue to the patient if a second injection is practiced before 
the cessation of the nega- 
tive phase. Wright him- 
self accentuates this dan- 
ger by expressing the 
opinion that, in typhoid 
inoculations, an excessive 
dose administered to a pa- 
tient in the physiological 
condition of the negative 
phase may be followed by 
a prolongation of this 
phase into a period of sev- 
eral months. 
In the case of succes- 
sive inoculations, as in vaccine treatment, a too rapid repetition— 
i. e., a repetition of injection during such a period of depression— 
leads to what YV right speaks of as a “summation of the negative 
phase,” which obviously may seriously aggravate the condition of the 
case. 
It is to such a cumulation of the negative phase that Wright 
attributes the failures attendant upon the use of tuberculin during 
the early days after its introduction, since injections at this time 
were carried out without any control of serum reactions in the 
patient and with comparatively large doses. 
The danger to be carefully avoided, therefore, is a too rapid 
succession of inoculations and too large a dosage, since both of these 
procedures may be followed by cumulation of the “ebb tide of im- 
munity,” and great harm may result. On the other hand, if the 
treatment is so spaced and measured that the successive inoculations 
are given just before the positive phase has ended—in other words, 
just before the apex of the curve is reached—a moderate negative 
phase may be then followed by a second positive phase still higher 
than the first, and corresponding improvement will result. It is even 
possible to occasionally obtain a summation or cumulation of the 
positive phase—in which the negative phase will be entirely sup- 
pressed. This is illustrated in the following curve, in the case of 
Prolongation of the Negative Phase Due 
to Too Vigorous Treatment with 
Typhoid Vaccine. 
(After A. E. Wright, Brit. Med. Jour., May 
9, 1903. Also from “Studies on Iin- 
munity,” p. 179.) FACTORS DETERMINING PHAGOCYTOSIS 
399 
the first and second inoculation indicated on the chart. This case, 
too, was a staphylococcus infection occurring in a laboratory at- 
tendant : 
Such a summation of positive phase, though of course the ideal 
to be aimed at, cannot be produced with regularity, however carefully 
we may attempt to control the treatment. It is worth mentioning, 
moreover, a fact which should become evident from the preceding 
and is too often overlooked, that a summation of the negative phase 
can certainly be attained by the frequent repetition of larger doses. 
Staphylococcus Index as Determined by Wright in a Case op Acne Treated 
with Staphylococcus Vaccines. 
Note summation of positive phase after third injection. (After A. E. Wright, 
“Studies on Immunity,” p. 348.) 
This is practiced not infrequently in the false hope of hastening the 
acquisition of immunity, and does harm more often than good. 
Ordinarily the opsonic index when raised to a level considerably 
above normal will gradually recede to the normal or even to a sub- 
normal condition. In isolated cases, however, especially in tubercu- 
losis, the index may remain high for periods as long as a month. 
This Wright speaks of as a sustained “high tide” of immunity. 
These laws of fluctuation are all of them entirely analogous to those 
long well known in the cases of other antibodies, for even in diseases 
in which the immunity following an attack—(typhoid fever, cholera, 
plague, and others)—is continued through life the antibodies disap- 
pear from the blood after varying periods and we are forced to seek 
the cause of the permanently high resistance, not in the circulating 
blood, but in the ultimate physiological units—the cells and tissues. 
According to Wright also, the treatment with vaccines may 
be either reenforced or entirely replaced by a process of autoin- 
qculation from the patient’s own lesion by increasing the local cir- 400 
INFECTION AND RESISTANCE 
dilation, thereby throwing more of the specific antigen into the 
blood stream. 
This reasoning has been applied, not only to the treatment of 
tuberculosis and other conditions, but has been utilized to explain 
fluctuations in the opsonic indices of untreated patients under the 
influence of unusual motion of the diseased parts—as in walking or 
other exercise. Wright’s meaning is well illustrated by the follow- 
ing curve of opsonins in a case of gonorrheal polyarthritis in which 
massage of the joints resulted in reactions similar to those ordinarily 
elicited by vaccine injections: 
A further modification of the vaccine treatment of Wright origi- 
nated in the observation that the exudate present in many infected 
Opsonic Curve in a Case of Gonorrheal Arthritis in Which Auto-Inocula- 
tion by Massage Was Practiced. 
(After Wright, Douglas, Freeman, Wells and Fleming, “Studies on Immunity,” 
p. 373.) 
foci is often very much less rich in opsonins than is the blood serum 
of the same patient. This is not unlikely to be due to an absorption 
of the antibodies by the bacteria—as well as by the tissue detritus in 
the lesion. But Wright has interpreted it as a purely specific ab- 
sorption by the bacteria, and has utilized it for diagnostic purposes. 
Thus, with Reid,76 he has examined in this way the comparative 
amounts of tubercle bacillus-opsonins in the blood, and in the local 
exudates (peritoneal fluid) in cases suspected of tuberculosis, and 
has determined the tuberculous nature of the condition by showing 
a discrepancy between the two. These results have not been uni- 
versally confirmed.77 But therapeutically, because of this supposed 
lack of opsonin in the fluid of lesions, Wright has advised the in- 
crease of the local flow of lymph by poulticing, heat, drainage, Bier’s 
cups, X-rays, Finsen light, and other means of accomplishing this 
purpose. 
All that has gone before (most of it taken directly from the 
78 Wright and Reid. Lancet, 1906; Proc. Roy. Soc., Vol. 77, 1906, 
77 Opie- Assoc, of Am, Phys., Washington, 1907, FACTORS DETERMINING PHAGOCYTOSIS 
401 
staphylococcus studies of Wright and his immediate followers) has 
tended to show a very close correspondence of clinical improvement 
with the increased opsonin contents of the blood. 
As applied to other infections, such as gonococcus arthritis, colon 
bacillus cystitis, localized pneumococcus lesions, and many other 
conditions of a localized character, observations of a similar general 
significance have been made. Such reports have been made, apart 
from the Wright school, by Emery,78 Potter, Ditman, and Bradley,79 
Potter,80 Tunnicliff,81 Whitfield,82 Cole and Meakins,83 and many 
others, and we may say with reasonable accuracy that, in localized 
infections particularly, there is much evidence to show that clinical 
improvement and rise of the opsonic index go hand in hand. 
There have been many exceptions to this—which, in view of the 
complicated factors involved in immunization, as well as the diffi- 
culty of the technique, is not surprising. 
In tuberculosis—in which many of Wright’s earlier studies were 
made—the parallelism has not been so consistent. Thus even the 
early work of Bullock 84 showed that, in contrast to similar staphy- 
lococcus investigations, the tuberculo-opsonic indices of patients may 
occasionally be higher than normal, and similar observations were 
made by Lawson and Stewart 85 in cases of acute pulmonary tubercu- 
losis. 
However we analyze the work done on tuberculo-opsonins—and 
the investigations on this subject are far too numerous to be here 
reviewed-—we are forced to the conclusion that in this disease the 
opsonic fluctuations are far more irregular than in most other condi- 
tions. Much,86 for instance, found no regular differences between 
the tubercle bacillus opsonins of healthy and of diseased individuals, 
and Koehlisch 87 obtained similar results, adding the important ob- 
servations that animals that show a high natural resistance to the 
human type of the tubercle bacillus invariously show an opsonic 
index much lower than that of man. 
We may question with much justice, therefore, whether in the 
case of this bacillus opsonic investigations can be looked upon as 
indicators of immunity with as much confidence as in cases of other 
bacterial invasions. It is true, indeed, that tubercle bacilli—as well 
as leprosy, rat leprosy, and other acid-fast bacteria—are eagerly 
78 Emery. “Immunity, etc.,” Lewis, London, 1909. 
79 Potter, Ditman, and Bradley. Jour. A. M. A., Yol. 47, 1906, p. 1722. 
80 Potter. Jour. A. M. A., Yol. 49, 1907, p. 1815. 
81 Tunnicliff. Jour. Inf. His., Yols. 4 and 5, 1907 and 1908. 
82 Whitfield. Practitioner, May, 1908. 
83 Cole and Meakins. Johns Hopkins Hosp. Bull., Vol. 18, 1907. 
84 Bullock. Transact, of Lond. Path. Soc., Yol. 56, 1905, and Lancet, 
1905, Vol. II, p. 1603. 
85 Lawson and Stewart. Lancet, 1905, Vol. II, p. 1406, 
86 Much. Munch, med. Woch., p. 496, 1908. 
87 Koehljsch. Zeitschr. f. Hyg., Ygl, 68, 1911, 402 
INFECTION AND RESISTANCE 
taken up by polynuclear leukocytes when they are injected into the 
peritoneal cavity of a guinea pig or rat or other experimental ani- 
mal. On the other hand, we have much evidence which seems to 
show that such phagocytosis is not in these cases a direct method 
of bacterial destruction. In another place we have cited the experi- 
ments of Tschemorutski,88 which showed that polynuclear leuko- 
cytes, though containing other ferments, were devoid of lipase. And 
Carey and the writer—experimenting with rat leprosy bacilli—found 
that these acid-fast bacteria were not disintegrated within leukocytes 
in the course of weeks, while they were often subject to rapid de- 
struction in the presence of living spleen cells in plasma. Further- 
more, in the discussion of the tuberculin tests we have reviewed 
evidence which points to the fact that in the reactions to 
tubercle bacilli we have probably to deal more particularly with 
sessile receptors on fixed tissue cells than with specific circulating 
antibodies. Bartel and Neumann 8a have concluded that the phago- 
cyte which takes up tubercle bacilli represents only a preliminary 
vehicle by which the micro-organisms are conveyed to the spleen and 
lymphatic tissues, in which actual destruction then takes place. 
While no final conclusions can be drawn from the available evidence, 
all these data render it uncertain whether the opsonic index as de- 
termined for polynuclear phagocytosis may be at all regarded as a 
reliable indication of increased or diminished resistance, and on 
this basis the control of therapy in tuberculosis by opsonin estima- 
tions is of course placed upon an uncertain basis. 
We have then very briefly traced the work done upon opsonin 
determinations from the purely practical point of view. There is 
of course no question about the scientific accuracy of the observa- 
tions upon which rests our knowledge of the opsonic properties of 
blood serum. There is also no doubt concerning our ability to in- 
crease the immunity of an individual by systematic treatment with 
vaccines made of pure cultures of bacteria. However, the work of 
Wright has concerned itself with two distinct questions which must 
be separately answered. Briefly stated these are: 1. What is the 
value of opsonic estimations in controlling the therapeutic vaccina- 
tions of patients? 2. To what degree and in which particular 
conditions may the process of vaccination (active immunization) be 
regarded as a hopeful method of therapy? 
The first question has, in part, been answered in the preceding 
paragraphs. Reasonably accurate comparative estimations of the 
opsonic properties of serum can unquestionably be made by Wright’s 
method, or some of its accepted modifications, in the hands of trained 
workers who look upon each estimation as an experimental problem 
88 Tschemorutski. Hoppe-Seyler’s Zeitschr. f. Phys. Chem., Yol. 75, 1911. 
89 Bartel and Neumann. Wien, klin, Wqc1i„ Nqs, 43 and 44, 1907 ; Cen 
tralbl, /. PnktXf YqI, 48,1909, FACTORS DETERMINING PHAGOCYTOSIS 
403 
and have time for control and repetition. That even in such cases 
the matter is difficult is amply testified to by such reports as that of 
E. C. Hort,90 who states that two of the most skilled experts 91 in 
London, working with samples of the same serum taken before and 
after vaccination, reported—“the one that the index was raised, the 
other that it was lowered by the treatment.” This, and similar ex- 
periments of other observers, do not, of course, invalidate the results 
obtained in special researches like those of Wright, Neufeld, and 
others, but they do indicate that the control of clinical cases by 
opsonic estimations is not a matter that can profitably be made a 
routine procedure by which the treatment of the cases can be regu1 
lated. As a problem of clinical research in a given series of patients 
opsonin studies are unquestionably valuable and the comparative data 
so obtained have proved, and will continue to prove, of great value. 
But we cannot hope as yet, it seems to us, to utilize this method, ex- 
cept in cases in which much time and care can be centered upon a 
few patients under the best conditions. Opinions essentially similar 
to this have been expressed by experienced clinicians (Potter,92 for 
instance), who have followed out series of cases on which systematic 
opsonin determinations were made. 
As to the opsonic index in tuberculosis, we believe that the ex1 
perimental evidence at present available does not show that such 
measurements are reliable measures of resistance, and, in this dis- 
ease, even when the index is taken with a degree of care which 
precludes gross error, it is doubtful whether its estimation is of as 
much value in controlling treatment as are the data obtained by 
skilled clinical observation. 
This leaves us, therefore, for the control of vaccine treatment in 
the routine work of the clinic only the information gleaned from 
such indications as alterations in any visible or palpable lesions, 
general systemic symptoms, temperature, leukocytosis, etc. Since 
these will present such manifold and variable pictures in different 
conditions, generalization is useless. 
90 Hort. Brit. Med. Jour., Feb., 1909, p. 400. 
91 Quoted from Adami, Trans. Amer. Phys. dfc Surg., Yol. 8, 1910. See 
also Pearson, Biometrica, 1911. 
92 Potter. Loc. cit. CHAPTER XVI 
ANAPHYLAXIS 
Historical Survey.—The fundamental principle of active im- 
munization is the fact that the treatment of animals with bacteria 
or bacterial products, carried out according to certain empirically 
determined methods, leads to increased tolerance or resistance. The 
limitations within which this statement is true, and the variable 
factors to which it is subject, we have considered in the foregoing 
discussions dealing with the antibody-antigen reactions. 
Although these reactions were studied at first purely from the 
point of view of increased resistance to infection, the most extensive 
studies of antibody formation have been made with such antigens as 
blood cells, serum, and other substances which are in themselves en- 
tirely harmless. For, in such reactions, great simplicity and ease of 
experimentation could be attained. For a time, therefore, the pri- 
mary problem of increased tolerance or resistance was relegated to a 
secondary position, or, at least, dealt with chiefly by analogy, and the 
phenomena of increased antibody formation and increased resistance 
to the antigen were assumed to maintain a more or less strict paral- 
lelism. 
That the problem is not as simple as this has gradually become 
obvious. We have come to recognize that the treatment of animals 
with any antigen, bacterial or otherwise, though leading to increased 
tolerance under certain conditions and within definite limits, may, 
under other conditions, give rise to the very opposite, that is, to an 
intolerance or increased susceptibility. 
The development of this knowledge, like much else that serum 
study has revealed in the last fifteen years, takes root in isolated 
observations scattered throughout the early literature, but often 
regarded as merely noteworthy accidents or technical errors. This 
particular problem, moreover, was confused by the fact that some of 
the earliest observations regarding hypersusceptibility were made in 
the course of experimentation with diphtheria and tetanus toxins, 
antigenic substances toxic in themselves and, therefore, as wre shall 
see, clouding some of the basic principles apparently involved in the 
phenomenon of which we now speak as anaphylaxis. We will for the 
present, therefore, limit our discussion to the development of the 
knowledge of anaphylaxis merely as it concerns the hypersuscepti- 
bility incited in animals and man by treatment with various antigens, 
404 ANAPHYLAXIS 
405 
such as animal sera and other proteins, which possess but slight 
native toxicity or no toxicity whatever in themselves. 
The special problem of toxin (“Giftuberemp- 
findlichkeit” of von Behring) we will deal with later in a separate 
section, since it is as yet very doubtful whether these phenomena 
may justly be incorporated with true anaphylaxis as we now define 
it, despite the admitted fact that attention was called to the prob- 
lems of acquired susceptibility largely because of these toxin in- 
vestigations. 
The earliest observation having direct bearing upon protein anaphylaxis 
is one which Morgenroth discovered in the writings of Magendie. Morgen- 
roth 1 mentions that, in his “Vorlesungen iiber das Blut,” published in 1839, 
Magendie describes the sudden death of dogs which had been repeatedly 
injected with egg albumen. Although Morgenroth, whose paper was written 
before the .present facts regarding hypersusceptibility were fully developed, 
attributes these results to the action of precipitins, there can be little doubt as 
to the anaphylactic nature of Magendie’s results. 
A clear statement of the fundamental phenomena was given, also, by 
Flexner,2 in 1894. In describing certain experiments he says: “Animals 
that had withstood one dose of dog serum would succumb to a second dose 
given after the lapse of some days or weeks, even when this dose was sublethal 
for a control animal.” 
One of the experiments cited to justify this statement is as follows: 
“Two rabbits received y2 of 1 per cent, and 1 per cent, of their body 
weight respectively of dog’s serum, twenty-four hours old, on January 19, 
1894. With the exception of hemoglobinuria, indisposition to move, and 
increased respiration, no ill effects were noted. The animals still showed 
hemoglobinuria on the following day. These symptoms disappeared and 
apparently the rabbits entirely recovered. On February 12, 1894, each re- 
ceived 1 per cent, of their body weight of dog’s serum intravenously. A 
control animal also received 1 per cent, of its body weight of the same serum. 
The two animals that had been previously inoculated died in two and twelve 
hours respectively; the control animal showed only hemoglobinuria which 
disappeared after a day or two.” 
The experiment here quoted is, as a matter of fact, a perfect example 
of what we now know as “active sensitization.” 
However, the isolated observations recorded above were neither correlated 
nor followed out to their logical developments, and a systematic and purpose- 
ful study of the problem was deferred until Richet and Portier 3 attacked it 
in 1902. 
Richet and Hericourt 4 had observed in 1898 that dogs treated with eel 
serum, which is toxic per se, could be killed by a second injection of an 
amount too small to injure normal untreated animals. Some years later 
Richet, in collaboration with Portier,5 determined a similar fact in the case 
of a poisonous substance, “actinocongestin,” which they isolated by extrac- 
tion of the tentacles of actinia. 
1 Morgenroth. “Ehrlich Gesammelte Arbeiten,” Transl., Wiley & Son, 
N. Y., 1906; p. 332 footnote. 
2 Flexner. Medical News, Yol. 65, p. 116, 1894. 
3 Richet and Portier. C. R. de la Soc. Biol., p. 170, 1902. 
4 Richet and Hericourt. C. R. de la Soc. Biol., 1898. 
5 Portier and Richet. C. R. de la Soc. Biol., p. 170, 1902. 406 
INFECTION AND RESISTANCE 
Some of the facts of Richet and Hericourt’s observations are as follows: 
Actinocongestin injected intravenously into dogs in quantities of 0.05 to 0.075 
gram per kilo weight may cause illness, with vomiting, diarrhea, and respira- 
tory distress, but does not kill. A dose of 0.002 gram per kilo causes no 
symptoms in a normal dog. If, however, 0.002 gram of the poison is in- 
jected into a dog which has previously received a sublethal dose and re- 
covered, the result is violent illness and often death. It was obvious, and 
this was clearly stated by Richet, that the first dose had induced a condition 
of markedly greater susceptibility to the poison. 
He, therefore, spoke of the phenomenon as “anaphylaxis” (“action 
anaphylactique de certains venins”) to express its antithesis to prophylaxis 
or protective effects. 
Although it has been disputed by a number of writers that Richet’s in- 
vestigations constitute the beginnings of our modern understanding of the 
anaphylactic phenomena, yet his recognition of the distinct dependence of 
the hypersusceptible condition upon a preceding inoculation with the same 
substance, and his conclusion that a definite incubation time must elapse after 
the first injection before susceptibility is developed, defined two of the most 
important criteria of the condition and initiated purposeful investigations in 
this field. It is true, on the other hand, that, like v. Behring and most of 
his other predecessors, he was working with primarily toxic substances, and 
the final recognition of the general biological significance of the anaphylactic 
phenomenon was necessarily deferred until a similar development of hyper- 
susceptibility was noted in animals injected with various antigens which of 
themselves were entirely harmless. In this the history of anaphylactic in- 
vestigations is similar to that of other reactions to antigen injections, lysin, 
agglutinin, and precipitin formation, in which the first observations were 
made upon pathogenic bacteria or their products, and in which subsequent 
extension of the investigations revealed that the response to inoculation with 
bacterial proteins represented merely a single phase of a general biological 
reaction on the part of animals to treatment with the large class of substances 
known as antigens. 
This generalization of Richet’s observations had really been foreshadowed 
by the observations of Magendie and by the experiments of Flexner quoted 
above, but this work had been lost sight of and the attention of investigators 
was again focused upon the problem mainly by the publication of Arthus 6 
in 1903 on the repeated injection of horse serum into rabbits, and some ob- 
servations made upon guinea pigs by Theobald Smith and communicated by 
him in 1904 to Ehrlich. 
Arthus 7 found that horse serum injected into rabbits by any of the usual 
paths of entrance is entirely innocuous. It is possible to inject 10, 20, or 
even 40 c. c. without harm. If, however, one repeatedly injects small amounts, 
5 c. c. or less, subcutaneously, at intervals of several days, eventually the 
later injections will give rise to infiltrations, edema, sterile abscesses, and 
even gangrene at the points of injection. He recognized that this was not 
due to cumulative action, and that it was not necessary to inject several times 
in the same place to produce the characteristic response. For instance, the 
early injections might be made into the peritoneum, the subsequent ones into 
the skin, and the local reactions to the later injections might nevertheless 
ensue. In other words, he recognized the systemic nature of the phenomenon 
and regarded it as analogous to the observations of Richet in that he spoke 
of the hypersensitive rabbits as “anaphylactises” by a series of preparatory 
injections. 
6 Arthus. C. R. de la Soc. Biol., Yol. 55, p. 817, Reunion biol., Marseille, 
June, 1903. 
7 Arthus et Breton. C. R. de la Soc. Biol., 55, p. 1478. ANAPHYLAXIS 
407 
The “phenomenon of Theobald Smith” is closely related to that of Arthus, 
and was made in the course of the standardization of diphtheria antitoxin 
in guinea pigs. It was noticed that guinea pigs which had been used for 
this purpose and had survived had acquired great susceptibility to subsequent 
injections of normal horse serum made several days or weeks later. 
With these observations as points of departure, together with the studies 
of v. Pirquet and Schick 8 upon the clinical manifestations of antitoxin injec- 
tions into human beings, a number of investigators took up the problem, chief 
among them Rosenau and Anderson, of the United States Hygienic Labora- 
tory, and R. Otto, of the Frankfurt Institute of Experimental Therapy. 
Although the paper of Otto 9 appeared in print a little earlier than did the 
first one of the American workers, the investigations were independent and 
almost synchronous. Their results, moreover, confirm each other in all essen- 
tials. Otto showed that the Theobald Smith phenomenon was entirely inde- 
pendent of the toxin or antitoxin contents of the injected serum, but could be 
produced (though somewhat less markedly) with horse serum alone. He 
also showed that, while a preliminary injection of horse serum “sensitized” a 
guinea pig to a subsequent dose given after an interval of 10 to 12 days, 
the repeated injection of considerable quantities at short intervals produced 
a condition of “antianaphylaxis” or immunity to the later injections. Otto, 
too, excluded from his results the direct relation of the anaphylactic state 
with the possible presence of serum precipitins, a thought suggested by Mor- 
genroth in his interpretation of the observations of Magendie mentioned 
above. 
Rosenau and Anderson 10 had attacked the problem with the primary 
purpose of throwing light upon the occasional accident of sudden death fol- 
lowing the injection of diphtheria antitoxin into human beings. Since the 
detailed description of their extensive investigations would tend to render 
more difficult the exposition of an already sufficiently complicated subject, 
it will be best to tabulate the chief results of this classical series of their 
earlier papers. Briefly, these are as follows: 
1. A single injection of horse serum into guinea pigs, harmless in itself, 
renders these animals hypersusceptible to a subsequent injection given after 
a definite interval or incubation time. 
2. This interval, with the ordinary dosages employed (about 1 to 2 c. c.), 
was about 10 days. Properly carried out injections after this period were 
usually fatal. 
3. The known antibodies, antitoxins, hemolysins, and precipitins, are 
not responsible for the reaction. 
4. The reaction is specific, injections of horse serum sensitizing to horse 
serum only. (The question of specificity will be further discussed below.) 
5. The sensitive condition is transmissible from mother to offspring,11 
the young of sensitized mothers being hypersusceptible to a first injection 
of horse serum. 
6. The reaction is extremely delicate. Rosenau and Anderson succeeded 
8 Yon Pirquet u. Schick. “Die Serumkrankheit,” Deuticke, Wien, 1906. 
9 Otto. “Das Theobald Smithsche Phaenomen, etc., v. Leuthold Gedenk- 
schrift,” Yol. 1, 1905; also Otto in Erganzungsband 2, “Kolle u. Wassermann 
Handbuch,” etc. 
10 Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab. 
Bull. 29, 1906; 30, 1906; 36, 1907; Jour. Med, Res., Yol. 15, 1906, Yol. 16, 
1907; also Jour. Inf. Dis., Yol. 4, 1907, Vol. 5, 1908. 
11 It is important practically, as Anderson points out, that a female guinea 
pig may transmit to its young sensitiveness to horse serum and immunity 
to diphtheria toxin. 408 
INFECTION AND RESISTANCE 
in sensitizing in one case with 0.000001 c. c. (one one-millionth) of horse 
serum. 
7. The hypersusceptible state is not a transient condition, but may last 
a long time. 
8. Sensitization, or the production of the hypersusceptible condition, 
can be carried out, not only with the various animal and vegetable proteins 
employed in the first experiment, but can be brought about by the use of 
extracts of various bacteria. In such cases also the reaction is specific. The 
first determinations with bacterial extracts carried out by Rosenau and Ander- 
son were made with colon, anthrax, typhoid, and tubercle bacilli. 
By these observations, then, the possibility of a direct relation between 
the phenomena of anaphylaxis and infectious diseases in animals was indi- 
cated. 
This, in essence, is the harvest of the two earliest purposeful re- 
searches into this problem. A large number of investigators now 
took up the question, and its further elucidation, as we shall see, has 
proved, not only the most directly fruitful of the phases of recent 
immunological studies, but has thrown much indirect light upon 
antigen-antibody reactions apart from the anaphylactic phenomena 
themselves. 
Varieties and Classification of Phenomena of Hypersuscepti- 
bility.—Before proceeding to considerations of detail, it may he well 
to make a brief survey of the many different varieties of altered 
reaction capacity which have been observed. It has been known 
for a long time that hypersensitiveness to a great many different 
substances in nature existed in the form of drug idiosyncracies, 
tuberculin reactions, hay fever, etc., etc. The relationship of these 
forms of increased reaction capacity to the phenomena described at 
a later date as “anaphylaxis,” is an important and, in many cases, 
an as yet unsolved problem. For this reason, Doerr 12 has made a 
very useful contribution to the study of hypersensitiveness by an 
attempt at a logical classification of these occurrences. Doerr adopted 
the term “Allergy” which was first used by von Pirquet in his studies 
on serum sickness and extended its meaning to include all forms of 
changed reaction capacity, whether the substance to which this altered 
reaction was evident, was an antigen or not. Von Pirquet, then, 
had used the term merely in the sense of signifying altered reaction, 
but he implied in his work that an antigen-antibody mechanism was 
at the basis of Allergy in general. Advances in the knowledge of 
these phenomena since von Pirquet’s earlier use of the term, justify 
the extension of the significance of the word by Doerr. Under 
Allergy, therefore, Doerr gathers together all phenomena of this 
nature. He subdivides Allergy into two main classes, on the one 
hand, altered reaction capacity to substances not antigenic in nature, 
and, on the other hand, a similar condition of the body toward sub- 
12 Doerr, Kolle & Wassermann. Handbuch, Second Edition, Vol. II, p. 
964. ANAPHYLAXIS 
409 
stances that are recognized as true antigens. Abbreviated some- 
what, the following scheme illustrates Doerr’s classification: 
ALLERGY 
1. Hypersusceptibility (and lessened 
susceptibility) to non-antigenic 
substances. 
2. Hypersusceptibility to antigenic 
substances. 
(a) Substances toxic in them- 
selves. 
(b) Protein antigens not pri- 
marily toxic. 
We need not in this place enter into a discussion of the minor 
subdivisions of Doerr’s scheme since these will appear as we discuss 
individual phenomena below. Kecently, changes in this classifica- 
tion have been proposed by Coca.13 Coca has brought a number of 
new points of view into the discussion which, whether we agree with 
his conclusions or not, should, in many of their aspects, be taken 
seriously, since they are thoughtfully formulated and are everywhere 
based on sound comprehension of prevailing conditions. Coca uses 
the word “hypersensitiveness” to include all of the conditions dis- 
cussed, and subdivides them into two main classes, namely, (a) Ana- 
phylaxis, and (b) Allergy. He limits the term “Anaphylaxis” very 
definitely to phenomena in which an antigen-antibody reaction has 
been proven, and applies the term “Allergy” to all conditions of 
hypersensitiveness in which this mechanism has not been shown to be 
the underlying cause of the symptoms. Under the second heading 
he classifies drug idiosyncracies, serum sickness in man (which he 
speaks of as serum allergy), and hay fever. He also separates from 
the conception of anaphylaxis, such reactions as the tuberculin reac- 
tion and toxin hypersusceptibility. It would lead us too far afield 
to discuss in this place the many individual points of possible dif- 
ference of opinion on the validity of Coca’s disagreements with 
Doerr’s classifications. We will take up these points as we discuss 
individual phenomena, and will, in general, hold ourselves to the 
Doerr scheme which we think is reasonable and, for the time being, 
logical. We deplore that Coca should have given a meaning to the 
word “Allergy” different from the original sense in which Von 
Pirquet used it, since this is likely to bring a certain amount of con- 
fusion into the literature without aiding much in clarifying his own 
conceptions of the actual differences in mechanism. 
For our own purposes we believe that the most comprehensive 
way of discussing these phenomena is to group together in one class, 
as true anaphylaxis, those manifestations of hypersensitiveness in 
which we know that an antigen-antibody mechanism is involved, and 
13 Coca. “Hypersensitiveness, Practice of Medicine,’’ F. Tice, New York, 
1920. 410 
INFECTION AND RESISTANCE 
take up in separate sections, then, other occurrences, discussing in 
each individual case how much we know of the mechanism and to 
what extent we have any right to believe that the process varies 
fundamentally from true anaphylactic phenomena. This is more 
or less the same method of treatment given to the subject in a recent 
review by Gideon Wells who also rigidly delimits the term “anaphy- 
laxis” to the protein phenomena referred to, incorporating in this 
class only antigen-antibody reactions, and deals separately with other 
conditions in which there is doubt or in which such reactions can 
be definitely ruled out, at least within our present understanding. 
While we lay the chief stress for purposes of classification upon 
the demonstrable occurrence of an antigen-antibody union as basic 
to the anaphylactic phenomena, we will see that other important 
differences enter. Thus, the question of inheritance, the possibility 
of passive transfer and of desensitization, all play roles which have 
been taken as fundamentally differentiating one variety of hyper- 
susceptibility from another. Our own opinion as we hope to express 
it a little further along, is very definitely that many of these differ- 
ences are more apparent than real, in that there are certain deep 
underlying processes which will bind all these phenomena of specific 
hypersensitiveness together. Indeed, other writers tacitly admit a 
similar opinion by the fact that they include all these matters under 
the heading of hypersusceptibility. Our scheme of treatment for the 
present will be as follows: 
(1) Protein hypersensitiveness or true anaphylaxis in which 
antigen and antibody reactions are involved, in which the inciting 
substance is unquestionably an antigen and in which, as far as we 
know, inheritance from both parents plays no role. 
In close relation to this we will discuss (a) serum sickness, in 
the interpretation of which analogy with anaphylaxis has been denied, 
particularly by Coca, for reasons that will be set forth, and the so- 
called (b) food idiosyncracies in which the inciting substance is 
clearly an antigen but in which the important factor of true in- 
heritance seems to be involved, and (c) Hay Fever in which the anti- 
genic nature of the inciting substance is reasonably an open question. 
(2) Idiosyncrasies to non-antigenic substances, drugs, etc., and 
allied phenomena, in which no antigenic properties can be demon- 
strated for the inciting substances, as such, in which the factor of 
inheritance seems to be a very important influence, and in which the 
possibility of passive transfer is at least doubtful. 
(3) Bacterial hypersensitiveness and specific reactions, such as 
the tuberculin, typhoidin, etc., with a consideration of Bacterial 
Anaphylatoxin. 
(4) Toxin hypersusceptibility. 
(5) The primary toxic action of normal serum. 
(6) Discussion of coordinating facts. ANAPHYLAXIS 
411 
HYPERSUSCEPTIBILITY TO PROTEINS, OR TRUE 
ANAPHYLAXIS 
Although some of the criteria by which True Anaphylaxis must 
be characterized will become clear only after a further study of 
succeeding sections, we believe it well to state them in this place, 
more or less as they are well stated in our opinion by Wells in the 
article referred to. The criteria which should, according to Wells, 
he met in order that a condition may be spoken of as True Anaphy- 
laxis are 
(1) The observed toxicity of the injected material must depend 
upon the sensitization of the animal, that is, the substance should 
not produce similar symptoms in the non-sensitized animal on first 
injection. 
(2) The symptoms produced must be those characteristic of 
anaphylactic intoxication as observed in the usual reactions with 
typical soluble proteins, being, therefore, the same for all antigens 
with the same species of animal, but differing characteristically in 
various species of animal. 
(3) It should be possible to sensitize a normal animal passively 
by serum transfer from a sensitized (or immunized) animal. 
(4) When guinea pigs are used, the typical uterine reaction 
should be demonstrable by the Dale method. 
(5) Desensitization under proper conditions should be possible. 
We omit some of the other criteria stated by Wells because we 
think the above are sufficient to characterize the condition and com- 
prise the fundamentally important ones. 
In discussing anaphylaxis at first in this limited sense, wTe wish 
to make entirely clear, however, that we do not mean to imply that 
our own opinion necessarily agrees with the assumption that there 
is absolutely no connection between protein hypersusceptibility and 
other forms, as maintained by Coca. As we proceed in the discus- 
sion of the various phenomena, it will become clear that, in spite 
of many apparent differences in mechanism, there may still be cer- 
tain fundamental physiological similarities between all forms of 
changed reaction on the part of the animal body to extraneous sub- 
stances. 
The Antigen.—Since, as we shall see, protein anaphylaxis is 
dependent upon a reaction between an antigen and its antibody, 
it is but natural that the sensitizing substances should be in every 
way similar to other antigens. We may state briefly, therefore, that 
any substance which can induce antibody formation by injection into 
an animal, is capable of being used as a sensitizing or anaphylactic 
antigen. Such a statement, as we have seen in other places, is syn- 
onymous with saying that all proteins may be regarded as possible 412 
INFECTION AND RESISTANCE 
anaphylactic antigens. It must not be forgotten, however, that there 
are certain chemically true proteins that have no antigenic prop- 
erties. Wells has gone into this matter thoroughly in the review of 
the subject to which we have referred above. The best known of the 
true proteins which have no antigenic properties is gelatin. Ac- 
cording to Starin,14 whose work Wells regards as sound, the chief 
difference between gelatin and the antigenic proteins depends upon 
the deficiency in gelatin of tryptophane and tyrosin, and a low con- 
tent of phenylalanine; and Wells believes it possible that the de- 
ficiency in aromatic amino acids may be the important factor. Wells 
and Osborne 15 have studied these questions particularly with zein 
from corn, in which glycin and tryptophane are lacking, and with 
gliadin and hordein in which glycin, lysine, arginine and histidine 
are deficient. Wells reasons that the fact that these proteins, so poor 
in diamino acids, are excellent antigens, while certain protamines 
which are composed to a large extent of diamino acids have no anti- 
genic action, may signify that the three diamino acids mentioned 
are of very little antigenic importance. He mentions this merely as 
a suggestion. 
As far as the question of the antigenic activity of protein cleavage- 
products is concerned, the problem in regard to anaphylaxis is en- 
entirely analogous to that bearing on the antigenic nature of such 
substances in general. The anaphylactic reaction has been the one, 
however, very largely used of recent years for the determination of 
antigenic properties with doubtful antigens, largely because of its 
extreme delicacy and, for this reason we discuss the subject in this 
place. Fink,16 working in Wells’ laboratory, studied particularly 
the cleavage products from hydrolyzed egg-white and found slight 
antigenic activity for the fractions precipitated at three-quarters and 
complete saturation with ammonium sulphate, but not for those 
obtained with lower concentrations of the salt. The question we 
are discussing here must not be confused with the problems of pep- 
tone shock, etc., which will be taken up later, since, at present, we 
are interested not so much in the ability of various protein cleavage- 
products to induce toxic, anapliylaxis-like effects upon injection into 
animals, but rather in the properties of such substances to sensitize 
animals to subsequently repeated injection. Among the most impor- 
tant contributions to this particular subject is the work of Zunz 17 
who obtained heteroalbumose and protoalbumose by peptic and tryptic 
digestion of fibrin, and found that both of these substances sensitized 
and intoxicated guinea pigs and rabbits. He was forced to use con- 
siderable dosea and his shock reactions were limited in degree. Also, 
14 Starin. Jour. Inf. Dis., 23, 1918, p. 139. 
15 Wells and Osborne. Jour. Inf. Dis., 8, 1911, p. 66. 
16 Fink. Jour. Inf. Dis., 25, 1919, p. 97. 
17 Zunz. Zeit. f. Immunities., 16, 1913, p. 580. ANAPHYLAXIS 
413 
specificity was not as marked as with proteins. Attempts to repeat 
Zunz’s work, by Friedberger and Joachimoglu 18 were unsuccessful, 
and they regard Zunz’s results as due to the primary toxic action of 
the large doses of beef serum used by him in testing beef-albumose 
sensitized animals. Hailer 19 carried out an extensive investigation 
on a similar subject with digestion products of beef and hog muscle. 
He obtained a certain amount of actual sensitization, but found 
that specificity was lacking and that the sensitization was never in- 
tense. Other investigations, such as those of Schmidt 20 were un- 
successful, and Wells 21 found that even digesting bovine serum as 
long as 16 months with trypsin did not result in a complete loss of 
coagulable material, in spite of the fact that it gave no Biuret reac- 
tion. The subject is an interesting and important one, and for a 
thorough discussion of much of the literature we refer the reader 
to Fink’s article cited above. We believe that in all probability there 
may be a certain amount of antigenic power in the higher cleavage- 
products of protein, but that, with the breaking up of the protein 
molecule, antigenic functions very rapidly diminish, as will be seen 
in our discussion of the tuberculin reaction, we agree with Land- 
steiner 22 that molecular size may be quite as definitely involved as 
chemical structure. This problem will be spoken of again when we 
describe our own work with bacterial antigens. 
There is a considerable literature on the intoxicating effect of 
protein-split products upon animals sensitized with the whole protein. 
Thus, Rosenau and Anderson 23 observed toxic effects when they 
reinjected guinea pigs that had been sensitized with horse serum, with 
partially digested horse serum. Gay and Robertson24 sensitized 
guinea pigs with casein solution, and obtained shock 23 days later 
by the injection of paranuclein, a partial peptic digestion product 
of casein. A number of other such experimental findings could be 
mentioned, but while they are extremely interesting and perhaps of 
great significance, have little bearing on the subject in hand. 
In regard to the antigenic properties of altered proteins, again, 
analogy with antibody reactions in general prevails in anaphylaxis. 
Work by Obermeyer and Pick, Schittenhelm and Stroebel and more 
recent work by Landsteiner has shown that proteins altered by 
iodizing or by the introduction of nitro- and other atom groups, may 
possess an antigenic action which is altered in its specificity, in that 
the chemical group introduced now determines the specific reactions. 
This work will be more particularly discussed in connection with 
18 Friedberger and Joachimoglu. Zeit. f. Immunitdts., 24, 1914, p. 522. 
19 Hailer. Arb. a. d. Kais. Ges., 1914, p. 527. 
20 Schmidt. Univ. of Calif. Pub., 2, 1916, p. 157. 
21 Wells. Jour. Inf. Dis., 6, 1909, p. 506. 
22 Landsteiner. Biochem. Zeit., 43, 1919, p. 106. 
23 Rosenau and Anderson. Hyg. Lab. Bull., 36, U. S. P. H. Service, 1907. 
24 Gay and Robertson. Jour. Exper. Med., 16, 1912, p. 470. 414 
INFECTION AND RESISTANCE 
drug idiosyncrasies, since Wolf-Eisner and others have suggested 
that so-called drug idiosyncrasies may be due to alterations of the 
body protein by the drug in such a way that the new complex antigen 
may act like an altered protein after the drug has been injected. 
Another point which is of great importance in connection with 
antigens in general, and particularly as it affects anaphylaxis, is the 
factor of racemization. Ten Broeck,25 working with proteins race- 
mized by Dakin’s method, that is, so changed in structure that they 
are no longer susceptible to enzyme action and are optically altered, 
found that they no longer were antigenic in anaphylactic reactions. 
This point is of considerable importance in working with protein 
extracts in which alkalin extraction is necessary. 
Here, as in other phases of the subject the question of the possible 
antigenic properties of lipoids has been raised, but again without 
definitely positive results. Pick and Yamanouchi 26 extracted beef 
and horse sera with alcohol, and evaporated and redissolved the solu- 
tions until neither contained coagulable protein nor gave the Biuret 
reaction. With this material they obtained a few positive anaphy- 
lactic experiments. Similarly curious are the results of Bogomolez,27 
who succeeded in sensitizing and producing shock with the lipoids 
extracted from egg yolks. Although such experiments would tend 
to persuade us that lipoidal substances may actually have sensitizing, 
and therefore antigenic, functions, this does not follow necessarily. 
As Pick and Yamanouchi themselves point out, it is practically im- 
possible to demonstrate with certainty the presence of slight traces 
of proteins as impurities in lipoid preparations, and we know espe- 
cially from Rosenau and Anderson’s work how minute are the quan- 
tities of antigen which still serve to sensitize. It is possible, moreover 
(a thought developed particularly by Pick and Schwartz 28 and by 
Landsteiner29), that we are dealing in many cases with combina- 
tions of protein and lipoid—a form of chemical substance of which 
very little is known analytically, but the existence of which many 
biological facts lead us to assume. 
Specificity.—That the anaphylactic reaction is specific we have 
mentioned in the brief summary we have given of Rosenau and 
Anderson’s work. These authors use the adjective “quantitative,” 
by which they simply mean to convey that the specificity here is not 
absolute, any more than it is absolute in the case of any of the known 
serum reactions. An animal sensitized with a certain variety of 
protein, animal serum, etc., reacts with disproportionately greater 
delicacy to a second injection of the same variety than of any other 
25 Ten Broeck. Jour. Biol. Chem., 17, 1914, p. 369. 
26 Pick and Yamanouchi. Zeitschr. f. Immunitatsforschung, Yol. 1, 1909. 
27 Bogomolez. Zeitschr. f. Immunitatsforschung, Vols. 5 and 6, 1910. 
28 Pick and Schwartz. Biochem. Zeitschr., 15, 1909. 
29 Landsteiner. Referat. “Weichhardt’s Jahresbericht,” 6, 1910. ANAPHYLAXIS 
415 
substance. In fact, apart from a few cases mentioned by Gay and 
Southard, there are not many instances of marked non-specific ana- 
phylactic reactions. Still we would expect here, as in other serum 
reactions, a certain limitation in the degree of specificity, and Otto 
recommends the less delicate subcutaneous method of testing for all 
experiments in which questions of specificity are involved. This 
point we will touch upon a little later. 
An interesting addition to our knowledge of such specificity was 
made by experiments of Rosenau and Anderson, which showed that 
a guinea pig could be rendered separately sensitive at one and the 
same time to blood serum, eggwhite, and milk, reacting specifically 
to each on second injection. 
In considering the specificity of antigens, it must be remembered 
that the ordinary substances used for anaphylactic and other antibody’ 
reactions, such as eggwhite, animal serum, etc., are really mixtures 
in which a number of different proteins may be represented and the 
various fractions of such a complex substance may possess individual 
specificity. An extremely interesting instance of this is found in 
the work of Dale and Hartley 30 who, using the guinea pig uterus 
technique of Dale, which is extremely sensitive, and using as antigens 
euglobulin, pseudo-globulin and albumin of horse serum, found that 
these horse serum fractions could be made to sensitize separately. 
Similarly, separate and individually specific antigens have been 
obtained from other proteins, such as egg albumen and plant proteins, 
particularly by Wells and Osborne. The last named authors, too, 
have obtained some extremely interesting results in their attempts to 
correlate chemical structure with specificity. Their opinion is based 
chiefly upon their experiments with vegetable proteins in which in- 
teraction between chemically related proteins could be demonstrated 
by anaphylactic reactions. Their experiments are of unusual im- 
portance, but the entire problem is still in its beginnings. In this 
connection, the more recent investigations of Landsteiner on proteins 
altered by the introduction of diazo bodies, and other atom groups, 
are extremely important in that they show, as Doerr also points out, 
that relatively slight changes in the molecule may considerably 
change and even completely mask specific reactions to the native 
protein. In appraising this work, Doerr in his most recent article, 
sums it up very well by stating that it seems quite certain at the 
present time that one and the same protein antigen may be altered 
into a considerable number of substances of individual specificity, 
and that by similarly altering a number of different proteins simi- 
larly, that is, in chemical treatment, a closely related specificity may 
be produced in them. Along lines, such as these, we may eventually 
hope to obtain a certain amount of insight into the hitherto mysterious 
problem of specificity. 
30 Dale and Hartley. Biochem. Jour., 10, 1916, p. 110. 416 
INFECTION AND RESISTANCE 
In anaphylaxis, again analogous to antibody reactions in general, 
the specificity, as a rule, is one of species. In other words, the pro- 
tein of any animal is specific for the proteins of its particular species 
generally, there being definitely similar characteristics in the body 
proteins of animals of like species, which, though chemically in- 
definable, are nevertheless delicately determinable by biological re- 
actions. In considering specificity of precipitins, however, we have 
seen that there are exceptions to the specificity of species expressed 
in the phenomenon of so-called organ specificity. The same thing 
has been shown for anaphylaxis. Kraus, Doerr and Sohma 31 were 
able to show that animals sensitized with protein from the crystalline 
lens were hypersusceptible to lens protein generally, whether this 
came from the species from which the original lens was taken, or 
whether some other variety of animal had furnished it. On the other 
hand, animals so sensitized, while hypersusceptible to lens protein, 
did not react to injections of homologous blood.32 In other words, 
this organ contains a characteristic variety of antigen (protein) 
peculiar to this kind of organ throughout the different animal species, 
but not common to other tissues and organs of the same animal. 
Results similar to these were obtained by von Dungern and Hirsch- 
feld 33 in the case of testicular protein, although here the phenomenon 
seemed to be less rigidly organ-specific than in the preceding case. 
These writers worked not with the systemic anaphylactic reaction, 
but with the localized (allergie) reaction described above as the 
phenomenon of Arthus. They injected extracts of the testicular 
materials into the ears of rabbits and incidentally made the very 
curious observation that pregnant females would not infrequently 
react to a first injection without previous sensitization. 
Of great importance also in connection with the subject of organ 
specificity is the further claim of Uhlenhuth and Haendel 34 that 
animals can be sensitized with their own lens protein, an observation 
which opens the possibility of other forms of “autosensitization” and 
consequently has given much opportunity for clinical speculation. 
Rosenau and Anderson,35 indeed, have found that guinea pigs can 
be sensitized by means of extracts of guinea pig placenta. They 
have applied this to the possible explanation of eclampsia, and similar 
reasoning, as we shall see, has been utilized in many other condi- 
tions. Attempts have also been made to show, by the anaphylactic 
reaction, that the tissue of malignant tumors possess such “tissue- 
specific” or “organ-specific” qualities. Yamanouchi,36 indeed, claims 
31 Kraus, Doerr and Sohma. Wien. Min. Woch., No. 30, 1908. 
32 Andrejew. Arb. a. d. kais. Gesundh. Amt., Vol. 30, 1909. 
33 Yon Dungern and Hirschfeld. Zeitschr. f. Immunitats., 4, 1910. 
34 Uhlenhuth and Haendel. Zeitschr. f. Immunitats, 4, 1910. 
35 Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab. 
Bull., 45. 
38 Yamanouchi. C. R. de la Soc. Biol., Vol. 66, 1909, p. 754. ANAPHYLAXIS 
417 
to have shown this, but his results were not confirmed by Apolant,37 
and the writer has carried out a series of entirely negative experi- 
ments upon the same subject. However, in view of the great diffi- 
culty of obtaining any kind of anaphylactic reaction in mice, the ani- 
mals in which the tumor experiments were carried out, there is little 
information to be obtained from negative results of this kind. 
The delicate quantitative method of studying problems of speci- 
ficity, which the reaction of anaphylaxis supplies, has further served 
to revive the unsettled question of the “organ-specific” properties of 
the tissues of such organs as the liver, spleen, kidney, blood, etc. 
Indeed, Pfeiffer,38 has published results which would seem to en- 
courage the belief of the existence of such specificity. However, 
Ranzi39 had previously obtained entirely negative results, and 
Pearce, Karsner, and Eisenbrey40 have recently made a careful 
inquiry into the same problem, failing to confirm Pfeiffer’s claims. 
By a similar process of reasoning Elschnig 41 has attempted to 
explain sympathetic ophthalmia. He claims to have shown that the 
laws of organ specificity apply to the proteins (especially the pig- 
ment) of the uveal tract. The destruction and absorption of injured 
uveal tissue, according to him, induce the formation of organ-specific 
antibodies by which the remaining uveal structures of the same, as 
well as of the opposite, eye are sensitized. The consequence is a 
“sympathetic” inflammation which “is to be regarded purely as an 
anaphylactic reaction.” 
These investigations, while regarded with much scepticism by 
immunologists in general, seem to be yielding confirmatory results 
in the hands of Allan Woods who has begun to carry out the prin- 
ciples of Elschnig’s reasoning on human beings suffering from sym- 
pathetic ophthalmia. Although nothing definite can be said about 
this work, as yet, it seems increasingly worth putting to practical 
test in view of Woods’ beginning. 
In this, then, as well as in other respects, the substances by which 
animals may be sensitized are entirely similar to antigens in general. 
Active Sensitization.—Active sensitization is a term analogous 
to active immunization in that it signifies that the animal is sensi- 
tized by the injection of the antigen, and develops its sensitiveness 
in physiological response to the injection. This is the original man- 
ner in which sensitization was first accomplished, and is in contrast 
to passive sensitization which we will discuss directly, in which the 
animal is rendered sensitive not by its own antibody formation, but 
by virtue of antibodies transferred to it from the actively sensitized 
37 Apolant. Zeitschr. f. Immunitdtsf orschung, Yol. 3, 1909. 
38 Pfeiffer. Zeitschr. f. Immunitdtsf orschung, Yol. 8, 1910. 
39 Ranzi. Zeitschr. f. Immunitdtsf orschung, Vol. 2, 1909. 
40 Pearce, Karsner, and Eisenbrey. Jour. Exp. Med., Vol. 14, 1911. 
41 Elschnig. Von Graefe’s Archiv. f. Ophthal., Yol. 75, p. 459; Yol. 76, 
p. 509; Vol. 78, p. 549. 418 
INFECTION AND RESISTANCE 
one. Thus, the terms are in complete analogy to the similar ones 
used in connection with immunization. 
Active sensitization, then, consists in the injection of the anti- 
genic substance into the animal, as a result of which it gradually 
becomes specifically hypersensitive to the injected substance. As 
far as the manner of injection is concerned, we can state, briefly, that 
the conditions are again exactly analogous to those prevailing in 
immunization in that any manner of administration in which the 
antigen may come in contact with the tissue cells of the body without 
previous alteration by digestive ferments (in other words, in its 
antigenic condition) may lead to sensitization. This is synonymous 
with saying that any manner of introduction by which the antigen 
may lead to antibody formation, may sensitize. Thus, sensitization 
may be accomplished by subcutaneous, intravenous, intracardial, 
intramuscular, or any other parenteral method of injection. It is 
probably possible, also, under unusual circumstances, that antigen 
may pass through mucous membranes like those of the eye and 
nasopharynx, and thereby sensitize. Whether or not sensitization 
through the intestinal canal is possible, is a question that has been 
much discussed, especially in an attempt to explain hvpersensitive- 
ness to various food substances in young infants and children, in 
whom it is assumed that the absorption of unchanged foreign protein 
may possibly take place through the intestinal mucosa, either during 
the very early days of life, or later, by reason of injury of the 
mucosa by the intestinal disease accompanying various forms of mal- 
nutrition and diarrhea. Recent work by Schloss and others seems 
to indicate that protein may pass into the body in its antigenic form 
in this way. Such a possibility becomes more plausible when we 
consider the very minute amounts of antigen which may lead to 
sensitization. 
Rosenau and Anderson, in their earliest paper, report success 
in sensitizing guinea pigs by the feeding of horse meat and horse 
serum. McClintoek and King 42 failed to confirm this, and the ob- 
servations of other writers seem to bear them out. However, when 
we consider that Ascoli, Oppenheimer, and others have shown that 
proteins fed to animals in large quantities may be subsequently 
demonstrated not only in the circulating blood but occasionally even 
in the urine by means of the precipitin reaction, there seems to be 
little room for doubting that antigen may enter the circulation un- 
changed, though possibly only under abnormal local conditions of 
the intestine. This, together with Rosenau and Anderson’s demon- 
stration of the extremely small amount of antigen necessary to sensi- 
tize, furnishes all the conditions necessary for anaphylaxis by way 
of the intestinal canal. 
42 McClintoek and King. Jour. Inf. Bis., 3, 1906. See section on normal 
antibodies. ANAPHYLAXIS 
419 
A study made by Lesne and Dreyfus 43 seems to us to have ex- 
plained the contradictory results of other workers on this phase of 
the problem. Without being able to associate the destruction of the 
sensitizing function with either the gastric or pancreatic secretions, 
they were nevertheless successful in showing that sensitization could 
be carried out regularly if the antigen were injected after laparotomy 
into the large intestine, whereas similar injections into the stomach 
or small intestine were negative. In these experiments we must take 
into consideration that the conditions following laparotomy, such as 
temporary intestinal atony and congestion, may have exerted con- 
siderable influence upon the positive outcome of their large intestine 
injections. Whereas they do not, therefore, permit us to assume the 
possibility of sensitization through the normal alimentary canal, they 
nevertheless confirm the assumption of the possibility of sensitiza- 
tion by this path under the influence of the abnormal local con- 
ditions incident to injury or disease. 
In this connection Besredka’s 44 experiments on the production 
of anti-anaphylaxis bv the intestinal administration of protein are 
of interest. He found that, if sensitized animals were given 5 c. c. 
of the antigen (milk) by rectum, they were thereby protected ironi 
the reaction following in controls upon a second injection. In his 
later experiments with egg white it appeared that the protection 
could also be conferred by mouth, but that in this case it developed 
more slowly, it being necessary to wait two days after injestion be- 
fore the anti-anaphylaxis had developed sufficiently to protect. Since 
attempts by mouth were not as rapidly successful as those per rec- 
tum, it is clear that these facts are in keeping with Lesne and Drey- 
fus’ results in showing that the antigen is probably absorbed chiefly 
or solely from the large intestine. In Lesne and Dreyfus’ experi- 
ments the sensitizing dose was given into the intestine, the toxogenic 
or second dose being administered intravenously, and since, as we 
shall see, minute doses may suffice to sensitize, whereas 100 or more 
times the sensitizing amount is necessary to produce intoxication, it 
is easy to understand why sensitization followed in Lesne and Drey- 
fus’ work, but no toxic effects followed in the experiments of Bes- 
redka. Furthermore, the slow absorption from the intestine in these 
experiments explains the development of anti-anaphylaxis in Bes- 
redka’s work, in that they are, in this respect, analogous to later 
experiments of Friedberger, cited below, in which it was shown that 
sensitized guinea pigs, which could (in controls) be killed by rapid 
intravenous injection of 0.1 c. c. of antigen and less, would withstand 
without symptoms many times this amount when it was gradually 
administered by slow injection covering an hour or longer. 
The quantities of antigen which are necessary to sensitize may 
43 Lesne and Dreyfus. C. R. de la Soc. Biol., Yol. 70, p. 136, 1911. 
44 Besredka, C. R. de la Soc. Biol., Yol. 65, 1908; Yol. 70, 1911. 420 
INFECTION AND RESISTANCE 
vary tremendously. The underlying fact governing this can prob- 
ably be stated best by saying that all that is necessary is to introduce 
an amount of antigen sufficient to initiate antibody production. This 
will, of course, vary within very wide limits for the several antigenic 
substances used, depending upon the concentration of the antigenic 
soluble protein in the preparation. When working with substances 
as antigenically potent as horse serum, Rosenau and Anderson 45 
obtained definite sensitization after the injection of as little as one- 
millionth of a cubic centimeter of horse serum. Their experiments- 
with horse serum generally range from a sensitizating dose of this 
minute quantity to as high as 10 c. c. in guinea pig experiments. 
Quantities almost as low as those reported by Rosenau and Anderson 
for guinea pig sensitization have been successfully used by Doerr and 
Russ,46 and still lower quantities have sensitized in the hands of 
Wells when using crystallized egg albumen. 
Such minute quantities are not, however, the most favorable for 
active anaphylactic sensitization, and are mentioned only to show 
how extremely delicate the reaction may be. It is also worth men- 
tioning, in passing, although we will refer to it again, that the second 
injection which elicits shock, must always be considerably larger than 
these small sensitizing amounts. To obtain constant and reliable 
results in guinea pigs, a single sensitizing injection, ranging in 
amounts from 0.01 to 0.25 c. c., is usually sufficient. These amounts 
as given, should be taken to apply only to guinea pigs. We will 
discuss variations as observed in other animals directly. 
Incubation Time.—The hypersensitive state may supervene at 
varying periods after the sensitizing injection, modified largely by 
the amount of antigen injected. Though often so stated, it is not 
likely that the route of injection matters very much, so long as the 
administration is parenteral. For, even though intravenous or intra- 
cardial injection means that the antigen comes into almost immediate 
contact with the tissues, subcutaneously injected antigen probably 
does not lag much more than 72 hours behind at most, and some of 
the antigen will probably get into the circulation in such cases almost 
immediately. The quantity, however, seems to influence the length 
of the incubation time considerably. At first there was an idea based 
upon the earlier observations of Rosenau and Anderson and of Otto 47 
that the length of incubation time was inversely proportionate to the 
size of the sensitizing dose; in other words, moderately small quan- 
tities would sensitize more rapidly than very large amounts. Later 
experiments of Rosenau and Anderson, however, seem to show that 
45 Rosenau and Anderson. TJ. 8. P. H. Bulletin, No. 29, 1906, and No. 45, 
1908. 
40 Doerr and Russ. Zeitschr. f. Immunitdts., 2, 1909. 
47 Otto. Cited from Kolle & Wassermajjn JJandbuch, Ergenzungsb., Np. 
2,4E ' ANAPHYLAXIS 
421 
this relation is not as definite as at first assumed. In the tables given 
by them guinea pigs receiving 0.01 c. c. reacted severely after 14, 17, 
and 155 days; others, receiving 1 c. c., after 14, 17 and 155 days; 
and, again, another series sensitized with 8 c. c. reacted severely after 
similar intervals. All of these series reacted but mildly after 245 
days, showing apparently that the anaphylaxis, contrary to general 
belief, does not last so much longer after the larger than after the 
smaller sensitizing doses. In general, we can say that, while very 
large doses of antigen may perhaps increase the length of the incuba- 
tion time, perhaps because of the prolonged presence of antigen in 
the circulation and the partial neutralization of the first formed anti" 
bodies by this residue, no such prolongation results from moderate 
doses. Below a certain limit, it is probable that the smaller the dose, 
the longer the incubation time, because of the slower development 
of antibody formation in consequence. Repeated injection of guinea 
pigs with the same antigen during the period of incubation, as one 
wrould expect, prolongs the incubation time because of the partial 
desensitization which follows each new injection. This will become 
clear on further discussion below. Thus, in general, we may say that 
guinea pigs actively sensitized with 0.01 c. c. and more of a serum 
antigen, may begin to grow definitely hypersensitive within 5 to 7 
days, not reaching their most highly sensitive condition, however, 
until the 8th to the 14th day, after the injection. 
As stated above, what has been said so far refers to guinea pigs. 
Strange as it seems, other animals cannot be sensitized in exactly 
the same way. 
Rabbits are best sensitized by repeated injections administered 
at intervals as in immunization. A single injection may occasionally 
produce the anaphylactic state in rabbits if a sufficient amount of 
antigen is injected in the first place, and the test injection made in 
between two and three weeks. However, such an experiment is not 
often successful. The best procedure in rabbits is to inject moderate 
quantities of antigen intraperitoneally or intravenously at intervals 
of 5, 6, or 7 days, as in the production of a high titer precipitating 
serum. It is not then very easy to know exactly when the best time 
for reinjection has come. Usually, when in their most sensitive 
condition, rabbits have at the same time a considerable amount of 
precipitating antibody in the serum and this is one of the reasons, 
as we shall see, for the earlier ideas that in rabbits the reaction took 
place entirely in the circulation. As a matter of fact, as many ob- 
servers have shown, the rabbit is not merely so sensitive to the ana- 
phylactic experiment as the guinea pig. 
In dogs, acute anaphylaxis cannot be produced by a single injec- 
tion, though protracted symptoms eventually ending in death, may 
be obtained after a single sensitizing injection. Richard Weil 48 did 
48 Weil, Richard. Jgtir. Immunol., 2, 1017, p. 525t 422 
INFECTION AND RESISTANCE 
finally succeed in producing acute shock in dogs by giving a first 
subcutaneous injection of 5 c. c. of the antigen, and following this 
after 3 days with a similar amount intravenously administered. 
Something over 2 weeks later, 20 c. c. of the serum brought on acute 
shock with death in one-half hour. 
Repeated studies on monkeys seem to show that the lower monkeys 
are certainly not easily sensitized under any conditions. Our own 
studies on this subject with Macaccus Rhesus and Ringtail monkeys 
seem to show that a single injection of horse serum may produce no 
anaphylactic reactions whatever, even though the tests were controlled 
by the delicate Dale uterine method. Mild anaphylactic symptoms 
may be elicited after repeated and prolonged treatment with the 
antigen.49 
Reinjection for the Production of Anaphylactic Shock.—The 
antigen may be administered for the purpose of eliciting anaphy- 
lactic shock in an animal in a number of ways. The best method 
of obtaining severe symptoms is the intravenous in which the maxi- 
mum concentration of the injected antigen comes into contact with 
the sensitized tissues suddenly. In animals that are not particularly 
sensitive to anaphylactic reactions, rabbits, dogs, etc., the intravenous 
method is the only one with which acute death can be obtained with 
any degree of regularity. Intracardial injections, as first practiced 
by Rosenau and Anderson, Lewis 50 and others in anaphylactic ex- 
periments, amount to about the same thing as intravenous injection, 
and has the disadvantage of a more difficult technique and some un- 
certainty in regard to getting the total quantity injected into the 
cavities of the heart. When the injections are made subcutaneously, 
considerably larger amounts must be given in order to obtain the same 
results. The subcutaneous injections into sensitized rabbits pro- 
duce marked local reactions consisting of oedema and sometimes 
eventual necrosis, the state of affairs spoken of above as the Arthur 
phenomenon. This does not occur in the same way in guinea pigs 
and other animals, though the explanation for the peculiarity of the 
rabbit in this respect is not clear. It is suggested bv Coca that per- 
haps the local reaction on subcutaneous injection in rabbits indicates 
a general capillary reaction in these animals similar to the vascular 
contraction demonstrated by him in the pulmonary system. Local 
reactions on subcutaneous injection may appear in guinea pigs, man, 
and rabbits, in the form of varieties of skin reactions which will be 
discussed in a subsequent section. The generalized reaction, how- 
ever, is identical with that following intravenous injection, only less 
severe, unless large doses are given. Shock may be obtained in guinea 
pigs also by intraperitoneal injection. Subdural and intracerebral 
injections were practiced extensively for a time in connection with 
49 Zinsser. Proc. Soc. Exper. Biol. & Med., 18, 1920, p. 57. 
50 Lewis. Jour. Exper. Med., 10, 1908. ANAPHYLAXIS 
423 
Besredka’s 51 earlier theory of anaphylaxis. Besredka and Stein- 
hardt 52 who began their studies soon after the first publication of 
Rosenau and Anderson,53 came to the conclusion that the most rapid 
and conclusive method of producing anaphylactic shock in animals 
consisted in such brain injections. Whether or not we have reliable 
evidence of anaphylactic symptoms supervening upon contact of the 
antigen with the mucous membrane of sensitized animals, is an im- 
portant question which will be discussed in a subsequent section in 
connection with asthma and hay fever. 
It is important to note that the quantities of antigen necessary 
for the production of anaphylactic symptoms or shock can never 
be reduced to the minute amounts with which sensitization has been 
possible. This is of considerable importance in judging some of the 
earlier theories, such as that of Besredka, and that of Gay and 
Southard, whose ideas involved the supposition that the substance 
in the antigen which sensitized was not identical with that which 
elicited shock on subsequent injection. The point will be taken up 
later, but it is worth noting in this place that while it is possible to 
sensitize with doses as small as one-thousandth of a cubic centimeter 
and less, considerably larger quantities are needed for the develop- 
ment of shock. 
Duration of Hypersensitive State.—The experiments of Rosenau 
and Anderson cited above have shown in a number of cases that mild 
but definite anaphylactic reactions may be elicited in guinea pigs 
sensitized with horse serum as long as 245 days after the injection, 
there being no particular difference in this respect between animals 
sensitized with moderate doses or those sensitized with large doses. 
A few guinea pigs sensitized with toxin-antitoxin mixtures by them, 
gave positive reactions after as long as 732 days, and in one case, 
more recently reported, they obtained a reaction after 1096 days. 
It is not at all unlikely that individuals once sensitized may remain 
so for life, the sensitiveness gradually fading unless new contact with 
the antigen occurs. Temporary desensitization, as we shall see, 
merely interrupts this condition for a limited time, the individual 
returning to eventual hypersusceptibility. 
Passive Sensitization 
Up to the present time we have confined ourselves to the descrip- 
tion of the basic anaphylactic experiment, which is spoken of as 
“active sensitization” in analogy to the expression “active immuniza- 
tion,” since, like the latter, it conveys the conception that the state 
51 Besredka. Ann. de VInst. Past., 1907, pp. 777, 950; 1908, p. 496; 1909, 
pp. 166, 801; Bull, de VInst. Past., 1908, No. 19, No. 20, No. 21; 1909, No. 
17; C. R. Soc. de Biol., 65, 1908, p. 478, and 67, 1909, p. 266. 
52 Besredka and Steinhardt. Ann. de VInst. Past., 1907, pp. 177, 384. 
53 Rosenau and Anderson. U. S. P. II. Bulletin, No. 50, 1909. 424 
INFECTION AND RESISTANCE 
of hypersusceptibility (like the immunity in active immunization) is 
here acquired by reason of physiological changes directly induced in 
the treated animal in reaction to the first injection of the foreign 
antigen. There is another method of inducing hypersusceptibility 
which, in continuance of the analogy to immunization, is spoken of 
as “passive anaphylaxis ” since it consists in transferring the hyper- 
susceptible condition to a perfectly normal animal by injecting into 
it serum from an actively sensitized one. The normal animal is thus 
merely the passive recipient of the reaction bodies produced in the 
sensitive animal by preliminary treatment. 
That such a passive transference of anaphylaxis is possible was 
shown by a number of investigators almost simultaneously and M. 
Hicolle,54 in February, 1907, published a study on the phenomenon 
of Arthus in which he showed that, if the serum of a hypersusceptible 
rabbit (sensitized with horse serum) was injected into a normal 
rabbit, the recipient was rendered sensitive, so that the subcutaneous 
injection of horse serum, made 24 hours later, produced typical 
infiltrations. Richet55 soon after this succeeded in transferring 
hypersusceptibility toward mytilocongestin (a mussel poison) from 
a sensitized to a normal dog by injecting considerable amounts of the 
blood from the former into the latter. In this case, too, the hyper- 
susceptibility of the second dog did not appear until one or two days 
after the injection of the blood. At almost the same time Otto 56 
and Friedemann 57 independently succeeded in transferring serum 
anaphylaxis from hypersusceptible to normal guinea pigs in a similar 
way. Experiments of Gay and Southard,58 published during the 
same year, may possibly be also interpreted as instances of passive 
anaphylaxis, although their experimental procedure renders this 
doubtful, even in their own opinions. They injected 0.1 c. c. of 
serum from both sensitive and refractory guinea pigs into normal 
animals and followed this, after 10 days, with injections of antigen. 
The fact that such animals reacted may be interpreted in a number 
of ways. They themselves regarded the hypersusceptibility which 
the injected animals developed as a “purely active one,” and it is 
more than likely that this was the case, the recipient animals being 
actively sensitized by traces of antigen remaining unassimilated in 
the blood of the actively sensitized donors. In the following year 
(1908) the facts of passive sensitization were rapidly confirmed and 
extended by Besredka,59 Lewis,60 and others,61 and information of 
54 M. Nicolle. Ann. de VInst. Past., Yol. 21, 1907. 
55 Richet. Ann. de VInst. Past., Yol. 21, 1907. 
56 Otto. Munch, med. Woch., No. 34, 1907. 
57 Friedemann. Munch, med. Woch., No. 49, 1907. 
58 Gay and Southard. Jour. Med. Res., Yol. 16, 1907. 
59 Besredka. Ann. de VInst. Past., Yol. 22, 1908. 
60 Lewis. Jour. Exp. Med., Vol. 10, 1908. 
61 Kraus and Doerr. Wien. klin. Woch., No. 28, 1908. ANAPHYLAXIS 
425 
the greatest value for the comprehension of the anaphylactic reaction 
was obtained. 
Otto showed that passive sensitization could be carried out with 
the serum of an actively sensitized animal 8 days after the antigen 
injection, at a period when this animal itself had not yet become 
hypersusceptible. He also showed that the passive transfer of ana- 
phylaxis need not be confined to animals of the same species, but that 
guinea pigs could be rendered passively anaphylactic with the blood 
serum of sensitized rabbits. From the work of Gray and Southard,62 
moreover, it appeared that not only by the blood of sensitive animals 
can anaphylaxis be transferred, but that this can also be done by 
injecting the blood of animals that have once been sensitive but have 
subsequently been rendered antianaphylactic or refractory. Analo- 
gous to this observation is the fact observed by these authors as well 
as by Friedemann that the young of antianaphylactic mothers are not 
refractory but hypersusceptible. This observation is unquestionably 
correct, and has been confirmed by several other works. It has had 
no inconsiderable bearing upon our theoretical understanding of 
anaphylaxis. 
It was soon found out, too, that hypersusceptibility was conveyed 
not only by the sera of sensitive and of refractory animals, but that 
it could likewise be transferred by the precipitating sera of animals 
systematically immunized with a foreign proteid. 
This method was later employed by Doerr and Russ 63 in their 
quantitative studies on the relations between anaphylactic antigen 
and antibody. We are confronted, then, with the facts that animals 
may be passively sensitized: 
(a) by the serum of a sensitized animal. 
(b) by the serum of an animal not yet sensitive—in the pre- 
anaphylactic period (8th day, Otto). 
(c) by the serum of an antianaphylactic animal. 
(d) by the precipitating serum of an “immunized” 64 animal. 
Lewis further showed that normal guinea pigs could be rendered 
hypersusceptible with the blood of congenitally sensitive animals. 
In order to transfer hypersusceptibility from one animal to an- 
other in this way, it is not necessary that both animals be of the same 
species. As a matter of fact, convenience in laboratory experiments 
has led to the almost general practice of using the serum of im- 
munized rabbits for the transference of passive anaphylaxis to guinea 
62 Gay and Southard. Jour. Med. Res., Yol. 18, 1908. 
63 Doerr and Russ. Zeitschr. f. Immunitdtsf orschung, Yol. 3, 1909. 
64 We must never forget that the term “immunized” as applied to animals 
treated with harmless protein is an analogy and not absolutely correct. Such 
animals, though probably capable of assimilating larger quantities of foreign 
injected protein than normal ones, and this more rapidly, may nevertheless 
be not a whit more tolerant of the antigen—sometimes even extremely sensi- 
tive and vulnerable. 426 
INFECTION AND RESISTANCE 
pigs. A great many different combinations, however, have been em- 
ployed and the antibody-containing serum of rabbits, dogs, man and 
horses have been successfully employed for the passive sensitization 
of guinea pigs. It is likely that any serum which contains antibodies 
and is not normally toxic for the injected species, may passively sensi- 
tize. Doerr notes definite exceptions which he summarizes from the 
literature, stating that transference was unsuccessful in attempts to 
use the serum of birds on mammalia, and vice versa, that of mam- 
malia on birds. He also records negative results in similar attempts 
to sensitize white mice with the serum of rabbits and guinea pigs. 
In the latter case, the difficulty of producing anaphylactic sensitive- 
ness in mice under any circumstances may account for this. In those 
instances in which species as far separated as mammals and birds 
have been used, it is cpiite possible that no relationship between any 
of the constituents of the foreign serum and the cells of the host is 
established. A few years ago we attempted to do some single cell 
anaphylactic experiments in which we tried to absorb anti-horse 
serum upon living paramcecia and amoeba3 with complete failure. 
The constitution of the proteins in species far removed from each 
other* seems to render difficult, or perhaps entirely preclude com- 
binations of cellular substances and the heterologous antibodies. 
Passive sensitization is carried by the blood serum purely, since, 
in ordinary cases, as Ilosenau and Anderson have shown, the blood 
corpuscles and tissues of g sensitive animal do not convey the hyper- 
susceptibility. An exception to this will be noted later when we 
come to discuss Bail’s experiments on the passive transfer of tuber- 
culin sensitiveness. 
Passive sensitization, once established, may persist for as long 
as 3 or 4 weeks, though Iiosenau and Anderson found that animals 
tested 26 days after treatment reacted but weakly. In the young 
of anaphylactic mothers Otto has observed positive reactions as long 
as 44 days after birth, though fatal results were obtained in pigs only 
a few days old. In general passive sensitization lasts longer if con- 
ferred with homologous rather than heterologous anti-serum. 
To summarize the matter briefly, we may state that passive sen- 
sitization may be accomplished with any serum that contains anti- 
bodies, and that the power of such a serum to convey passive sensiti- 
zation is in direct proportion to its antibody concentration; in short 
the process of sensitization consists in the introduction of antibodies. 
This fact, which is, of course, of the greatest theoretical importance, 
will be further discussed in the succeeding chapter. 
Throughout the earlier investigations upon passive sensitization 
the curious fact recurs in the experiments of successive workers that 
a definite period must elapse between the injection of the sensitive 
blood and that of the antigen. 
Both Friedemann and Otto found that when the serum of a sensi- ANAPHYLAXIS 
427 
tized animal was injected subcutaneously the best results were ob- 
tained by administration of the antigen 24 to 48 hours after this. On 
intraperitoneal injection of the sensitizing serum Doerr and Russ 65 
obtained the best results by permitting an interval of 24 hours to 
elapse, and the same investigators still further shortened this period 
to 4 hours by injecting the sensitive serum intravenously. Beyond 
this, the interval could not be shortened with success. Indeed, some 
writers, notably Gay and Southard, have claimed that the maximum 
hypersusceptibility in guinea pigs treated with sensitive serum is 
reached only after 10 or more days, and Rosenau and Anderson, 
Lewis, and others have obtained results which seemed to point in 
the same direction. However, as we have already indicated, the 
testing of animals so long after the injection of sensitive serum 
leaves us in doubt whether we are dealing with true “passive” trans- 
ference of anaphylaxis or with active sensitization due to traces of 
antigen carried over with the serum of the sensitive animal.66 
From these observations the natural deduction was made that the 
anaphylactic symptoms were the result of cellular occurrences, and 
that the antigen could act only after the sensitizing substance (how- 
ever conceived) had become attached to certain cells, probably to 
those of the central nervous system. It was thought that a meeting 
of antigen and the sensitized body in the circulation would result in 
no reaction; that, in other words, the effective reaction was not a 
direct, but an indirect, one after the anaphylactic “antibody” of the 
sensitive serum had become bound to the cells. It will be neces- 
sary to recur to this problem when we discuss the various theories 
of anaphylaxis, where we will see that this point has been one of the 
crucial ones in the controversy between the two main directions of 
thought on anaphylaxis. 
To summarize what we have so far said, then, about passive ana- 
phylaxis, the facts elucidated by all these investigations were, in the 
first place, that the state of hypersusceptibility cun be specifically 
transferred to a normal animal with serum which contained anti- 
bodies, whether it was derived from an animal in the condition of 
hypersensitiveness, or from one in which no actual hypersensitiveness 
could be elicited, but which had been treated actively with the serum 
protein in question. Such transfer could take place not only within 
65 Doerr and Russ. Zeitschr. f. 1mmunitatsforschung, Yol. 3, p. 181, 
1909. 
66 An exception to this, contradicting the then prevailing opinion, were 
the researches of Weill-Halle and Lamaire (C. R. de la Soc. de Biol., Yol. 
65, July, 1908, p. 141), who showed that, under certain conditions, guinea 
pigs would react with typical, often fatal, anaphylaxis if injected simul- 
taneously with the serum of sensitized rabbits and the antigen horse serum. 
According to them, the success of such experiments depended entirely upon 
the condition of the sensitive serum—that is, the time at which the rabbits 
treated with horse serum were bled. These researches, with others of a simi- 
lar bearing, will be discussed in a later section. 428 
INFECTION AND RESISTANCE 
the same species, hut by the injection of the blood serum of an animal 
of one species to that of another. The other fundamental fact is that, 
with very few exceptions, it was noticed that between the time of the 
injection of the antiserum and the development of hypersuscepti- 
bility, a definite interval must elapse. The first fact indicated defi- 
nitely an antigen-antibody union in the body, as the true mechanism 
of the reaction, the second suggested a cellular seat for the anaphy- 
lactic occurrence. This last conception was early suggested by Doerr 
and Russ, and others, and the interval was supposed to represent the 
time necessary for the union of what we shall for the present call the 
anaphylactic antibody, with the cells. This point of view of the 
necessity for the union of antibody with cells as a necessary criterion 
of anaphylaxis as we ordinarily see it, has since been definitely con- 
firmed, hut soon after the time when this mechanism was first sug- 
gested, the issues were somewhat clouded by apparently contradictory 
experiments of Friedmann and a number of others which led to the 
formation for a time of two schools of investigators, one adhering 
to what was spoken of as the “cellular” theory of anaphylaxis, the 
other maintaining that the entire process took place in the circulation 
and of which, therefore, we may speak of as the “humoral” school. 
The researches upon which a final solution of this problem was based 
and a summary of what we believe the proper point of view should 
be at the present time, will follow in another section. 
We have briefly noted above that the offspring of hypersensitive 
guinea pigs may be found anaphylactic. Rosenau and Anderson 67 
noted this in their early experiments, and the fact has been variously 
confirmed since then. According to Lewis 68 such inherited hyper- 
susceptibility disappears within the first few months, and may vary 
in intensity in offspring of the same litter. It appears from every- 
thing that we can find in the experimental literature that such in- 
herited hypersusceptibility is based merely upon a passive transfer of 
antibodies from mother to child. Experimental evidence, also, seems 
to point to transfer by way of the placenta and not through the 
colostrum and milk, a fact which hears out recent investigations by 
Kuttner and Ratner in our own laboratory, who showed that this is 
entirely the case in connection with the transfer of diphtheria anti- 
toxin from mother to child in human beings, although, as we have 
stated in discussing their work, it is rash to draw conclusions of this 
kind from one animal to another, owing to the diverse histological 
anatomy of the placental structures in different species of animals. 
The Antibody Involved.—Doerr and Russ were the first to de- 
velop quantitative methods of studying the so-called anaphylactic 
antibody. They obtained the important result that there was an 
apparently complete parallelism between the ability of an immune 
67 Rosenau and Anderson. Loc. cit. 
68 Lewis. Jour. Exper. Med., 10, 1908. ANAPHYLAXIS 
429 
serum to sensitize passively, and its content in precipitins. Their 
method consisted in producing precipitating sera in rabbits in the 
usual way. With these they then passively sensitized guinea pigs, 
subsequently testing them with antigen 24 hours later. To arrive 
at quantitative results they developed two reliable methods. These 
consisted in: 1. Intraperitoneal sensitization of guinea pigs with 
constant quantities of titrated precipitating serum. Twenty-four 
hours later intravenous test with diminishing amounts of specific 
antigen. 2. Intraperitoneal sensitization with diminishing quanti- 
ties of the titrated precipitating serum, and 24 hours later intrave- 
nous tests with constant amounts of antigen. 
In this way they showed that there was a direct relationship be- 
tween the power of a serum to convey anaphylaxis passively and its 
contents of precipitins. We may elucidate this by an example from 
their work. They possessed a rabbit serum which gave precipitation 
with sheep, goat, beef, pig, human, and horse sera, but not with 
chicken serum. The precipitation titre of this serum for the sera 
mentioned varied from 1 in 20,000 in the case of sheep and goat 
sera, to 1 in 100 in the cases of the human and horse sera. When 
guinea pigs were injected intraperitoneally with 1 c. c. of this serum, 
and after 24 hours were intravenously injected with the various anti- 
gens mentioned above, in decreasing quantities, the sera which were 
precipitated in the highest dilutions gave anaphylactic shock in the 
smallest quantities. Those sera for which no precipitin or little had 
been present gave little or no reaction by this method even where con- 
siderable quantities were used. Thus, in animals prepared with 1 
c. c. of the antiserum (which precipitated sheep serum in dilutions 
as high as 1 to 20,000), death was caused by reinjection of sheep 
serum in amounts as small as 0.006 c. c.; whereas, in similar animals, 
horse serum which was precipitated in concentrations not higher than 
1 to 100 by the passively sensitizing serum, elicited slight symptoms 
only even when injected in amounts as high as 2 c. c.; and when such 
animals were reinjected with chicken serum which was not precipi- 
table at all by the antiserum, no reactions whatever could be elicited. 
In this, then, we have a definite quantitative analysis which shows 
a direct relationship between the concentration of antibodies in a 
serum and its ability to sensitize passively. Doerr and Russ at this 
time were inclined, therefore, to assume that precipitin and anaphy- 
lactic antibody were identical,69 and this was in keeping with Fried- 
berger’s purely theoretical idea that precipitins and the anaphylactic 
antibody might be identical.70 It is true that other investigators, 
notably Hintze 71 found also a parallelism between complement fix- 
ing properties and ability to sensitize passively, but in this connection 
69 Doerr and Russ. Zeitschr. f. Immunitats., 3, 1909, pp. 181 and 706. 
70 Friedberger. Zeitschr. f. 1mm., Yol. 2, 1909, p. 208. 
71 Hintze. Quoted from Doerr and Russ, loc. cit. 430 
INFECTION AND RESISTANCE 
we may refer the reader to a preceding section in this book which 
deals with the essential identity of antibodies. In that section it may 
be seen that we ourselves very strongly favor the view that there are 
two main classes of antigens in nature, one represented by the true 
toxins and exotoxins, or other active substances, like ricin, aborin, 
snake venoms, spider poisons, enzymes, etc., which induce a neu- 
tralizing antitoxin-like antibody in animals; that inactive antigens, 
like the proteins, however, give rise to the formation of, in each case, 
a single sensitizing antibody which either agglutinates, precipitates, 
fixes complement, opsonizes or sensitizes in the anaphylactic sense, 
according to the physical and chemical conditions and other environ- 
mental factors under which the reaction takes place. We believe, 
therefore, that the antibody which sensitizes and which, by the way, 
has a heat stability just like the other protein antibodies, is identical 
with the general protein sensitizing antibody. If this is not entirely 
clear, it can easily be understood, at least as far as our opinion is con- 
cerned, by referring back to the section alluded to. Most observers 
have confined themselves to an attempt to either disprove or prove 
the identity of anaphylactic sensitizing substances and precipitins. 
Richard Weil 72 found, incidentally, that a specific precipitate ob- 
tained by the precipitation of serum with an antiserum could be 
washed free from serum, and then produced passive sensitization 
when injected into a guinea pig. His experiments, which exclude 
the possibility of antiserum having been transferred with the pre- 
cipitate, could easily be explained by a dissociation of the precipitat- 
ing antibody from the precipitate after injection. Since such pre- 
cipitates were found by him both to transfer active as well as passive 
anaphylaxis, the supposition is that the antigen and antibody united 
in the precipitate may have dissociated and in this way produced both 
effects. The further experiments of Weil in which he removed the 
precipitating function of a serum by heating at 70° (or by extrac- 
tion of the antibody from the precipitate with sodium carbonate), 
without removing its sensitizing functions, are not in our opinion 
contradictory to the general identification of these antibodies, since 
it is well known that the actual floculating properties of an antiserum 
may be physically interfered with by heating, and that an alkalinity 
above a Ph of 8.5 to 9 definitely prevents agglutination and pre- 
cipitation. 
Symptoms of Anaphylactic Shock.—If a properly sensitized 
guinea pig receives a second injection of an antigen after a suitable 
incubation time a very characteristic train of symptoms ensues. 
There is usually a short preliminary period—lasting either a fraction 
of a minute or several minutes according to the violence of the reac- 
tion and the mode of administration—during which the pig appears 
normal. At the end of this time the animal will grow restless and 
72 Weil, Richard. Jour. Immunol., 1, 1916, p. 19. ANAPHYLAXIS 
431 
uneasy, and will usually rub its nose with its forepaws. It may 
sneeze and occasionally emit short coughing sounds. At the same 
time an increased rapidity of respiration is noticeable and the fur 
will appear ruffled. In light cases the animals may remain in this 
condition, with further irregularity and difficulty of respiration, 
possible discharges of urine and feces; then gradual slow recovery 
may set in, with complete return to normal in from 30 minutes to 
several hours. In more severe cases these preliminary stages are 
rapidly followed by great apparent weakness. The animals fall to 
the side, the legs and trunk muscles twitch irregularly, and the respi- 
ration becomes slow and shallow; the thorax never entirely contracts, 
but remains in a more or less expanded condition. The very evident 
dyspnea is of an inspiratory character. The excursions of the lung 
itself seem to grow shallower and shallower in spite of apparent 
strong inspiratory efforts—the volume of the thorax and lung re- 
maining in the expanded condition. At this stage evidences of motor 
irritation may appear, in that the animal may arise and attempt to 
run. More often, however, in this phase general convulsions set in, 
often several times repeated, and in these the animals usually die. 
On the other hand, after cessation of convulsions they may lie 
perfectly still on the side as though paralyzed, the breathing becom- 
ing gradually slower and more shallow, finally ceasing entirely. The 
heart may continue to beat for a considerable time after the breath- 
ing has stopped. 
If such an animal is immediately autopsied a very characteristic 
condition is found—to which, in the essentials, attention was first 
called by Gay and Southard.’73 They speak of finding “pulmonary 
emphysema as a constant feature at autopsy,” and attribute the 
anaphylactic death in guinea pigs to cessation of respiration in the 
inspiratory phase under the influence of respiratory central intoxi- 
cation. 
The lungs of such guinea pigs after death are found distended 
and completely filling the thorax. They are usually pale and blood- 
less and do not collapse as the pleurae are opened. On microscopic 
examination the alveoli are seen to be distended and small hemor- 
rhages may appear upon the serous surfaces. According to Gay 
and Southard, furthermore, histological study of the other organs 
shows also hemorrhages in the brain, stomach, heart, cecum, and 
spleen—more rarely in other organs—and there are local fatty 
changes in the capillary endothelium which they regard as causa- 
tively related to the hemorrhages. 
That the respiratory symptoms are the most striking feature of 
the clinical picture of guinea pig anaphylaxis had, as a matter of 
fact, been noticed by Rosenau and Anderson. A detailed physiologi- 
73 Gay and Southard. Jour. Med. Res., Yol. 16, 1907; Vols. 18 and 19, 
1908. 432 
INFECTION AND RESISTANCE 
cal study of the mechanism of the respiratory death in these cases 
was first made, however, by Auer and Lewis74 in 1909. 
These investigators showed that, during the later respiratory 
symptoms, little or no air enters the lungs, although the animal 
makes violent respiratory efforts. This is due, as they found, to a 
tetanic contraction of the small bronchioles, which practically oc- 
cludes the air passages. That the origin of this contraction is not, as 
previously supposed, of central origin, but is referable to peripheral 
cause, they proved by showing that the same phenomena occur in 
the guinea pigs even after the cord and medulla have been destroyed 
and the vagi divided. In such cases, of course, with the cord and 
medulla destroyed, artificial respiration had to be done, and when 
the symptoms set in it was found that the lungs could no longer be 
expanded by the same force of artificial respiration which before this 
had been sufficient. 
They showed also that the non-collapsible expansion of the lungs 
after death was due to imprisonment of the air in the alveoli by the 
contracted musculature of the small bronchioles, and further con- 
firmed their opinion of the peripheral origin of this contraction by 
the important discovery that atropin will markedly protect, often 
preventing death or hastening recovery. It is noteworthy, too, that 
Auer and Lewis speak of occasionally finding slight pulmonary 
edema, a feature which Biedl and Kraus consider incompatible with 
true anaphylaxis. 
Anderson and Schultz,75 who have confirmed much of the work 
of Auer and Lewis, find that not only atropin will prevent asphyx- 
iation in these cases, but methane, chloral hydrate, adrenalin, and 
pure oxygen will exert a similar effect. The animals may he saved 
from suffocation in this way, but may nevertheless die, probably as 
the result of lowered blood pressure. 
The observations of Auer and Lewis have been further confirmed 
especially by Biedl and Kraus,76 who regard it as well established 
that anaphylactic death in guinea pigs is caused primarily by suffo- 
cation, due to tetanic spasms of the musculature of the small bronchi. 
These spasms are not of central origin, but are peripherally initiated, 
possibly by direct action upon the smooth muscle itself. The fact 
that atropin is not effective in preventing death in all severe cases is 
no argument against this, since such an effect would naturally de- 
pend upon the relation between the amount of atropin given and the 
severity of the attack. In this connection the studies that have been 
made upon the irritability of smooth muscle fibers in normal and in 
74 Auer and Lewis. Jour. A. M. A., Vol. 53, p. 458, 1909; Jour. Exp. 
Med., Yol. 12, 1910. 
75 Anderson and Schultz. Proc. Soc. Exp. Biol, and Med., 7, 1909, p. 32. 
76 Biedl und Kraus. Zeitschr. f. Immunitatsforschung, Yol. 7, 1910 j 
Centralbl. f. Physiol., 1910; Wien. klin. Woch., No. 11, 1910, ANAPHYLAXIS 
433 
sensitized animals are of great interest, Schultz,77 following out an 
observation made by Rosenau and Anderson, studied the intestinal 
muscle of normal sensitized guinea pigs excised and suspended in 
Howell’s solution. In this way be showed that during the period of 
hypersusceptibility the smooth muscle is abnormally sensitive to 
treatment with the antigen. The contraction which normally occurs 
in smooth muscles under the influence of serum is markedly aug- 
mented if the preparations are taken from sensitized animals. 
In addition to these predominant features of the anaphylactic 
symptomatology in guinea pigs, there are a number of secondary re- 
actions which, though less prominent, are nevertheless of considerable 
interest and theoretical importance. The conditions in the circula- 
tion are probably, to a great extent, dependent upon the respiratory 
condition, and the fall of blood pressure in guinea pigs is regarded by 
some investigators as merely a secondary manifestation just preceding 
death. The fall of temperature first described by H. Pfeiffer,78 
however, seems to be an occurrence which, though standing in no 
causative relation to the symptoms as a whole, is so constant and well 
marked that it has been taken by a number of workers as one of the 
necessary criteria for the characterization of the anaphylactic con- 
dition. 
There is, indeed, an almost regular drop of several degrees in the 
rectal temperature, and a close observation of this may be of much 
aid in determining the occurrence of mild reactions, when other 
symptoms of shock are not strongly marked. Pfeiffer79 himself 
goes so far as to claim that by this symptom alone delicate anaphy- 
lactic reactions may he determined when all other symptoms are 
lacking. 
Friedberger,80 too, has found the sudden drop of temperature 
a very regular occurrence, and has employed this method of study 
for the analysis of the intensity of anaphylactic shock. He calls 
attention to the apparent difference between infection and anaphy- 
laxis in this respect in that in the former there is fever, in the latter 
there is depression of body heat; hut, at the same time, he points out 
that this discrepancy is an apparent one only, and determined bv 
quantitative differences, for when he treated sensitized animals with 
varying doses of antigen he found that quantities which produced 
other anaphylactic symptoms of noticeable degree would regularly 
depress the temperature as Pfeiffer had shown. It was possible, 
however, to determine a minimal dose necessary for temperature 
reduction. Quantities just below this left the temperature un- 
77 Schultz. Jour. Pharm. and Exp. Therap., 1, 1910; 2, 1910. 
78 H. Pfeiffer. Wien. Jclin. Woch., No. 1, 1909. 
79 Pfeiffer u. Mita. Zeitschr. f. Immunitatsforschung, Yol. 4, 1910. 
80 Friedberger. Deutsche med. Woch., No. 11, 1911. Friedberger und 
JVlita. Zeitschr. f. Immunitatsf orschung, Vol. 10, 1911.. 434 
INFECTION AND RESISTANCE 
changed, and still smaller quantities produced fever or even in- 
creased the temperature. This fact is extremely significant in that, 
as we shall see, it has an important bearing upon views which inter- 
pret bacterial infection as a series of anaphylactic poisonings, the 
multiplying bacteria furnishing the constant supply of minute 
amounts of antigen. This thought, indeed, based also on the study of 
temperature curves in animals, was expressed by Vaughan 81 as early 
as 1909, and was developed by him with Cumming and Wright 82 in 
an extensive study upon what he called “protein fever.” It was 
shown in these experiments that continued fever, not unlike that of 
infectious diseases, could be produced in rabbits by repeated subcu- 
taneous injections of primarily harmless substances, such as egg 
white and vegetable proteins. The conditions observed and the con- 
clusions drawn from them in this work, as well as in the similar in- 
vestigations of other workers, were clearly foreseen by Vaughan in 
his early investigations on proteid split-products studies, which we 
will find occasion to discuss in a later section. 
The rigidity of the diagnostic value of the temperature relations 
for anaphylactic shock in particular, as advanced by Pfeiffer, was 
somewhat weakened by Ranzi’s 83 observations that foreign serum 
may produce temperature depression when injected into perfectly 
normal animals and that, injected into sensitized animals, the same 
reaction may follow if other proteins than the original antigen were 
administered. 
Although these objections of Ranzi are perfectly just, yet there 
is such a marked quantitative difference between the reaction in nor- 
mal and in sensitized animals that, in principle, Pfeiffer’s claim is 
not invalidated. Friedberger 84 very logically remarks that, after 
all, the phenomena of sensitization as well as those of immunity are 
merely an exaggeration of normal physiological conditions, and in 
experiment he has shown that, whereas noticeable depressions of tem- 
perature will follow in the normal animal only upon quantities of 
antigen exceeding 0.5 c. c., the temperature of the sensitized animal 
may be depressed by amounts as small as 0.0005 c. c. 
Apart from the symptoms so far discussed, there are other less 
apparent characteristics of anaphylaxis in guinea pigs, all of which, 
however, possess considerable importance theoretically. The most sig- 
nificant of these is the reduction in the amount of alexin or comple- 
ment, first noticed by Sleeswijk,85 which occurs after the injection 
of the second or toxogenic dose—during the development of shock. 
81 Vaughan. Zeitschr. f. Immunitdtsforschung, Vol. 1, 1909. 
82 Vaughan, Cumming, and Wright. Zeitschr. f. Immunitdtsforschung, 
Vol 9, 1911. 
83 Ranzi. Zeitschr. f. Immunitdtsforschung, Vol. 2, 1909; Wien. klin. 
Woch., No. 40, 1909. 
84 Friedberger u. Mita. Loc. cit. 
85 Sleeswijk. Zeitschr. f. Immunitdtsforschung, Vol. 2, 1909. ANAPHYLAXIS 
435 
This phenomenon is so closely interwoven with the later theoretical 
aspects of anaphylaxis that we will defer its discussion until we have 
completed a more general survey of the field. 
In guinea pigs, as in dogs, Friedberger and others have also seen 
a lowered coagulability of the blood and a temporary diminution of 
the polynuclear leukocytes (leukopenia) during shock. 
Differences Between Physiological Reactions of Guinea Pigs 
and other Animals.—We have described the occurrences in guinea 
pigs at some length because protein anaphylaxis has been most thor- 
oughly studied in these animals. We have already seen, however, 
that in discussing active sensitization, there are marked differences 
in the anaphylactic phenomena as they occur in different species of 
animals. The physiological reactions noted in guinea pigs, there- 
fore, as a result of anaphylactic insult, cannot, indiscriminately, be 
applied to other animals. 
In sensitized rabbits the injection of a further dose of antigen 
is usually followed, after a short but definite incubation time, by 
great weakness with, often, discharge of urine and feces. The ani- 
mals sink down until the abdomen touches the ground, the legs are 
stretched out weakly but not paralyzed, and the head may drop 
forward or to one side. After this, the animal may gradually fall 
upon its side and lie motionless except for labored and irregular 
breathing and occasional twitching of the legs and head. Sometimes 
this gradual relaxation may be interrupted by a sudden motor irri- 
tation, the rabbit suddenly getting up and running a short distance 
but soon falling down again apparently from a sudden return of the 
muscular weakness. During these running spells it seems as though 
there was no sense of direction or purpose—the animals running into 
obstructions or off tables as the case may be. During this period gen- 
eral convulsions and a drawing back of the head by a tetanic spasm 
of the muscles of the neck are not uncommon. Death may occur 
within a few minutes, or it may follow a gradually increasing weak- 
ness in the course of several hours. The fall of blood pressure here 
seems to be purely secondary to the general failure of all the func- 
tions. 
Unlike guinea pigs, rabbits do not show the distention of the 
lungs found in guinea pigs. Characteristic, however, is the disten- 
tion of the right side of the heart and, as Auer 86 first noted, there is 
a definite pathological change in the muscles of the right ventricle 
which he believed was directly connected with anaphylactic death. 
This supposition has been recently confirmed by an important ob- 
servation by Coca 87 who perfused the vascular pulmonary system 
of normal rabbits and of rabbits dead of anaphylactic shock. He 
86 Auer. Centralbl. f. Physiol., 24, 1911, p. 957. 
87 Coca. Proc. Soc. Exper. Bio. & Med., 16, 1919, p. 47. Coca. “Ana- 
phylaxis,” Tice’s Modem Medicine. 436 
INFECTION AND RESISTANCE 
found that while a pressure of 10 cm. will easily perfuse the normal 
pulmonary vascular system of the rabbit, a similar experiment on 
rabbits that have just died of anaphylactic shock failed, even under 
a pressure of as much as 90 cm. When he cut the lung near the 
pleural surface in such rabbits, while this pressure was being applied, 
no fluid oozed through, proving that the obstruction is on the arterial 
side of the pulmonary circulation. He justly concludes that death in 
the anaphylactic rabbit is probably due to a tonic spasm of the muscu- 
lar coat of the pulmonary arterioles. Coca suggests that a similar 
spasm in the systemic arterioles may take place in the anaphylactic 
rabbits, and that such a localization of the anaphylactic reaction in 
the arterial walls may account for the local manifestation spoken of 
above as the Arthus phenomenon which is peculiar to rabbits so far, 
and has not yet been satisfactorily explained. 
Anaphylaxis in dogs has been very extensively studied, especially 
by Biedl and Kraus,88 and by Pearce and Eisenbrey.89 The symp- 
toms in dogs are characterized by a rapid progressive fall in the 
blood pressure, followed by the symptoms of cerebral anemia. Ana- 
phylactic dogs, after injection, will at first grow restless, vomit, and 
pass urine and feces. They then grow rapidly weak, fall to the 
ground, and continue to twitch and vomit and the respiration be- 
comes labored and irregular. There is general weakness of the 
muscles, but no paralysis. The marked, constant, and characteristic 
feature of the condition in these animals is the fall of blood pressure. 
There is also a lessened coagulability of the blood, much more 
strongly developed than in guinea pigs and rabbits. 
According to Biedl and Kraus this may amount to almost a pre- 
vention of the coagulation in anaphylactic dogs. 
As in other animals the blood picture is changed in that there is 
a falling off of the total number of leukocytes with a relative diminu- 
tion of polynuclear cells. 
Quantitative measurements by Calvary,90 moreover, have shown 
that anaphylaxis in dogs is accompanied by a marked increase of the 
lymph flow (1 times the amount observed in normal dogs in the same 
time) and, by controlling the blood pressure with barium chlorid, 
that this lymphagogue action is not directly dependent upon the low 
pressure. This observation is of especial interest in connection with 
the similarity of anaphylaxis to peptone poisoning in which Heiden- 
heim 91 noticed a similar increase of the lymph. 
Pearce and Eisenbrey found, at autopsy of dogs dead of anaphy- 
lactic shock, subserous petechial hemorrhages in the rectum and gall 
88 Biedl and Kraus. Loc. cit.; also in “Kraus u. Levaditi Handbuch,” 
Erganzungsband 1. 
89 Pearce and Eisenbrey. Proc. Soc. Exp. Biol, and Med., 7, 1909, p. 30; 
Trans. Cong. Am. Phys. and Surg., Yol. 8, 1910. 
90 Calvary. Munch, med. Woch., No. 13,1911. 
91 Heidenheim. Pfliiger’s Archiv, 49, 1891. ANAPHYLAXIS 
437 
bladder, hemorrhagic spots on the gastric and duodenal mucosa, and 
in the colon. According to these workers, in agreement with Biedl 
and Kraus, the fall of blood pressure is not due to centra] causes but 
depends upon influences exerted upon the peripheral vasomotor sys- 
tem. Biedl and Kraus believe that this action is exerted upon the 
muscle cells themselves rather than on the nerve endings. They 
admit the inconclusiveness of their experimental data, but take the 
above standpoint because of the fact that adrenalin, which acts by 
stimulation of the vasomotor nerve endings particularly, does not 
raise the low pressure in dogs during anaphylaxis while barium 
chlorid, which acts upon the smooth muscle fibers themselves, strongly 
raises the blood pressure in such animals. Pearce and Eisenbrey are 
inclined to believe that the action is chiefly upon the nerve endings, 
though both factors, nerve and muscle, may be involved. They 
worked with apocodein, a substance which, in large doses, paralyzes 
the vasomotor nerve terminals.92 
When a sensitized dog was treated with apocodein and the anti- 
gen then injected, no further drop of pressure was obtained. Appar- 
ently a paralysis of the vasomotor nerve endings had removed the 
point of attack upon which the anaphylactic poison could act. 
In addition to the symptoms already enumerated Weichhardt and 
Schittenhelm 93 claim that anaphylaxis in dogs is invariably accom- 
panied by a severe local reaction in the gut. The intestinal mucosa 
is swollen and contains miliary hemorrhages and the lumen is often 
filled with a mucus mixed with blood. In the further analysis of the 
anaphylactic reaction in dogs, Manwaring94 has recently reported 
observations of great interest. He investigated the participation in 
anaphylactic shock of the various organs and determined that shock 
did not occur when the abdominal vessels were ligated just above 
the diaphragm. In further localizing the source of shock he found 
that exclusion of the spleen, stomach, kidneys, suprarenals, and 
ovaries from the circulation had no effect upon the occurrence of 
anaphylactic shock. However, when he operated in such a way that 
the liver was thrown out of circulation, none of the seven dogs that 
he used reacted with anaphylactic shock to the injection of serum. 
He concludes from this that the liver is directly responsible in some 
way for the production of anaphylaxis. The intestines, too, were 
found, by a similar procedure, to take part, though to a less important 
extent than the liver. 
It would seem, in view of a number of confirmations of Man- 
waring’s work, that the great importance of the liver in anaphylactic 
dogs cannot be questioned. Richard Weil, who also investigated these 
conditions, believed that most of the circulatory disturbances in the 
92 Brodie and Dixon. Jour. Phys., 30, 1904. 
93 Weichhardt and Schittenhelm. Deutsche med. Woch., 19, 1911, 
94 Manwaring. Zeitschr. f. Immun., Yol. 18, 1911. 438 
INFECTION AND RESISTANCE 
anaphylactic dog could be attributed to obstruction in the portal 
circulation. 
Recognizing as fairly well established that in the three animals 
most thoroughly studied in anaphylaxis, namely, the guinea pig, 
rabbit and dog, physiological reactions and localization in organs 
vary distinctly, many attempts have been made to explain such dif- 
ferences. In guinea pigs, as we have seen, the cause of death resides 
very largely in the musculature of the bronchioles, in rabbits a similar 
cause is found in the circulation of the lung, and in dogs the hepatic 
and splenic circulations seem to be particularly involved. A most 
ingenious and plausible explanation is suggested in Wells’ review as 
the result of anatomical studies by a number of different workers. 
He states that histological study of guinea pigs has shown an astonish- 
ing development of the bronchial musculature, the finer bronchioles 
consisting almost entirely of muscular tubes. In rabbits “the pul- 
monary arteries present a remarkable degree of muscular development 
analogous to that of the guinea pig bronchioles.” As far as dogs are 
concerned, he quotes Simonds 95 as stating that the walls of the 
hepatic veins of the dogs are different from those of any other animal, 
again, in showing a very high development of musculature. And, 
Wells concludes, it is at least a reasonable hypothesis that the differ- 
ences in these species, in reaction to one and the same mechanism, de- 
pend upon fortuitous differences in the distribution of non-striated 
muscle. Interesting in connection with Wells’ suggestion, is the fact 
that Huber and Koessler 96 have shown that asthmatic individuals 
show a hypertrophy of the musculature of the bronchioles which 
transforms them histologically into a condition not unlike that nor- 
mally prevailing in the guinea pig, and it is well known that in such 
individuals severe attacks of an asthmatic nature can be elicited on 
serum injection. As a matter of fact, it is in individuals of this sort 
that one avoids or exercises unusual care in connection with protein 
injections. 
Other animals than those mentioned have been little used for 
anaphylactic experiment. Observations incidental to other work, 
however, have shown that horses and goats are particularly sensitive. 
In goats the writer has observed both serum and bacterial anaphy- 
laxis, and the symptoms here were those of general trembling, weak- 
ness, labored respiration, and involuntary evacuation of urine. 
The lower monkeys are exceedingly difficult to sensitize. Our 
own experiments on this subject have already been cited. When, 
after repeated injection, sensitization is accomplished, the symptoms 
are usually mild, and not unlike those prevailing in human beings, in 
fact, we have cited a single case in which oedema of the face and a 
picture somewhat similar to serum sickness in man resulted. 
95 Simonds. Jour. A. M. A., 73, 1919, p. 1437. 
96 Huber and Koessler. Arch. Int. Med., 1921. ANAPHYLAXIS 
439 
Mice can be sensitized but not as easily as guinea pigs. The best 
method consists in giving several relatively large preliminary doses 
a few days apart. Anaphylactic reactions in man will be discussed 
in a subsequent section. 
Rats are refractory to sensitization. Longcope’s work with iso- 
lated rat uteri was entirely negative. F. and J. T. Parker in our 
laboratory have obtained mild uterin reactions, but so far no uterin 
response. 
In the discussion of the anaphylactic reaction given above, we 
have dealt chiefly with the very severe symptoms of acute shock. It 
must not be forgotten, however, that these conditions are the mani- 
festations of a very acute and profound injury, and the localizations 
in various organs with which we have dealt are chiefly those which 
are responsible for death, perhaps to a large extent, or entirely, be- 
cause these particular sites of injury affect the tissues upon which 
the immediate continuation of life depended. It is not at all out of 
question, however, that many other parts of the body may suffer, 
although their injury might have no immediately noticeable effects, 
but expresses itself, eventually, especially if the insult is repeated, 
in degenerations and chronic disease. Chiray97 and Longcope98 
have attempted to approach this problem by repeated administrations 
of protein in moderate doses to rabbits and investigation of subse- 
quent changes in the kidney. The lesions found by Longcope were 
interpreted by him as perhaps signifying a subacute anaphylactic 
injury. This conclusion, however, has been questioned, since then, 
by a number of workers who believe that similar lesions are too fre- 
quently found in untreated animals to permit definite assumption 
that the repeated protein injections had caused them. Boughton 99 
has made observations similar to those of Longcope, and though the 
question must remain an open one for perhaps some time to come, it 
is, nevertheless, reasonable to suppose that, whether or not these 
particular experiments of Longcope and of Boughton can be accepted 
as proof, the likelihood is strong that repeated anaphylactic reactions 
of whatever severity, may cause injury to many organs in which they 
cannot be immediately observed. 
Desensitization or Antianaphylaxis.—In properly sensitized 
animals the result of a sufficient dose of antigen, given at the proper 
time, is very often death. When the time and quantity are so chosen 
that instead of death there is merely a more or less severe anaphy- 
lactic shock, the animals are immediately thereafter in a refractory 
condition. That is, they are no longer sensitive to further injections 
of the antigen. This observation was made by Otto and by Rosenau 
and Anderson in their pioneer investigations, was confirmed by Gay 
and Southard, and was subsequently very thoroughly studied by 
97 Chiray. These de Paris, 1906, cited from Longcope, loc. eit. 
98 Longcope. Jour. Exper. Med., 22, 1915, p. 793. 
99 Boughton. Jour. Immunol., 4, 1919, p. 213. 440 
INFECTION AND RESISTANCE 
Besredka and Steinhardt.100 The last-named workers named this 
refractory or immune condition “antianaphylaxis” There is ob- 
viously a great deal of both practical and theoretical significance in 
this fact, and methods were sought by which such an antianaphy- 
lactic state might be induced in sensitized animals without subjecting 
them to the dangers of actual shock. It was found that this could 
be accomplished in a number of ways. According to Besredka and 
Steinhardt the injection of moderate quantities of the antigen at a 
time just preceding the development of hypersusceptibility, in the 
preanaphylactic period, will render them refractory to later injec- 
tions. This preventive administration, however, must be given 
during the later days of the anaphylactic incubation time. If given 
too soon after the first injection it does not prevent eventual sensitiza- 
tion, though it may occasionally delay its development, acting then 
simply as though a larger dose had been given in the first place. Thus 
if antigen is given by a method of introduction and in a quantity 
which would justify us in expecting hypersusceptibility to be de- 
veloped at the end of 12 days, we can render the animal “antianaphy- 
lactic” by a second administration given, say, on the 8th, 9th, or 10th 
day. If we give it on the 2d, 3d, or 4th day after the first injection, 
it is very likely that sensitization will proceed nevertheless. Rosenau 
and Anderson have also investigated the repeated injection of antigen 
during the incubation time, and their results would also seem to 
emphasize the necessity of making the preventive injection close to 
the time at which hypersuscepfibility may be expected. If quanti- 
ties of 2 c. c. were injected 10 times in the course of 17 days, and 
15 to 17 days thereafter 6 c. c. of horse serum were given, the animals 
showed symptoms proving that antianaphylaxis was but partial. If 
amounts of 0.001 c. c. were given 5 times in a period of 8 days, and 
the animals were tested 23 days later, death often ensued. It is also 
possible, as a number of investigators have shown, to produce the 
antianaphylactic state by the injection of sublethal doses, even after 
the time has set in at which the animals are hypersusceptible. This 
method can be carried out successfully according to Besredka by in- 
jecting very small amounts into the brain (1/50 to 1/400 of a cubic 
centimeter). Within a few hours after such an injection the animals 
may withstand an otherwise fatal dose with slight or no symptoms; 
although it is generally stated that intraperitoneal injections, carried 
out after hypersusceptibility has set in, must be of considerable quan- 
tity (large enough to cause symptoms) in order to induce antianaphy- 
laxis. Besredka 101 states, in a recent resume, that 1/50 to 1/100 
cubic centimeter injected intraperitoneally and giving “practically 
no symptoms” in a sensitized guinea pig, after the anaphylactic state 
has set in, may render the animal entirely refractory after 5 hours. 
100 Besredka and Steinhardt. Loc. cit. 
101 Besredka in “Kraus u. Levaditi Handbuch,” Erganzungsband I. ANAPHYLAXIS 
441 
On the other hand, Rosenau and Anderson,102 working with sub- 
cutaneous injection, obtained results which differ considerably from 
those of Besredka. They sensitized a series of guinea pigs with 
mixtures of toxin and antitoxin, and 48 days later, at a time when 
the animals were hypersusceptible, gave 20 subcutaneous injections 
of 0.001 c. c. daily. Two days after the last injection, 0.2 c. c. of 
horse serum was given intracerebrally, and all of the animals showed; 
symptoms, and many of them died. They conclude, therefore, that 
the repeated injection of small amounts of antigen into sensitized ani- 
mals has no appreciable effect. The same worker has shown by ex- 
periment that the introduction of large amounts of antigen into the 
previously cleansed rectum of sensitive animals is entirely without 
danger and will produce an antianaphylaxis, which becomes evident 
after 12 hours. This is probably dependent upon the very slow pene- 
tration of small amounts of antigen into the circulation from the gut, 
and has, therefore, an effect similar to the repeated injection of small 
amounts directly, or the very slow and gradual method of intrave- 
nous injection advocated by Friedberger for the prevention of serum 
sickness in man. This phase of the subject is considered in greater 
detail in a subsequent discussion of serum sickness. 
Antianaphylaxis produced in this way is specific,103 although, as 
we shall see, there are other methods by which it is claimed that a 
nonspecific antianaphylaxis can be produced. 
Such specificity readily appears from the fact that if an animal 
is sensitized, either actively or passively, to two or more antigens, 
separate desensitization can be accomplished with each individual 
antigen. Yet, as we shall see, in such experiments there is a certain 
degree of non-specific protection as a consequence of the first de- 
sensitization, a matter to which we will recur presently. 
It is plain, therefore, that since it is pretty well established that 
antianaphylaxis means merely a desensitization by a partial or com- 
plete saturation with antigen of the antibodies present in the sen- 
sitized animal, quantitative relations must be extremely important 
in this reaction. For, if a dose of antigen is injected into a highly 
sensitive animal which is insufficient completely to saturate the anti- 
bodies, partial desensitization only could be expected. And, con- 
versely, probably the most thoroughly desensitized animal is one that 
has recovered from a dose of antigen, almost but not quite fatal. 
Moreover, since the suddenness of union of antigen with antibody 
seems to determine the severity of the reaction, any method in which 
a sufficient amount of antigen can be introduced in a manner so 
gradual that the union must needs be slow, may serve to desensitize 
102 Rosenau and Anderson. Loc. cit., U. S. Pub. Health and M. H. S. 
Hyg. Lab. Bull. 45, 1908. 
103 Pfeiffer has recorded an exception to this in that he claims to have 
rendered a horse-serum-sensitive animal refractory by an injection of swine 
serum. 442 
INFECTION AND RESISTANCE 
without severe symptoms. This, then, should be the principle in all 
methods of desensitization. Besredka, as we have seen, attempted 
to accomplish this by the repetition of minute doses, whereas, Fried- 
berger and Mita 104 accomplished the same purpose by administering 
the antigen very slowly, by a gravity feed apparatus and in a dilu- 
tion; we have already mentioned attempts to desensitize through the 
gut, a method based on the same principle. 
This mechanism of desensitization explains why the refractory 
condition should supervene almost immediately after the injection 
of the antigen, whether or not shock has taken place. Also, such an 
understanding will indicate that there is no fundamental difference 
in the desensitization of actively or passively sensitized animals. 
Although it is somewhat anticipating the theoretical considera- 
tions to be taken up in the next chapter, we may mention in this 
place that the now generally accepted conception that the anaphy- 
lactic reaction takes place largely between the antigen and antibodies 
which are in some way in union with tissue cells, has led to the very 
natural supposition that perhaps shock can be prevented by supplying 
a sufficient concentration of homologous antibodies to the circulation 
just before administering the shock dose of antigen. The reasoning 
here is identical with that on which we explained the protective value 
of antitoxin, in that the circulating antibody, by uniting with the 
antigen, prevents the latter from reaching the susceptible cells. And 
for our present purposes we may conceive the sensitized cell as readily 
amenable to injury by the protein antigen, as the normal cell would 
be to toxins of the diphtheria or tetanus bacillus. Weil, as far as 
we know, was the first to subject these conditions to careful analysis, 
and, indeed, found that he could protect actively or passively sensi- 
tized guinea pigs by introducing into the blood large amounts of 
homologous immune body previous to the shock injection. Although 
the theoretical conception was thus upheld, he found that unex- 
pectedly large amounts of antibody were needed for this purpose, a 
fact which he explains by “an increased avidity of the ‘anchored’ 
antibodies, as compared with the free antibodies, for antigen.” Our 
own opinion which differs only in the principle of mechanism from 
that of Weil, is based upon the observation noted in another chapt 
that antigen and antibody in the circulation are prevented from 
uniting with any degree of speed, probably by the colloidal protective 
action of other serum constituents; and that, in consequence, a con- 
siderable excess of antibody in the circulation is necessary in order to 
render an experiment of this kind successful. Nevertheless, from a 
practical point of view, an excess of antibodies in the blood might 
be a safeguard of considerable value, when the antigen is at the same 
time injected slowly and with caution. Some of the other points 
involved in these very interesting researches of Weil will be dis- 
104 Friedberger and Mita. Zeitschr. f. Immunitdt., 10, 1911. ANAPHYLAXIS 
443 
cussed in the following chapters in which we will deal with theories 
of anaphylaxis. 
Of practical importance is the duration of antianaphylaxis. As 
one would expect, a dose of antigen which renders a sensitive animal 
refractory, would, after the first reaction, act merely as a further 
antibody producing stimulus. In consequence, then, one would ex- 
pect the animal to return to the sensitive condition after a relatively 
brief period. In guinea pigs, according to a number of observers, 
the refractory condition may last two weeks or longer. In rabbits, 
according to Scott105 the return to the sensitive condition is very 
rapid, perhaps because of the energetic antibody production ob- 
served in these animals. 
Non-Specific Desensitization.—The complete or partial desensi- 
tization of a sensitive animal by injection of the antigen appears 
obviously to be due to the partial or complete saturation of the sessile 
antibodies by the antigen. Less clear is the unquestionable fact that 
partial desensitization can be non-specifically accomplished in a num- 
ber of ways. Dale made the observation that the uterus of a guinea 
pig sensitized against two different antigens and then desensitized 
completely to one of them, loses a certain amount of sensitiveness to 
the other antigen subsequently instilled into the bath. Biedl and 
Kraus,106 moreover, in their studies of peptone shock in dogs, re- 
ported that the injection of peptone into a hypersensitive dog pro- 
duced a considerable resistance against subsequent injection of the 
antigen. Similar conditions have been noted in connection with 
certain poisons which we shall speak of later as “anaphylatoxins” as 
prepared by Friedberger. These, as we shall see, are toxic products 
obtained in various ways from bacteria, agar suspensions, etc., which 
give, on injection into guinea pigs, symptoms simulating anaphy- 
lactic shock. Bessau 107 passively sensitized guinea pigs with 1 c. c. 
of anti-horse serum intraperitoneally, and on the following day in- 
jected them intravenously with 1 c. c. of horse serum. He gauged 
his dose so that the animals should have severe shock but survive. 
One or two days later he injected the amount of typhoid anapliy- 
latoxin which was fatal for normal pigs, and found that those which 
had been treated as described were now able to withstand the anaphy- 
latoxin. These experiments of Bessau would indicate that anti- 
anaphylaxis was to a certain extent due to tolerance to the poison, 
and that it was non-specific. Friedberger,108 together with Szyma- 
nowski, Kumagai, Odaira and Lura, later studied this problem and 
came to the conclusion that antianaphylaxis is strictly specific, de- 
pending, as Friedberger had suggested, upon the diminution of 
specific antibodies rather than upon tolerance to the poison. They 
105 Scott. Jour. Path. Bad., XIV, 1909, and XV, 1910. 
106 Biedl and Kraus. Wien. Kl. Woch., No. 11, 1909. 
107 Bessau. Centralbl. f. Bakt., 60, 1911. 
108 Friedberger and Szymanowski, etc. Zeitschr. f. Immunitdts., 14, 1912. 444 
INFECTION AND RESISTANCE 
claimed that animals that had been sensitized and then had survived 
the “shock” dose of homologous protein showed no tolerance for 
anaphylatoxin, and that animals that had been treated with the 
sublethal dose of anaphylatoxin are, for 24 hours, as sensitive to 
anaphylatoxin, however prepared, as are normal animals. Recent 
studies along the same lines by Zinsser and Dwyer 109 have yielded 
results differing from these conclusions. Working with typhoid ana- 
phylatoxin they found that guinea pigs treated with a sublethal dose 
of anaphylatoxin develop a tolerance which enabled them to resist 
to 2 units of poison, the tolerance developing within three days 
and lasting, to a slight degree, for as long as two months. It seemed 
to them that animals treated with a second dose of anaphylatoxin 
within 24 hours after the first, if the results of this first injection 
have been severe, as they usually are, still weak and generally de- 
pressed in vitality so that a developed tolerance may be clouded by 
this condition. The tolerance did not seem to be strictly specific in 
that typhoid anaphylatoxin seemed to produce a moderate tolerance 
to prodigiosus anaphylatoxin. 
It would seem, therefore, that in antianaphylaxis we might have 
two very important elements. The one strictly specific depends upon 
the depletion of antigen from the body, a true “desensitization.” The 
other non-specific, and probably of secondary importance since so 
far it has not been shown to any very powerful degree, consists of 
the development of tolerance by the body cells for the anaphylactic 
poison. 
Of similar bearing are the observations of Vaughan 110 who, work- 
ing with his toxic split products of the Colon bacillus protein, found 
that repeated injections into guinea pigs produced a tolerance which 
permitted the animals to resist about double the fatal dose, but did 
not protect against larger multiples. 
The preventive influence of atropin in anaphylaxis has already 
been noted in connection with the work of Auer and Lewis. Besredka, 
as we have also seen before, claims to have prevented fatal shock by 
ether anesthesia on the theory that most of the reactions took place 
in the central nervous system. It is more than likely that Besredka 
slightly underdosed his animals, and by the ether anesthesia merely 
prevented convulsive symptoms. Banzhaf and Famulener 111 pre- 
vented shock by large doses of chloral hydrate. Rosenau and Ander- 
son,112 carrying on similar investigations with urethan paraldehyde, 
chloral hydrate and magnesium sulphate, came to the conclusion that 
109 Zinsser and Dwyer. Reported at the meeting of Am. Ass. of Path. 
& Bact., Toronto, April, 1914. 
110 Vaughan. “Protein Split Products, etc.,” Lea & Febiger, Phila., 1913, 
p. 139. 
111 Banzhaf and Famulener. Stud. N. Y. Dept. Health, Labor., 1908, p. 
107. 
112 Rosenau and Anderson. Jour. Med. Res., No. 21, N. S. 16, 1909. ANAPHYLAXIS 
445 
none of these drugs had any noticeable effect on anaphylactic guinea 
pigs. The interesting observation of Friedberger 113 that injections 
of moderate amounts of hypertonic salt solution would diminish the 
severity of anaphylactic shock was explained by him on the basis 
that hypertonic salt prevented complement or alexin action, and 
thereby diminished the production of anaphylatoxin, a poison which, 
we shall see, he believes is produced by the action of alexin on the 
antigen antibody complex, and is responsible for anaphylaxis. 
Zinsser, Lieb and Dwyer,114 however, were able to show that the 
hypertonic salt solution probably acts by diminishing the irritability 
of smooth muscle in the animal body and thereby reduces the severity 
of the spasms in the bronchioles and other smooth muscle tissues of 
the guinea pig, preventing death in otherwise fatal experiments. 
113 Friedberger and Hartoch. Zeitschr. f. Imm., Yol. II, 1909. 
114 Zinsser, Lieb and Dwyer. Jour. Exper. Biol, and Med., 12, 1915. CHAPTER XVII 
ANAPHYLAXIS (Continued) 
In view of the rapidity with which the surprising facts of pro- 
tein hypersusceptibility were uncovered, it is but natural that many 
of the earlier theories conceived regarding the phenomena developed 
along lines which now seem fantastic. Yet, we must remember that 
there was much conflicting evidence about some of the fundamental 
facts, and it was some time before the recognition that an antigen- 
antibody reaction was basic to the occurrence, was firmly estab- 
lished. Among the most prominent of these earlier theories were 
those of Gay and Southard and of Besredka, and while these views 
have now been definitely discarded, the discussion to which they 
led, nevertheless, stimulated much interesting investigation. 
These two theories were premised upon the view that the injected 
antigen (horse serum, egg white, etc.) contained two separate sub- 
stances, one of them responsible for the sensitizing process, the other 
giving rise to shock on subsequent injection. 
The views of Gay and Southard 1 are summarized in the following para- 
graphs given, as nearly as space permits, in their own words: 
Increased susceptibility in the sensitized animal is due to the continued 
presence in the circulation of an unneutralized element of the antigen (in 
their case horse serum), which they call “anaphylactin,” which acts as an 
irritant or stimulant to the body cells, and, in some way, causes them to 
assimilate over rapidly certain other elements of horse serum. These assimi- 
lated or toxic elements are the same as those eliminated without producing 
intoxication during the incubation period following the first dose. This over- 
assimilation after anaphylaxis is the cause of the intoxication. 
Gay and Southard find much support for their contentions in the results 
of experiments done with the so-called “passive” transfer of hypersuscepti- 
bility. As mentioned above, hypersusceptibility may be transferred to a 
normal animal with the blood serum not only of a sensitive animal, but even 
more surely and effectually with that of a refractory, or “antianaphylactic,” 
animal. They believe that such transfer is not “passive” but “active” sensiti- 
zation, being accomplished by the transfer of “anaphylactin” to the normal 
animal. The refractory animal has received more horse serum than the 
merely sensitive one, since antianaphylaxis is produced by massive injections. 
Therefore its blood contains more anaphylactin and is consequently more 
active in transferring sensitiveness. The fact that a considerable incubation 
time is necessary in active sensitization they attribute to the gradual action 
of the anaphylactin. 
1 Gay and Southard. Jour. Med. Res., Vol. 16, 1907; Vol. 18, 1908; 
Vol. 19, 1908. 
446 ANAPHYLAXIS 
447 
In passive sensitization, therefore, they assumed a similar gradual irrita- 
tion of the vulnerable cells by the anaphylactin and, as we have seen, obtained 
their reactions in animals so treated, usually 10 to 14 days after the sensitive 
serum had been given. This conception of the mechanism of passive anapny- 
laxis was, of course, rendered unlikely by the demonstrations by Friedemann, 
Otto, and others that shock could be elicited in passively sensitized animals 
within 24 hours or less after transfer of the anaphylactic serum. 
To this, however, Gay and Southard 2 answer by implying that this greater 
speed of development of sensitiveness in the experiments of Otto is due to the 
larger doses used by him. They say “if the doses are sufficient it (trans- 
mitted sensitiveness) may be shown in a single day (Otto).” However, it is 
very likely that the sensitiveness noted by them in animals two weeks after 
the transference of anaphylactic serum was actually positive sensitization 
with antigen rests, entirely comparable to the usual “Theobald Smith” 
phenomenon. 
Gay and Southard’s definite objections to the possibility of an antigen- 
antibody reaction are found in the following arguments based on experi- 
mental observations: 
1. Sensibility persists for a long time, antibodies disappear rapidly. 
2. In the serum of animals sensitive to horse serum antibodies to this 
serum are not demonstrable by complement fixation. 
3. Although sensitiveness can be transferred to a normal animal, never- 
theless a definite period of incubation must elapse before the recipient be- 
comes sensitive. 
To the first of these arguments Besredka 3 objects by saying that, while it 
is true that sensitiveness persists for a long time, the power to transmit 
anaphylaxis passively disappears rapidly, as Otto, Richet, and others have 
shown. 
The second contention is contradicted by the work of Nicolle and Abt.4 
But since these workers made their observations upon rabbits their experi- 
ments do not necessarily contradict those of Gay and Southard. This point 
at best is a difficult one to determine, especially as recent investigations have 
shown us that under certain circumstances antigen and antibody may be 
found side by side in the same serum without uniting and without therefore 
fixing alexin or complement. 
The point of their third argument has been discussed above. 
It is clear that Gay and Southard separate distinctly the substance in the 
antigen which sensitizes from that which exerts the toxic action on second 
injection. 
Another theory which is based on such a separation of a sensi- 
tizing and shock-producing element in the original antigen is that 
of Besredka. 
Besredka 5 assumes that in the injected antigen (serum) there are two 
separate substances. One of these, the sensibilisinogen, induces, during the 
time of incubation, a specific antibody (sensibilisin). This antibody remains 
in part attached to issue cells and in part circulates freely in the blood. 
The other substance in the antigen he calls “antisensibilisin.” This, at the 
second injection, reacts with the sensibilisin and anaphylactic shock results. 
The nature of the symptoms is explained by the fact that the antibody or 
sensibilisin is attached to cells of the central nervous system, and shock can 
2 Gay and Southard. Jour. Med. Res., p. 427, 1908. 
3 Besredka. Bull, de Vlnst. Past., 6, 1908, p. 826. 
4 Nicolle and Abt. Ann. de Vlnst. Past., Yol. 22, p. 132, 1908. 
5 Besredka. Loc. cit. 448 
INFECTION AND RESISTANCE 
result only when such attachment is present. Thus, in passive transference 
of sensitization, the property of hypersusceptibility is bestowed upon the 
normal animal by the sensibilisin or antibody present in the circulating blood, 
but the significance of this body for anaphylaxis is not in evidence until a 
connection with the central nervous system has been established. 
There is much in Besredka’s theory which is at variance with prevailing 
conceptions of biological phenomena of this category. The fact that an 
antigen should give rise to an antibody which reacts not with the substance 
that induced it, but with a third body, is quite out of keeping with experience. 
However, it is clear that in both theories, that of Gay and Southard, as 
well as that of Besredka, the cardinal point is this separation in the antigen 
of two substances, a sensitizing and a toxic or shock-producing, and, since 
this forms the chief argument against an antigen-antibody conception of 
anaphylaxis, it will be necessary to examine the experimental evidence on 
which it is based. 
Gay and Adler 6 attempted to show such a dual function of the original 
antigen by chemical methods. They report that, by fractional precipitation 
of horse serum with ammonium sulphate, the successive protein fractions 
obtained, as saturation is increased, are found to be less sensitizing and 
more toxic as more and more ammonium sulphate is added. The first frac- 
tion (euglobulins) obtained by 1/3 saturation is as sensitizing as whole serum 
and corresponds to anaphylactin, but is nontoxic when injected into sensitive 
animals. The last fraction, while distinctly less sensitizing than either the 
whole serum or the first fraction, is at least as toxic as the whole serum. 
In these experiments, therefore, we have a strong argument in favor of 
the separate presence in an anaphylactic antigen of two bodies, the one 
sensitizing and the other toxogenic. However, this assertion has not been 
borne out by later work. 
Pick and Yamanouchi,7 whose extensive investigation cannot be fully 
reviewed here, were unable to obtain such a separation; in fact, they con- 
clude that the same substances which sensitize are also toxic, and, working 
with a large variety of methods, find that both the sensitizing and toxogenic 
properties of proteins show no differences either in chemical condition or in 
resistance to chemical agents or heat. 
The work of Pick and Yamanouchi, however, was done with rabbits and, 
therefore, as bearing on the theory of Gay and Southard, the work of Doerr 
and Russ 8 is more directly to the point. These workers using guinea pigs, 
and both horse and beef sera, obtained results which are practically dia- 
metrically opposed to those of Gay and Adler. They found that the euglobu- 
lins, obtained by 1/3 saturation with ammonium sulphate, are the most 
strongly sensitizing and, at the same time, the most toxic of the fractions 
of the sera. As saturation with the salt is increased, the proteins which 
come down decrease progressively and in parallelism, both as regards the 
power to sensitize and the faculty of exerting toxic action on second injec- 
tion. The albumin, which finally comes out on total saturation, is devoid 
both of sensitizing and of toxic properties. Similar results were obtained by 
Doerr and Russ with the precipitation of serum proteins with C02. 
The weight of evidence, therefore, seems to point against a chemical 
separation of the two functions in the antigen. 
Besredka’s contentions in favor of such a separation were based chiefly 
upon a difference in resistance to heat. 
His experiments showed that the sensitizing properties of serum are not 
lost even if it is heated to 120° C., while the toxogenic powers are destroyed 
6 Gay and Adler. Jour. Med. Res., Vol. 13, 1908. 
7 Pick and Yamanouchi. Zeitschr. f. Im-munitatsfarschung, Vol. 1, 1909. 
8 Doerr and Russ. Zeitschr. f. Immunitatsforschung, Vol. 2, 1909. ANAPHYLAXIS 
449 
by much lower temperatures. The results of Besredka as to the differences 
in thermostability between the two properties have found confirmation by 
Kraus and Volk 9 and others, and there can be little doubt that the sensitizing 
function is extremely heat-resistant, since this has also been shown by Wells,10 
Rosenau and Anderson, and many others. However, researches by Doerr 
and Russ,11 and notably by Wells, have shown that, though not destroyed 
by high temperatures, even moderate heating markedly diminishes the sensi- 
tizing function, and that larger doses have to be given as the temperature 
is increased; and since the smallest quantities of antigen necessary for in- 
ducing shock at the second injection must be anywhere from 100 to 1,000 
times as large as the smallest sensitizing doses, it is quite likely that a com- 
bination of such conditions might stimulate an actual difference in heat re- 
sistance. In fact, this is the view expressed by Wells 12 and borne out by 
experiments carried out by Doerr and Russ. 
Wells, too, confirms the identity of sensitizing and toxic substance by his 
experiments on the influence of tryptic digestion upon these properties of 
the antigen. He concludes that both sensitizing and intoxicating properties 
are attacked and slowly decrease as the coagulable protein disappears. 
As to that aspect of Besredka’s theory which deals with the indirect par- 
ticipation of the central nervous system, his arguments are based mainly on 
the fact that ether narcosis seemed, in his experiments, to prevent anaphy- 
lactic shock when animals were deeply anesthetized during the second injec- 
tion, and also upon the regularity, severity, and speed with which anaphy- 
lactic symptoms follow injections directly into the brain. The former con- 
tention regarding narcotics cannot, by any means, be accepted as yet, since 
Rosenau and Anderson failed to confirm it and claim that ether narcosis 
merely masks the symptoms but does not prevent death. If we admit the 
beneficial effects of ether, moreover, it may well be that this is accomplished 
by relaxation of the bronchial spasms, known, since Auer and Lewis, to be 
the cause of death in guinea pigs, and the action of ether could hardly be 
utilized, therefore, to argue in favor of a central localization of the anaphy- 
lactic process. 
That phase of the two theories so far mentioned, therefore, which depends 
upon the assumption of two separate substances in the original antigen does 
not seem established nor even sufficiently likely to warrant the formulation 
of a theory upon it. 
The second premise is the necessary participation of the body cell, in 
that the reaction cannot take place unless the cells are rendered vulnerable 
by preliminary alteration. In Gay and Southard’s theory this is assumed to 
occur by irritation exerted by the “anaphylactin,” in Besredka’s scheme it 
is attributed to the antisensibilisin which is attached to the nerve cells. 
It is plain, therefore, that both Gay and Southard and Besredka admit a 
preliminary preparation of the cells of the body, and this is, as we shall 
see, an important factor in anaphylaxis, though not exactly in the sense of 
any of the observers named. 
In contradistinction to the more complex theories of Gay and 
Besredka which we have just discussed, the early views of von 
Pirquet, of Rosenau and Anderson, and of most other observers who 
worked upon this subject, upheld the true antigen-antibody nature of 
the reaction with the antigen. 
9 Kraus and Yolk. Zeitschr. f. Immunit'dtsf orschung, Yol. 3, 1909. 
10 Wells. Jour. Inf. Bis., Yol. 5, 1908. 
11 Doerr and Russ. Loc. cit. 
12 Wells. Jour. Inf. Dis., Yol. 6, p. 521, 1909. 450 
INFECTION AND RESISTANCE 
Site of the Reaction.—Subsequent investigation was for some 
time largely aimed at analysis of the pathological physiology of the 
reaction, and one of the most important points of controversy which 
developed was that which concerned the question of whether the 
harmful antigen-antibody union took place in the circulation or in 
or upon the cells of the body. With the antigen-antibody mechanism 
accepted, it is but natural that earlier views favored the probability 
that the harmful union of antigen and antibody took place in the 
circulating blood. It will be seen, however, that even in the earlier 
theories mentioned above, the possibility of a participation of the 
tissue cells in the reaction was hinted at. Friedberger 13 was prob- 
ably the first to definitely suggest such a theory, although his subse- 
quent work carried him farther away from the cellular ideas than 
any of the other workers. The cellular conception was very largely 
based from the beginning upon the observation that passive sensitiza- 
tion in guinea pigs regularly demanded an interval of six or more 
hours between the injection of the antibodies and the development 
of the hypersensitive condition. Exceptions to this and their inter- 
pretation we will discuss presently. However, the regularity with 
which this interval was necessary in guinea pigs, the animals most 
frequently used for experiments at that time, demanded an explana- 
tion. The simplest explanation presenting itself was to the effect 
that a certain length of time was necessary for attachment of the 
injected antibody to the cells, and that hypersensitiveness implied 
such “sessile receptors.” 
The idea, itself, was not new. Wassermann had first suggested 
it in an attempt to explain the peculiar hypersusceptibility of toxin 
possessed by some of Behring’s toxin immunized animals. He as- 
sumed in such cases that the formation of antitoxin may, indeed, 
have been stimulated, but not to such a degree that the antitoxic 
substances had been to any extent distributed into the circulation. 
This he argued would leave the cell with a very much increased num- 
ber of antitoxin complexes, capable of uniting with toxin, but still 
united to the cell plasma in some way. In consequence, the cell would 
be excessively vulnerable during the stage just preceding the concen- 
tration of the antitoxin in the circulation. As Behring expressed 
it, in the case of antitoxin, the very substance which when free in 
the circulation has protective functions because of its ability to unite 
with the toxin and deflect it from the cells, may represent an actual 
cause for increased injury when still attached to the cell.14 The 
conception is quite plain and reasonable if one follows closely 
Ehrlich’s point of view, and it is mentioned here chiefly because it 
was upon this reasoning that Friedberger first based his idea that the 
13 Friedberger. Zeitschr. f. Immunitatsforschung, Yol. 2, 1909, p. 208. 
14 See discussion of Side Chain Theory on p. 143, etc. ANAPHYLAXIS 
451 
anaphylactic reaction might be a sort of cellular precipitin phe- 
nomenon. 
Friedberger, however, because of a new turn taken by his in- 
vestigations at this time, soon abandoned the cellular point of view, 
and in the course of subsequent development, two definite schools of 
opinion grew up, one which we may speak of as the “cellular” school 
of anaphylaxis, the other the “humoral.” 
In order to avoid further confusion, we may anticipate the 
analysis of these various opinions at once by stating that, as we now 
understand protein anaphylaxis, there can be little doubt about the 
fact that the most important phases of true protein anaphylaxis are 
the results of a reaction which takes place between the antigen and an 
antibody which is in some manner attached to, or in relation with 
body cells. To how great an extent an intravascular antigen-antibody 
reaction may play a minor role, we will discuss below; but, as to the 
primary importance of the cellular reaction in determining the 
anaphylactic phenomenon, there is little question. 
The Cellular Site of the Anaphylactic Reaction.—The reasons 
why it is generally believed that such a peremptory statement con- 
cerning the cellular site of the anaphylactic reaction can be made, 
are the following: 
In the first place, as wre have stated above, particular attention 
was called to the possibility of a cellular site for the anaphylactic 
reaction by the interval which was found to be absolutely necessary 
between the injection of an antiserum and the development of the 
hypersusceptible condition in the injected animal. Both Friedemann 
and Otto found that when the serum, containing antibodies, was 
injected subcutaneously, the best results were obtained by adminis- 
tration of the antigen 24 to 48 hours after this. On intraperitoneal 
injection of the sensitizing serum, Doerr and Russ 15 obtained the 
best results by permitting an interval of 24 hours to elapse, and these 
investigators attempted to determine the minimum period necessary 
when fairly strong antisera were used, but could not shorten the 
interval successfully to less than 4 hours, even when the serum was 
injected intravenously. There were, indeed, certain exceptions to 
this observation in the earlier literature, such, for instance, as those 
published in 1910 by Biedl and Kraus 16 who claimed that they ob- 
tained immediate and severe symptoms in guinea pigs when they 
injected, intravenously, mixtures of horse serum, together with the 
serum of sensitized animals. Gurd 17 also, more recently, claimed 
he obtained reactions in guinea pigs upon the intravenous injection 
of anti-sheep rabbit serum, followed immediately by sheep serum. 
Exceptions to the observation of an interval in rabbits, we will take 
15 Doerr and Russ. Zeitschr. f. Immunitats., Yol. 3, 1909, p. 181. 
16 Biedl and Kraus. Zeitschr. f. Immunitats., Yol. 4, 1910. 
17 Gurd. Jour. Med. Res., Yol. 31, 1914, p. 205. 452 
INFECTION AND RESISTANCE 
up a little later in connection with a discussion of the humoral views, 
but all observations of this type are probably of minor significance 
and perhaps dependent upon a mechanism quite distinct from that 
of true anaphylaxis, relations which we will discuss under “anaphy- 
latoxins.” Except for these isolated instances, the literature is con- 
sistent in regard to the necessary interval and the fact is beyond 
question. Complementary to the observation of the interval is the 
determination by Hoerr and others that, as the sensitive state de- 
velops, the antibody disappears from the circulation, and Doerr and 
Pick determined that anaphylactic shock could be obtained at a 
period when the circulation no longer contained any antibody. More- 
over, earlier observations on active anaphylaxis by many investi- 
gators, showed that complete absence of demonstrable antibody in the 
circulating blood was not at all inconsistent with the existence of 
the sensitive state. 
Another important plan of experimentation which had much in- 
fluence in convincing investigators of the correctness of the cellular 
theory, were transfusion experiments. Man waring 18 was one of 
the first to approach this problem in this way, by bleeding a sensi- 
tized dog extensively, and reinfusing him with the blood of a normal 
dog until there had been an almost complete replacement of blood. 
He found that the sensitized animal, in spite of the replacement of 
his blood, remained sensitive. Pearce and Eisenbrey 19 working wTith 
two normal and one sensitized dog, transfused the blood of one of 
the normal animals into the sensitized one, transferring the blood 
of the latter to the normal dog. “At the proper moment the normal 
dog containing the blood of the sensitized animal and the latter con- 
taining the blood of the normal dog, each received intravenously the 
toxic dose of horse serum.” The normal dog having the sensitized 
blood did not react; the sensitized dog having the normal blood 
showed typical fall of blood pressure. Pearce and Eisenbrey drew 
the conclusion “that the phenomenon of anaphylaxis is due to a reac- 
tion in the fixed cells and not either primarily or secondarily in the 
blood.” Similar experiments have since been done by Weil. 
Coca 20 further fortified the cellular point of view by a similar 
method which is more than mere confirmation of the dog experi- 
ments of Pearce and Eisenbrey because he employed guinea pigs. 
He succeeded in perfusing actively and passively sensitized guinea 
pigs with the defibrinated blood of normal guinea pigs, until the 
original blood of the sensitized animals was reduced to a slight 
residue. Animals so treated could be kept alive for as long as 6 hours 
after the transfusion, in spite of the blood substitution, and remained 
sensitive. Again of considerable importance, and complementary to 
18 Manwaring. Zeitschr. f. Immunitats., Vol. 8, 1910, p. 1. 
19 Pearce and Eisenbrey. Cong. Am. Phys. and Surg., Yol. 8, 1910, 
so Coca. Zeit. f. Immunitats,, Vol, 20, 1914. ANAPHYLAXIS 
453 
the transfusion experiments, is the work of Weil 21 on the influence 
of large amounts of circulating antibody upon anaphylactic shock. 
Weil found that guinea pigs that had been either actively or passively 
sensitized, could be protected against anaphylactic shock by the in- 
troduction into their blood of large amounts of immune body just 
previous to the shock injection. It is true that Weil obtained no 
results when he used moderate amounts of immune serum only, and 
large amounts had to be introduced in order to give him this protec- 
tion by circulating antibodies. He explained this by an avidity of 
the antigen greater for the sessile than for the circulating antibodies. 
Our own explanation, in keeping with the work of the writer with 
Young reported in an earlier chapter, would be that the phenomenon 
is not a question of avidities or chemical affinities, but rather at- 
tributable to the fact that in the circulation there is a colloidal pro- 
tective action which prevents the sudden union of antigen and anti- 
body, thus constituting perhaps the fundamental automatic protec- 
tion against intravascular anaphylactic shock. However this may 
be, the fact that a sufficiently large amount of antibody present in 
the circulation will protect a sensitized guinea pig against shock 
seems to us very strong evidence in favor of the cellular conception. 
Although it would hardly seem that with all this evidence further 
proof of the cellular site of the reaction would be necessary, neverthe- 
less further incontrovertible proof was brought by the development 
of the isolated tissue technique, first by Schultz and later by Dale. 
In 1910 Schultz began to work with what is now spoken of as 
the physiological method. He determined that smooth muscle— 
freshly excised from various animals—will react with contraction 
when brought into contact with serum. When such muscle was 
taken from anaphylactic animals after being thoroughly washed free 
of blood, it would react more energetically and to smaller amounts of 
the homologous serum than would similar tissue from normal ani- 
mals. There are many interesting by-products of Schultz’s work, 
such as the differences between fresh arterial blood and blood serum 
in their abilities to stimulate contraction, hut this and other points 
will not be discussed at present. The important and incontrovertible 
fact established by Schultz is the changed reaction-energy or, in 
truth, “allergie” of the smooth muscle of anaphylactic animals to 
the stimulus of the sensitizing antigen. Dale 22 confirmed and ex- 
tended these observations of Schultz. He removed the uteri from 
guinea pigs after thoroughly perfusing them with Ringer’s solution 
to remove all blood. He then suspended them in baths of Ringer’s 
solution and by the customary physiological methods measured the 
contractions following the addition of various amounts of foreign 
protein in the form of—among other things—horse serum and beef 
21 Weil. Jour. Med. Res., New Series, Yol. 22, 1913, p. 497. 
22 Palg. Jour, Pharmacol, and Exper. Therap., 1913, iv. 454 
INFECTION AND RESISTANCE 
serum. He found that the uterus of an animal sensitized to horse 
serum would react to this substance in dilutions of 1:2,000 or 
1:10,000, while the organ taken from a normal guinea pig reached 
its limit of reactionability at dilutions often less than 1:20. A 
uterus that had reacted strongly was found to be subsequently desen- 
sitized. A normal uterus could not—strangely—be passively sen- 
sitized by immersion into a solution containing serum antibodies. 
Richard Weil 23 has fully confirmed the principles laid down bv 
Schultz and Dale. He incidentally also answered an objection to 
the conclusions of Dale and Schultz (never indeed a very valid ob- 
jection) brought forward by Larson and Bell ,24 namely, that the 
reaction of the muscle tissue of a sensitized animal might be in part 
due to the fact that the blood, i.e., the antibodies, had not been en- 
tirely washed out of the tissue spaces by perfusion. Weil performed 
the very simple and ingenious experiment of injecting a normal 
guinea pig with large amounts of immune serum (anti-horse serum) 
and, after a few minutes, killing the animal. He then suspended 
the uterus in Ringer’s solution in the usual manner without washing 
it completely free of blood. Contact with the homologous antigen 
produced no response. We may accept as definitely established by 
these researches of Schultz, Dale, and Weil that the fixed cells of 
anaphylactic animals possess an increased reaction-ability toward 
the antigen which is in no sense secondary to processes involving the 
circulating antibodies. Moreover, the work of Weil seems to indi- 
cate that desensitization of a passively prepared guinea pig deprives 
the uterus of its power to respond and that the gradual spontaneous 
diminution of hypersusceptibility on the part of the guinea pig is 
accompanied by an entirely parallel loss of reaction-capacity on the 
part of the isolated uterus. 
Weil 25 also carried out a very ingenious series of experiments 
in which he tried to show that both antigen and antibody could be 
present on the tissue cell at one and the same time. Again using 
the isolated uterus method of Dale, he partially desensitized a 
guinea pig in the following way. The guinea pig was first actively 
sensitized with normal rabbit serum. After the lapse of a proper 
interval, the animal was reinjected, but this time not with normal 
rabbit serum, but with the serum of a rabbit immunized with horse 
serum, in other words, with rabbit serum containing antihorse 
antibodies. Thus, the effect of the second injection would be, in 
the first place, to act as antigen as far as its nature as rabbit serum 
is concerned, and in this capacity would partially desensitize the 
guinea pig to the rabbit serum element; on the other hand, this 
23 Weil, R. Jour. Med. Research, 27, 1913; 30, 1914; Proc. Soc. Exper 
Biol, and Med., 1914, xi, 86. 
24 Larson and Bell. Jour. Inf. Bis., Yol. 24, 1919, p. 185. 
25 Weil. Jour. Med. Res., New Series, Yol. 25, 1914, p. 299. ANAPHYLAXIS 
455 
rabbit serum administered at the second injection, also represented 
horse serum antibodies by virtue of which it could passively sensitize 
the animal to horse serum. On the day following this second in- 
jection, Weil tested the uteri of the guinea pigs and found them 
partially sensitive to rabbit serum, but also sensitive to horse serum. 
From this he drew the conclusion that since rabbit serum antigen 
was at the same time a conveyor of antibodies, the passive sensitiza- 
tion of the guinea pig against horse serum indicated that rabbit 
serum antigen must be attached in some way to the uterine tissue 
cells. The only possible objection to this line of very ingenious rea- 
soning on the part of Weil is the development of our knowledge on 
the production of apparently protein-free antibodies, more recently 
developed by Huntoon in connection with pneumococcus antibodies. 
It is still possible that the presence of horse serum antibodies in the 
cells by virtue of the passive sensitization, might not necessarily 
indicate the presence of rabbit serum antigen as such. Yet, there 
is as much to be said on Weil’s side as on the other, and the experi- 
mental evidence is all on his side. The reader may have a little 
difficulty in understanding this experiment unless he draws himself 
a simple diagram of the conditions in reading it. In describing Weil’s 
work in a book of this kind it is difficult for a contemporary to 
omit deploring the fact that his untimely death deprived American 
anaphylaxis investigations of one of its most earnest and brilliant 
workers. 
There is, thus, an incontrovertible mass of evidence available 
which proves without question that the site of the reaction which 
carries in its train the symptoms which we speak of as anaphylaxis, 
is predominantly on the cells of the body. Whether or not the 
intravascular reaction of antigen and antibody, which unquestion- 
ably to a certain extent takes place, has any, if even a minor im- 
portance, we will discuss directly. 
In discussing cellular anaphylaxis, the experimental analysis 
has necessitated the investigation of tissues like smooth muscle, in 
which a definitely noticeable reaction was produced, and we have 
seen that it is probably the smooth muscle tissue, in guinea pigs 
in the bronchioles, in rabbits in the pulmonary vessels, in dogs in 
the herpatic circulation, which lead to the severe injuries probably 
responsible for death. This does not, however, mean necessarily, 
that other tissues of the body are not either acutely or subacutely 
injured. Indeed, many facts would indicate that sensitization by 
attached antibodies and subsequent injury is not by any means 
limited to individual tissues. The sudden drop of body tempera- 
ture, the peculiar intestinal symptoms (vomiting, defecation and sero- 
sanguinous exudation of the intestinal mucous membrane) noticed 
by Weichardt and Scliittenhelm26 in dogs dying of protracted 
26 Weichardt and Schittenhelm. Dent. med. Woch., 1902. 456 
INFECTION AND RESISTANCE 
anaphylactic shock, the increased secretion of many glands and the 
symptoms referable to the nervous system, may, of course, be to 
some extent entirely secondary to changes in other organs. But 
it is also possible, and we may even say likely, that such occurrences 
indicate a generalized anaphylactic effect which in the case of many 
organs, of course, cannot be subjected to experiment with the same 
definiteness with which we can observe contractile tissues like 
the smooth muscle. 
Similar in effect are the observations which we may classify as 
local anaphylaxis. In the general historical review we have already 
mentioned the phenomenon of Arthus, where a subcutaneous in- 
jection in a sensitive rabbit gives rise to (edema and eventual local 
necrosis. This phenomenon, however, has not been observed either 
in guinea pigs or in dogs with any degree of regularity. Lewis,27 
indeed, has described such a phenomenon in guinea pigs, but ex- 
tensive experiments in our own laboratory, as yet unfinished, have 
failed to give anything like regular results in attempts to produce 
the Arthus phenomenon in these animals. Coca’s suggestion that 
perhaps the conditions in rabbits differ from those in other animals in 
regard to this phenomenon because of a more severe and general 
effect upon the capillary walls, is worth considering, although other 
possibilities cannot be excluded. Whether or not the many local 
manifestations apparent in man on, let us say, the injection of 
antitoxic horse sera, or the quite severe and progressively increasing 
local reactions observed by anyone who has given the Pasteur treat- 
ment for rabies, can be regarded as evidences of local anaphylaxis, 
is a matter that cannot as yet be definitely determined. We, our- 
selves, are inclined to believe that this is the case. Coca, on the 
other hand, excludes such phenomena from classification with anaphy- 
laxis for reasons which we will have occasion to discuss under the 
heading of serum sickness. 
Auer 28 more recently made some very interesting and signifi- 
cant observations on local anaphylaxis. Auer was carrying out tests 
for hypersusceptibility on dogs which had been treated with horse 
serum some years before, and employed heavy doses of horse serum 
for reinjection. At the site of operation in the inguinal region, he 
observed two days later an extensive, thick oedema. In explaining 
this he assumed that the local reaction was due to the fact that a 
foreign protein, horse serum, was circulating in the blood and a 
certain amount of this protein, therefore, passed into the tissues 
along the wound during the development of the ordinary injury 
following operation. This was probably further increased by oozing 
of blood serum and plasma into the wound. Since the dogs were 
sensitized, his idea was that the skin and adjoining tissues, being 
27 Lewis. Jour. Exp. Med., Yol. 10, 1908, p. 1. 
28 Auer. Jour. Exper. Med., Yol. 32, 1920, p. 42<. ANAPHYLAXIS 
457 
sensitive, would respond locally to the protein so applied. In fol- 
lowing this up he carried out analogous experiments in rabbits as 
follows: He reinjected sensitized rabbits with the homologous anti- 
gen and applied xylol, which is a powerful skin irritant, to the 
ears. In such cases he often obtained severe inflammation with 
the formation of crusts, and, occasionally, gangrene of the entire 
tips of the ears. Xylol applied in the same dosage and in the 
same way, to the ears of normal rabbits injected once with horse 
serum, and of sensitized but not reinjected rabbits, caused only a 
minor inflammation with a little oedema which disappears com- 
pletely in two or three days. His explanation is that the lesion 
on the ears of the sensitized and reinjected rabbits after xylol, was 
anaphylactic in nature as the result of a local autoinoculation of the 
ear with the circulating antigen. The local autoinoculation was 
brought about, he thought, by the irritant action of the xylol which 
caused an inflammation and oedema at the site of application. An 
anaphylactic reaction now occurred because inflamed tissues were 
more active metabolically than the normal tissues, and the in- 
flamed cells, therefore, affected more by the antigen than the normal 
cells. He concludes that this process may theoretically occur in 
any tissue of a sensitized animal under analogous conditions. These 
experiments are of considerable interest in perhaps pointing out pos- 
sible anaphylactic injury where some mechanical or other cause 
of inflammation coincides with the presence of antigen in a sensitized 
subject. 
The Factor of Injury in Anaphylaxis.—However well we may 
believe we understand the mechanism of anaphylaxis, and the site 
of the reaction, it is still obscure what the factor of injury actually 
consists in. Two chief lines of thought have been followed in the 
explanation of this, one of them attributing the injury to mechanical 
factors due largely to colloidal changes brought about by the antigen- 
antibody union, the other regarding the process as an intoxication 
with a poison formed in consequence of the union. 
To the former category belong several theories in which the 
formation of a precipitate in or upon the cells has been suggested. 
Similar, also, is the idea of Xolf 29 who assumes a colloidal change 
by which fibrin is formed and becomes in some way attached to 
cellular elements. Because of utter lack of experimental data, or 
even methods by which this subject can be approached, we can pay 
very little attention to the physical theories. Moreover, most views 
involving physical processes, such as that of Jobling and Petersen,30 
imply the secondary formation of a toxic substance. Most of the 
theories implying the formation of toxic products will be discussed 
in our consideration of “anaphylatoxins.” We wish merely in this 
29 Nolf. Arch. Internat. de Phys., Yol. 10, 1910, p. 37. 
30 Jobling and Petersen. Jour. Exper. Med., Yol. 20, 1914, p. 37. 458 
INFECTION AND RESISTANCE 
place to discuss briefly the attempts to chemically define the poison 
possibly involved. Stimulus for investigations of this kind was 
given largely by the investigations of Biedl and Kraus 31 on pep- 
tone shock. DeWaele 33 had called attention to the similarity of 
the symptoms following the injection of peptone, to those of anaphy- 
lactic shock. Biedl and Kraus studied the phenomenon very carefully 
in dogs, and found complete analogy between anaphylactic mani- 
festations and peptone shock. We have already called attention to 
the fact that they, furthermore, maintained the partial desensitiza- 
tion of sensitized dogs by preliminary peptone injections. While 
these observations have been amply confirmed, the chemically un- 
defined nature of Witte’s peptone prevents any degree of accuracy 
in chemical reasoning from these facts. Of more definite importance 
is the observation of Bale that histamin, a protein cleavage product, 
gives rise to symptoms in guinea pigs almost identical with those 
observed in anaphylaxis. Bale and -Laidlaw 33 in their extensive 
studies on this substance, have worked out an almost complete analogy 
between the two phenomena. Moreover, as Wells points out, 
histamin, which is a histidin derivative, is present in every known 
complete protein. While these considerations are of great in- 
terest and should be held in reserve for possible bearing on future 
investigations, there are many points about the assumption that 
histamin is the actually effective toxic product developed in the 
course of anaphylaxis which are unproven and unlikely. We do 
not believe that Bale, himself, meant to draw any definite conclu- 
sions in this regard, but merely recorded the interesting observed facts 
in an objective manner. The suddenness with which anaphylactic 
shock supervenes on the injection of adequate amounts of antigen, 
alone renders it extremely difficult to assume that such an extensive 
cleavage of either the antigen or the cell substance could take place 
with the speed observed in shock. Moreover, Boerr 34 points out 
that histamin can be obtained from cellular materials only by ex- 
tensive acid hydrolysis, and he seems to us quite logical in question- 
ing whether it is possible to assume such a rapid and destructive 
cleavage as the result of a simple antigen-antibody union upon a cell. 
Moreover, he also points out that the production of histamin in a 
cell is quite incompatible with subsequent life of the cell, and from 
Bale’s uterus experiments it is quite plain that a desensitized uterus 
may retain its contractile properties for a long time, sometimes not 
only returning to normal, but being unusually irritable. 
The various opinions upon the production of toxic substances as 
31 Biedl and Kraus. Wien. Min. Woch., No. 11, 1909. 
32 De Waele. Bull, de VAcad. Royale de Med. de Brux, 1907, cited from 
Doerr. 
33 Dale and Laidlaw. Jour. Physiol., Yol. 52, 1919, p. 355. 
34 Doerr. “Review of Anaphylaxis,” Erg. d. Hyg. & Bakt., etc., Yol. 5, 
1922, p. 226. ANAPHYLAXIS 
459 
a consequence of antigen-antibody union are extensively discussed 
in connection with the humoral theories and anaphylatoxin formation. 
The Humoral Theories and the Development of the Anaphyla- 
toxin Ideas.—We have seen that the most important point of de- 
parture for the conception of the cellular site of the anaphylactic 
reaction was the observation of the necessity of an interval between 
the injection of an immune serum in passive sensitization, and the 
development of the anaphylactic condition. From the beginning, a 
certain number of exceptions to this rule were noticed. Weill-Halle 
and Lamaire 35 described experiments in which it was found that 
guinea pigs would react with typical anaphylactic symptoms if in- 
jected simultaneously with the serum of sensitized rabbits and the 
antigen. According to them, successful results in such experiments 
depended entirely upon the time at which the rabbits sensitized with 
the horse serum were bled. Friedemann 36 at about the same time 
published experiments on anaphylaxis in rabbits in which he obtained 
apparently typical anaphylactic symptoms on the simultaneous in- 
travenous injection of antigen and the serum of sensitive animals. 
According to Friedemann, this was the best way to elicit an anaphy- 
lactic reaction in rabbits for he claimed that if the sensitive serum 
was injected 24 hours before the antigen, the reaction became in- 
distinct. He concluded from these experiments that the anaphylactic 
poison, whatever it might be, is, in rabbits at least, formed "in the 
circulating blood. In the same year, namely, 1909, Kichet 37 ob- 
tained similar results with crepitin with which he sensitized dogs, 
bled them during the hypersusceptible period, mixed the serum with 
a harmless dose of crepitin, and injected the mixture into a normal 
dog. Violent anaphylaxis-like symptoms resulted immediately. He 
speaks of his experiments as “reaction anaphylactique in vitro.” His 
experiments cannot be interpreted as having very much bearing on 
the present question because he worked with a substance primarily 
toxic, and of doubtful antigenic nature. However, Biedl and Kraus 38 
shortly afterwards obtained immediate and severe symptoms in 
guinea pigs on intravenous injection of mixtures of horse serum, 
together with the serum of sensitized guinea pigs. Similarly, Briot,33 
reported similar observations in young rabbits, into which he had 
injected mixtures of horse serum and antihorse serum. Gurd 40 
carried out similar experiments in guinea pigs with sheep and anti- 
sheep serum on simultaneous injection. Scott41 reported similar 
observations on rabbits, and certain observations reported by Man- 
35 Weill-Halle and Lamaire. Compt. r. d. Soc. Biol., Yol. 65, 1908, p. 141. 
36 Friedemann. Zeitschr. f. Immunitats., Yol. 2, 1909. 
37 Richet. 
38 Biedl and Kraus. Zeitschr. f. Immunitats., Vol. 4, 1910. 
39 Briot. Compt. r. d. Soc. Biol., Vol. 68, 1910, p. 402. 
40 Gurd. Jour. Med. Res., Vol. 31, 1914, p. 205. 
41 Scott. Jour. Path. & Bacter., Vol. 15, 1910, p. 31. 460 
INFECTION AND RESISTANCE 
waring 42 indicate that analogous observations can be made in dogs. 
To this not inconsiderable mass of evidence we may add observa- 
tions of our own which, though showing considerable irregularity, 
pointed in the same direction as the work reported by Friedemann. 
indeed, as regards our own experiments on this subject, we may 
say that while we occasionally obtained severe unmistakeable anaphy- 
laxis-like symptoms in rabbits on simultaneous or immediately con- 
secutive injection, we could never be sure of successfully duplicating 
the result. If we summarize all the evidence concerning the ques- 
tion of the interval from the available literature, we believe that 
the following statement would come pretty close to a just appraisal. 
In guinea pigs there seems to be no question whatever about the 
necessity of some sort of an interval, and observations to the con- 
trary can probably be explained upon some peculiarity of the ex- 
perimental conditions which are not clear; but the overwhelming 
mass of evidence points to the necessity of some interval, probably 
not less than 4 hours. In rabbits and dogs, when large amounts of 
the two reacting substances are injected, passive sensitization may 
occur almost immediately, or within 5, 10 or 15 minutes. But this 
occurrence is irregular and cannot be obtained, in our opinion, with 
anything but a considerable dosage. We, ourselves, believe that the 
irregularity and the necessity for large dosage can be explained by 
the facts brought out in our studies on the simultaneous presence of 
antigen and antibody in the same animal from which we concluded 
by analogy, that the circulating blood normally contains a substance, 
probably associated with the albumins of the plasma, which, by a 
mechanism of colloidal protection, prevents the rapid union of antigen 
and antibody in the blood. This we regard as an automatic pro- 
tective device which prevents the probably harmful effects of such 
sudden unions, making them gradual and slow, but is broken down 
by the use of large amounts of both substances and perhaps in in- 
dividual animals may be, for some reason or other, defective. 
However this may be, these investigations, coming early in the 
development of anaphylactic studies, gave rise to the assumption 
in the minds of many workers, that anaphylaxis might be the result 
of a meeting of antigen and antibody in the circulation, and many 
investigations were initiated in attempts to explain the mechanism 
by which this occurred. 
The Anaphylatoxin Conception.—The concept underlying most 
of the work now undertaken in attempts to prove a humoral mechan- 
ism was the following: When a specific antigen meets its anti- 
body the reaction between them gives rise to a toxic product, and 
this causes the characteristic symptoms. A similar idea, it will 
be remembered, is found in the original endotoxin theory of Pfeiffer. 
According to this, the action of the specific lysin liberated from 
42 Manwaring. Zeitschr. f. Immunitats., Yol. 8, 1910. ANAPHYLAXIS 
461 
bacteria a preformed poison, the endotoxin. In 1902 Weiehhardt,43 
bearing this conception in mind, subjected syncytial protein of rabbit 
placenta to the action of specific antisera and obtained substances 
toxic for normal rabbits. This work was done long before the days 
of anaphylaxis studies, and the results were interpreted in keeping 
with Pfeiffer’s theory. However, as Weiehhardt himself now claims, 
it is not unlikely that he was dealing with a phenomenon analogous 
to the ones we are now discussing. A similar opinion of the pro- 
duction of toxic substances by specific cytolysis was expressed by 
Wolff-Eisner 44 in 1904. 
The idea is really a morphological one in which the “endotoxin” 
is regarded as something present in the antigen which is set free 
by disintegration of the cell. In applying this to serum anaphylaxis 
Wolff-Eisner 45 preserved this morphological simile in that he spoke 
of the dissolved protein antigen (serum, etc.) as “nur scheinbar 
gelost” and “dass es erst durch die Lysine wirklich resorbierbar 
wird.” 
Indeed the sudden liberation of endotoxins by immune sera had 
been regarded by Pfeiffer and others as the cause of the rapid death 
often ensuing in immunized guinea pigs wdien more than a definite 
maximum of cholera spirilla or other organisms was injected. In 
all these opinions the basic conception was that certain bacteria con- 
tained a characteristic preformed poison (endotoxin) upon the 
pharmacological properties of which the peculiar symptoms caused 
by each organism depended. 
Probably the most important of the earlier investigations along 
these lines, at least in its direct bearing on anaphylaxis, was the 
work of Vaughan and Wheeler,46 published in 1907. Their con- 
ception of anaphylaxis takes root in the earlier investigations of 
Vaughan 47 and his pupils upon the extraction of a poisonous group 
from the protein molecule. 
Vaughan and Wheeler48 believed that the sensitizing and the toxogenic 
properties of the anaphylactic antigens are in truth contained within the 
self-same protein molecule; but can be chemically separated from each other. 
They were able to split egg albumen and other proteins by treatment with 
absolute alcohol (containing 2 per cent. NaOH) into 2 fractions—a toxic 
alcohol-soluble and a non-toxic alcohol-insoluble one. The former fraction 
gave protein reactions, and they regarded it as a true protein—while Wells,49 
43 Weiehhardt. Deutsche med. Woch., 1902, p. 624. 
44 Wolff-Eisner. Centralbl. f. Bakt., Yol. 37, 1904. 
45 Wolff-Eisner. “Handbuch der Serum Therapie,” p. 24, Lehmanns, 
Miinchen, 1910. 
46 Vaughan and Wheeler. Jour. Inf. Dis., Yol. 4, 1907. 
47 Vaughan. Trans. Ass’n Am. Phys., Vol. 16, 1901; Jour. A. M. A., Yol. 
36, 1901; Am. Med., 1901; Jour. A. M. A., Vol. 43, 1904. 
48 V. C. Vaughan, Jr. Jour. A. M A., Vol. 44, 1905, p. 1340; Boston 
Med. and Surg. Jour., Vol. 155, 1906. 
49 Wells. Jour. Inf. Dis., Vol. 5, 1908. 462 
INFECTION AND RESISTANCE 
considering the hydrolytic nature of the cleavage resorted to, considers this 
fraction as possibly a soluble peptone or polypeptid (the positive protein 
reactions being possibly due to amino acids). The non-alcohol-soluble, non- 
toxic fraction also gives protein reactions. Injections into guinea pigs of 
the toxic fraction produced symptoms not unlike anaphylaxis—but did not 
sensitize against protein. The alcohol-soluble portion was non-toxic and sensi- 
tized against protein in doses of 0.001 to 0.005 gm. Based on these results, 
their views of mechanism of anaphylaxis were as follows: 
At the first injection a slow lysis (cleavage) of the injected protein grad- 
ually liberates a fraction, corresponding to the alcohol-insoluble substance— 
and this by its antigenic action gives rise to the formation, in excess, of an 
enzyme (lysin), which on reinjection brings about the rapid cleavage of the 
injected protein—with an explosive liberation of the toxic fraction and conse- 
quent symptoms.50 
ISicolle believes that the injection of a protein into an animal induces 
the production in the subject of antibodies. These are preeminently two— 
albuminolysins, which cause its cleavage, and albuminocoagulins or precipi- 
tins, which coagulate and prevent the action of the lysin. At the time at 
which an animal is hypersusceptible or anaphylactic there has been a pro- 
duction of albuminolysins which cause cleavage of the protein, with the rapid 
liberation of toxic substances; but the albuminocoagulins or precipitins have 
not yet adequately developed. In a refractory animal the neutralizing action 
of the albuminoprecipitins prevents the harm which the lytic action might 
otherwise accomplish. The relative amounts of these two antibodies present 
in the circulation of the animal at any particular time determine whether the 
animal is anaphylactic or refractory or immune. This theory assumes arbi- 
trarily the protective nature of precipitation, an idea which has no foundation 
in experiment and, in fact, is rendered extremely unlikely by more recent 
developments of our knowledge of the precipitating antibodies. 
Given, then, a reasonable, hypothesis in which anaphylaxis is 
associated with the cleavage of protein by lysis, given, in other words 
an antigen-antibody conception, it is but natural that experimenters 
should ask themselves: What is the relation of the alexin to this 
cleavage? For in all known lytic reactions, of course, the union of 
antigen and antibody leads to the absorption of alexin, by means of 
which, then, the lysis has been assumed to take place. This problem 
suggested itself to a number of the earlier investigators who attempted 
to approach it by determining whether or not the sera of sensitive 
animals, added to antigen, would fix alexin. Gay and Southard, 
Sleeswijk,51 and others obtained negative results, while Ricolle and 
Abt,52 and Doerr and Russ 53 obtained positive results. As far as 
this particular method is concerned, therefore, no conclusions can 
be drawn. Sleeswijk, however, has approached the question in an- 
other way and examined whether or not there is a diminution of 
alexin in the blood of an animal immediately after anaphylactic 
50 For the sake of completeness it is well also to mention Nicolle’s theory, 
which, though attractive, is not borne out by recent knowledge concerning 
the nature of precipitins. See Ann. de VInst. Past., Vol. 22, 1908. 
51 Sleeswijk. Zeitschr. f. Immunitdtsforschung, Yol. 2, 1909. 
52 Nicolle and Abt. Ann. de VInst. Past., Yol. 22, 1908. 
53 Doerr and Russ. Zeitschr. f. Immunitdtsf orschung, Vol. 3, 1909. ANAPHYLAXIS 
463 
shock. He found that this was indeed a regular occurrence, and his 
results have been confirmed by Friedberger and Hartoch 54 and a 
number of others. 
It was shown by these workers that, both in active and passive 
anaphylaxis in rabbits and dogs, as well as in guinea pigs, there is a 
definite and considerable diminution of complement immediately 
after anaphylactic shock. 
The question now arises: What is the significance of this dimi- 
nution of alexin? Do the animals die because of a sudden loss of 
circulating, physiologically necessary alexin, or does the alexin take 
an active part in producing the conditions which cause death ? 
Either of these possibilities might follow from the mere fact of 
alexin diminution, but the former—the possibility that complement 
depletion is the cause of death—was ruled out by Friedberger and 
Hartoch.55 They showed that, by supplying fresh complement to 
sensitive animals at the time of reinjection, shock cannot be pre- 
vented. They now proceeded to demonstrate the active participation 
of complement in the production of anaphylaxis. They did this in an 
ingenious wray which depended on utilization of the fact observed by 
Holf,56 Hektoen and Ruediger,57 and others that hypertonic salt 
solution (1.5 to 2 per cent.) will prevent the combination of comple- 
ment with its sensitized cells. By slowly injecting into sensitized 
guinea pigs 0.3 cubic centimeter of concentrated HaCl solution 
just before the injection of antigen they were able to markedly 
diminish anaphylactic shock—saving animals from injections which 
invariably killed the controls. The force of this experiment we think 
has been largely eliminated by work of the writer with Dwyer and 
Lieb 58 in which it was shown that the effect of the salt is upon the 
smooth muscle which is rendered less irritable by the salt and there- 
fore less susceptible to the antigen. 
An ingenious attempt to demonstrate the important role played by com- 
plement in anaphylaxis was that of Loeffler. Loeffler,59 using guinea pigs sen- 
sitized with horse serum, completely depleted their complement by injecting 
intraperitoneally considerable quantities of sensitized sheep corpuscles. Tested 
by injection of horse serum one hour later no anaphylaxis occurred, while 
controls regularly succumbed.60 
It seemed thus possible the complement or alexin might play 
an important active part in the production of anaphylaxis, and the 
54 Friedberger and Hartoch. Zeitschr. f. Immunitatsforschung, Yol. 3, 
1909. 
55 Friedberger and Hartoch. Loc. cit. 
56 Nolf. Ann. de Vlnst. Past., 1900. 
57 Hektoen and Ruediger. Jour. Inf. Dis., Yol. 1, 1904. 
58 Zinsser, Lieb and Dwyer. Proc. Soc. for Exp. Biol. & Med., Vol. 12, 
May, 1915, p. 123. 
69 Loeffler. Zeitschr. f. Immunitatsf orschung, 8, 1910. 
60 For additional evidence pointing in the same direction see also Uhlen- 
huth and Haendel, Zeitschr. f. Immunitatsf orschung, Vol. 3, 1909. 464 
INFECTION AND RESISTANC E 
next logical step was to attempt to produce the anaphylactic poison 
by the action of alexin upon an antigen-antibody complex in vitro. 
This was first done, with direct reference to anaphylaxis, by Ulrich 
Friedemann.61 Friedemann chose as his antigen-antibody complex 
the sensitized red blood cell after he had demonstrated by preliminary 
experiment that the basic anaphylactic experiment could be carried 
out in rabbits with washed beef corpuscles. He found that if 3 c. c.. 
of such corpuscles were injected into rabbits and the injection re- 
peated after 3 weeks anaphylactic symptoms were regularly elicited. 
He then allowed alexin to act upon sensitized beef blood in vitro, 
interrupted the action by cooling at a time just preceding the occur- 
rence of hemolysis (to exclude the supposed toxic action of hemoglo- 
bin), and injected the supernatant fluid of such mixtures into normal 
rabbits. The result was marked illness resembling anaphylaxis, and 
Friedemann thus had succeeded in producing the anaphylactic poison 
in vitro under conditions as nearly as possible similar to those occur- 
ring in the circulation of the anaphylactic rabbit. In the conclusions 
drawn from his experiments he expresses the opinion that the poisons 
were not preformed in the red blood cells, but were formed by the 
proteolysis exerted by “amboceptor” and complement. In this state- 
ment he sets down the basic conception of the production of anaphy- 
lactic poisons now generally held. 
Friedemann, then, in attempts to apply the same methods to the 
study of serum anaphylaxis, attempted to produce similar poisons 
by the action of rabbit alexin upon the washed precipitates formed 
by mixtures of antigen and precipitating sera. In this he failed— 
probably because of his choice of rabbits as subjects for experiment. 
Where he had failed, however, Friedberger 62 succeeded by using 
guinea pigs. Doerr and Russ 63 had previously shown that feeble 
symptoms of shock could be produced by the injection of serum pre- 
cipitates into normal guinea pigs. With this additional evidence 
in favor of his reasoning, Friedberger proceeded as follows: 
One c. c. of a rabbit serum which precipitated sheep serum in a 
dilution of 1 to 10,000 was mixed with 30 c. c. of a 1 to 50 sheep 
serum dilution. This was kept one hour at 37.5° C. and over night 
in the ice-chest, when a heavy flocculent precipitate had formed. 
This precipitate was washed to remove all traces of serum, and to it 
were added 2 c. c. of fresh normal guinea pig serum—as comple- 
ment. This was again allowed to stand for 12 hours and then the 
supernatant fluid was injected into a guinea pig intravenously. In 
most cases the pigs so treated showed marked symptoms soon after 
the injection and died within a few hours. 
61 Ulrich Friedemann. Zeitschr. f. Immunitatsforschung, Yol. 2, 1909. 
62 Friedberger. Berl. klin. Woch., 32 and 42, 1910; also Zeitschr. f. 
Immunitatsforschung, Yol. 4, 1910. 
63 Uoerr and Russ, Zeitschr, f. Immunitatsforschung, Yol. 3, p. 181, 1909, ANAPHYLAXIS 
465 
Friedberger concludes, therefore, that anaphylactic shock is a 
true intoxication due to a poison produced from the products of a 
precipitin-precipitinogen reaction by the action of a complement; 
he speaks of the formed poison as anaphylatoxin. This hypothesis 
although very attractive does not entirely meet with the facts as they 
have been developed since Friedberger’s first work. The premises 
on which it is based assume in the first place that the poison or “ana- 
phylatoxin” is formed out of the matrix of the antigen; further, it 
is definitely assumed that in the production of the poison after the 
antigen and antibody have met, the complement or alexin, plays an 
active part. Friedberger’s hypothesis as stated by him, moreover, 
assumed that the entire process took place intravascularly. How- 
ever, even his own early experiments aroused some misgivings con- 
cerning the matrix of the poisons produced, for he found that they 
could be obtained as well when boiled antigen was used as when the 
fresh, unheated substances were employed, and the poisons were 
easily obtained from such organisms as the tubercle bacillus, which 
is extremely insoluble and unamenable to serum influence. It was 
also doubted whether one could truly assume the participation of the 
specific antibody or sensitizer in the production of Friedberger’s 
poisons, since it soon developed that from bacteria, at least, the poison 
could be produced when the organisms were directly exposed to the 
action of fresh guinea-pig serum without the presence of any immune 
serum. 
Experiments which soon cast a definite doubt on the assumption 
that the poisons were produced by a decomposition of the antigen 
were reported by Keysser and Wassermann.64 These workers sub- 
stituted insoluble substances like barium sulphate and kaolin for the 
antigen; that is, the precipitates or bacteria used in Friedberger’s 
experiments. They found that if kaolin were treated with horse 
serum and then exposed to the action of guinea-pig serum or comple- 
ment, poisons were produced identical in every respect to those pro- 
duced by Friedberger’s method. The conclusions they drew were 
that the poisons were produced, not by action of the complement on 
the antigen, but by its action on the horse serum absorbed by the 
kaolin. In other words, they transferred the matrix of the poison 
from the antigen to constituents in the serum itself, possibly the 
sensitizer or amboceptor. Bordet 65 also was able to show that poisons 
similar to those of Friedberger could be produced by the action of 
fresh guinea-pig serum on agar and recently Bordet has further 
shown that this is the case even when the agar has been by special 
methods deprived entirely of its nitrogenous components and repre- 
sents simply a complex of carbohydrates. Agar-guinea-pig serum 
64 Keysser and Wassermann. Folia serol., 1911, vii; Ztschr. f. Hyg. u. 
Jnfectionskrankh., 1911, lxviii. 
65 Bordet, Compt. rend. Soc. de biol,, 1913, Ixxiv, p, 877, 466 
INFECTION AND RESISTANCE 
mixtures of this kind show an increase in total nonprotein nitrogen 
which would prove that the proteolytic action of the guinea-pig serum 
must have been active against its own proteins. 
An interesting further development of this work has recently appeared 
in the experiments of Jobling and Petersen.66 They showed that when bac- 
teria are mixed with fresh active serum they adsorb the antienzymes normally 
present in blood. They have shown this experimentally and have proved that 
similar antienzyme removal can be accomplished by the addition of kaolin 
or agar, and by treatment with chloroform. Serums so treated become toxic, 
the actions of the poisons formed showing great similarity to that produced 
by Friedberger’s anaphylatoxins. According to them, the poisons are formed 
because of the fact that antienzymes are adsorbed by the antigen, thus setting 
the normal ferments in the fresh serum free to act on their own serum protein. 
It should be recalled that Friedemann, who was really the first one to 
show that the toxic substances could result from the interaction of fresh 
serum and sensitized antigens, although he succeeded only in doing this with 
red blood cells, suggested rather early that the success of such an experiment 
does not necessarily mean that the antigen furnishes the matrix entirely. 
He had studied the metabolism in anaphylactic poisoning and with Isaac has 
shown that the nitrogen output following reinjection in a sensitized animal 
is far in excess of that which could be derived solely from the injected anti- 
gen, and in this he has been confirmed by many other workers, notably by 
Vaughan. Moreover, recent investigations of Jobling and others have seemed 
to show that proteolysis is not necessarily a function of antibodies but is 
accomplished by non-specific protease in the blood. 
Among the most elaborate studies of anaphylatoxin formation are those of 
Novy and de Kruif.67 These workers produced toxic substances of this 
nature by incubating guinea pig serum with trypanosomes and with agar, 
and found that with this serum, as well as with rabbit and rat serum, poison 
production took place with great speed. In working with agar they found 
that the physical condition of the agar had important influence upon the 
speed of the reaction. When the agar was very finely divided in the gel 
state, toxic products were often obtained in less than five minutes. . The 
amount of agar necessary was so slight that 0.0025 mg. of dried agar was 
enough to toxify a cubic centimeter of serum. This confirmed many of the 
observations cited above which completely excluded the production of the 
toxic substance from the substrate. In other words, these investigations, as 
well as other similar ones, showed that in all likelihood substances, such as 
bacteria, agar, etc., added to sera, acted indirectly in bringing about processes 
in the serum rather than in furnishing a material upon which the serum 
enzymes could act. The observations of Novy and de Kruif were extended 
to the well-known phenomenon first, we believe, noted by Doerr and Moldovan, 
that in the course of interrupted coagulation, blood may become primarily 
toxic. Blood apparently becomes toxic in the stages just preceding coagula- 
tion. According to Novy and de Kruif when the blood is drawn, catalytic 
substances form which change the fibrinogen into fibrin and produce anaphy- 
latoxin-like substances in the process. These observations are of consider- 
able importance in invalidating such experiments as those of Friedberger, 
Thiele and Embleton 68 and others, who tried to prove the existence of toxic 
substances in the circulation of animals, in various connections, by producing 
66 Jobling and Petersen. Jour. Exp. Med., 1914, xix, No. 5. 
67 Novy and de Kruif. Jour. Inf. Dis., 1917, 20, p. 499. 
68 Thiele and Embleton. Zeitschr. f. Immunitats., 19, 1913, pp. 643 and 
666. ANAPHYLAXIS 
467 
toxic symptoms when they injected fresh blood from one animal into another. 
Unless this is done with great speed, poisons may result even when the blood 
of normal animals is injected. According to Novy and de Kruif (we quote 
from their summary in the American Medical Association Journal for May 
26th, 1917), the two reactions, both coagulation and toxification, go on hand 
in hand with the production of insoluble fibrin on the one hand and soluble 
anaphylatoxin on the other.69 The disturbances which can bring about such 
transformations may be readily produced by the addition of almost any alien 
substance to sera, such as bacteria, trypanosomes, pepton, agar, starch, kaolin, 
etc., and identical reactions may be produced in vivo by the injection of 
these totally unlike substances into the circulation of an animal. Novy and 
de Kruif explain anaphylaxis by the fact that such a disturbance occurs when 
antigen and antibody meet in the body. 
Along similar lines of reasoning are studies made upon the Abderhalden 
reaction by Bronfenbrenner70 and others who found that in this reaction 
when the fresh serum of pregnant or immunized animals is brought together 
with the corresponding boiled protein, poisons appear, but that these poisons 
originate from the serum as a result of auto-digestion, and not from the 
substratum. 
As a result of all the foregoing studies on the so-called anaphy- 
latoxic substances we are forced to the conclusion that a great many 
processes associated with the contact of blood plasms with foreign 
materials may give rise to toxic products. This occurs easily in vitro, 
and may probably occur in vivo as well, and the poisons thus pro- 
duced give rise to symptoms not unlike those of anaphylaxis. More- 
over, the production of such substances by contact of fresh serum with 
specific precipitates in the hands of Friedberger, would render it not 
unlikely that the meeting of an antigen and antibody in the circula- 
tion could bring about such poison formation; and the experiments 
of Novy and de Kruif seem to indicate that the minutest amount of 
antigen and antibody union might bring about the result. There- 
fore, we must conclude that anaphylatoxin investigations, in them- 
selves, are of the greatest importance in general pathological physiol- 
ogy, and that many phenomena in the body may perhaps be related to 
such poison formation. We would not, therefore, deny for one mo- 
ment, the scientific importance of anaphylatoxin observations from 
which we may perhaps expect not a little sound scientific progress. 
On the other hand, there is much which opposes the explanation 
of anaphylactic phenomena on the basis of anaphylatoxin formation. 
First and foremost among such objections is the fact of the cellular 
site of the reaction discussed in the preceding section. All workers 
who have inclined to an anaphylatoxin theory, have necessarily as- 
sumed the site of the reaction to be in the circulation. There is still 
a possibility, of course, that the union of antigen and antibody in or 
upon the cell many produce substances analogous to anaphylatoxins, 
either from the action of the plasma in close contact with the cells 
69 Conf. with Theory of Nolf. 
70 Bronfenbrenner. Jour. Exper. Med., 21, 1915, p. 480. 468 
INFECTION AND RESISTANCE 
or in the cell protoplasma, itself, but this is so purely speculative at 
the present time that it cannot alter the probability that true protein 
anaphylaxis cannot in any definite way, at the present time, be cor- 
related with the formation of anaphylatoxins. 
The extensive work which has been done on the formation of 
anaphylatoxin from bacteria, and the theories regarding the relation- 
ship of the poisons so produced to the explanation of the phenomena 
of infectious diseases will be dealt with in the chapter on hyper- 
susceptibility in bacterial infection. CHAPTER XVIII 
SERUM SICKNESS, FOOD IDIOSYNCRASY, HAY FEVER, 
DRUG IDIOSYNCRASY 
Serum Sickness.—We have mentioned that Ilosenau and An- 
derson attacked the problem of hypersusceptibility primarily in the 
hope of casting light upon the nature and cause of the distressing 
symptoms which in human beings often ensue upon the injection of 
diphtheria antitoxin. It has been one of the staple objections of lay 
opponents to the use of antitoxins that the injections are apt to cause 
severe illness and occasionally death, and indeed a few cases are on 
record in which sudden death has followed the first injection of 
diphtheria antitoxin. Since it was known by accumulated clinical 
experience as well as by experiments like those of Bertin,1 of Johan- 
nesen,2 and others that the harmful effects were not dependent upon 
the antitoxin contents, hut could he produced by injections of normal 
horse serum, it was but natural to bring these ill effects into analogy 
with the phenomena of hypersusceptibility. A large number of refer- 
ences to such antitoxin illness or serum sickness have appeared in 
the literature since the first beginnings of the therapeutic use of sera, 
yet no careful analysis of the condition was made until von Pirquet 
and Schick,3 in 1905, published their studies. 
As a rule the results of serum injection have been mild and with- 
out danger, though sufficiently frequent and troublesome to call for 
thorough study and attempts to discover the prophylactic measures. 
As stated above, a few cases are on record in which sudden death 
has followed a single first injection. There are no reports in the 
literature known to us, however, of fatalities after second injections, 
although not infrequently such cases have taken on alarmingly seri- 
ous aspects. 
The percentage of incidence and the variety of symptoms have 
been the subjects of many reports. The most frequent and striking 
single occurrence has been an urticarial rash. Rolleston,4 in a large 
series of cases, found urticaria in all but 17 of 289 cases of serum 
eruptions occurring between the first and tenth days after injection, 
and in all but ten of ninety-four later eruptions. 
1 Bertin. Gaz. Med. de Nantes, 1895. Quoted from Levaditi. 
2 Johannesen. Deutsche med. Woch., No. 51, 1895. 
3 Yon Pirquet u. Schick. “Die Serum Krankheit,” Deuticke, Leipzig, 
1905. Also Munch, med. Woch., 53, p. 67, 1906. 
4 Rolleston. The Practitioner, Yol. 74, 1905. 
469 470 
INFECTION AND RESISTANCE 
Bashes occurred in from 69.4 to 81.9 per cent, of the 600 anti- 
toxin cases which Kolleston reports. 
Joint pains commonly accompany the appearance of the rash, 
and these joint pains may offer considerable diagnostic difficulties. 
They are not often severe in actual discomfort, but may lead to a 
stiffness and difficulty in motion in cases in which tetanus antitoxin 
has been administered, giving rise to a considerable amount of worry. 
The writer has seen two cases in which the joints of the jaws, of the 
back of the neck and of the legs were so affected that slight rigidity 
and in one instance the difficulty of easily opening the mouth caused 
a good deal of uncertainty as to whether serum sickness or early 
tetanus were involved. 
Albuminuria is occasionally present, and with it oliguria. Fever 
may accompany the appearance of the rash, but is rarely very severe. 
(Edema around the point of injection supervening upon a wheal, was 
noticed not uncommonly by Rolleston in cases of second injection of 
horse serum, and in such cases he occasionally saw vomiting, rigours 
and collapse within a few hours after injection, these severe symptoms 
being more apt to follow upon large than upon small doses. The 
lymph nodes in the area of injection may be enlarged and tender, 
and even glands not directly connected with the area of injection 
may enlarge. This symptom is quite regular and early. 
While we have not desired to spend much time on the clinical 
description of this condition, we have sufficiently characterized it to 
show that there is a very real and definite sequence of events which 
may occur in human beings upon the injection of horse serum or 
other antisera.5 
Von Pirquet and Schick have studied the condition with careful 
reference to a comparison between the symptoms occurring in sub- 
jects after a first injection of serum and those following upon re- 
peated treatments. Their studies revealed the very important fact 
that the ill effects following a second injection were not only more 
severe than those occurring after the first injection, but developed 
after much shorter periods of incubation. In the ordinary “first 
injection” case the symptoms appear usually in from one to twelve 
days. After a second injection this incubation period may be con- 
siderably shortened and symptoms may appear in from five to seven 
days, the local and general reactions being much more marked than 
those subsequent to a first injection. Indeed, in some of the cases re- 
ported they may attain very alarming degrees of severity. This is 
the so-called accelerated (“beschleunigte”) reaction of von Pirquet 
and Schick, and is different from the “first injection” symptoms only 
5 The statement made by Kraus, from observations in Argentine that 
bovine serum does not produce serum sickness has been contradicted by Euro- 
pean workers. Doerr suggests the possibility of different dietetic habits 
to explain the discrepancy. Ref. See Doerr. “Ergebnisse der Imm., etc.,” 
Vol. 5, 1922, p. 96. SERUM SICKNESS, FOOD IDIOSYNCRASY 
471 
in its greater severity and speedier onset. In addition to this, how- 
ever, the “second injection’’ cases may show a train of immediate 
symptoms 6 (sofortige Reaktion), which occur within twenty-four 
hours after injection, and are characterized by marked local erythema 
and edema with often urticaria and constitutional disturbance. Both 
reactions may occur in the same individual, the “accelerated” reac- 
tion setting in as the “immediate” reaction subsides. 
Again, one reaction or the other may occur alone. The analogy 
between the immediate reaction and the anaphylaxis of animal ex- 
periment is obvious. The cases may be classified on the basis of 
these reactions, according to von Pirquet and Schick, the nature of 
the reaction being, within certain limits, determined by the interval 
ensuing between the first and the second injection. Thus, when the 
interval was twenty-one days or less the immediate reaction alone 
was noticed. When the interval was between two and six months 
both the immediate and accelerated reactions were present, and when 
the interval was still longer (seven months or more) the accelerated 
reaction alone was present. Isolated exceptions to this are noted in 
the series of sixty-one cases so reported. 
Currie,7 who has made similar studies, confirms the results of 
von Pirquet and Schick in all essentials, and agrees with their state- 
ment that the nature of the reaction is chiefly dependent upon the 
interval between injections. 
Percentage analyses of the various occurrences of serum sickness 
have been made by a number of writers, and can be found presented 
in compilations of the subject, like those of Doerr and of Coca re- 
ferred to many times above. Longcope,8 too, has analyzed this 
information with some thoroughness in his Harvey Lecture. Ap- 
parently, there are a number of different factors that determine the 
severity of serum sickness. According to studies of Weaver 9 culled 
from the literature of 801 reported cases, only about 10 per cent, 
developed the disease when quantities lower than 10 cubic centimeters 
were used, and in these cases the condition was usually a mild one. 
When 90 cubic centimeters, or over, were used, the disease was more 
severe, and over 75 per cent, of the cases developed it. Variations 
depending upon the source of the serum from different individual 
horses are too vague to be discussed, although a number of observers 
have reported such differences. The effect of the second injection 
varies, as we would expect from analogy with anaphylaxis, according 
to the interval elapsing between the first and second administration. 
6 Rankin in the Lancet, Dec., 1911, reports a case of “immediate” reac- 
tion 15 minutes after injection. 
7 Currie. Jour. Hyg., Vol. 7, 1907. See also Goodall, Jour. Hyg., Yol. 
7, 1907. 
8 Longcope. Harvey Lectures, Ser. 11, 1915-1916, Jour. Med. Sc., 152, 
1916, p. 625. 
9 Weaver. Arch. f. Int. Med., Yol. 3, 1909, p. 485. 472 
INFECTION AND RESISTANCE 
Just as in guinea pigs, if one reinjects within a short period after 
the first administration, say a week or less, a refractory condition is 
produced, according to Longcope; and the development of sensitive- 
ness is considerably prolonged. According to Goodall10 however, 
eventually sensitiveness supervenes and is then quite severe. It is 
well to keep these facts in mind for purposes of future discussion, 
since they constitute analogies to the condition of true protein ana- 
phylaxis in guinea pigs and other animals. If the second injection 
is made two weeks or later, after the first, the patients are likely to 
have the local immediate reaction which has been mentioned above, 
within less than an hour after the injection, and the general reac- 
tions may come on within less than a day, and then be accompanied, 
on occasion, by severe respiratory difficulties. In such cases the 
“Beschleunigte” reaction of von Pirquet and Schick is often noticed 
a few days after the injection. It is well again to note here the 
analogy between the immediate local reaction and the Arthus phe- 
nomenon observed in rabbits. 
The period after the first injection at wdiich the immediate reac- 
tions are most likely to occur are stated by, Longcope from his study 
of the literature and from his own observations to lie between the 
35th and 80th days after the primary inoculation, in which instance 
it occurs in 60 per cent, of the cases, if enough serum is injected. 
The accelerated or “Beschleunigte” reaction occurs in injections made 
three months after the first injection. From a practical point of view, 
it is important to consider just how serious the danger of death may 
he in such cases. Specific statistical studies have been made on this 
question by Park,11 and by Cuno,12 according to whom subcutaneous 
injection never causes death, though the symptoms may be extremely 
severe. But isolated cases of death have occurred after intravenous 
use. One such case was reported by Koch 13 in 1915. In view of 
the frequency with which serum is injected intraspinally in the 
treatment of meningitis and some other conditions at the present 
time, Longcope’s summary of results of this procedure is worth 
quoting. According to Auer who summarized these cases in 
Forscheimer’s System, and whom we quote from Longcope, 
death after second injection into the ' spinal canal has been 
reported by at least three observers.14 Auer regards these as results 
of a local immediate reaction in the meninges, analogous to the 
Arthus phenomenon. 
The analogy between serum sickness and anaphylaxis, is, there- 
10 Goodall. Jour. Hyg., 7, 1907, p. 607. 
11 Park. Trans. Ass. Amer. Phys., 28, 1913, p. 95. 
12 Cuno. Deut. med. Woch., 40, 1914, p. 1017. 
13 Koch. Wien. klin. Woch., 52, 1915, p. 685. 
14 Hutinel. Presse Med., 18, 1910, p. 497. Grysez and Dupuich, cited 
from Longcope, loc. cit. Archard and Flandin. Compt. rend. d. 1. Soc. Biol., 
73, 1912, p. 419. SERUM SICKNESS, FOOD IDIOSYNCRASY 
473 
fore, a pretty close one. It was early shown that the antitoxin con- 
tents of the sera had nothing to do with the occurrence of the serum 
sickness, but that this depended upon the foreign proteins rather than 
upon its antibody contents. Moreover, the shortening of the incuba- 
tion time upon second injection and the greater severity of the symp- 
toms following second and third injections, makes analogy with the 
protein sensitization of guinea pigs a very persuasive one. If serum 
sickness is truly an anaphylactic phenomenon, however, it is still by 
no means clear why symptoms should at all ensue after the first in- 
jection. The views of von Pirquet and Schick upon this were based 
upon the fact that in many cases the incubation period of serum sick- 
ness after first injection is a long one, namely, 8 to 12 days. Their 
idea was that, as the antigen is gradually brought into contact with 
the cells of the body, antibody begins to form, but that antibody 
formation precedes considerably the removal of all the antigen from 
the body. That this was a possibility was indicated by the studies 
wre have many times referred to of von Dungern and others, that 
antigen and antibodies may be present in the circulation of an animal 
at one and the same time, especially after the injections of consider- 
able quantities. The newly formed antibodies might, then, react 
with the antigen residue still present. In supporting this idea, they 
investigated the formation of precipitins in the blood of a rabbit in- 
jected with horse serum, finding that precipitins occurred as early 
as the eighth or ninth day. Wells 15 investigated this condition in 
human beings in which he found that the incubation of precipitin 
production varied, but was not dependent particularly upon the 
amount of antigen injected; nor was there any regularity in the incu- 
bation period of the serum sickness, which was very variable and 
sometimes not longer than four days. The matter has been exten- 
sively studied of recent years by Longeope and Rackemann 16 and 
by Longeope, Mackenzie and Leak.17 Longeope and Rackemann 
concluded from their experiments that the blood serum of most pa- 
tients who suffer from an attack of serum disease following injection 
of horse serum, show precipitin for horse serum, and that such serum 
could passively transfer the anaphylactic state to guinea pigs, thus 
bearing out the work of Hamburger and Moro,18 and of Wells. They 
found, also, that these antibodies could not be demonstrated in the 
blood serum of patients treated with horse serum who did not later 
present symptoms of serum sickness, and that the appearance of the 
antibodies shortly preceded recovery from the disease. The continua- 
tion of this work by Longeope and his collaborators on 19 further 
cases led to the conclusion that there is a definite relationship be- 
15 Wells. Jour. Inf. Dis., 16, 1915, p. 63. 
16 Longeope and Rackemann. Jour. Exp. Med., 27, 1918, p. 341. 
17 Longeope, Mackenzie and Leak. 
18 Hamburger and Moro. Wien. klin. Woch., 16, 1903, p. 445. 474 
INFECTION AND RESISTANCE 
tween antibody formation and the severity of the symptoms and 
that severe reactions come chiefly and “perhaps exclusively” in those 
individuals who develop a high concentration of precipitin in the 
circulation. They found that in such patients the horse serum is apt 
to disappear from the circulation soon after a large amount of pre- 
cipitin appears. They differentiate between individual patients 
by stating that some individuals seem to be relatively insusceptible 
to serum reaction, and that these are the cases in which little or no 
precipitin is formed. The tables given by the last named group of 
observers show a number of cases in which a failure of the develop- 
ment of serum disease was accompanied by a failure of antibody 
production, and also a considerable number of cases in which a num- 
ber of symptoms attributable to serum disease preceded the observa- 
tion of any antibodies in the blood. Indeed, in one case symptoms 
seemed to appear 24 hours after the injection, in several cases after 2 
days, and in a number of cases between the fourth and seventh day, 
which would be in keeping with Wells’ observation of a frequent in- 
cubation time as short as 4 days, the height of the serum disease 
usually, however, coming on later than this. 
We can, thus, state without going into extensive tabulation of 
statistics that in analogy with anaphylaxis, serum disease comes on 
after the injection of an antigenic substance; that when it conies on 
after first injection the incubation period may be, but is rarely less 
than 5 days, and is usually longer than this; that, as von Pirquet, 
Goodall,19 and Currie 20 have shown, second and subsequent injec- 
tions usually give rise to a shortening of the incubation period; 
and there is at least considerable reason to believe that antibody 
formation and serum disease are in some way related. 
As against these observations, Coca has entered objections, sepa- 
rating serum disease definitely from true anaphylaxis, and classifying 
it with idiosyncrasies, such as drug allergies. One of his arguments 
is the clinical parallelism between the symptoms of serum disease 
and drug allergies in regard to fever, skin rashes, the polymorphous 
form of these rashes, oedema, joint pains, leukopenia, etc. As far as 
this clinical argument is concerned, we are inclined to give it very 
little weight. As Doerr points out, there are at least as many simi- 
larities between human serum disease and anaphylaxis in animals as 
there are similarities between drug idiosyncrasies and human serum 
disease. He cites such occurrences as the secretions of tears, in- 
creased secretion of the nasal and gastrointestinal mucous membranes, 
and difficulties of respiration; and as far as the eruptions are con- 
cerned, he calls attention to the apparent discomfort of the skin and 
the scratching of guinea pigs in non-fatal anaphylactic shock. To 
this we may add that, after all, the symptoms of true anaphylaxis 
19 Goodall. Jour. Hyg., 7, 1907, p. 607. 
20 Currie. Jour. Hyg., 7, 1907, p. 35. SERUM SICKNESS, FOOD IDIOSYNCRASY 
475 
are so various in the manifestations as observed in different species 
of animals in which we can experimentally determine the anaphy- 
lactic mechanism, that it would be surprising if man reacted to the 
anaphylactic condition with symptoms entirely similar to those pre- 
vailing in lower animals. 
Coca criticizes the attempts to show parallelism between antibody 
reaction and the onset of serum disease by stating, in the first place, 
that the symptoms generally preceded the appearance of precipitin 
and sometimes disappeared while precipitin and horse serum were 
still present together in the patient’s blood, and in the second place 
because the symptoms sometimes continued after precipitin was no 
longer demonstrable, and occasionally occurred in patients in whom 
no precipitin production whatever could be shown. These are in- 
teresting suggestions, but do not seem to us convincing. In the first 
place, we must assume that antibodies are formed in or on the cells 
of the body, and that these cells may, therefore, be sensitive a con- 
siderable time before demonstrable precipitins occur in the circula- 
tion ; and, indeed, our own point of view about antibodies is such that 
we would expect a not inconsiderable amount of antibody actually 
to occur in the blood before a visible precipitin reaction could be 
elicited. Thus, our own view in which the identity of precipitins 
with other protein antibodies is a factor, would incline us to expect, 
as found by Longcope and his collaborators, that sensitiveness of the 
cellular complexes of the body would precede successful demonstra- 
tion of precipitins in the blood. That symptoms may continue after 
precipitins have disappeared, and that symptoms could occur in in- 
dividuals in whom no precipitins could be demonstrated, does not by 
any means signify that in such individuals there could not have been 
antibodies sessile on the cells capable in this site of reacting with the 
antigen. Indeed, the very arguments Coca uses in this case would 
seem to indicate that in this instance he favored an intravascular 
antigen-antibody reaction, in spite of the fact that his own work has 
been of considerable contributory value in locating the anaphylactic 
reaction upon the cells in lower animals. 
A type of reaction which Coca mentions in his argument, which 
cannot be explained by any such reasoning as that which we have 
employed above, is that in which symptoms appear almost imme- 
diately upon first injection. This, as he rightly remarks, seems to 
offer serious obstacles to the harmonizing of serum disease with ana- 
phylaxis. However, these cases are not very frequent, and while we 
do not wish to imply that we possess a positive explanation for this 
phenomenon, nevertheless, we believe it worth considering that it is 
not at all impossible that certain individuals may be accidentally 
sensitized to various proteins, either through the intestine or through 
some other mucous membranes as suggested by Richet, and by 
Rosenau. Again it is important to remember that patients sys- 476 
INFECTION AND RESISTANCE 
tematically sensitive to horse serum usually show immediate 
cutaneous reactions to this antigen (Longcope and Rackemann), and 
that the writer, with J. T. Parker, noted the development of similar 
skin sensitiveness in horse serum sensitized guinea pigs coincident 
with the development of the general anaphylactic state. Our own 
observations on monkeys, 1 think, have some bearing on this subject 
in that we found that it was extremely difficult to produce anaphy- 
laxis in several of the lower monkeys, but that when any symptoms 
at all were observed (these symptoms in one case being strikingly 
similar to human serum disease), they were associated with some 
precipitin formation. 
A line of reasoning in which it is not so easy to reply to Coca is 
the apparent difficulty of desensitizing human individuals. Doerr 
replies to this by stating that these difficulties must be interpreted 
with proper consideration of quantitative relations and the degrees 
of sensitiveness, factors which have not been sufficiently considered 
in desensitizing experiments on man. As a matter of fact, Doerr, 
who discusses Coca’s objection in this respect at considerable length, 
cites a number of cases in which desensitization has been success- 
fully reported by clinicians, both in conditions of horse serum sensi- 
tiveness, as well as in food idiosyncrasies. Schloss particularly has 
reported cases of this kind, and while the desensitization of human 
beings does not seem to be as simple as it is in animals, perhaps be- 
cause we cannot treat as drastically and have no measure of the 
degree of sensitiveness, there is sufficient evidence available to at least 
make Coca’s argument in this respect a questionable one. 
We are, therefore, inclined to believe that for the present, the 
parallelism between anaphylaxis and serum disease in man is suffi- 
ciently striking to incline us to favor the fundamental identity of 
both processes, admitting, however, that man seems to be an animal 
in which sensitiveness to the anaphylactic experiment is fortunately 
considerably less than in other animals, for reasons perhaps asso- 
ciated wTith peculiarities of antibody formation, or with the relative 
distribution of antibodies between cells and circulation. It is our 
opinion that this relationship has not been sufficiently considered in 
reasoning about anaphylaxis; for instance, the equilibrium estab- 
lished between cells and circulation in regard to the relative amounts 
of cellular and circulating antibodies must be a matter dependent 
both upon the species of animal and upon the temporary conditions 
prevailing respectively in the circulation and in the cells. As far as 
the influence of animal species is concerned, we know that while 
guinea pigs are the most delicate animals to anaphylactic reactions, 
thus arguing for a powerful concentration of antibodies in or upon 
the cells, they are much less favorable for the production of circulat- 
ing precipitins than are, for instance, rabbits. 
Desensitization in Serum Sickness.—Although the danger of SERUM SICKNESS, FOOD IDIOSYNCRASY 
477 
fatal accident in the subcutaneous administration of antitoxic sera 
is not a very great one, yet the increasing frequency of intravenous 
administration in, for instance, the use of pneumococcus serum and 
the isolated reports of cases of fatal result, render it desirable not 
only to gauge the sensitiveness of the patient, but also to apply some 
of the principles of desensitization, determined for animals, in the 
procedure. In view of the apparently reasonable parallelism between 
skin reactions and general hypersensitiveness, it has become a custom 
before giving such cases an intravenous dose of the serum, to apply 
an intradermal skin test with about 0.01 to 0.02 c. c. of the horse 
serum diluted 1-10 with salt solution, that is, a total amount of 
0.002 c. c. of the actual serum. A control with salt solution should 
be made to guard against errors from urticarial tendency in the 
patient. If a positive skin reaction, that is, the rapid development 
of an urticarial wheal over the injected spot appears which gradually 
fades in the course of one or two hours, great care in injecting should 
be used. Desensitization as advised by Cole and his collaborators 
can be attempted by starting injections of the patient with 0.025 c. c. 
of serum, and repeating this subcutaneously at one-half hour in- 
tervals, doubling the size of injection. If in this way 1 c. c. has been 
given without reaction, they continue with 0.1 c. c. at one-half hour 
intervals, intravenously. The process, of course, is based on the 
observations of Friedberger, Besredka and others, cited in a pre- 
ceding section, and made upon guinea pigs, and can be gauged in 
speed and amounts only by careful observation of the patient. In 
addition to this safeguard, the procedure first applied by Fried- 
berger should be used, namely, the dilution of the therapeutic serum 
to at least 50 per cent, of its original concentration with salt solution 
and the very slow intravenous injection by gravity, in such a way 
that the first 10 c. c. occupies at least ten minutes. 
Differentiation between True Anaphylaxis and Idiosyncrasies. 
—Before going on to the consideration of other forms of hyper- 
susceptibility, we insert this brief section for the purpose of making 
our opinions concerning the inter-relationship of the various forms 
of hypersusceptibility perfectly clear. In the preceding sections we 
have dealt with protein anaphylaxis and with serum sickness, a condi- 
tion which we believe to be more closely allied to protein anaphylaxis 
than to any other form of hypersensitiveness. We do not think that 
there is sufficient ground for excluding serum sickness from the pro- 
tein anaphylactic phenomenon, if all the conditions and analogies are 
properly taken into consideration. In the following sections, we shall 
take up conditions of hypersusceptibility to various food substances, 
the disease spoken of as hay fever, and the so-called drug idiosyn- 
crasies. It is impossible to state any specific criteria by which we 
can sharply separate all of these conditions from true anaphylaxis. 
To base this partly on clinical symptoms as Coca has done we do 478 
INFECTION AND RESISTANCE 
not believe is valid, nor can the entire class be separated from ana- 
phylaxis on the basis of the claim that the inciting substances are non- 
antigenic, for, in the food idiosyncrasies and even in hay fever, in 
our own opinion, the possibility of the participation of a true anti- 
genic substance is not only possible, but likely. We think the prob- 
lem may be justly stated in the following way: In true protein 
anaphylaxis we have a well defined group of phenomena in which the 
inciting substance is always an antigen; in which it can be proved 
that the reaction depends upon the sudden union of antigen and anti- 
body within the animal body; which can be easily and always elicited 
by artificial active sensitization; and can be passively transferred 
with the serum containing the antibodies. The site of the reaction 
can be proved to be predominantly cellular, and specific desensitiza- 
tion is easily possible. Of great importance, moreover, is the fact 
that true anaphylaxis, as far as we know, is not inheritable, except 
in so far as placental transmission of antibodies from mother to off- 
spring may occur, thereby conferring a temporary passive sensitiza- 
tion ; but true bioplastic inheritance in true anaphylaxis is not known. 
In contrast to this, let us characterize the so-called idiosyncrasies: 
In these conditions the inciting substance may or may not be a recog- 
nizable antigen. In many of the food idiosyncrasies the antigenic 
nature of the inciting substance is unquestionable; in hay fever there 
is an antigen in the pollen, but we are not ready to state that this 
antigen is the responsible ingredient; in drug idiosyncrasies, the 
original inciting substance, as such, is non-antigenic. Most impor- 
tant, moreover, is the fact that for these so-called idiosyncratic condi- 
tions, a definite possibility of inheritance has been established, 
although again let us emphasize that inheritance is not a factor in 
all cases. These various differences have long been recognized by a 
considerable number of observers. Doerr has again and again called 
attention to many of them, and we, ourselves, have discussed some 
of them, particularly in relationship to bacterial hypersusceptibility, 
but they have been brought together most clearly of recent years by 
Coca, who bases much of his arguments upon the important studies 
of Cooke and Vander Veer. 
We do not ourselves, however, believe that in these phenomena 
we can establish a sharp line of separation between protein anaphy- 
laxis and the idiosyncrasies, but that the conditions shade into each 
other by many gradual transitions, dependent upon a number of 
factors, some of which will be discussed in our concluding section. 
In this discussion we have omitted bacterial hypersensitiveness, 
a matter which can be most easily discussed in a separate section. 
Food Idiosyncrasies.—Hypersensitiveness to the ingestion of 
various kinds of meat, of milk, of eggs, sea food, strawberries, etc., 
has again and again been reported. In this general category, also, we 
may include hay fever, and, with it, such conditions as horse asthma SERUM SICKNESS, FOOD IDIOSYNCRASY 
479 
and the hypersusceptibility observed in many individuals to the 
emanations from various animals with which they come in contact. 
In such patients even temporary and brief contact with the re- 
sponsible animal may call forth reactions of extreme violence, clini- 
cally quite similar to the severe seizures of hay fever. The inclusion 
of horse asthma and allied conditions with food idiosyncrasies and 
hay fever, instead of with serum sickness, though in both cases an 
antigen derived from an animal is responsible, is based very largely 
upon the similarity of clinical pictures and upon the extremely 
minute dosage which in these conditions can call forth the most vio- 
lent reactions. Also, as stated in the preceding section, these condi- 
tions, like food idiosyncrasies and hay fever, seem in many cases to 
be governed by inherited predispositions. We believe it quite well 
worth noting in addition to this, that food idiosyncrasies and such 
conditions as hay fever and horse asthma, have another common 
feature which may be more important than hitherto supposed, namely, 
they are all conditions in which the sensitization seems to take place 
by the path of a mucous membrane, the intestinal canal in the one 
group, the upper respiratory passages in the other. In the former 
the sensitization seems entirely an intestinal one, and the symptoms 
usually, if not always, present some form of gastro-intestinal seizure. 
In the latter, hay fever, horse asthma, and allied conditions, the 
sensitization is always, as far as we know, one involving the upper 
respiratory tract, and while systemic symptoms are to some extent 
noticed in the attacks, nevertheless, the disturbances of the mucous 
membranes of the nose, throat and bronchial passages predominate, 
as though a condition of local sensitization played an important role 
in the mechanism. It is the problems of local sensitization about 
which we know least in the analysis of hypersusceptibility. 
The frequency of such forms of hypersensitiveness is much 
greater than formerly suspected, for there are probably cases of 
latent idiosyncrasy in which accidental freedom from the harmful 
contact has prevented recognition of the condition and which, there- 
fore, have been revealed only by the experimental testing of human 
beings by the skin and conjunctival reactions. 
The relative frequency of the various kinds of sensitization have 
been studied by a number of observers. Hay fever will be discussed 
in a separate section. In regard to other forms, however, it is inter- 
esting to note that Cooke and Vander Yeer 22 in a study of 318 cases, 
found 37 sensitive to horse, 35 to strawberries, 28 to shellfish, and 
27 to fish. 
Schloss23 has particularly studied the hypersensitiveness oc- 
curring in children to eggs, cows’ milk and other forms of protein; 
and similar studies have been made by many other observers. In 
22 Cooke and Yander Veer. Jour. Immunol1, 1916, p. 201. 
23 Schloss. Amer. Jour. Dis. Child., 3, 1912, p. 341. 480 
INFECTION AND RESISTANCE 
such individuals a variety of apparently specific reactions may occur. 
Schloss24 analyzes his cases into a number of different clinical 
groups. The most severe are those in which the food is followed by 
so rapid and violent a reaction that the child does not swallow it. 
Immediate discomfort leads to spitting and vomiting within a few 
minutes, aaid this is followed by swelling of the lips, tongue and 
mouth, with which severe general symptoms develop, sometimes suffi- 
ciently serious to be termed collapse. During recovery from this, 
general eruptions of an urticarial type may develop. In other cases, 
again, conjunctival congestion and attacks of sneezing not unlike hay 
fever, may occur, and in the more severe ones of this type, typical 
attacks of bronchial asthma supervene. Urticarial eruptions may 
occur alone as the only evidence of the hypersusceptibility, or may 
accompany any other form of the disease. Agioneurotic oedema oc- 
casionally results, either with or without urticaria. It is of impor- 
tance that urticaria is not the only form of skin eruption which is 
noticed, but Schloss has been able to trace various forms of erythema 
and even eczema to food intoxication. An important group studied 
by him, furthermore, is one in which the disturbances are chiefly 
gastro-intestinal, with vomiting and diarrhea, and the frequency 
with which these symptoms occur alone or together with other symp- 
toms, is important in view of the remarks we have made above con- 
cerning local sensitization. 
All this evidence gathered by Schloss, Brack 25 and many others 
show that a considerable variety of clinical disturbances may be 
specifically related to hypersensitiveness to many different foodstuffs, 
and these studies have again brought out the important fact that 
general sensitiveness is associated with cutaneous hypersusceptibility 
which can be determined by the intracutaneous injection of minute 
amounts of the inciting substances. 
Another fact of considerable importance is that a patient may be 
susceptible to a number of different foodstuffs at the same time. One 
of Schloss’s cases was, at one and the same time, susceptible to eggs, 
almonds and oatmeal, and he and others have mentioned similar cases 
in which a multiplicity of proteins would produce symptoms in one 
patient. Cooke and Vander Veer’s study of 551 cases, revealed 318 
which were sensitive to one protein only, 172 which were sensitive 
to two, 42 which were sensitive to three varieties, and 19 to more 
than 3 inciting substances; 42.3 per cent, of their cases, therefore, 
were sensitive to more than one substance. 
How may such sensitiveness be explained ? From the careful 
studies of many clinicians it appears that two classes of patients 
occur. In a definite proportion of these there is reasonable evidence 
for complete analogy with the anaphylactic experiment, in that sensi- 
24 Schloss. Amer. Jour. Dis. Child., 19, 1920, p. 433. 
15 Bruck. Arch. f. Dermat. & Syph., 96, 1902, p. 241. SERUM SICKNESS, FOOD IDIOSYNCRASY 
481 
tization is developed a certain length of time after the first contact 
with the responsible foodstuff, the first ingestion causing no symp- 
toms whatever. Such cases again raise the problem of the possibility 
of sensitization by way of the intestinal canal. Ordinarily, of course, 
we know that sensitization is not easily accomplished unless the anti- 
genic agent is administered to the body by some path which avoids 
digestion in the intestine and under ordinary circumstances we know 
that proteins, as such, will not pass through the intestinal mucosa. 
However, Rosenau and Anderson 26 have shown that sensitization 
of this kind can be accomplished in guinea pigs, and Wells 27 has 
made similar observations. Schloss 28 accomplished the same thing 
by feeding guinea pigs for from 10 to 32 days with egg-white. He 
not only found that in some of his guinea pigs, animals so fed and 
subsequently tested by intraperitoneal injection with egg-white be- 
came definitely ill, hut, in other experiments, also determined that 
intraperitoneally sensitized guinea pigs might show many of the 
symptoms of anaphylaxis if fed considerable amounts of egg-white 
13 days after the sensitization, following a short period of starvation 
to empty the bowel. In these animals it is interesting to notice that 
one of the immediate symptoms was the development of watery stools. 
The experiments done by Schloss on the basis of a large experience 
and great experimental care, would seem to permit no question con- 
cerning the fact that guinea pigs can be sensitized by feeding, and 
that sensitized guinea pigs can be intoxicated by feeding. That 
similar conditions may occur in human beings seems more than likely, 
in view of the fact that the intestinal history of all human beings 
from the earliest age, will show a great many periods of abnormal 
intestinal function due to inflammatory conditions of various kinds, 
coincident or perhaps sometimes secondary to over-feeding in which 
proteins are copiously ingested. In regard to these cases, then, 
analogy with protein anaphylaxis can be taken to be a close one, and 
the burden of proof would rest definitely upon those who doubted 
such a classification. 
On the other hand, there are the so-called Congenital cases in 
which there may be absolutely no traceable previous contact with the 
harmful foodstuff. This causes its symptoms upon first ingestion 
and, in many of these cases, a definite family history of hypersensitive 
conditions in the immediate ancestors of the patient can be traced. 
This class of so-called congenital cases is not at all infrequent. In 
1912 Schloss reported an interesting case of a boy of eight, in whom 
urticaria and other symptoms followed the ingestion of egg, almonds 
and oatmeal. The oatmeal sensitiveness appeared some time after a 
preceding feeding with oatmeal, the first ingestion having caused 
26 Rosenau and Anderson. U. S. Hyg. Lab. Bull., 29, 1906. 
27 Wells. Jour. Inf. Bis., 8, 1911, p. 66. 
28 Sehloss. Loc. cit. 482 
INFECTION AND RESISTANCE 
no symptoms at all. In the case of the almonds and the egg, however, 
no relationship with previous contact could be determined. The child 
showed positive skin reactions to the three proteins, both in their 
native and purified forms, which was so intense that on some occasions 
positive reactions were obtained by mere contact of the protein with 
the unbroken skin. Schloss was able to transfer the egg sensitiveness, 
passively, to guinea pigs by injections of the patient’s serum. It is 
of great interest to note that careful feeding of ovomucoid desensi- 
tized the patient not only to egg, but, to some extent, to the other two 
substances involved. This is of great interest in indicating the bare 
possibility that an antigenic group reaction may, to some extent, ex- 
plain cases of this kind. This however may also be explained by 
experiments such as those of Dale on non-specific desensitization (see 
previous section on this subject). 
Such an explanation, however, cannot be applied to a great many 
other cases. It can hardly be questioned any longer that true in- 
heritance may play a role in food idiosyncrasies, hay fever, etc., and 
is an important characteristic of these conditions. The problem has 
been studied by a great many observers, but most extensively by 
Cooke and Vander Veer. Their work was carried out by a careful 
experimental plan with a clinical material so extensive that their 
conclusions deserve the most serious consideration. They studied 504 
cases of demonstrable hypersensitiveness. In 2G0 of these there was 
no history of hypersensitiveness in the immediate ancestors. In 205 
of them there was a positive history on the one side, and in 39 a 
positive history on both sides. Thus, 48.4 per cent, of the sensitive 
cases had a traceable history of inheritance, as against 14.5 per cent, 
of ancestral sensitiveness for normal individuals. This discrepancy 
in itself seems significant, and becomes more so if we consider the 
possibility, referred to above, of latent hypersusceptibilities which 
could not be revealed in any manner in the last few generations be- 
cause of the relatively recent introduction of the skin and conjunc- 
tival tests. 
According to them, the child is not born hypersusceptible, but 
develops this susceptibility some time after birth. In those with a 
double inheritance, the susceptibility developed, in a considerable 
proportion of the cases, within the first five years. With a single 
hereditary history, the most frequent age for the development of 
the condition was between ten and fifteen years, and in those with 
a negative history, twenty to twenty-five years was the most frequent 
age for the development of symptoms. From this they conclude that 
the inheritance is not that of a specific hypersensitiveness, but rather 
represents merely the transmission of a tendency to hvpersuscepti- 
bility. This, too, is indicated by the fact that the specific sensitive- 
ness of the child may be for a substance quite different from that 
afflicting the parent. This curious fact might incline one to argue SERUM SICKNESS, FOOD IDIOSYNCRASY 
483 
that such a supposed inherited tendency might mean nothing more 
or less than the variable reaction capacity for the production of anti- 
bodies noticeable in laboratory animals; for anyone who has ever 
extensively immunized rabbits or other animals with various anti- 
gens, has found very marked ditferences in antibody production in 
individual subjects, irrespective of physical condition, feeding, etc. 
However, Cooke and Vander Veer think they excluded this possi- 
bility by observing reactions to horse serum injections in antitoxin 
treated individuals, some of whom were naturally idiosyncratic to 
food or pollen, the others being normal. They found practically no 
difference between the two groups in regard to the development of 
skin tests to horse serum after antitoxin administration. Thus, it 
does not appear that idiosyncratic individuals are more easy to sensi- 
tize artificially than are normal ones. 
Similar results were obtained by Scliloss in a smaller series. 
Schloss summarizes his own studies in stating that he was able to 
obtain satisfactory family histories in 80 cases. In 40 there was a 
definite history of allergy in parents, brothers and sisters. In 7 
other cases no history of allergy in the parents could be obtained, but 
it was positive for the grandparents. In the remaining 33 cases, no 
family history of allergy could be obtained, but of these 16 had 
developed symptoms on the first ingestion of the responsible food, 
though the history had been negative. In 17 of the 80 cases no family 
history was obtained, and the characteristic symptoms did not appear 
until the food had been once or several times partaken of, previous 
to the development of symptoms without causing ill effects. In this 
group, therefore, it seems reasonably sure that the sensitiveness was 
acquired. Schloss’s figures give a relative estimate of the numerical 
relationship between the congenital and the acquired cases. 
That the inheritance follows Mendelian principles in general is 
also apparent from the researches of Cooke and Vander Veer, though 
the difficulty of obtaining complete data and of finding families with 
sufficiently large numbers of children, hampered them considerably. 
We have stated the arguments in favor of the acceptance of in- 
heritance in a group of so-called congenital cases, in food idiosyn- 
crasy, hay fever, etc., at considerable length because we wish to be 
just to all circumstances, and because such a view has been appar- 
ently accepted by investigators as experienced as Doerr, Coca and 
others. We still believe, however, that fundamentally such so-called 
congenital cases will probably be found to be very closely related 
to the acquired ones. We may, and experimental and clinical ob- 
servations force us to admit, that there is an inheritance factor in 
hypersusceptibility. On the other hand, it is by no means impossible 
in our opinion that such an inheritance factor may consist largely in 
the bioplastic inheritance merely of a capacity for being easily sen- 
sitized. Such individual variations in reaction capacity to antigens 484 
INFECTION AND RESISTANCE 
could be well assumed in human beings in whom inherited biological 
differences seem to be more common than in lower animals, such as 
guinea pigs. Such a simple inherited capacity for sensitization, 
therefore, would not only be possible, but likely; would then, of 
course, follow the general laws of inheritance, and would probably 
signify the inheritance of a capacity for antibody formation. Even 
in animals, as we have mentioned above, individually varied capacity 
for antibody formation is noticeable. Granted such an inherited 
predisposition to sensitization in general, it is still possible that every 
individual that actually becomes sensitive, does so only after pre- 
liminary contact with an antigen. Such a view would explain why 
the sensitiveness in the offspring is often for substances, different 
from those to which his ancestors were susceptible; and we have seen 
in analyzing the work of Cooke and Vander Yeer, that, in the large 
majority of cases, hypersnsceptibility does not develop until at 
least a year after birth. In such cases, granting the possibility of 
gradual sensitization through mucous membranes of the upper 
respiratory and intestinal canal, it would seem to us next to impossible 
to exclude antigenic contact; and in those cases in which sensitiveness 
develops very soon after birth, and actual personal contact can be 
excluded, prenatal sensitization through the mother, or a similar 
process during the period of lactation, is not entirely out of question 
in view of the work which has shown the possibility of the passage 
of proteins and antibodies through the placenta and perhaps through 
the milk. Let us not forget, moreover, that in the human species 
there is only one layer of cells intervening between the maternal and 
the fetal blood.29 
Of importance is the question of passive transfer of these condi- 
tions. From the experiments of Schloss cited above, we have seen 
that in guinea pigs sensitized by mouth with the various foodstuffs, 
passive transfer was possible, and in some of his cases, as well as 
those of others, the serum of food idiosyncratic individuals conveyed 
specific hypersusceptibility to guinea pigs. As Doerr points out, 
however, this is insufficient to prove the true anaphylactic nature of 
the condition since it does not indicate definitely that the antibodies 
to the protein are involved in the process in the individual himself. 
We believe that perhaps this is an exaggerated point of view, but, 
nevertheless, it is, strictly speaking, logical. The transfer of the 
idiosyncrasy from one human being passively to another (the crucial 
experiment) is naturally a process for which the available literature 
is small. Eamirez, however, has published such a case in which a 
large amount of the serum of a horse-sensitive man was transfused 
to another individual who became typically horse sensitive, and 
developed cutaneous hypersusceptibility. 
29 Grosser. “Eihaute und der Placenta,” Braumiiller, Wien u. Leipzig, 
1909. SERUM SICKNESS, FOOD IDIOSYNCRASY 
485 
Desensitization in these cases may be successfully carried out by 
a number of ways, according to pediatricians who have studied it. 
Schloss states that he has completely immunized three egg sensitive 
patients with injections of ovomucoid, starting with 0.0001 mg. of 
the ovomucoid in saline solution, and injecting it at intervals of 5 
days, doubling the injection with each dose. He found a parallelism 
between the acquisition of immunity and the fading of cutaneous 
reactions. In one of his patients therapeutic injection caused mod- 
erate general symptoms in addition to a local reaction. A consider- 
able number of his egg and milk cases he improved by gradually 
increased feedings of the responsible protein. The initial doses as 
given by him in capsule form were extremely small, namely, amounts 
of 2 to 5 mg. diluted in milk sugar or starch. Increases of dose were 
gauged by the results, and by the age of the patient. One infant, 
one year of age, was so sensitive to milk that swelling of the tongue 
and lips resulted from one drop of milk diluted with water. He 
started, in this case, with the administration of 1/20 of a drop three 
times a day. 
Thus, again analogous with anaphylaxis, it seems possible to de- 
sensitize specifically in these cases, but we must remember, as 
repeatedly stated by Schloss and other pediatricians, that once de- 
sensitized, the ingestion of the responsible food must be continued, 
otherwise, again in analogy with anaphylaxis, the sensitive condition 
will reappear. 
Before going on to hay fever, we may, therefore, summarize the 
so-called food idiosyncrasies by stating that here we have a series of 
conditions in which close analogy to anaphylaxis exists in the anti- 
genic nature of the inciting substance, and in the existence of well- 
defined cases in which the condition was an acquired one. Additional 
corroborative analogy exists in experiments like those of Schloss, in 
which guinea pigs were sensitized by mouth. Moreover, specific de- 
sensitization further brings these cases into line with protein anaphy- 
laxis. However, in these conditions the inheritance factor is a very 
distinct one, and separates them, in this respect at least, from true 
protein anaphylaxis; also the dependence of the condition upon 
specific antibodies has not been sufficiently demonstrated. 
Hay Fever and Asthma.—The peculiar conditions which are 
grouped together under the term of “hay fever” were first associated 
with the inhalation of pollen by Blackley 30 as early as 1873.31 Black- 
ley also recognized a peculiar urticarial skin reaction when the mate- 
rial was rubbed upon the scarified epidermis, and observed catarrhal 
reactions upon experimental instillation of pollen upon the mucous 
30 Blackley. Cited from Longcope, loc. cit. His original work in “Ex- 
perimental Researches on the Cause and Nature of Hay Fever,” London, 
1873. 
31 For historical literature see Cooke and Vander Veer, loc. cit. 486 
INFECTION AND RESISTANCE 
membrane of the nose, and on the conjunctive. The supposition of 
Blackley concerning the relationship of pollen grains to the condition, 
was very thoroughly confirmed by Dunbar 32 who regarded the con- 
stituent in the pollen responsible for the reactions as primarily toxic. 
This idea has been pretty well eliminated, however, by subsequent 
researches, and it is quite definitely established that the disease is 
due to a specific hypersensitiveness to some material in the pollen 
and analogous to other forms of idiosyncrasy. According to Scheppe- 
grell,33 it has been estimated that in most parts of the United States 
from 1 to 2 per cent, of the population suffer from the disease at 
some period of the year. A considerable variety of plant pollens 
may be responsible. There was at first considerable difficulty in 
identifying the various pollen weeds which caused the disease. This 
was due, according to the writer quoted above, to the fact that many 
pollens are wind-borne, and often a common weed like the golden 
rod, because of its profusion in the neighborhood of the patient, was 
held responsible, in spite of the difficulty with which its pollen grains 
are distributed, while the more commonly responsible rag-weed was 
neglected. Scheppegrell characterizes hay fever weeds particularly 
as those which are wind-pollinated, are very numerous, with incon- 
spicuous flowers, without bright color, or scent, in which pollen is 
formed in great quantities. All such weeds, he says, are suspicious 
from a hay fever point of view. The most common of these are the 
common rag-weed, and the giant rag-weed, in both of which the pollen 
is wind-borne, and is produced in such abundance that very slight 
winds will dislodge them and carry them in clouds to considerable 
distances. According to investigations cited by Scheppegrell, these 
weeds are responsible for probably 85 per cent, of all cases of autum- 
nal hay fever in the United States. Many grasses, however, can also 
give rise to hay fever. For a botanical survey of these, we refer the 
reader to Scheppegrell. 
People who are susceptible to hay fever are usually attacked in 
the autumn, or the summer at a time when the particular weed to 
which they are sensitive has reached the pollen stage. The disease 
usually appears gradually, with the general signs of a cold in the 
head, with itching and burning of the conjunctiva, and a watery 
catarrhal discharge from the nasal passages. In rare cases there may 
be fever and general signs of systemic malaise, fatigue and depres- 
sion. While autumnal cases are the most common, this is not by any 
means the only season of the year in which the disease may occur, 
but as one would expect from the variety of plants which may be 
responsible, there are cases of so-called “rose-cold” which may occur 
32 Dunbar. Deut. med. Woch., 29, 1903, p. 149. 
33 Scheppegrell. “Hay Fever, etc.,” U, S. Pub, Health. Rep., Rep. No. 
349, July, 1916, SERUM SICKNESS, FOOD IDIOSYNCRASY 
487 
in the summer, and other cases are characterized by onset in the 
spring of the year. 
In view of the minute dosage of responsible pollen material which 
causes violent symptoms in the susceptible individual, as compared 
with the apparently complete harmlessness of considerable quantities 
of the same material for normal individuals, it was quite natural 
that, as Wolf-Eisner 34 first suggested, that anaphylaxis-like hyper- 
sensitiveness should be considered as a possible mechanism in expla- 
nation of the condition. In consequence of this, many observers tried 
to produce artificial sensitization and antibodies in animals by in- 
jection of various pollen preparations. This has been attempted, 
especially by Cooke, Flood and Coca who failed entirely in attempts 
to produce precipitating and complement fixing antibodies by the 
injection of pollen extracts into rabbits, nor could they induce ana- 
phylactic conditions in guinea pigs by sensitization with the pollen 
extracts. This failure to show antigenic properties is taken by Cooke 
and his associates, and by Coca, as one of the arguments on which 
they definitely rule hay fever out of the category of anaphylactic 
phenomena. As far as this particular point is concerned, experiments 
in our own laboratory, undertaken by J. T. Parker 35 because of 
this claim that pollen was not antigenic, definitely contradict the 
observations of Cooke and others. Because of their failures, the 
more delicate method of testing for anaphylactic sensitization by 
means of the Dale method was undertaken. Guinea pigs were sensi- 
tized by injection preparations of high rag-weed pollen extracted, 
with weak sodium hydrate and injected in twelve doses on successive 
days. The fluid, which gave Millon and Xanthroproteic tests, was 
given in large amounts to guinea pigs, a total of 70 c. c. being ad- 
ministered in the course of 7 weeks. This method was undertaken 
because in our work on sensitization with bacterial proteins, we had 
found it advantageous. For the same reason, injections were given 
every day for several weeks. About a month later, the uteri were 
tested and found highly and specifically susceptible to the extracts. 
We were careful in the publication of this research from our labora- 
tory, to state that we did not believe that this proved the anaphy- 
lactic nature of hay fever, but the experiments were so clean-cut and 
unquestionable, that we have no hesitation in at least concluding 
from them that hay fever pollen does contain an antigenic substance. 
It is important, however, in this case, as we have found in our work 
on bacterial antigens which will be mentioned later, that sensitization 
of the guinea pigs follow the painstaking method we used, and that 
the production of extracts, if extracts are used, shall be made with 
originally alkalin solutions. 
As to other work which may possibly indicate the intervention 
34 Wolf-Eisner. “Das Heufieber,” Munich, 1906. 
35 Parker, J. T. Proc. Soc. Exper. Biol. & Med., 18, 1921, p. 237. 488 
INFECTION AND RESISTANCE 
of antibodies in the hay fever complex, there are isolated claims in 
which the negative results of Cook and his associates are contra- 
dicted. Clowes 36 reported the detection of precipitins and comple- 
ment fixing antibodies in the sera of patients, and Koessler 37 claims 
that he passively sensitized guinea pigs to rag-weed pollen with the 
blood of a hay fever patient. 
The fact that hay fever pollen does contain antigenic substances, 
together with the work cited above, does not, of course, prove the 
anaphylactic nature of hay fever which we, ourselves, consider un- 
certain ; but these facts should not be completely thrown out of court 
in the discussion of hay fever, as this is done by Coca. 
In hay fever, as in other conditions discussed, specific hyper- 
sensitiveness of the skin and of the conjunctiva run parallel with the 
degree of sensitiveness to the disease, and has been used by all ob- 
servers who have studied the disease as an index for the diagnosis of 
the specific pollen responsible, and as an index for treatment. Vander 
Veer and Cooke recommend the intradermal injection of 1/50 to 
1/100 of a cubic centimeter of a properly prepared pollen extract 
which in positive cases results within 5 to 20 minutes in a urticarial 
wheal, surrounded by a hyperemic zone from 1 to 3 cubic centimeters 
in diameter. This is usually accompanied by local itching. Their 
pollen extract consists of a simple extraction of the pollen with 0.8 
per cent, sodium chloride solution, and are standardized into three 
strengths, according to the amount of nitrogen per cubic centimeter. 
They use three solutions, one containing 0.01 mg. of nitrogen, an- 
other with 0.1 mg., and a third with 0.5 mg. per cubic centimeter. 
They first determine by intradermal test which pollen the indi- 
vidual is susceptible to. Then by the ophthalmic reaction, using the 
weakest solution first, they test the degree of sensitiveness. When 
a positive reaction is obtained with a solution containing 0.01 mg. 
per cubic centimeter, they give their first dose in an amount of 0.005 
mg. of nitrogen per cubic centimeter. For a discussion of the treat- 
ment and its results, we refer the reader to the excellent article of 
Cooke and Vander Veer. From the point of view of the theoretical 
conception of the condition, the important matter is that consistent 
treatment with the specific pollen does desensitize the patients. The 
results have been so favorable that some improvement may be ex- 
pected in 90 per cent, of the cases of real hay fever. The late hay 
fever cases, however, they state are much more difficult. As in other 
cases of desensitization, the effects are temporary, the desensitization, 
according to Freeman 38 rarely extending completely from one sea- 
son to the other, and hardly ever to the second season. 
30 Clowes. Proc. Soc. Exper. Biol. & Med., 10, 1913, p. 69. Cited from 
Cooke and Vander Veer. 
37 Koessler. “Forcheimer’s Therapeusis, etc.,” Vol. 5, p. 671. 
38 Freeman. Lancet, September, 1911; April, 1914. SERUM SICKNESS, FOOD IDIOSYNCRASY 
489 
While, therefore, we must again call attention to the important 
difference between true anaphylaxis and hay fever because of the 
probable role of inheritance, we can also emphasize the many points 
of definite analogy between the two conditions. There is an antigenic 
substance in the pollen, though it is not easy to demonstrate. Actual 
presence of antibodies in the blood of the patients, and the possi- 
bility of passively sensitizing with such blood, we regard as ques- 
tionable, and still calling for further investigation. Phenomena of 
desensitization are analogous to those prevailing in anaphylaxis. 
Drug Idiosyncrasies.—Changed reaction of the body to various 
drugs, alkaloids, glucosides, organic compounds, such as many hyp- 
notics, to salicylic acid, and even to such simple substances as iodin 
and the iodides, arsenic, bromides, mercury, etc., has been a curious 
phenomenon that has puzzled investigators for a good many years. 
Acquired resistance to drugs, so-called drug tolerance has been known 
for a long time, and every physician knows, especially in the case of 
the alkaloids, the enormous quantities of highly toxic substances that 
can be supported by individuals who habitually use them. The same 
is true even of such simple substances as arsenic. In a preceding 
section dealing with immunization, we have discussed this problem 
briefly, the brevity of our treatment of the subject being enforced, in 
spite of its importance, by the incompleteness of our knowledge. 
Since all of these materials are non-antigenic in themselves and, there- 
fore, develop no antibodies, since all efforts to demonstrate neutraliz- 
ing substances in the blood of the tolerant individuals have so far been 
unsuccessful, and since there is almost complete unanimity as to the 
impossibility of transferring drug tolerance from one individual to 
another, there are very few experimental data from which theory can 
take its departure. The only thing that seems definitely established 
is that drug tolerance is a function of the tissue cells, rather than of 
any mechanism comparable to antibody formation. Just as there 
exists a drug tolerance which is to some extent analogous to immunity 
to antigens, so is there a changed reaction capacity in the opposite 
direction, namely, a drug hypersusceptibility, a class of conditions 
grouped together under the term of drug idiosyncrasy. 
The earliest observations on drug idiosyncrasies were made 
largely by skin specialists who noticed peculiar forms of dermatitis 
and urticaria in connection with the administration of certain drugs. 
Jadasohn 39 was perhaps one of the first who systematically studied 
these phenomena, which he spoke of as “Arzneidermatosen,” and 
which he classified into two groups of cases, one group in which the 
individuals react abnormally to the very first contact with the drug, 
and others in which repeated administration is necessary to develop 
the idiosyncratic susceptibility. The supposedly “acquired” cases 
seem to be relatively rare and in most cases the idiosyncrasy becomes 
39 Jadasohn. Quoted from Sauerland, Berl. klin. Woch., 49, 1912, p. 629. 490 
INFECTION AND RESISTANCE 
apparent without preceding habitual use. This is perhaps one of 
the most important arguments against classification with anaphy- 
lactic phenomena. Attempts to induce drug hypersusceptibility arti- 
ficially in animals and in man have been generally unsuccessful 
as far as we can analyze the literature, although there are a certain 
number of cases in which salvarsan and quinin therapy have seemed 
to bring about hypersusceptibility. 
The skin eruptions are not the only manifestations of this condi- 
tion. Fever, systemic symptoms and a variety of other disturbances 
may characterize the attacks. It is important to note in this connec- 
tion that the symptoms of the hypersusceptibility are not those of 
an increased development of the physiological action of the drug, but 
•take forms quite easily distinguishable from the ordinary pharmaco- 
logical action. Furthermore, the specific clinical pictures for one 
drug may be consistent and different from that induced by other 
drugs. 
The problem is a complicated one, and calls for a considerable 
amount of investigation. The relationship of the sensitiveness to 
the chemical constitution of the drug is also a peculiar one. In some 
cases it is specific for the entire molecule, such as the idiosyncrasies 
occasionally noticed against some of the hypnotics. Bruck 40 reports 
a case of apparently acquired antipyrin susceptibility in a physician, 
who, after a considerable series of antipyrin administrations, began 
to suffer from aphthous eruptions of the tongue which always re- 
turned when he took antipyrin in the course of ten years. All anti- 
pyrin derivatives, pyramidon, etc., had the same effect. In other 
cases, such as iodoform hypersusceptibility, the hypersusceptibility 
seems to be specific for the methyl radical since, as Doerr states, such 
individuals react also to bromoform, etc. Doerr, however, also men- 
tions cases such as quinin idiosyncrasies in which closely related 
substances gave no reaction whatever. 
A great many investigations have been carried out with the pur- 
pose of bringing drug idiosyncrasies into close analogy with anaphy- 
laxis. Such are the investigations of Bruck, Klausner41 and 
others. Their experiments have consisted chiefly in attempting to 
transfer the sensitiveness of man, passively, to animals. Bruck in- 
jected guinea pigs with 5 c. c. of the serum of the antipyrin case 
spoken of above, and 24 hours later administered 0.3 grams of anti- 
pyrin subcutaneously. About three-quarters of an hour after the 
antipyrin injection, one of the guinea pigs that was supposedly 
passively sensitized, showed uneasiness, dyspnea and finally cramps 
and paralysis of the hind legs. It died 5 hours later. In criticizing 
this apparently clear experiment, it should be borne in mind that the 
amount of antipyrin administered was only 25 per cent, less than the 
40 Bruck. Berl. klin. Woch., 47, 1910, p. 517. 
41 Klausner. Munch, med. Woch., 27, 1910. SERUM SICKNESS, FOOD IDIOSYNCRASY 
491 
minimum toxic dose for a normal pig, and that of two pigs sensitized 
in exactly the same way, only one reacted. Cruveilhier 42 similarly 
claims to have produced antipyrin hypersusceptibility, both actively 
and passively, in guinea pigs. With these and a few other exceptions, 
all other attempts in a similar direction seem to have been negative. 
Auer,43 encouraged by the observation that repeated injections of 
salvarsan occasionally give rise to anaphylaxis-like reactions 44 in 
man, attempted to throw light upon this condition by animal experi- 
ments. He injected salvarsan, both in acid and alkalin solutions in 
concentrations of 0.1 per cent., into guinea pigs and rabbits. In none 
of them was he able to produce any degree of hypersusceptibility. 
It will be seen from the foregoing that our knowledge of the laws 
governing idiosyncrasies is defective and available evidence is con- 
tradictory. There is very little therefore on which reliable opinion 
can be based. The only line of research which has shown possible 
promise of eventual light upon the subject is that initiated by Ober- 
meyer and Pick,45 and recently further developed by Landsteiner.46 
Obermeyer and Pick, as we have seen in another part of this book, 
treated proteins with various chemical substances such as Lugol’s 
solution, and succeeded in so altering the serum protein that it 
changed in specificity without losing its antigenic properties. Thus, 
an iodized protein now acted as antigen with sera produced against 
many other iodized proteins, the insertion of the iodin atom in the 
serum apparently having changed its specific antigenic nature. Land- 
steiner has done exceedingly interesting work along similar lines in 
azotyzing various proteins by treating them with metanilic acid (an 
amino-sulphone-benzoic acid), and immunizing rabbits with the 
altered protein. The antibody produced by the rabbits now precipi- 
tated only azo-proteins, and not the native proteins from which the 
azo-protein had been produced. Apparently, the specificity of an 
antigen, therefore, may be considerably altered by coupling it with 
other chemical groups, the added group determining the specificity. 
Wolf-Eisner has suggested this as an explanation for many drug 
idiosyncrasies, assuming that the administered drug may in some 
way form a compound with the serum protein, permitting the final 
product to react within the body as a foreign antigen. Swift 47 
working along this line of reasoning, treated guinea pigs with sal- 
varsanized guinea pig serum. With this mixture he sensitized guinea 
pigs and obtained a few reactions which might be interpreted as truly 
anaphylactic in nature. He naturally expressed himself in a very 
conservative manner. 
42 Cruveilhier. Comp. rend. d. 1. Soc. Biol., 69, 1910. 
43 Auer. Jour. Exp. Med., 14, 1911, p. 497. 
44 Weischelman. Deut. med. Woch., 1912, p. 1174. 
45 Obermeyer and Pick. 
46 Landsteiner. See section on Antigens. 
47 Swift. Jour. A. M. A., 59, 1912, p. 1236. 492 
INFECTION AND RESISTANCE 
While this general direction of research is attractive, it cannot be 
said to have so far had very much effect upon our conception of drug 
allergies, largely because so many contradictions prevail in the litera- 
ture. We do not feel at the present time that any concise opinion 
concerning the nature of drug allergies or their relation to anaphy- 
laxis is warranted. All we know is that we have here both tolerance 
and hypersusceptibility, in many ways analogous to immunity, on the 
one hand, and anaphylaxis on the other; but since the inciting sub- 
stances are diffusible and can easily penetrate into the cells, no sys- 
tematic investigation of the mechanism has been possible. It may 
well be that the reasoning of Swift based on such unquestionably 
definite experiments as those of Landsteiner may explain certain 
types of acquired drug allergies. The remainder in which the process 
is probably intracellular, must await further investigation. 
It is worth noting, moreover, that in most instances in which 
passive transfer of a drug idiosyncrasy has been claimed, the drug 
employed has been one of those with which human beings occasionally 
acquired their hypersusceptibility as a consequence of habitual use. 
Thus, Bruck’s case was one of antipyrin idiosyncrasy in a physician 
who did not develop it until he had been taking the drug for a long 
time. With the serum of this patient he claims to have passively 
sensitized guinea pigs. Cruveilhier’s work was done with antipyrin, 
and Swift’s with salvarsan, and in the case of salvarsan, not only 
have we definite records of occasional cases in which severe reaction 
followed after several harmless preliminary injections, but we have 
a number of reasons to believe that the drug undergoes some form 
of combination with the serum. This is apparent from the work of 
Swift and Ellis on the use of salvarsanized serum in spinal cases, 
and also from the fact that the treponemacidal action of salvarsan 
is not a function of the pure solution, in which the treponemata do 
not seem to be particularly injured, but becomes evident only 
when serum is present. Thus, it is not at all out of question that 
drug idiosyncrasies are not all alike, and although most of them are 
so far inexplicable on the basis of pure analogy with anaphylaxis, 
there may be others which depend upon the mechanism of union be- 
tween the drug and the serum constituents, thereby producing an 
altered antigen within the body. And this possibility becomes par- 
ticularly important in view of the work of Landsteiner 48 in which he 
succeeded in altering rabbit sera by treatment with alcoholic solutions 
of sulphuric acid, so that an altered antigen resulted which would 
produce antibodies, even in the rabbits from which the original 
protein was obtained. The species specificity was destroyed, and a 
“sulphon” specificity produced which then acted on the same species 
from which the serum was derived. 
48 Landsteiner and Jablons. Zeitschr. f. Imm. u. Exp. Ther., Vol. 20, 
1914, p. 618. CHAPTER XIX 
BACTERIAL ANAPHYLAXIS. TUBERCULIN AND ALLIED 
REACTIONS. BACTERIAL ANAPHYLATOXIN THEO- 
RIES. TOXIN HYPERSUSCEPTIBILITY. TOXICITY 
OF NORMAL SERUM. GENERAL CORRELATION. 
In the case of most serum reactions the original observations 
were made upon the sera of bacteria-immune animals, and later ex- 
panded into generalizations applicable to antigens as a class. This 
was the case with the phenomena of lysis, agglutination, and precipi- 
tation. In the case of anaphylaxis the reverse was true. The fun- 
damental observations were made with non-bacterial antigens, but 
the thought that analogous observations could be made with bac- 
terial proteins was an obvious one, and since the problem was one of 
altered susceptibility there was great promise that investigation of 
this subject might prove of profound significance for our knowledge 
of the pathology of infectious diseases. 
Accordingly Rosenau and Anderson,1 in one of their earliest 
researches, carried out experiments upon the sensitizing properties 
of bacterial proteins. They were successful in sensitizing guinea 
pigs with extracts of colon, tubercle, anthrax, and typhoid bacilli, 
with Bacillus subtilis extracts, and with those of yeast. In most 
cases they used considerable quantities of bacterial extracts and 
obtained but slight or moderate symptoms. However, their results 
were conclusive in showing that the anaphylactic experiment could 
be carried out with bacterial proteins and was, in every detail, 
analogous to the similar phenomena of serum anaphylaxis. 
Not only could the basic experiment of active sensitization be 
carried out with these substances, but it was found that the reaction 
here, as in other cases, was specific, and that shock was followed by 
a period of “antianaphylaxis” or “immunity.” Rosenau and An- 
derson suggested that the incubation time of many infectious dis- 
eases may be represented by the period necessary for the development 
of susceptibility after a first injection, and that the crisis of pneu- 
monia might possibly find an explanation in the analogy with anaphy- 
laxis. 
The criteria governing the successful production of bacterial 
anaphylaxis were then studied especially by Ivraus and Doerr,2 Holo- 
1 Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab 
Bull. 36, 1907. 
2 Kraus and Doerr. Wien. Jclin. Woch., No. 28, 1908. 
493 494 
INFECTION AND RESISTANCE 
but,3 Delanoe4 and others, and the essential points of Rosenau and 
Anderson’s experiments were confirmed. Although Kraus and 
Doerr succeeded in frequently sensitizing guinea pigs with a single 
injection of bacteria, this was not found to be the most favorable 
method for sensitization. Braun 5 obtained entirely negative results 
by such a procedure, but this may well have been because in the first 
place single sensitization with bacteria is evidently irregular in 
result, and because Braun carried out his intravenous test-injection 
slowly, a technique by which Briedberger found later that shock 
could be avoided. Delanoe, in the main, confirmed the fact that 
bacterial sensitization was possible, but denied the specificity of the 
resulting anaphylaxis, in that he succeeded in producing shock in 
tubercle-sensitized guinea pigs with comparatively large amounts 
of typhoid, paratyphoid, and other bacilli, and conversely found 
typhoid-sensitjzed guinea pigs hypersusceptible to tubercle-injec- 
tions. Other workers, however, notably Kraus and Doerr, Holobut, 
and Kraus and Admiradzibi,0 agree that the reaction is specific, at 
least in the same limits within which other serum reactions may be 
called specific. 
Holobut then developed a technique of sensitization with bac- 
teria more reliable than any which had been previously employed by 
other workers. He found that the most regularly successful results 
were obtained when he injected small quantities of bacteria (1/100 
loopful) daily for ten days, subcutaneously, and tested with fairly 
large amounts (1-2 c. c. of an emulsion of the bacteria) intrave- 
nously about 3 weeks after the last sensitizing injection. This is in 
keeping with later experience, and in our own work with typhoid 
immunization in young goats we have found that anaphylactic 
reactions were not observed unless the goats had previously received 
several injections. A second injection never elicited symptoms. 
It is hardly worth while to recount all the various methods recom- 
mended for the successful sensitization of guinea pigs and other ani- 
mals with bacterial proteins. The general principle that it is not as 
easy to sensitize with bacterial substances as with proteins, like egg- 
white and the various animal sera, may be regarded as well estab- 
lished. In our own work with tubercle bacillus protein we attained 
a considerable regularity of results by injecting ten and twelve times 
intraperitoneally either daily or every other day and testing between 
two and three weeks after the last injection. Quantities used for 
injection were considerable, consisting of one or more cubic centi- 
meters of a suspension or extract of ground dead tubercle bacilli. 
3 Holobut. Zeitsclir. f. Immunitdtsf orsch., Yol. 3, 1909. 
4 Delanoe. C. R. de la Soc. de Biol., Yol. 66, 1909, pp. 207, 252, 348, 389. 
5 Braun. Quoted by Bail and Weil, Zeitschr. f. Immunitdtsf orsch., Vol. 
4, 1910. 
6 Kraus u. Admiradzibi. Zeitschr. f. Immunitdtsf orsch., Vol. 4, 1910. BACTERIAL ANAPHYLAXIS 
495 
Thus, while active sensitization is perfectly definite in the case 
of bacterial materials, care and attention to this principle is neces- 
sary in demonstrating it. The reasons for this are probably to be 
found in the constitution of the bacterial protein. The bacterial cell 
body contains a relatively small amount of coagulable albumin or 
globulin like substances, and it is, therefore, necessary to administer 
very considerable amounts of bacterial material in order to simulate 
anything like the quantities of coagulable protein obtained even in 
small amounts of serum. Moreover, it is not at all impossible that 
most of the bacterial substances ordinarily injected are not in solu- 
tion, and are actively phagocited with consequent loss of antigenic 
stimulation of the body cells. The major part of the bacterial cell 
bodies seems to consist of nucleo-protein-like materials, and a residue 
which will be described below, the chemical nature of which is not 
as yet quite certain. This residue is antigenic in the sense that it 
reacts with antibodies, but it seems to be either extremely difficult, 
or impossible to incite antibody formation in the animal body by 
injecting it. These substances which we have been studying for 
some time, probably represent bodies like the “Haptenes” foretold by 
Landsteiner. It seems plain, then, that the most likely explanation 
for the difficulties one encounters in sensitizing with bacterial pro- 
teins is due to the peculiar chemical constitution of the bacterial cell 
body. A small amount of Bence-Jones protein has been found in 
bacteria by us in our tuberculin studies, and Bence-Jones protein 
has been found by Bayne-Jones to have antigenic properties. The 
antigenic constitution of the bacterial cell, then, is a complex one, 
and this must be taken into consideration when drawing conclusions 
about anaphylaxis or other immunological reactions against bacteria. 
Passive sensitization against bacterial substances has also been 
obtained, in all respects analogous to the passive sensitization pro- 
duced in serum anaphylaxis. The earlier work on this was done bv 
Kraus and Doerr,7 and by Kraus and Admiridzibi,8 and their find- 
ings have been variously confirmed by Holobut,9 Yamanouchi,10 
Delanoe 11 and others. Some of these writers, especially Yamanouchi 
and Delanoe, suggested the utilization of passive anaphylaxis for 
the diagnosis of tuberculosis and typhoid fever. However, their 
experiments were entirely too indistinct to warrant such procedures, 
and even now that we know more about passive sensitization, the 
method would be of little practical usefulness. 
In order to eliminate any possible sources of misinterpretation, 
the writer and Parker 12 some years ago reinvestigated passive sen- 
7 Kraus and Doerr. Wien. klin. Woch., 21, 1918, p. 1008. 
8 Kraus and Admiridzibi. Zeitschr. f. Immunitats., Orig., 4, 1909, p. 607. 
9 Holobut. Zeitschr. f. Immunitats., 3, 1909, p. 639. 
10 Yamanouchi. Compt. Rend, d, l. Soc. Biol., 66, 1909, p. 531. 
11 Delanoe. Compt. Rend. d. 1. Soc. Biol., 66, 1909, pp. 207, 252, 348, 389. 
12 Zinsser and Parker. Jour. Exp. Med., 1917, p. 411. 496 
INFECTION AND RESISTANCE 
sitization of guinea pigs to bacteria by the isolated uterus method of 
Dale. Working with typhoid bacilli and young passively sensitized 
guinea pigs, it was found very easy to sensitize these animals with 
1 c. c. of anti-typhoid serum of an agglutinating titre of about 
1-10,000, the sensitiveness developing a little more slowly than in the 
experiments reported by Kraus and his collaborators, being most 
thoroughly developed after about 3 to 5 days. At this time typical 
shock could be elicited in some of the animals by intravenous injec- 
tion of a cubic centimeter of simple extracts of typhoid bacilli, and 
typical uterine contraction could always be obtained by the Dale 
method. 
These experiments also enabled us to investigate the extremely 
interesting question raised by Friedberger’s anaphylatoxin work, 
namely, whether typical anaphylactic reactions in guinea pigs could 
be produced by the intravascular production of toxic substances from 
the bacteria. It opened the problem of the mechanism by which 
organisms like the typhoid bacillus intoxicate an animal, and whether 
or not this has any relation to anaphylaxis. Our experiments showed 
the following points, to which we will refer later in connection with 
the discussion of Friedberger’s anaphylatoxin work. In the first 
place, it was found that the normal unsensitized guinea pig uterus 
showed no anaphylactic contractions when brought into contact with 
considerable amounts of typhoid extract. This would mean that the 
anaphylactic mechanism of injury does not enter into the acute death 
of unsensitized guinea pigs into wThich one injects large quantities of 
dead or living typhoid bacilli sufficient to kill them in a few hours. 
The mechanism of injury here must depend upon some toxic product 
probably formed intravascularly in which anaphvlaxis-like processes, 
and probably antibodies, play no role; it may be due to the liberation 
of endotoxin-like substances in the sense of Pfeiffer, or a toxic split 
produce in the sense of Friedberger’s interpretation. The uteri of 
sensitized guinea pigs, however, are delicately susceptible to similar 
extracts, showing that in animals or human beings in which typhoid 
bacilli have been present long enough to give rise to antibody produc- 
tion, the anaphylactic mechanism of injury may be superadded to 
the other. It is, furthermore, we think, of considerable interest that 
the sensitized uterus, though delicately susceptible to the extracts of 
typhoid bacillus, was hardly, or not at all affected by whole washed 
typhoid bacilli. In other words, in order to give rise to anaphy- 
lactic symptoms, the bacterial antigen must be in solution, and such 
processes can take place only in the infected animal body as there 
is a destruction of the morphological integrity of the bacterial cell in 
the circulation. Rapid phagocytosis would, therefore, protect against 
it. Moreover, these experiments demonstrated the essential cellular 
nature of bacterial anaphylaxis. 
It is, thus, plain that all the analogies with serum and protein BACTERIAL ANAPHYLAXIS 
497 
anaphylaxis have been fulfilled with bacterial materials, and that 
if this particular mechanism plays any part in infectious disease, 
it must be under conditions in which sessile antibody formation has 
taken place and the bacterial protein has gone into solution in some 
way in the body. 
There is, however another form of hypersensitiveness in the ani- 
mal body to bacterial products about which there has been a consider- 
able amount of discussion, and the relationship of which to true pro- 
tein anaphylaxis is not entirely clear. This is the type of hyper- 
sensitiveness exemplified by the tuberculin, typhoidin, etc., reactions. 
Tuberculin and Similar Reactions.—There is one class of phe- 
nomena, however, which calls for further discussion in this connec- 
tion, since its dependence upon anaphylaxis, while generally assumed, 
is still opposed by many authorities. This consists of the various 
reactions in which micro-organisms are injected, or brought into 
contact with the skin or conjunctiva of infected subjects. Such are 
the various forms of the tuberculin reaction, the typhoidin reaction of 
Chantemesse, the one of Gay, and the mallein and abortin reac- 
tions.13 In the tuberculin reaction the conditions have been thor- 
oughly studied, and we may make a detailed consideration of this 
example serve to bring out the general principles involved. 
In all forms of the tuberculin reaction there is a very evident 
hypersusceptibility to various forms of antigen derived from the 
bacillus. When the tuberculin is injected subcutaneously the reac- 
tion is systemic and also localized to a certain extent in any tubercu- 
lous foci which may be present. When the v. Pirquet or Moro skin 
reactions are carried out, or the Calmette ophthalmic test is made, the 
reactions are almost purely local. In all cases reactions are induced 
by quantities of antigen wdiich cause no effect whatever in normal 
individuals. 
The basic observation leading to the diagnostic use of tuberculin 
was made by Koch 14 upon guinea pigs. He describes his observa- 
tion as follows: 
“Tuberculin may be injected into normal guinea pigs in consid- 
erable quantities without causing noticeable symptoms. Tuberculous 
guinea pigs, on the other hand, react to comparatively small doses 
in a very characteristic manner.” 
Since, in Koch’s experiments upon tuberculin, it was desirable 
for his particular purposes at the time to obtain very sharp reac- 
tions, he did not content himself with the production of moderate 
symptoms by the injection of slight amounts of tuberculin into in- 
13 We do not include Noguchi’s luetin reaction since we do not believe that 
this reaction is specific. Our original view on this, based on studies of cul- 
ture treponemaba, has been borne out by clinical studies. See Kolmer & 
Greenbaum, Jour. A. M. A., 79, 1922, p. 2063. 
14 Koch. Deutsche med. Woch., No. 43, 1891. 498 
INFECTION AND RESISTANCE 
fected animals, but increased his dosage until the guinea pigs were 
killed. He showed that guinea pigs having a moderately advanced 
infection—4 to 5 weeks after inoculation—could be killed by doses 
of 0.2 to 0.3 gram, while animals in very advanced stages would suc- 
cumb within 6 to 30 hours to quantities as small as 0.1 gram sub- 
cutaneously. In the animals so studied he determined not only a 
systemic effect, but a very marked local reaction as well in the skin, 
areolar tissues, and adjacent lymph nodes. 
Koch’s observations upon guinea pigs were applied by him, Gutt- 
stadt,15 Beck,16 and others to man, and the result was the develop- 
ment of the present important diagnostic test. The fundamental 
fact in this as well as in other tests of this kind, then, is the appear- 
ance of local and systemic reactions in infected subjects to contact 
with specific antigenic material which, at least in the same quantities, 
produces no effects in normal individuals. The analogy with the 
phenomena of anaphylaxis is thus indicated. 
Koch’s original interpretation of the phenomenon was of course 
unaided by any of the later observations on anaphylaxis. According 
to him the tuberculin contained substances which caused tissue 
necrosis. The necrotizing action was particularly powerful upon 
tissues which were tuberculous, and therefore already saturated with 
the toxic material. The destruction of such tissues resulted in sys- 
temic symptoms. 
Very similar to this view is the one later expressed by Babes and 
Broca,17 who attribute the systemic symptoms to a sudden lighting 
up of the existing lesions by the small amount of extra tuberculin 
added to that already present in these foci. 
The first suggestion of the possible association of the tuberculin 
reaction with the union of an antigen and its antibody was made 
by Wassermann and Brack.18 They accepted Ehrlich’s assumption 
that certain cells of the tuberculous foci (those situated just below 
the periphery and already affected by the tubercle toxin, though still 
resistant) were possessed of an increased receptor apparatus for the 
tubercle antigen. For this reason the injected tuberculin was concen- 
trated in these foci, attracted out of the circulation by the increased 
avidity of these cells, the consequence being increased activity of 
the lesions and systemic symptoms. The tuberculin reaction, 
according to these writers, therefore, would be caused bv the union 
of the tuberculin with the “sessile receptors” upon the diseased 
tissues—a point of view which would specify the diseased tissues and 
their products as the sources from which emanated the toxic factors 
inciting the systemic symptoms. 
15 Guttstadt. “Klin. Jahrbuch” Erganzung-sband, 1891. 
16 Beck. Deutsche rued. Woch., No. 9, 1899. 
17 Babes u. Broca. Zeitschr. f. Hyg., Vol. 23, 1896. 
18Wassermann and Brack. Deutsche med. Woch., No. 12, 1906. BACTERIAL ANAPHYLAXIS 
499 
The theories of Koch and of Babes do not, as Meyer points out, 
explain the frequent absence of the tuberculin reaction in very ad- 
vanced cases of human tuberculosis, as contrasted with its frequency 
and regularity in the earlier cases. For, according to both of these 
views, the more severe the existing lesions the more actively would 
the injected tuberculin initiate tissue necrosis and consequent symp- 
toms. The theory of Wassermann and Brack avoids this objection 
since it presupposes the acceptance of Ehrlich’s view that the in- 
creased receptor apparatus is present and free only in those cells in 
which necrotic destruction has not yet set in. In the necrotic areas 
the receptor apparatus is already saturated or satisfied as to its 
affinities, and extensive areas of necrosis, therefore, are unaffected 
by contact with further quantities of tuberculin. 
These theories have been mentioned for the purpose of complete- 
ness, but can be dismissed at the present time as no longer fitting in 
with more recent observations on this type of reaction. The most 
important advances made in the general understanding of reactions 
of this type have been made since the development of the so-called 
skin reactions by von Pirquet 19 and the opthalmo reactions, by 
Calmette 20 and by Wolff-Eisner.21 Von Pirquet found that when a 
small amount of a 25 per cent, solution of old tuberculin was applied 
to a slightly scarified area on the forearm in tuberculous patients, a 
typical, now generally well known, reaction occurred. The actual 
diagnostic value of the reaction has lost considerable weight, since this 
discovery by the recognition that in human beings the test is such a 
delicate one that over 70 per cent, of adults react positively, and even 
in children above two years of age, a positive reaction does not by 
any means signify an active tuberculosis. The same is true to about 
the same extent of the opthalmo reaction which consisted in the in- 
stillation of a 1 per cent, solution (of an alcoholic precipitate of old 
tuberculin) in distilled water into the eye. However, that these vari- 
ous tuberculins are specifically related to infection with the tubercle 
bacillus has been subsequently worked out with considerable accuracy 
in animals. The form in which the reactions have been mainly done 
in recent years in man and in animals is by the intracutaneous in- 
jections of minute amounts of the respective substances. 
Since similar facts were elicited in the case of other infections, 
such as glanders, typhoid, B. abortus, etc., it was apparent that we 
had here a specific hypersusceptibility of animals, and man, to a bac- 
terial product which, though it differed in important respects from 
the phenomena of true protein anaphylaxis, nevertheless, manifested 
19 Yon Pirquet. Berl. klin. Woch., 20, 1907, p. 644 and 22, 1907, p. 699. 
“Klinische Studien iiber Vaccination,” Deuticke, Wien, 1907. 
20 Calmette. Compt. rend. d. VAcad. de Scien., June, 1907. 
21 Wolff-Eisner. Berl. klin. Woch., 1907, p. 1055, discussion of paper by 
Citron. 500 
INFECTION AND RESISTANCE 
a great many analogies. Yon Pirquet, accordingly, included the 
reaction under the phenomena of what he called allergy, or altered 
reaction capacity. 
He assumed that the reaction depended upon the presence in the 
system of antibodies, which formed a union with the applied tubercu- 
lin, the result being the formation of poisons and a reaction. This 
assumption, according to the anaphylatoxin theory of Friedberger, 
would imply the participation of alexin—acting upon the united 
tuberculin-antituberculin complex, though v. Pirquet does not express 
himself positive as to this. 
This, moreover, is the clearly expressed opinion of Friedberger.22 
Consistently with his general theory of anaphylaxis he assumes that 
the injected tuberculin comes into relation with specific antibodies 
with which it unites, the alexin then splitting off anaphylatoxin from 
the complex. He bases this view upon his experimental demonstra- 
tion, mentioned above, of the production of anaphylatoxin from 
tubercle bacilli by in vitro digestion with guinea pig complement. 
In principle the view of v. Pirquet is similar to that previously 
expressed by Wolff-Eisner 23 that the union of tuberculin with its 
lytic antibody, present in the tuberculous animal, gave rise to poisons 
as the result of lysis. Both of these theories simply apply, to the 
special case of the tuberculin reaction theories of mechanism applied 
by these workers to anaphylactic reactions in general. 
Calmette,24 who separates the tuberculin reactions distinctly from 
anaphylaxis, nevertheless assumes that tuberculous animals and man 
develop a lytic principle, presumably in the nature of a bacteriolysin, 
which enters into reaction with the specific antigen in the tuberculin 
preparations. This, again, is fundamentally an anaphylatoxin view 
in Friedberger’s sense. 
However, more recent careful analysis of tuberculin and similar 
reactions in animals have yielded a great many important facts which 
show that while there is much fundamental analogy between these 
reactions and anaphylaxis, there are, on the other hand, also, impor- 
tant differences which will not permit us to assume exactly the same 
mechanism in both. 
Study of the reactions from a theoretical point of view was begun 
by Homer 25 in 1909, and by Baldwin 26 in 1910. Baldwin’s work is 
fundamental, in showing that, in spite of previous assertions, guinea 
pigs could not be rendered skin-sensitive by implantation of porous 
filter capsules or celloidin capsules containing tuberculoprotein, or 
living tubercle bacilli. He showed that skin sensitiveness could never 
22 Friedberger. Munch, med. Woch., Nos. 50 and 51, 1910. 
23 Wolff-Eisner. Berl. klin. Woch., Nos. 42 and 44, 1904. 
24 Calmette. “L’infeetion bacillaire et la tnbereulose,” Paris, 1920. 
25 Romer. Berl. klin. Woch., 46, 1909, p. 813, and Deutsch. med. Woch.. 
35, 1909, p. 245. 
26 Baldwin. Jour. Med. Bes., 22, 1910, p. 189. BACTERIAL ANAPHYLAXIS 
501 
be produced without actual infection with living organisms. Animals 
treated with tuberculoprotein, however, often showed reactions to 
intravenous inoculation of the homologous preparation which could 
be recognized as anaphylactic. His conclusions can be summarized 
as follows: Tuberculous animals become sensitive to anaphylactic 
test, but not uniformly so. There is no absolute relation between the 
degree of sensitiveness and the stage of the disease. Injections of the 
tuberculoprotein may sensitize normal guinea pigs. Sensitized 
guinea pigs, however, do not react to the ordinary tuberculin test, 
though some respond slightly to the intradermal test. He adds: “This 
difference between anaphylactic sensitization and tuberculin reac- 
tivity need not be fundamental, as it is possibly due to another factor 
as yet undetermined.” His experiments on the transfer of passive 
anaphylaxis to tuberculoprotein were inconclusive, but it has been 
shown since then, by Austrian 27 and others, that passive sensitiza- 
tion can be attained. From a theoretical point of view the most im- 
portant observation of Baldwin is the fact that there seemed to be a 
discrepancy between skin sensitiveness and general anaphylaxis. 
Krause,28 following out the work of Baldwin, confirmed and extended 
the above observations and established an interesting relationship 
between skin sensitiveness and the progress of infection. He asserts 
that skin sensitiveness develops simultaneously with the development 
of the initial focus, increases progressively with the lesions, varies 
directly with the extent and intensity of the infection, and diminishes 
with healing. It is blunted by a general tuberculin reaction which 
suggests analogy to saturation, such as that which occurs in connec- 
tion with anaphylaxis. 
In dealing theoretically with the tuberculin reaction, especially 
as regards the possibility of its being an anaphylactic phenomenon, 
it is of great importance to determine whether or not the reaction 
can be transmitted passively. The evidence on this is contradictory. 
Only a few writers have reported positively in this connection. In 
1909, Bail 29 injected finely divided tissue mash of tuberculous organs 
of guinea pigs into normal guinea pigs, and 24 hours later gave the 
animals so treated 0.5 c. c. of old tuberculin, or a preparation which 
in this quantity had practically no effect upon normal animals. The 
animals prepared with the tuberculous tissue died in some cases, while 
controls treated with normal tissue suspensions showed no symptoms 
whatever. 
Helmholz 30 in the same year reported positive skin reactions in 
normal guinea pigs 2 to 6 days after he had injected them intra- 
peritoneally with the defibrinated blood of tuberculous guinea pigs. 
27 Austrian. Bull. Johns Hopkins Hosp., 23, 1912, p. 1. 
28 Krause. Amer. Rev. Tuberc., 1, 1917-1918, p. 65. 
29 Bail. Zeitscher. f. Immunitats., Orig., 4, 1909, p. 470. 
30 Helmholz. Zeitschr. f. Immunitats., Orig., 3, 1909, p. 371. 502 
INFECTION AND RESISTANCE 
Both of these observations would be of fundamental importance if 
they could be confirmed. 
In considering the mechanism of the tuberculin reaction, it will 
be well to examine also the work that has been done on the typhoidin 
reaction. Gay and Claypole 31 believed that positive skin reactions 
in rabbits were parallel with the degree of immunity of the animal. 
They succeeded in transferring the susceptibility to typhoidin from 
an immune to a normal animal by inoculation of 20 c. c. of typhoid- 
immune serum, testing 24 hours later. These experiments were re- 
peated and confirmed by Meyer and Christiansen; 32 and in their 
first work with rabbits, these last observers, using what they called a 
typhoid autolysate (by which they mean an alcohol precipitate of a 
heated distilled water suspension of a 48 hour agar culture), con- 
cluded that “the typhoidin and similar reactions in rabbits are ana- 
phylactic in nature and the result of an interaction of antigen and 
antibody.” They stated that “the logical assumption from these facts 
is that cutaneous hypersensitiveness is the result of bacterial protein 
sensitization.” Later Meyer 33 found that injected rabbits react with 
typhoidin more intensively than do immunized rabbits, and drew the 
conclusion that cutaneous hypersensitiveness does not indicate that 
the rabbits are particularly immune, and that no definite relationship 
existed between agglutinins and complement-fixing antibodies and 
skin sensitiveness. From these first two papers of Meyer’s we gather 
that he believed that in rabbits skin sensitiveness to typhoidin is a 
sign of infection, rather than of immunity, but that as stated in his 
own words “cutaneous hypersensitiveness of rabbits . . . is, in all 
probability, the result of sensitization with typho- or similar bac- 
terial proteins.” Nichols 34 also considered the typhoidin reaction as 
he observed it in human beings as a protein sensitization. 
The apparent discrepancies between the results of Baldwin with 
the tuberculin reaction and those of the workers just mentioned with 
the analogous typhoidin reaction, probably depend upon the fact that 
Baldwin used guinea pigs and the other observers used rabbits. When, 
subsequently, Fleisclmer, Meyer and Shaw33 studied cutaneous 
hypersensitiveness in guinea pigs treated with repeated intraperi- 
toneal injections of B. abortus, B. typhosus, and old tuberculin, and 
carried out parallel experiments with animals infected with living 
organisms the conclusions that they reached coincided with those of 
Baldwin in the case of the tuberculin reaction. They found, in other 
81 Gay and Claypole. Arch. Inter. Med., 14, 1914, p. 671. 
32 Meyer and Christiansen. Collected Repr. G. W. Hooper Foundation 
for Med. Res., 2, 1916, p. 1. 
33 Meyer. Collected Reprints G. W. Hooper Foundation for Med. Res., 
2, 1916, p. 68. 
34 Nichols. Jour. Exp. Med., 22, 1915, p. 780. 
35 Fleischner, Meyer and Shaw. Amer. Jour. Din. Child., 18, 1919, p. 577. BACTERIAL ANAPHYLAXIS 
503 
words, that guinea pigs treated with dead bacterial proteins might 
become anaphylactic, bnt did not give skin reactions. 
It will be seen, thus, that while results in rabbits are confusing, 
the conditions in guinea pigs are more or less analogous to those pre- 
vailing in human beings, and there seem to be at least two funda- 
mental differences between protein anaphylaxis and reactions like 
the tuberculin reaction. On the one hand, these reactions, unlike true 
anaphylaxis, cannot be elicited by the treatment of animals with 
the dead bacterial products but occur only in animals that have been 
infected with the living micro-organisms. In the second place, pas- 
sive transfer of typical tuberculin sensitiveness, in spite of the 
isolated observations to the contrary cited above, cannot be accepted 
as established. 
Our own work 36 dealing with this matter confirmed these points, 
and was suggestive of a possible explanation. Studying particularly 
guinea pigs in regard to the relationship between skin reactions, ana- 
phylaxis and infection, it was first found that, in these animals, two 
types of skin reaction could be elicited, just as this was possible in 
the human being. In typical protein anaphylaxis, with antigens like 
horse serum, during the period of sensitiveness, it was possible to 
elicit the urticaria-like, evanescent skin reactions noticeable in horse 
serum sensitiveness in human beings by the intracutaneous injection 
of small amounts of the antigen. When guinea pigs were treated for 
long periods with extracts of ground tubercle bacilli, they became 
typically anaphylactic as shown by the uterine reaction, but never 
developed typical tuberculin skin reactions. Such reactions, how- 
ever, could always be developed within 8, 9, or 10 days in guinea pigs 
infected with living tubercle bacilli. Typical anaphylaxis, also, oc- 
curred in tubercle infected guinea pigs, but did not develop in less 
than 3 weeks or more; so that it was definitely possible to separate 
the typical, slowly developing and necrotic tuberculin skin reaction 
from true anaphylaxis, in that infected animals would show only the 
tuberculin reaction for considerable periods and no signs of uterine 
anaphylaxis; whereas, animals treated with dead materials would 
eventually develop true anaphylaxis but not skin reactiveness. 
In following up these phenomena, it was found that the material 
which elicited the typical tuberculin reactions were chemically quite 
different from the coagulable proteins of the bacterial body, or of ani- 
mal sera. The materials which gave typical tuberculin reactions 
(and only in isolated instances anaphylactic response in the uterus) 
were prepared as follows : 
From powdered tubercle bacilli of the human type we produced extracts 
by shaking in a 0.02 per cent, sodium hydroxide solution in physiological 
salt solution. Three or four hours’ shaking and perhaps a day or so in the 
ice chest sufficed to bring a considerable amount of the material into solution. 
36 Zinsser. Jour, of Exp. Med., Yol. 34, 1921, p. 495. 504 
INFECTION AND RESISTANCE 
This extract was centrifugalized until all the particles had been removed and 
a moderately opalescent supernatant fluid was decanted. 
This material, upon being acidified to approximately Ph 5 to 6, with 2 
per cent, acetic acid in the cold, became turbid and soon precipitated in 
large flakes. Further acidification up to Ph 4 did not redissolve these flakes. 
The precipitate represented the bulk of the dissolved substance in the ex- 
tracts. The precipitate could be redissolved in a slightly alkalin salt solution 
and reprecipitated with acid. Because of their precipitability by acetic acid 
in the cold, we designated these substances as nucleoproteins or phospho- 
proteins, although we do not wish to commit ourselves, chemically, since we 
are aware of the indefiniteness of these biochemical terms and realize fully 
our incompetence to deal authoritatively with a chemical problem of such 
difficulty, without further intensive study. 
After removal of these acid-precipitable substances by centrifugation and 
filtration through Berkefeld candles, the fluid was brought to a boil in the 
acid condition, and sometimes a very faint turbidity developed which was 
taken to represent the presence of coagulable protein, albumin or globulin, 
or both. These precipitates were so fine and slight that only Berkefeld 
filtration would remove them, and even this was not always completely 
successful. 
The fluid was then neutralized; that is, brought to an approximate Ph of 7. 
When this neutral fluid was filtered hot, we observed on numerous occa- 
sions that a slight precipitate developed over night in the ice chest, and 
that this precipitate redissolved on heating, a point which indicated the 
possibility that the tubercle bacilli may contain Bence-Jones protein. This 
point, however, will need further chemical analysis, a task which we have 
not yet had time to undertake. The water-clear fluid which was left after 
removal of all these substances gave in all cases a very definite precipitate 
with alcohol. The precipitate could be thrown down, collected and redis- 
solved in water with salt solution, and like the similar precipitate obtained 
from crude tuberculin was astonishingly soluble. 
This final water-clear material gave no biuret reaction, and usually gave 
no sulfosalicylic acid reaction, though occasionally this was very faint; in 
most cases it gave no Millon reaction, was not clouded on boiling with acid, 
and usually gave no xanthro-protein reaction, though in some cases a slightly 
yellowish color appeared when the ammonia was added in the second part 
of the test. This material, for want of a better name, we speak of as the 
“proteose” residue. 
Materials of this kind have since been prepared by us from a num- 
ber of other bacteria, and have been found to be excellent antigens 
in the sense that they react by precipitation and complement fixation 
with anti-sera; they have not so far produced antibodies after in- 
jection into animals, although extensive efforts to do this are still 
being carried on. It is not impossible that they represent the 
“Hoptenes” of Landsteiner. Whether or not they represent non- 
coagulable proteins—like some of the plant proteins studied by Wells, 
or whether they are molecularly simpler nitrogenous substances we 
cannot say as yet but hope to ascertain shortly. 
Thus, to summarize the relation of reactions like the tuberculin 
reaction to anaphylaxis, the chief differences are the association of 
the tuberculin type of reaction with actual infection in which the 
body is constantly under the influence of substances diffusing out BACTERIAL ANAPHYLAXIS 
505 
from the growing focus; the quantitative and time factor of sensitiza- 
tion, therefore, is different. That the sensitization must be derived 
from a specific stimulus emanating from the bacteria seems unques- 
tionable because of its specificity in each type of infection. More- 
over, the failure of passive sensitization in these reactions, together 
with the more slowly developing and necrotic type of skin reaction, 
would influence one to consider that these reactions are intimately 
associated with the cell protoplasm, rather than merely with the cell 
surfaces. Differences in the chemistry of the inciting substance, 
namely, the probability that in the tuberculin and similarly pre- 
pared antigens from other bacteria, we may be dealing with a smaller 
molecule, certainly not with a coagulable protein, would make it seem 
possible that an important factor in these reactions might be found 
in differences of chemical structure, and perhaps of diffusibility be- 
tweqp these antigens and the coagulable proteins which constitute the 
antigen in true anaphylaxis. It is interesting to note that antigens 
of this type, moreover, can be easily obtained from young cultures 
of the various bacteria in broth, and can be washed off the surfaces 
of bacteria like the pneumococcus, and influenza bacilli, making it 
more than likely that the sensitization is accompanied by a constant 
diffusing out from the focus of these materials. In comparing the 
production of these substances by living tubercle bacilli with extrac- 
tion from dead tubercle bacilli, we found that when relatively large 
amounts of dead tubercle bacilli were infused in glycerin broth and, 
in parallel flasks, less than one-fifth of these amounts were planted 
on the broth in the living condition, the fluid surrounding the living 
tubercle bacilli was powerful in “tuberculin” substance as early as 
the third day when a slight amount only was present in the broth in 
which the dead tubercle bacilli had been infused. 
Without going further into theoretical deductions from this, we 
are inclined to believe that reactions like the tuberculin, etc., reaction, 
are, indeed, phenomena in which antigen-antibody-like mechanisms 
are involved, but that here the process is probably transferred to the 
interior of the cell, and that because of the chemical and physical 
differences between the inciting substances involved, the manner of 
sensitization, both in time and quantity factors, is a different one. 
The speed with which tuberculin sensitiveness develops in guinea 
pigs after infection with living organisms, that is, the 8th or 9th day, 
and the seriousness of the injury, inclines us to believe that such 
sensitization to these diffusible products may constitute an important 
factor in the injury of the body during bacterial infection. 
Moreover, we have tried to simulate the conditions going on dur- 
ing infection in the body by injecting animals frequently with large 
amounts of dead tubercle bacillus material, and with Petroff have 
found that skin reactiveness can be developed if large amounts of 
dead tubercle bacilli are repeatedly administered. 506 
INFECTION AND RESISTANCE 
This is particularly important in view of the fact that Krause 
has shown that skin hypersensitiveness and resistance to super-in- 
fection with tubercle bacilli, go hand in hand. Experiments on the 
artificial production of such increased resistance with dead tubercle 
products are going on. 
While, therefore, this type of reaction typified bv tuberculin, 
typhoidin, mallein and other bacterial skin reactions, cannot be prop- 
erly spoken of as true protein anaphylaxis it, nevertheless, has many 
points in common with it, but also definite differences, the funda- 
mental basis for which will probably in our opinion be found in the 
chemical and physical nature of the antigen involved. 
The Anaphylatoxin Theory Applied to Bacterial Infection.—As 
one of the results of the attempts to explain anaphylactic phenomena 
by an intravascular reaction between antigen and antibody, a con- 
siderable amount of work was done with bacterial antigens, some 
of which took its departure from attempts to eliminate Pfeiffer’s 
ideas of endotoxins. 
Indeed the sudden liberation of endotoxins by immune sera had 
been regarded by Pfeiffer and others as the cause of the rapid death 
often ensuing in immunized guinea pigs when more than a definite 
maximum of cholera spirilla or other organisms was injected. In 
all these opinions the basic conception was that certain bacteria con- 
tained a characteristic preformed poison (endotoxin) upon the 
pharmacological properties of which the peculiar symptoms caused 
by each organism depended. 
The earliest unambiguous statements of a conception differing 
from this original view of the nature of bacterial endotoxins, and 
approaching the later conceptions of Friedberger, are found, we 
believe, in the work of Vaughan.37 In an article by him, published 
in 1908, Vaughan, after describing the incubation time occurring in 
man and animals after inoculation with typhoid bacilli, says: “The 
sickness begins when the animal body becomes sensitized and begins 
to split up the bacilli.” By “splitting up” he means here, as in his 
other work 38 on protein split products, not a mere liberation of pre- 
formed poisons, but a chemical (enzymotic) proteolysis by which a 
poisonous group of the bacterial protein-molecule is set free. 
The essential difference of this point of view from the endotoxin 
theory at first sight seems a trivial one—in the one case liberation of 
a preformed poison molecule, in the other liberation of a poison by 
the breaking up of a molecule. The difference, however, is a funda- 
mental one. For, in the earlier theory, the specific element of the 
toxemia was in the nature of the different poisons—whereas in the 
view of Vaughan the lysin which breaks up the protein molecule is 
alone the specific element, the formed poisons being concerned as 
37 Vaughan. Am. Jour, of Med. Sci., Sept., 1908. 
38 Vaughan. Zeitschr. f. Immunitatsforsch., Vol. 1, 1909. BACTERIAL ANAPHYLAXIS 
507 
non-specific and alike, whether produced from colon bacilli, tubercle 
bacilli, or eggwhite. 
Friedberger,39 finally, in 1910, repeating with bacteria his -ex- 
periments upon “anaphylatoxin” liberation from specific precipitates, 
succeeded in obtaining such poisons in the test tube by allowing fresh 
guinea pig complement to act upon sensitized bacteria. 
These results were confirmed by extensive experiments carried 
out soon after this by Friedberger40 himself with a number of 
collaborators. 
The results of these investigations may be summarized as follows: 
1. The action of alexin upon sensitized or unsensitized bacteria 
yields toxic substances which, injected into normal guinea pigs, 
produce the characteristic symptoms of anaphylaxis, with frequent 
death and typical autopsy findings. 
2. These poisons (“anaphylatoxins”) may be produced from 
any variety of bacteria, pathogenic and non-pathogenic.41 (The or- 
ganisms used in the earlier experiments were Vibrio metchnihovi, 
the bacillus of tuberculosis, the typhoid, prodigiosus, and subtilis 
bacillus, and Aspergillus fumigatus.) 
3. The successful production of the poisons depends intimately 
upon the relative amounts of antigen (bacteria) and alexin used, and 
upon the time and temperature conditions under which the ex- 
posures are made. 
4. The poisons can be produced from boiled as well as from 
native bacteria. 
Although unsuccessful with none of the bacteria with which ex- 
periments were carried out, different species yielded the poison with 
varying degrees of intensity, and qualitatively the poisons were simi- 
lar. Bacillus prodigiosus, though non-pathogenic, seemed to be one 
of the most favorable micro-organisms for such experiments. 
The experiments of Friedberger and his associates were rapidly 
confirmed by Neufeld and Dold,42 Kraus,43 Ritz and Sachs,44 and 
39 Friedberger. Berl. klin. Woch., Nos. 32 and 42, 1910. 
40 Friedberger; Friedberger and Goldsehmid; Friedberger and Szymanow- 
ski; Friedberger and Schiitze; Friedberger and Nathan. Zeitschr. f. Im- 
munitdtsforsch., Yol. 9, 1911. 
41 Neufeld and Dold, comparing virulent and avirulent strains of pneu- 
mococcus in this regard, have found that the virulence of the race has no 
relation to its yield of anaphylatoxin. Indeed the anaphylatoxins from 
various bacteria seem to be qualitatively entirely alike. 
Note.—We wish to note here that we are giving Friedberger’s theories 
and experiments at some length because they have had considerable influence 
on our conceptions of infection. We do not wish to create the impression 
that we accept them, however, and will discuss their fallacies later in the 
same chapter. 
42 Neufeld and Dold. Berl. klin. Woch., No. 2, 24, 1911; Arb. a. d. kais. 
Gesundheits Amt., Yol. 38, 1911. 
43 Kraus. Zeitschr. f. Immunitatsforsch., Vol. 8, 1911. 
44 Ritz u. Sachs. Berl. klin. Woch., No. 22, 1911. 508 
INFECTION AND RESISTANCE 
many others,45 and, though the conditions under which the anaphy- 
latoxin formation took place were defined with slight variation by 
different workers, the essential features of Friedberger’s claims were 
upheld. 
It was further shown by Friedberger and Nathan that the con- 
ditions prevailing in the test-tube experiment in truth represent the 
processes taking place within the animal body. This they accom- 
plished by injected bacterial emulsions into the peritoneal cavities of 
guinea pigs, killing the animals after several hours and examining 
the peritoneal exudates for their toxic properties. Centrifugalized, 
cleared of bacteria, and injected intravenously into other guinea pigs, 
these exudates produced the typical acute symptoms characteristic 
of the poisons obtained in test-tube experiments. 
Friedberger now suggested that we may regard bacterial infec- 
tion, after all, as the presence in the body of a living foreign protein 
—in this case varying in distribution and quantity by reason of the 
particular invasive properties of the given germ and the balance 
between these and the resistance of the host. It would not be neces- 
sary, therefore, to assume that the character of the disease is deter- 
mined by the existence of different preformed “endotoxins.”' He 
believed that we may justly assume that the toxic substances appear 
only after protein cleavage of the bacterial bodies has been initiated 
by the action upon them of the serum components, and that the ap- 
parent specificity of the poisons, or differences between the toxemic 
manifestations of various diseases, may depend, not on differences 
in the pharmacological actions of these poisons, but rather upon 
variations in the invasive properties of the bacteria, both as concerns 
their quantitative distribution and their accumulation and localiza- 
tion in the infected body. 
To support this assumption Friedberger pointed out the simi- 
larity in the clinical manifestations of several diseases in which the 
inciting bacteria are biologically very different, but in which the dis- 
tribution and invasive properties are alike. For instance, lobar pneu- 
monia caused by the pneumococcus is clinically very similar to that 
caused by the Friedlander bacillus, though the micro-organisms in- 
citing them are extremely unlike each other. He drew a similar 
parallel between true cholera and cholera nostras, and we may add 
another striking example in the great similarity existing clinically 
between the various forms of acute and subacute septicemia in which 
a definite bacteriological diagnosis can rarely be made except by blood 
culture. 
Conversely, the same micro-organism may call forth diseases 
which clinically, apart from the purely local manifestations, are very 
dissimilar, according to the localization and distribution of the 
bacteria. 
45 Izar. Zeitschr. /. Immunitatsforsch, Yol. 11, 1911. BACTERIAL ANAPHYLAXIS 
509 
These views of Friedberger constituted a serious objection to the 
“endotoxin” theory of Pfeiffer for the explanation of the toxic effects 
accompanying infection with organisms such as the typhoid bacillus, 
cholera spirillum, etc. However, the mechanism by which he ex- 
plained the formation of the so-called “anaphylatoxins” was soon 
shown to be incorrect by investigations which indicated that the bac- 
terial cell could not be justly regarded as the matrix of these poisons.48 
The possibility that the bacterial cell might not furnish the actual 
chemical source of the poisons was first rendered questionable by the 
fact that Friedberger’s anaphylatoxins could be produced as well 
from boiled as from living bacteria. 
Such vague suspicion became a very definite doubt, in the light 
of the experiments of Keysser and Wassermann.47 Keysser and 
Wassermann utilized the fact that certain serum elements may be 
absorbed out of serum if this is shaken up with indifferent suspen- 
sions such as barium sulphate or kaolin (aluminium orthosilicate).48 
They therefore substituted these insoluble substances for antigen, 
allowed them to absorb serum constituents, assumed by them to be 
amboceptor, out of normal and inactivated immune sera, and thpn 
allowed complement or alexin to act upon the “sensitized” kaolin. 
In this way they obtained active and powerful anaphylatoxin, 
and claimed, in consequence, that the matrix of the poison was not 
in the bacterial antigen, but in the sensitizer or amboceptor, which 
was mechanically absorbed by the bacteria (as by the kaolin), and 
thus made amenable to the alexin action. 
The experiments of Keysser and Wassermann found confirmation 
in the hands of other investigators, although the results of Heufeld 
and Dold, as well as our own, with this method were far more irregu- 
lar than were those of Keysser and Wassermann. The many experi- 
ments on the production of anaphylatoxins (better Sero—or Proteo 
toxins) by placing serum in contact with pepton—with agar and a 
variety of other substances in fine suspension, followed. (See pre- 
ceding chapter.) 
Jobling and Petersen believe that, by the ordinary technique of 
anaphylatoxin production with bacteria and serum, most of the toxic 
substances originate from the serum proteins. The bacteria act 
merely by removing the antiferments from the serum, thereby setting 
free the ferments normally present in the serum, and permitting 
them to act upon the serum proteins. The result is cleavage and the 
production of toxic split products. This would explain such results 
46 See also Section on Anaphylatoxins in a preceding chapter. 
47 Keysser and Wassermann. Folia Serologica, Yol. 7, 1911; Zeitschr. 
f. Byg., Yol. 68, 1911. 
48 Kaolin emulsions will absorb amboceptor only out of diluted serum. Out 
of concentrated serum complement is completely absorbed. Friedberger u. 
Salecker, Zeitschr. f. Immunitatsforsch., Yol. 11, 1911; Zinsser, from Jour. 
Exp. Med., Yol. 18, 1913. 510 
INFECTION AND RESISTANCE 
as those of Keysser and Wassermann. Jobling and Petersen sup- 
ported their contention by experiments in which they have obtained 
typical anaphylatoxins by removing serum antiferments with chloro- 
form, kaolin, and agar. They further showed that emulsions of bac- 
teria actually do remove antiferments from fresh serum, and that 
the bacteria used in the process become more resistant to tryptic 
digestion in consequence. 
Experiments of similar general import are those of Bronfen- 
brenner with the Abderhalden reaction. 
It is thus clear that the interaction of bacteria and serum may 
produce toxic effects, certainly in the test tube, perhaps in the ani- 
mal body. Whatever the mechanism of this process may be, it is 
not out of question that this phenomenon may have important bear- 
ing on the toxemia of infectious disease in which bacteria, as fine 
suspension, are in contact with plasma and tissues. This forces 
caution in further acceptance of the endotoxin theory—but the extent 
to which these processes are effective in the animal body is, as yet, 
quite unclear. We refer the reader to the preceding chapter in which 
anaphylatoxins are dealt with at considerable length. 
Toxin Hypersusceptibility.—Toxin hypersusceptibility, which 
has occasionally been acquired by animals in the course of immuniza- 
tion with diphtheria and tetanus toxin, is usually classified with ana- 
phylaxis, indeed is often cited as the earliest observation of this 
phenomenon. However, it is by no means clear that the two condi- 
tions are actually analogous, since in the case of the toxins we are 
dealing with antigens which are not only toxic in themselves, but 
against which neutralizing antibodies are formed in the reacting 
animal. This last fact alone would separate toxin hypersusceptibility 
sharply from true protein-anaphvlaxis in that entirely different re- 
acting-mechanisms seem to he called into play by the two varieties 
of antigen. It will he necessary, therefore, to discuss toxin liyper- 
susceptibility at some length. 
Probably the earliest authentically recorded observation is that 
of von Behring,49 who determined, both for diphtheria and tetanus 
toxins, that animals once inoculated with these poisons were oc- 
casionally more sensitive to them subsequently than were normal 
animals. He spoke of “Gift Ueber empfindlichkeit” as a property 
acquired by reason of a preceding injection, and the observation was 
further developed by Knorr 50 in 1895, and by von Behring himself, 
in collaboration with Kitashima 51—a few years later. These writers 
showed that guinea pigs which are treated repeatedly with small 
doses of diphtheria toxin may, under certain circumstances, not only 
fail to show immunity, but may even develop a susceptibility in- 
49 Yon Behring. Deutsche med. Woch., 1893. 
50 Knorr. Quoted from Otto, “Dissertation,” Marburg, 1895. 
51 Von Behring u. Kitashima. Berlin, klin. Woch., 1901. BACTERIAL ANAFHYLAXIS 
511 
creased to such an extent that doses far too small to injure a normal 
animal will cause their death. Again, in the case of diphtheria toxin 
similar observations were made upon horses by both Salomonsen and 
Madsen,52 and by Kretz.53 The last-named worker observed that 
horses that had been immunized with diphtheria toxin would often 
react to neutral mixtures of toxin and antitoxin by which normal 
horses were unaffected. This so-called “paradox phenomenon” was 
much discussed, and many theories advanced to explain it, a most 
ingenious adaptation of the side-chain theory being applied to it by 
Kretz 54 and by Wassermann.55 They assumed that the partial im- 
munization in such treated animals had in truth induced the forma- 
tion of excessive receptors; that, in the stages of hypersusceptibility, 
however, these receptors had not yet been cast off from the cells. In 
consequence there was an excess of “sessile receptors”—by means of 
which the cell was rendered more exposed to toxin action than it was 
normally—it being still unprotected by the presence of freely cir- 
culating “antitoxin” receptors. The difficulties arising from the 
observation of similar hypersusceptibility in animals whose blood 
contained free antitoxin were disposed of by Wassermann by the 
convenient assumption of variations of affinity. 
He assumed that the treatment with toxin, i. e., the intoxication, 
may induce a condition of higher affinity for the poison on the part 
of the sessile cell receptors, leading to a selective toxin-absorption by 
the cells and consequent greater susceptibility to injury. With 
Behring, he speaks of this as a “histogenic hypersusceptibility,” 
implying an increased vulnerability of the tissue cells. 
The analogy between these early observations and the phenomena 
which we now classify as anaphylaxis is unquestionably a striking 
one. However, it is doubtful, as Friedemann suggests, whether the 
two processes depend upon similar mechanisms. For, as we have 
seen in the case of the sensitiveness to toxin, we are dealing with 
primarily poisonous substances against which in the reacting animal 
neutralizing antibodies are found—a combination of conditions quite 
different from those with which we are confronted in hvpersuscepti- 
bility against primarily harmless proteins. It is, of course, possible 
that the toxin hypersusceptibility is a true anaphylaxis against the 
toxin-protein—independent of the specifically poisonous nature of 
this substance. However, this is unlikely, since Lowi and Meyer 
have shown that with tetanus toxin, the symptoms of such hypersus- 
ceptibility are not those of anaphylaxis, but of increased but charac- 
teristic tetanus poisoning. Lowi and Meyer regard tetanus toxin 
52 Salomonsen et Madsen. Ann. de VInst. Past., 1897. 
63 Kretz. Quoted from Otto in “Kolle u. Wassermann Handbuch,” Er- 
gangzung'sband 2, p. 232. 
54 Kretz. Zeitschr. f. Heilkunde, 1902. 
65 Wassermann. “Kobe u. Wassermann Hundbuch,” Vol. 4, 479. 512 
INFECTION AND RESISTANCE 
hypersusceptibility as a “summation”—meaning thereby that it de- 
pends upon an alteration of the cells of the spinal cord because of 
traces of the poison retained in them. When the toxin was given 
intraneurally no antitoxin formation occurred, but the animals de- 
veloped a marked hypersusceptibility in the course of several weeks. 
This has been used as an argument that antibodies play no role in 
the process. However, the absence of circulating antitoxin would 
not seem to us to prove necessarily that no antitoxin was present in 
or on the cells. Similar reasoning has been found faulty in connec- 
tion with protein anaphylaxis. The experiment, we believe, leaves 
us as uncertain as we were before. 
A striking difference between toxin hypersusceptibility and pro- 
tein anaphylaxis is the apparently uniform failure to produce toxin 
hypersensitiveness passively by the injection of antitoxic sera. 
Thus we may summarize the conditions somewhat as follows: 
In conformity with the general laws of anaphylaxis, toxin hyper- 
susceptibility is a process in which the inciting substance is an anti- 
gen, against which hypersusceptibility can be specifically induced 
by the successive injection of sub-lethal amounts. Differing from 
protein anaphylaxis, however, are the facts that the antigen in this 
case is an antitoxin producing toxic substance quite different from 
the sensitizer-producing ones represented by the proteins; that the 
symptoms elicited are not those of a general anaphylaxis, but rather 
the specific ones represented by the pharmacological action of the 
toxin, and that passive transfer with antitoxic serum has been un- 
successful. 
It would be, therefore, merely a question of words to assign to 
these occurrences $ definite place in the classification of phenomena 
of hypersusceptibility. We still believe that the most likely explana- 
tion is that the preliminary treatment gives rise to a certain degree of 
antitoxin production which remains sessile upon the cells and, in the 
animals in which the hypersusceptibility is noticed, thereby renders 
the tissues more susceptible to toxin injury. Passive sensitization 
may possibly have failed for the reason that it is very difficult to 
attain a balance in the passively injected animal in which the sessile 
antitoxic antibodies sufficiently outweigh the neutralizing ones in 
the circulation and we do know, from much work on antitoxin prop- 
erties of circulating blood, and, conversely, from the experiments 
done, especially by Weil, on attempts to protect sensitized animals 
by the injection of antisera, that the union of toxin and antitoxin in 
the circulation takes place with much greater ease than does that of 
protein and sensitizer. Moreover, we know almost nothing about the 
absorption of antitoxin from the circulation by the body cells, and 
have no means at the present time of testing whether, after such 
absorption, if it occurs at all, the antitoxin is functionally intact, or 
whether it is destroyed. All these matters must be taken into con- BACTERIAL ANAPHYLAXIS 
513 
sideration when we deal with toxin hypersusceptibilities, and here 
again, as emphasized in our section on bacterial anaphylaxis, it is 
the nature of the antigen both chemically and physically, and the 
laws governing the union of the varieties of antigen with the anti- 
bodies wdiich characterizes and varies the manifestations of various 
types of hypersusceptibility. 
Toxic Action of Normal Sera.—There are a number of well- 
defined phenomena of acquired hypersusceptibility or sensitiveness 
which, in nature, seem closely analogous to true anaphylaxis as we 
understand it to-day, hut regarding the mechanism of which the 
opinions of experimenters are still to some extent at variance. 
Among the most important of these is the toxic action of normal 
sera when injected into animals of another species—a phenomenon 
which is now generally accepted as belonging in principle to the true 
anaphylactic phenomena, though this opinion is of comparatively 
recent formulation. The subject is of sufficient theoretical and 
practical importance to be considered in some detail. 
The older studies of phenomena belonging to this category fol- 
lowed closely in the footsteps of experiments on transfusion, and as 
early as 1666 a commission of the London Royal Philosophical So- 
ciety reported deaths following transfusion, alleging intravascular 
coagulation as the probable cause of death. 
The cause of death following injections of foreign whole blood, 
blood cells, and serum has, since that time, occupied the attention of 
many workers whose studies need not be reviewed for our present 
purposes. Chief among them were Morgagni, Brown-Sequard, Ma- 
gendie, and, more recently, Haunyn, Landois, and Ponfick.56 
The work of Landois is of special interest in that he worked 
with blood serum free from cells, and attempted to correlate the 
occurrences after the injection of animals witli the action of the 
serum upon the cellular blood elements in vitro. Landois observed 
both the solution of hemoglobin and hemagglutination, and was 
led to regard the action of serum upon erythrocytes as the primary 
cause of death after transfusion. His conception of the mechanism 
is apparently twofold. On the one hand, he believed that when 
small quantities of blood were transfused, a formation of fibrin 
(stroma-fibrin) was initiated in the stroma of the injured erythro- 
cytes which led to coagulation and thrombosis in the capillaries of 
the central nervous system and lungs. In the case of the transfusion 
of rabbit’s blood into dogs he attributed death to embolism in the 
pulmonary vessels due to “Massenhafte Verklebung der Kaninchen- 
zellen im Hundeblut”—or, in other words, to hemagglutination. 
Ponfick and others have disputed the validity of Landois’ con- 
clusions, but the basic principles of his explanations have been up- 
56 A brief historical review of this work can be found in the paper of 
Coca (1), Virchow’s Arch. f. path. Anat., 1909, Vol. 196, p. 92. 514 
INFECTION AND RESISTANCE 
held within recent years by workers who have gone over the same 
ground with the aid of more modern methods. Two careful re- 
searches have appeared during the last two years in which the prob- 
lem has been approached by different routes, but in which the gen- 
eral conclusions show much agreement. Coca,57 investigating the 
cause of death following the injection of washed blood cells into ani- 
mals of different species, concludes that in these cases death is due to 
mechanical obstruction of the pulmonary circulation owing to ag- 
glutination of the injected cells. It is important to note, however, 
that he adds in his conclusions the following paragraph: “The 
mere presence of specific agglutinins does not suffice, in the injection 
of ‘toxic’ erythrocytes, to occlude the pulmonary circulation. The 
cooperation of another factor must be assumed—-a factor probably 
found in the capillary walls.” 
Loeb, Strickler, and Tuttle,58 investigated the cause of death fol- 
lowing the injection of normal dog and beef sera into rabbits. They 
correlated their animal experiments carefully with the action of the 
sera in vitro upon the blood elements of rabbits, and utilized the 
property of hirudin to inhibit the coagulation of blood, finding, in 
the case of dog serum, that injections of hirudin, while not alwajs 
preventing death, at any rate prolonged life or necessitated an in- 
crease in the lethal dose. The conclusions of these authors are as 
follows: “Death following the injection of foreign serum is brought 
about by obstruction of the pulmonary circulation either by heaps of 
agglutinated erythrocytes or by fibrinous plugs. Dog serum and beef 
serum represent two different types. In the case of dog serum he- 
molysis of the blood cells of the recipient liberates substances attached 
to the stromata, which hasten coagulation. In consequence fibrin is 
formed which is carried into the pulmonary vessels and occludes 
them. In the case of beef serum death is due to hemagglutination.” 
The more recent understanding of the liberation of toxic bodies 
from blood cells by immune hemolytic sera, especially bv the experi- 
ments of Friedemann cited above, have rendered it likely that a 
similar anaphylatoxin formation from the cells of the recipient may 
play a role in the toxic action of normal sera. And it is a fact, in- 
deed, that such toxic sera are always hemolytic for the corpuscles of 
the susceptible animal. 
An analysis of the toxic action of certain normal sera from this 
point of view has been made by Uhlenhuth and Haendel,59 who, in 
studying the necrotizing action of beef serum injected into guinea 
pigs, attribute this action of the serum to a “complex process de- 
pending upon the cooperation of complement,” but not identical 
with the hemolytic mechanism. The toxic action of such serum, how- 
57 Coca. Virchow’s Archiv, Vol. 19(3, 1909. 
58 Loeb, Stricklev, and Tuttle. Virchow’s Archiv, Vol. 201, 1910. 
59 Uhlenhuth and Haendel. Zeitschr. f. Immunitatsforsch., Vol. 7, 1910, BACTERIAL ANAPHYLAXIS 
515 
ever, they separate from the necrotizing action, concluding that this 
is independent of complement, and more thermostable than either 
the mechanism causing necrosis or that responsible for hemolysis. 
Studies of the writer 60 on the toxic action of goat serum for rab- 
bits have shown that, contrary to Loeb, Strickler, and Tuttle, hemag- 
glutination and blood coagulation can be excluded as causes of death 
in some instances and that, in agreement with fjhlenhuth and Haen- 
del, the toxic action may be due to action of the serum not neces- 
sarily identical with the hemolysins. Unlike Uhlenhuth and Haen- 
del, however, it seemed clear that the participation of alexin was 
definitely necessary—the process being probably entirely analogous 
to Friedemann’s results with immune hemolytic (cytolytic) sera. 
The poisonous action of dissolved hemoglobin could be excluded. 
Thus, it would seem that foreign serum can be injurious to any 
given species of animals in a number of ways. Before any other 
mechanism is suggested, it would seem of the greatest importance 
always to exclude thrombosis by hemagglutination. In addition tc 
this, however, processes of hemolysis with the formation of toxic sub- 
stances must be considered as possibilities. From our own wTork we 
would further assume that a serum which has hemolytic properties 
for the red blood cells of another species is, at the same time and for 
the same reasons, probably capable of injuring the tissue cells of the 
injected animal by processes not visually observable, for our experi- 
ments have led us to believe that if normal hemolytic substances are 
present in such a serum, similar, perhaps antibody-like substances, 
for the remaining components of the cell protein may also be 
present. This is hard to prove, but is suggested by the wTork re- 
ported above, and would appear to us a logical possibility. Finally, 
it must not be forgotten that in such complex processes, alterations 
of the coagulation mechanism of the injected animal are not at all 
unlikely, and, as we have seen in the discussion of the anaphylatoxin 
theories, toxic action is also associated with processes which initiate 
coagulation. 
An Attempt at the Coordination of Various Phenomena of 
Hypersusceptibility.—In concluding the section of this book dealing 
with the phenomena of hypersusceptibility, we cannot resist the 
temptation of attempting briefly to present our views as to the pos- 
sible coordination of the various conditions. In protein anaphylaxis, 
itself, we have a condition in which the animal body becomes deli- 
cately susceptible to non-diffusible protein antigens to which specific 
antibody production occurs whenever the inciting substance pene- 
trates the physiological interior of the body. We have suggested 
that the physical inability of these substances to penetrate into the 
cell may possibly have causal relationship to their antigenic func- 
60 Zinsser. Jour. Exp. Med., Yol. 14, 1911. 516 
INFECTION AND RESISTANCE 
tions, the production of antibodies representing the mechanism by 
which the body cells eventually react with the non-diffusible proteins. 
This conception thus suggests a relationship between antigenic prop- 
erties, non-diffusibility and molecular size. The antigen-antibody 
mechanism, it can no longer be doubted, lies at the bottom of the 
mechanism of protein anaphylaxis, and that the reaction which 
causes injury is predominently localized upon the cellular elements 
of the body, is also beyond question. 
In serum sickness, we have a condition in which the antigenic 
nature of the inciting agent, the probable relation with antibody pro- 
duction, the coincidence of local and systemic symptoms, form such 
a striking analogy to anaphylaxis in lower animals that nothing but 
the most decisive reasons could persuade us that it is an unrelated 
process. Indeed, the occasional rapid appearance of the reaction 
on first injection, finds explanations which have at least as much basis 
as the complete throwing out of all the other analogies because of 
this one discrepancy; and the appearance of serum sickness without 
antibodies in the circulation is, as we have also seen, easy to explain. 
(Vide infra.) Moreover desensitization and parallelism with typical 
skin hypersensitiveness, further strengthen the analogy of serum 
sickness with protein anaphylaxis. 
In the food idiosyncrasies we have another set of conditions in 
which the inciting substance is antigenic, and in which a considerable 
proportion of the cases can be very definitely explained by sensitiza- 
tion, inasmuch as sensitization through the intestinal canal has been 
proved to be possible in guinea pigs. The peculiarities of this con- 
dition, we think will find a clearer comprehension, when we take 
more seriously into consideration the possibility of a localization of 
the sensitizing process in the intestine, due to a more prolonged con- 
tact of the antigen with the cells at the point of entrance. These cells 
are perhaps thereby much more highly sensitized than the body as a 
whole when antigen is injected intravenously, and reaches all the cells 
in a diluted condition. This is merely speculation; but views that 
tend to separate this condition from protein anaphylaxis are also very 
largely based on speculative reasoning from complex experimental 
results, and a trellis of speculation seems to us very useful in this 
subject, if not too dogmatically presented. In food idiosyncrasies, 
again, desensitization and a gradual return after this to the sensitive 
condition, link the process with anaphylaxis. Our views of the 
hereditary influence of this condition have been stated in a previous 
section. (See section on food idiosyncrasies.) 
If we look upon hay fever in the same way, practically all the 
remarks we have made about food idiosyncrasies are pertinent. Here 
the localized sensitization, both as to point of sensitization and as to 
development of symptoms, indicates a powerful influence of the 
Localization of the process, and in other respects, such as the heredi- BACTERIAL ANAPHYLAXIS 
517 
tary influence, skin reactions, and desensitizations, analogy is estab- 
lished. Here we are dealing, however, with a substance, the antigenic 
nature of which has been questioned. While the responsible sub- 
stances here may in the end be shown to be non-antigenic, we must 
regard this still as undecided, in view of the experiments of Parker 
in our own laboratory, in which it was proven that an antigenic 
material could be extracted from pollen, investigations which have 
been entirely neglected in the discussion of the subject by other 
writers. 
In bacterial hypersensitiveness, we see two processes, a true ana- 
phylaxis to the protein constituents of the bacteria, and another one, 
also specific, such as tuberculin, typhoidin, etc., reactions, which 
differ in some of their manifestations from true anaphylaxis, espe- 
cially in the more destructive and intracellular nature of the injury 
produced in the sensitized animals, and from the fact that infection 
with a constant feeding into the animal of certain bacterial products, 
seems to be a prerequisite for the acquisition of the sensitiveness. 
While this is probably not true in the extreme, yet in general spon- 
taneous occurrence it is the case. Here as we have shown, ourselves, 
the inciting substance involved is a material which, if it finally 
turns out to be a true protein, nevertheless differs in many important 
respects from the ordinary proteins. It is obtained by extraction 
of the bacteria, and left as a residue after boiling with acid, even in 
some cases after autoclaving in an acid state, in an attempt to remove 
all proteins, and we believe that these substances of which the tuber- 
culin extracts studied by us are an example, may perhaps represent 
the “Haptenes” foretold by Landsteiner. They react with antibodies, 
but, so far, have not incited them. With such fundamental differ- 
ences from ordinary proteins in the inciting substances, manifesta- 
tions and laws governing sensitization may well differ from those 
governing protein hypersensitiveness, and still be based upon a 
reaction of these substances with a cellularly placed antibody. It 
may be, though we have not yet been able to prove it, that the intra- 
cellular nature of the injury in these cases, as evidenced in tuberculin 
reactions in contrast with the ordinary urticarial protein cutaneous 
reactions, is related in some way to the diflfusibility of the materials 
in question. This, again, however, is largely speculative. 
In the drug idiosyncrasies, we have a group of reactions in which 
the inciting substance is clearly not antigenic in itself. There are 
two possibilities here; either that these substances have close analogy 
to protein anaphylaxis in that the drug in question undergoes a defi- 
nite chemical reaction with the blood and tissue proteins, producing 
an altered antigen, such as the methylated and azotized antigens of 
Pick, Landsteiner and others, and then act just like an antigen, sub- 
ject, perhaps, to laws varying in accordance with the intracorporal 
formation of the antigen; or the whole process may take place intra- 518 
INFECTION AND RESISTANCE 
cellularly, in that the drug diffuses into the cells, there entering into 
combinations which then become antigenic. This would be to some 
extent indicated by the apparent impossibility, hitherto, of passively 
transfering any of the drug hypersensitivenesses. However, we admit 
that the drug idiosyncrasies are the most vague, so far, and the most 
distant from true protein anaphylaxis in analogy. Nevertheless, the 
fundamental analogy remains, namely, that individuals may become 
either hypersusceptible or resistant, that is, relatively immune to 
the primary injurious action of the drugs; and it is important to 
note that when hypersusceptibility occurs, this hypersusceptibility is 
not evidenced by increased physiological action of the drug, but by a 
general type of reaction which for the species of animal, man, shows 
great similarity for a variety of drugs absolutely independent of the 
physiological action of the drug itself. 
Thus, while we are faced with a large variety of hypersensitive 
states and while classifications, such as those of I)oerr and Coca are 
of considerable preliminary value, it would be a pity, we believe, if 
the erection of such fences of division at the present time should 
influence investigators to overlook fundamental similarities. CHAPTER XX 
THERAPEUTIC IMMUNIZATION 
(Passive Immunization and Serum Treatment) 
Therapeutic Use of Diphtheria Antitoxin 
It is not consistent with the purpose of this brief treatise to dis- 
cuss extensively the therapeutic benefits obtained by serum therapy in 
diphtheria. We can convey briefly an adequate idea of this by citing 
some of the tables given by Northrup in NothnagePs “Encyclopedia 
of Practical Medicine,” American Edition, Volume on Diphtheria, 
etc., p. 143. These figures are taken from the statistics of the New 
York Board of Health, which began treatment of diphtheria with 
antitoxin in January, 1895. Dr. Northrup states, however, that 
serum treatment cannot be considered to have been in general use 
until some time later. 
Without Antitoxin 
Year 
Cases reported 
Deaths 
Mortality, per cent. 
1891 
5,364 
1,970 
36.7 
1892 
5,184 
2,106 
40.0 
1893 
7,021 
2,558 
36.4 
1894 
9,641 
2,870 
29.7 
Total 
27,210 
9,504 
Avg. 34.9 
With Antitoxin 
1895 
10,353 
1,976 
19.0 
1896 
11,399 
1,763 
15.5 
1897 
10,896 
1,590 
14.5 
1898 
7,173 
919 
12.8 
1899 
8,240 
1,085 
13.1 
1900 
8,364 
1,176 
14.0 
Total 
56,425 
8,509 
Avg. 15.0 
Table taken directly from Northrup, loc. cit. 
From this table there appears a reduction of 58 per cent, in 
mortality and a similar drop is evident from the German statistics 
of Dieudonne,1 from those of Welch, and many others. 
1 Dieudonne. Arb. a. d. kais. Gesund., Yol. 13, 1897. 
519 520 
INFECTION AND RESISTANCE 
It should be considered, moreover, in reading such statistics that 
they are made on gross mortality reports without elimination of the 
many cases that have not come under observation until too severely 
diseased to react to any form of treatment. The reason for the fail- 
ure to obtain results with antitoxin when the cases have proceeded 
beyond a certain stage of intoxication will become evident when we 
consider the manner of absorption of the poison in a succeeding para- 
graph. The mortality sinks to between 8 and 9 per cent., when such 
cases are omitted, as is shown by the collective investigations of the 
American Pediatric Society in 1896—figures which we take also 
from Horthrup’s comprehensive study. This purely statistical evi- 
dence, however good, is further reenforced by the unquestionable 
and considerable diminution of emergency operations,2 such as intu- 
bation and tracheotomy, since introduction of the antitoxin. More- 
over, there is the manifold clinical evidence of benefit, after the 
serum treatment, familiar to every practicing physician. 
Although the injection of antitoxin is of benefit by whatever 
route and in whatever quantity it may be given, nevertheless recent 
experimental investigations have taught us much regarding the 
proper use of this therapeutic agent. Especially interesting are the 
investigations of Meyer,3 who showed the extreme importance of an 
early use of the antitoxin. Apparently, as we have mentioned in 
another place, like tetanus antitoxin, the diphtheria poison may be 
in part absorbed directly by the nerves.4 
There is apparently a great difference in therapeutic efficiency, 
according to the method by which the serum is administered, a differ- 
ence probably depending upon speed of absorption. Berghaus 5 
showed that intravenous injection is 500 times more potent therapeu- 
tically than the subcutaneous, and 80 to 90 times more so than the 
intraperitoneal injection. Schick, for this reason, in discussing this 
problem from the clinical point of view, lays special stress upon the 
speed of administration. He says: “Hot only days but hours are 
of great importance.” He bases this opinion largely upon the fact 
that the toxin which has already united with the nerve substance can 
no longer be neutralized by antitoxin injections—perhaps owing to 
its union with constituents of the cells. 
According to the experiments of Meyer and Ranson diphtheritic 
paralysis may follow even when vigorous serum treatment has been 
employed. For, according to them, only the toxin which has reached 
the central nervous system through the circulation can be influenced 
2 Siegert. “Jahrbuch f. Kinderheilkunde,” Yol. 52, cited after Wernicke. 
3 Meyer. Berl. klin. Woch., Nos. 25, 26, 1909; Arch. f. exp. Path. u. Ther., 
Yol. 60, 1909, and Berl. klin. Woch., No. 45, 1911. 
4 For a thorough discussion of these conditions see Schick, Centralbl. f. 
Bakt., Rev. Vol. 57, 1913, “Report of 7th Meeting of the Mikrobiol. Gesell.,” 
Berlin, 1913. 
5 Berghaus. Cited from Schick, loc. cit. THERAPEUTIC IMMUNIZATION 
521 
by the serum, but no effect is possible upon the fraction which has 
been absorbed from the nerve endings directly. 
Schick,6 on the basis of extensive experiments, comes to the con- 
clusion that the subcutaneous injection of 1,000 to 2,000 units in 
diphtheritic cases has an immunizing value which protects the tissues 
from further injury and leads to cure only if, at the time of injec- 
tion, the lethal dose has not yet united with the sensitive cells. “If,” 
he states, “we wish to obtain antitoxic action upon toxin which has 
already gone into action before the injection of the serum, then re- 
sults can be obtained both in man and in animals only if a great deal 
of antitoxin is injected intramuscularly or intravenously.” 7 
Interesting also from a clinical point of view are the studies of 
Schick,8 Hahn,9 and others 10 upon the presence of antitoxin in the 
blood of normal, untreated individuals at different ages. These 
investigations were carried out by the intracutaneous method of 
toxin and antitoxin determination described in greater detail in a 
later section. The following table, taken from the article of Hahn, 
illustrates the experience, in such investigations, both of Schick and 
of Hahn himself. The determinations were carried out upon indi- 
viduals who had never had diphtheria, as far as could be learned. 
Cases with 
Cases without 
Highest antitoxin 
Age 
antitoxin serum 
antitoxin serum 
value in 1 c. c.* 
Schick.. 
•H 
r Newborn 
0—1 year 
11 
1 
0 
3 
under 1.5 units 
0.11 unit 
r 2-10 years.. 
7 . 
5 
1.0 unit 
11-20 years.. 
8 
9 
0.75 unit 
Hahn... 
21-30 years.. 
9 
5 
2.5 units 
31-40 years.. 
5 
1 
0.25 unit 
41-65 years.. 
2 
8 
2.5 units 
* Table taken directly from Hahn, loc. cit. 
The table shows that in newborn children there is almost regu- 
larly a definite and sufficient protective value in the serum which 
diminishes up to the first year, so that at the end of the first year 
three out of four individuals have no antitoxin in their serum. In 
subsequent years up to the age of 40 an increasing percentage of 
people have sufficient amounts of diphtheria antitoxin in their blood. 
After the age of 40 an increasing percentage is without such protec- 
tion. The first observation, that newborn children usually possess 
considerable amounts of antitoxin, is very probably due to passive 
8 Schick. Loc cit. 
7 Schick. Loc. cit., p. 32. 
8 Schick. “Uber Diphtherimmunitat,” Wiesbaden, 1910. 
9 Hahn. Deutsche med. Woch., Yol. 38, 1912, No. 29, p. 1366. 
10 Karasawa and Schick. Zeitschr. f. Kinderkrankheiten, 1910, and “Jahr- 
buch. f. Kinderheilkunde,” 1910. 522 
INFECTION AND RESISTANCE 
immunization by the blood of the mother, a fact which we have men- 
tioned in another place. The original method by which such meas- 
urements can be made is described in a subsequent section on the 
intracutaneous method of determining toxin and antitoxin. 
Another convenient method of titrating antitoxin in human 
serum, colostrum, milk, etc., has been described by Kellogg.11 This 
method has been used with much convenience and accuracy in our 
laboratory by Kuttner and Ratner. The preliminary measurements 
on which Kellogg’s method is based are the following: In guinea 
pigs 1/300 of a M L D gives a definite intracutaneous reaction which 
is at its height in about 48 hours; 1/40 of a M L D is the smallest 
amount of toxin that will produce necrosis. Since the L -j- dose of 
toxin is the amount that will neutralize one standard unit of anti- 
toxin, leaving one minimal lethal dose of toxin free, it follows that if 
we make a series of mixtures in which the fraction of the L -f- dose is 
accurate, known and constant with varying amounts of patient’s 
serum, milk, colostrum or other materials and with such mixtures 
carry out skin reactions on guinea pigs we can determine, wfith fair 
accuracy, the antitoxic strength of the specimen in units of antitoxin. 
Kellogg makes a mixture of toxin in which the amount contained 
in 1 c. c. equals 1/30 the L + dose. Now, if 1 c. c. of such a toxin 
mixture is mixed with 1 c. c. of serum containing 1/30 unit of anti- 
toxin, this will leave a balance of 1/30 minimal lethal dose, and mix- 
tures of 0.1 c. c. of each will have a surplus of 1/300 minimal lethal 
dose of toxin. The volume of mixture injected is 0.2 c. c., which, 
therefore, represents 1/300 of the L -}- dose of antitoxin, plus what- 
ever antitoxin may be contained in 0.1 c. c. of the serum. 
Readings can be made from 48 to 72 hours. If the serum con- 
tains exactly 1/30 a unit of antitoxin per c. c., which amount, it will 
be remembered, is the arbitrarily determined one that is supposed to 
still give protection in human beings, 1/300 minimal lethal dose of 
toxin will remain free, and will still give the mild reaction deter- 
mined for this. 
If the amount of antitoxin is greater than this, no reaction at all 
will appear. If the amount of antitoxin is less than this, necrosis 
will be evident. By titrating in one direction or another, a fairly 
accurate idea of the antitoxin content of the serum can be obtained. 
Appropriate controls are, of course, always made. 
The work of Schick, that of J. Henderson Smith, and recent 
studies by Park and Biggs promise to alter considerably the methods 
of antitoxin therapy as at present in use in diphtheria. Smith meas- 
ured the speed of absorption of antitoxin injected subcutaneously into 
the abdominal wall of a healthy man. His results are shown in the 
following table, which we take from his communication (page 213) : 
11 Kellogg. Jour. A. M. A., June 10, 1922. Kellogg’s method is really a 
convenient modification of the method of Roiner which is described on p. 536. THERAPEUTIC IMMUNIZATION 
523 
Table Y 
One c. c. of the patient’s serum contained: 
Before injection No demonstrable antitoxin 
5 hours after injection 0.1 unit antitoxin 
14 hours after injection 0.225 unit antitoxin 
32 hours after injection 0.68 unit antitoxin 
44 hours after injection 1.0 unit antitoxin 
3 days after injection 1.3 units antitoxin 
4 days after injection 1.3 units antitoxin 
6 days after injection 0.68 unit antitoxin 
13 days after injection 0.17 unit antitoxin 
15 days after injection 0.14 unit antitoxin 
20 days after injection 0.08 unit antitoxin 
27 days after injection No demonstrable antitoxin * 
* J. Henderson Smith. Jour. Hyg., Yol. 7, 1907, p. 205. 
Park and Biggs12 have made similar studies and have contrasted 
the speed of absorption after subcutaneous administration with that 
after intravenous injection, basing their curves upon careful measure- 
ments of the sera of the treated patients. We reproduce their charts 
as given in their recent publication. 
It is apparent from these charts, as well as from the work of Hen- 
derson Smith, that antitoxin, subcutaneously given, is slowly ab- 
sorbed, and does not reach its maximum concentration in the blood 
stream until forty-eight hours or more after the injection. It fol- 
lows that, as Park and Biggs point out, it is more rational to inject 
a single adequate dose than to divide the dosage and inject at inter- 
vals. They have obtained results in animal experiment which 
graphically illustrate this principle. A rabbit which had received 
ten fatal doses of toxin intravenously was given a total of 50 anti- 
toxin units in divided doses as follows: 100 units after twenty min- 
utes, 100 after 40 minutes, and 150 units each after 60 and 80 min- 
utes. This rabbit died. Another animal given the same dose of 
toxin received 200 units of antitoxin twenty minutes later and lived. 
The amount necessary to save life in rabbits receiving ten fatal doses 
intravenously was as follows: 
Given after 10 minutes 5 units antitoxin 
Given after 20 minutes 200 units antitoxin 
Given after 30 minutes 2,000 units antitoxin 
Given after 45 minutes 4,000 units antitoxin 
Given after 60 minutes 5,000 units antitoxin 
Given after 90 minutes No amount 
These extremely important experiments of Park and Biggs hear 
ont the opinion of Schick and show beyond question that the proper 
way to give antitoxin is to give a single adequate dose as early as 
12 Park and Biggs. Collected Studies from the N. Y. Department of 
Health, Bureau of Laboratories, Yol. 7 1912-1913, p. 27. 524 
INFECTION AND RESISTANCE 
Chart I.—Showing the extent and rapidity of absorption of 10,000 units of 
antitoxin given subcutaneously. Each line represents the antitoxin content 
of 1 c. c. of blood at different intervals of time. (From Park and Biggs, 
loc. cit.) THERAPEUTIC IMMUNIZATION 
525 
Chart II.—The antitoxic power of human blood after an intravenous injection 
of 10,000 antitoxic units. (From Park and Biggs, foe. cit.) 
possible. They emphasize the fact that probably the most important 
single point in the specific therapy of diphtheria is the speed with 
which the diagnosis can be made and the antitoxin given. At the De- 
partment of Health the dosage now employed, as given by Park and 
Biggs, is the following: 
Units in Cases 
Mild 
Moderate 
Severe 
Very Severe 
Infants under 1 year 
2,000 
3,000 
10,000 
10,000 
Children 1 to 5 years 
3,000 
5,000 
10,000 
10,000 
Children 5 to 9 years 
4,000 
5,000 
10,000 
15,000 
Persons over 10 years 
5,000 
10,000 
10,000 
20,000 
Practical Considerations Connected with Diphtheria Anti- 
toxin Production and Standardization 
The conditions which govern the active production of toxins by 
bacteria in culture media are not only of great theoretical interest 
but possess unusual practical value in that the most important factor 
for the successful production of a strong antitoxin consists in the 
preliminary preparation of a potent toxin. The bacterial true toxins 
are all “exotoxins” in that they are soluble, moderately diffusible 
substances which pass readily from the bacterial bodies to the en- 
vironment, and for this reason can be obtained most readily by the 
cultivation of the bacteria upon fluid media and subsequent filtration 
of the cultures through earth or porcelain filters. 
The choice of culture or strain is an important element in this 
procedure, since within the same species of toxin-producing micro- 526 
INFECTION AND RESISTANCE 
organisms there is much variation in the speed and energy of toxin 
production. Thus for unknown reasons some strains of diphtheria 
bacilli will far outstrip others in this respect. An excellent illustra- 
tion of this is the experience of Park and Williams 13 with two diph- 
theria cultures—a very virulent and a very weak one. Of the 
former, 0.002 c. c. of a forty-hour bouillon culture killed a guinea 
pig, while of the latter 0.1 c. c. of a similar culture was necessary 
for the same result. 
In the case of tetanus, cultural differences do not seem to be as 
common. Individual strains also may gain or lose in toxin-pro- 
ducing powers, according to the method of handling them which is 
practiced. It is stated,14 for instance, that a diphtheria culture will 
lose in energy of toxin production if permitted to grow without suffi- 
ciently frequent transplantation. However, transplanted on solid 
media with reasonable frequency, these bacteria show a remarkably 
constant toxin production. A well-known strain, the Park-Williams 
No. 8, now in use in many antitoxin laboratories throughout the 
world, has persisted for over 15 years in producing a strong toxin. 
There are occasional strains among toxin-forming species which are 
entirely devoid of this property. Diphtheria bacilli which were 
virulent while possessing all the other cultural characteristics of the 
group have been described, but appear, from the experience of the 
writer, to be rather uncommon.15 Of tetanus bacilli little is known 
in this respect. 
Given a powerfully toxic strain of the proper bacteria the 
method of cultivation is also of great importance in influencing the 
eventual yield of poison. These relations have naturally been stud- 
ied with the greatest care in the case of diphtheria and tetanus 
bacilli, since in these cases there has been the greatest practical ap- 
plication for such knowledge. 
In the case of diphtheria, though toxin will be produced on all 
media on which the bacillus grows easily, the most favorable medium 
for this purpose is a slightly alkaline broth made of lean beef or 
veal infusion and containing peptone. Since acid formation hinders 
the production of toxin, Martin 16 has suggested fermentation of the 
muscle sugar with yeast, while Theobald Smith 17 recommends pre- 
liminary fermentation with Bacillus coli. 
Park and Williams 18 regard this as unnecessary. They recom- 
mend a 2 per cent, peptone broth made of veal. This is neutralized 
to litmus and 7 to 9 c. c. of normal NaOH solution to the liter are 
13 Park and Williams. “Pathogen. Micro-organ.,” New York, 1910. 
14 Park and Williams. Loc. cit. 
15 Zinsser. Jour. Med. Res., N. S., Vol. 12, 1907. 
16 Martin. Ann. de VInst. Past., 1896. 
17 Th. Smith. Jour. Exp. Med., Yol. 4, 1899, p. 373. 
18 Park and Williams. Jour. Exp. Med., Yol. 1, 1896. THERAPEUTIC IMMUNIZATION 
527 
added. In such a medium at 37.5° C. the production of toxin begins 
within 24 hours and reaches its highest point in from five to ten 
days. When at its height the process must be stopped and the cul- 
tures exposed to a lower temperature, otherwise rapid deterioration 
takes place because of the instability of the toxin. Even when kept 
cold and in the dark this deterioration proceeds steadily though 
slowly. At first, however, even under these conditions a compara- 
tively extensive loss of toxin goes on—a process sometimes spoken 
of as “maturing of the toxin”—after which the poison strikes a 
fairly constant and very gradual rate of weakening, and is, com- 
paratively speaking, stable. 
Apparatus Arranged for the Sterile Filtration of Diphtheria Cultures 
in Toxin Production. 
(After Rosenau, U. S. Hyg. Lab. Bull. 21, 1905, p. 38.) 
In the United States Hygienic Laboratory in Washington, ac- 
cording to Rosenau, the recommendations of Theobald Smith are 
largely followed in the production of toxin. The procedure is as 
follows: 
The culture medium, “Smith’s Bouillon,” is prepared from 
chopped beef from which fat and tendon have been cut out. This is 
adjusted by plenolphthalein titration to 0.5 per cent, acidity. It is 
then placed into Fernbach flasks and inoculated on the surface with 
a Park-Williams bacillus Ho. 8. The flasks are incubated for 7 days 
at 37.5° C. The reaction of the medium after such incubation is 
determined, and flasks showing an acidity of 1.5 or over are dis- 
carded. The usual reaction at the end of incubation is 0.6 to 0.8 
per cent, acidity. This broth is filtered through Berkefeld filters or 
porcelain candles. 528 
INFECTION AND RESISTANCE 
Toxin so prepared is now tested and its Lo and L+ doses deter- 
mined by the methods described above. Rosenau 10 states that poi- 
sons are discarded as containing too large a proportion of toxon if 
the difference between L0 and L+ is greater than 15 M L D. The 
toxin is now set aside in flasks for the process which Rosenau calls 
“seasoning.” At intervals of about a month it is retested and finally 
it is found that the rate of toxoid formation decreases and the poison 
reaches a period of equilibrium. It can now be used for accurate 
determination of the L+ dose, and this is done from careful measure- 
ments on a large number of guinea pigs. 
Examples 20 of such measurements, abbreviated for the sake of 
simplicity, are given in the following tables: 
Toxin Determinations of M L D or “T” 
Dose in o. c. Result 
0.03 = death in V/i days 
0.02 = death in days 
0.01 = death in 2 days 
0.008 = death in 3 days 
0.006 = death in days 
0.005 = death in 4 days M L D 
0.004 = death in 6 days 
0.003 — death in 8 days 
0.002 = late paralysis 
0.001 = well in 16 days. 
Toxin Determination of L+ Dose 
1 Antitoxin unit + 0.2 c. c. = 0 
1 Antitoxin unit + 0.21 c. c. = 0 = L0 
1 Antitoxin unit + 0.22 c. c. = local infiltration 
1 Antitoxin unit + 0.23 c. c. = fatal in 17 days 
1 Antitoxin unit + 0.24 c. c. = fatal in 14 days 
1 Antitoxin unit + 0.26 c. c. = fatal in 9 days 
1 Antitoxin unit + 0.28 c. c. = fatal in 6 days 
i Antitoxin unit + 0.29 c. c. = fatal in 4 days = L+ 
1 Antitoxin unit + 0.3 c. c. = fatal in 3 days 
Tlie production of antitoxin is carried out by the graded injec- 
tion of antitoxin into horses. Young, healthy horses are chosen, 
tested for freedom from glanders, and the first injections are made 
either with toxin attenuated by the addition of Lugol’s solution or 
terchlorid of iodin, or, as in the New York Health Department, the 
first injections consist of mixtures of toxin and antitoxin. We take 
our description largely from the account given by Park.21 The first 
injection consists of 12 c. c. of toxin (M L D 1/400 c. c.), together 
with 100 units of antitoxin. After the reaction from such an injec- 
tion has completely subsided—after 3 to 5 days—a second injection 
19 Rosenau. Hyg. Lab. Bull. No. 21, April, 1905. 
20 Examples are taken from measurements reported by Rosenau, loc. cit. 
21 Park and Williams. “Pathogenic Bacteria,” p. 213. THERAPEUTIC IMMUNIZATION 
529 
is given of toxin without antitoxin; then 15 c. c., 45 c. c., 55 c. c., 
65 c. c., 80 c. c., 95 c. c., 115 c. c., 140 c. c., etc., the intervals be- 
tween injections being about three days and depending upon the 
reaction of the horse and the speed with which it entirely recovers 
from the preceding injection. In a particular case cited by Park 
675 c. c. of toxin could be given by the 60th day; in this case by the 
28th day the horse was yielding 225 units to the c. c.; on the 40th 
day, 850 units; on the 60th day, 1,000 units. 
The determination of the antitoxin unit, carried out from time 
to time on the serum of such a horse against the L+ dose described 
in our preceding table, would be carried out as follows: 
In all such standardization great care must be taken in employ- 
ing accurately standardized glassware. Rosenau recommends em- 
ploying “capacity instruments” rather than “outflow instruments.” 
Dilutions of unknown antitoxin are made in 0.85 per cent, sterile 
salt solution. As a basic dilution one part of the antitoxic serum to 
nine of the salt solution gives 1/10 c. c. to each cubic centimeter, 
and from this initial dilution further dilutions may be easily made 
as follows: 1 c. c. of dilution I. -f 9 c. c. salt solution = 1-100, etc. 
A series of mixtures is then made in each of which the quantity of 
toxin equals the L+ dose, and in which the quantity of antitoxin 
varies within a wide margin of the limits of strength to be expected. 
This is illustrated in the following table: 
L+ (0.29 c. c.) -f 1/500 c. c. of antitoxic serum = lives 
L+ (0.29 c. c.) + 1/600 c. c. of antitoxic serum = lives 
L+ (0.29 c. c.) -j- 1/700 c. c. of antitoxic serum = lives 
L+ (0.29 c. c.) + 1/800 c. c. of antitoxic serum = dies in 8 days 
L+ (0.29 c. c.) + 1/900 c. c. of antitoxic serum = dies in 4 days 
L+ (0.29 c. c.) + 1/1,000 c. c. of antitoxic serum = dies in 2 days 
In the above tables, according to onr previous definition of the 
antitoxin unit, the serum would contain 900 units to the cubic centi- 
meter, since 1/900 c. c., injected together with the L+ dose of the 
standard toxin, resulted in the death of the guinea pig in four days. 
In order to allow a margin of safety Rosenau and others have sug- 
gested that the unit should be determined, not by the quantity of 
antitoxin, which delays death by the L+ dose for four days, but 
rather by the quantity which, with the L+ dose, results in saving 
the life of the guinea pig. According to this latter standard the 
serum employed in the table would be spoken of as containing 700 
units to the cubic centimeter. Of course the tabulated measurements 
are rough, leaving an undetermined zone of 100 units. The exact 
number of units to the cubic centimeter could, of course, be deter- 
mined with greater accuracy by now carrying out another series of 
tests in which the amount of serum varied between 1/700 and 1/900 
of a cubic centimeter. 530 
INFECTION AND RESISTANCE 
In carrying out such a standardization the toxin is diluted so 
that the L+ dose is contained in 2 c. c. This can easily be done. 
For instance, in the above the L+ dose being 0.29, it merely necessi- 
tates adding to each 0.29 c. c. of toxin 1.71 c. c. of salt solution, 
to each 2.9 c. c., 17.1 c. c. The antitoxin also is made up in such a 
way that the required dilution is contained in two cubic centimeters, 
since a total volume of 4 c. c. has been agreed upon as standard for 
these tests, the injected volume having much influence upon the speed 
of absorption. In using the so-called Rosenau syringe, shown in the 
figure for the standardization of antitoxin, the antitoxin is made up 
to 1 c. c. in each case, so that 1 c. c. of salt solution may be added to 
wash out the syringe after injection of the mixture. The mixtures 
Battery of Rosenau Syringes Prepared for Antitoxin Standardization, 
(Taken from Rosenau, U. S. Hygienic Bulletin 21, 1905.) 
can be made directly in these syringes or in test tubes, and are 
allowed to stand one hour at room temperature, so that there 
may be time for complete union. If the mixtures are made directly 
in the syringes the needles are dipped into sterile vaselin, which 
closes them and prevents leakage while standing. The mixture is 
then forced out of the syringe with a rubber bulb, thus ensuring 
complete injection of all the fluid. As Rosenau states, much depends 
on the guinea pigs. They must be of standard weight, about 250 
grammes, well fed and cared for, and must not be descendants of 
pigs that have shown marked or unusual resistance to diphtheria 
toxin. This, as Theobald Smith has shown, occasionally happens. 
The antitoxic serum as obtained from the horse directly may 
be concentrated in a number of ways, representative of which is the 
method developed at the Hew York Department of Health by Gib- 
son,22 Banzhaff, and others.23 The original method consisted in 
22 Gibson. Jour. Biol. Chem., Yol. 1, 1906. 
23 Gibson and Collins. Jour. Biol. Chem., Yol. 3, 1907. THERAPEUTIC IMMUNIZATION 
531 
heating horse serum to 56° C. for 12 hours, by which some of the 
pseudoglobulin was converted into euglobulin, the antitoxin remain- 
ing in the pseudoglobulin fraction. After this an equal volume of 
saturated ammonium sulphate solution is added and the globulin 
precipitated. After several hours the precipitate is filtered off and 
again taken up in water corresponding in amount to the original 
volume of serum. After' filtration this solution is precipitated with 
ammonium sulphate and this precipitate is treated with saturated 
solution of ]STaCl in quantity twice that of the original serum. After 
standing for 12 hours the supernatant fluid containing the antitoxin 
is decanted, and this is precipitated with 0.25 per cent, acetic acid. 
The resulting precipitate is dried by pressing it between filter papers 
and is placed in a parchment dialyzing bag, after neutralization with 
sodium carbonate. At the end of seven or more days of dialyzation 
against running water, the globulin solution remaining in the dial- 
yzer is filtered and made isotonic. 
More recently the method as modified by Banzhaff is as follows: 
The serum, as obtained from the horse, is diluted by one-half the 
volume of water, and to this a saturated solution of ammonium 
sulphate is added up to 30 per cent, saturation. This is heated to 
61° C. for two hours. It is then filtered and the residue on the filter 
paper, which contains the antitoxin, is thoroughly dried by pressing 
between filter papers and is directly dialyzed. 
Observations by Park and Throne 24 have shown that this con- 
centrated antitoxin which, according to Gibson, represents a yield of 
about 70 per cent, original antitoxic power of the serum, is equally 
efficient for therapeutic purposes as an unconcentrated preparation 
and has the advantage of introducing less foreign protein into the 
human body. It retains its potency, according to Park and Throne, 
as long as does the whole serum. 
Active Immunization in Diphtheria with Mixture of Toxin 
and Antitoxin 
Recently Behring 25 has advocated the immunization of human 
beings with mixtures of diphtheria toxin and antitoxin. This 
method represents essentially active immunization with toxin ren- 
dered harmless by neutralization with antitoxin. The use of such 
mixtures had previously been studied with considerable care, in the 
case of the toxin of symptomatic anthrax, by Schattenfroh and 
Grassberger,26 and the procedure had been used in the New York 
Department of Health for some years in the initial treatment of 
24 Park and Throne. Amer. Jour. Med. Sci., Yol. 132, 1906. 
25 Behring. Deutsche med. Woch., Yol. 39, No. 19, 1913. 
26 Schattenfroh and Grassberger. Deuticke, Wien, 1904; see also Schat- 
tenfroh, Wien. klin. Woch., No. 39, September, 1913. 532 
INFECTION AND RESISTANCE 
antitoxin horses. Theoretically considered on the basis of Ehrlich’s 
opinions, one would be inclined to wonder at the fact that relatively, 
neutral mixtures of toxin and antitoxin should possess any antitoxin- 
inciting properties. Behring explains the immunizing value of such 
mixtures by the reversible nature of toxin-antitoxin union in the 
animal body. Ho calls attention to the fact that our analyses of 
diphtheria toxin-antitoxin mixtures have been made entirely with 
guinea pigs as indicators. In studying such mixtures in other ani- 
mals Behring has come to the conclusion that complete detoxication 
of the poison in vitro does not occur. He found, for instance, that a 
toxin-antitoxin mixture that was entirely innocuous for guinea pigs 
produced an active febrile reaction in an ass. In monkeys (Maca- 
cos rhesus) he finally found an animal in which lie obtained evidence 
satisfactory to him that toxin may be powerfully active in the animal 
body, even if it has been previously mixed with antitoxin. If, for 
instance, he gave a monkey a mixture in which as much as 20 to 40 
antitoxin units were mixed with one toxin unit, and repeated the 
injection two or three times, the animal died of subacute diphtheria 
toxin poisoning. The mixture ceased to be poisonous for monkeys 
only when the relation of antitoxin to toxin became one of 80 to 100 
antitoxin units to one toxin unit. This final detoxication when suffi- 
cient amounts of antitoxin were used, it seems to us, may be taken 
as sufficient evidence that Behring’s monkeys did not die of ana- 
phylaxis. 
We gather from Behring’s writings that he attributes these dif- 
ferences in susceptibility to toxin-antitoxin mixtures in various ani- 
mals to differences in the reversibility of the toxin-antitoxin com- 
plex in the bodies of the individual species. 
Human beings are less susceptible to such mixtures than are 
monkeys, but nevertheless more so than guinea pigs. It also appears 
that diphtheria bacillus carriers or such persons who, because of a 
previous infection, have antitoxin in their blood are much more sus- 
ceptible to these mixtures than are others. Newborn children are 
less susceptible than are children from 4 to 15 years. Mixtures 
which are entirely neutral for the newborn may incite febrile reac- 
tion in older children. In all cases the injection of such mixtures is 
followed by a more or less active production of antitoxin. 
The mixtures which von Behring advocated were so prepared 
that the toxin action upon guinea pigs is practically nil; in other 
words, the mixture was completely neutralized. 
The method represents in purpose, and apparently in achieve- 
ment, a safe process of actively immunizing against diphtheria. 
Heretofore the method of protecting human beings prophylactically 
against diphtheria has consisted in the injection of antitoxic serum. 
This, unquestionably a wise procedure, has nevertheless the disad- 
vantage of bringing about an immunity of short duration only. THERAPEUTIC IMMUNIZATION 
533 
Within 20 to 30 days the antitoxin injected may have completely 01 
almost completely disappeared from the blood stream. Prophylactic 
immunization with the toxin-antitoxin mixtures, however, repre- 
senting as it does an active immunization, is likely to be more pro- 
longed in its effects. According to Behring a human being possess- 
ing 0.01 antitoxin unit in 1 c. c. of blood may be regarded as still 
moderately protected against diphtheria. According to his estima- 
tion a decline to this amount, in a person actively immunized by the 
mixtures (an estimation based upon curve measurements of treated 
cases), would take about two years. He has observed that horses 
that had been actively immunized by him, and subsequently used in 
agricultural work, retained measurable antitoxin values in their 
blood after five years without treatment. 
Schreiber 27 and others state, also, that this method of active 
immunization with mixtures of toxin and antitoxin has the advan- 
tage of avoiding the anaphylactic dangers incident to the injection 
of antitoxin alone. Their opinion is probably erroneous, since it is 
most likely that whatever anaphylactic dangers there are result from 
the injection of horse serum rather than from the antitoxin con- 
tained in the injected substance. Moreover, the recent studies of 
Park have shown satisfactorily that the danger of anaphylaxis in the 
injection of antidiphtheritic sera is practically nil. Among 330,000 
cases on record there were but five deaths. 
The chief value of this new method of immunization is that it 
represents a safe technique for the prophylactic treatment of indi- 
viduals exposed to the disease and possibly for the general prophv- 
lactic immunization of school children, nurses, physicians, etc. In 
the case of children during the ages at which they are most suscep- 
tible to the disease, the prolonged immunity resulting from the treat- 
ment should strongly recommend it as a method of promise for the 
gradual eradication of epidemics. Behring also suggests it as a 
hopeful method of treatment in the case of bacillus carriers. 
Schreiber and others have reported upon the effects of treatment 
when carried out with Behring’s mixtures. In the earlier experi- 
ments of Hahn, mixtures were used in which there was a slight excess 
of toxin. The later experiments were made with mixtures which 
were completely neutralized for guinea pigs. In Sclireiber’s cases 
from two to six injections were made at intervals of three to five 
days, most of them subcutaneously, and some of them intramuscu- 
larly. In no case were there serious reactions, although occasionally 
there were slight swelling of regional lymph nodes and a little fever. 
The effects of immunization were noticeable about 23 to 25 days 
later. When two injections only had been made, at least 0.075 of an 
antitoxin unit to the cubic centimeter was present. The highest value 
obtained after two injections was one unit to one cubic centimeter. 
27 Schreiber. Deutsche med. Woch., Yol. 39, No. 20, 1913. 534 
INFECTION AND RESISTANCE 
In nine patients who had been treated by four to seven injections 
with gradually increasing doses, as much as 10 to 75 antitoxin units 
to the cubic centimeter resulted. It appears, therefore, that in med- 
ical practice this method is safe, and that with as little as two injec- 
tions antitoxin values may be obtained which entirely suffice for the 
protection of human beings against the ordinary dangers of diph- 
theria infection, an immunity which, as far as we can judge at 
present, may last about two years. 
Another advantage which Behring claims for his method is the 
production of homologous antitoxin in human beings for the passive 
immunization of other human beings. Mathes has tried this in 
children with the idea of thereby avoiding the dangers of anaphy- 
laxis. Incidentally it was claimed in this case that the passive im- 
munization, when carried out with homologous serum, lasted longer 
than did that conferred by horse serum. However, one case is 
hardly enough to establish such a fact. 
The Mixtures Actually Used.—The most extensive use of the 
Behring toxin-antitoxin immunization in diphtheria has been made 
by Park and his associates at the Hew York Department of Health. 
In carrying out this work, it has been found advisable from time 
to time to change the actual proportions of the mixtures used, as 
experience indicated the desirability of this. For a time the direc- 
tions given for the mixtures used in this treatment were given as 
follows: the toxin and antitoxin used should be at least three months 
old, from the day of planting and the day of bleeding, respectively. 
The L+ dose is determined with reference to the United States stand- 
ard antitoxin. The unitage of the antitoxin to be used should be 
determined against this antitoxin, so that one unit of the antitoxin 
to be used, when mixed with the L+ dose of this toxin, should permit 
the survival for 96 hours of 25 per cent, of the guinea pigs injected. 
When a tentative mixture of the product has been made, no further 
toxin should be added. Any changes made should consist, not in the 
addition of toxin, but in the reduction of toxicity by the addition of 
antitoxin, if necessary. Approximately, one unit of antitoxin is, 
therefore, added to start with, for each L+ dose in the bulk of toxin, 
and with each addition of antitoxin to the mixture, it shall be im- 
mediately thoroughly shaken for 20 minutes. By repeated test and 
fractional addition of antitoxin, 0.1 of a unit at a time, the mixture 
is brought to a point where 5 c. c. injected into a pig, permits sur- 
vival for 96 hours, and at this point the mixture is ready for filtering. 
After filtering, the mixture is preserved until at least one-half of the 
5 c. c. guinea pigs on each bulk bottle, survive 10 days, but show 
definite paralysis. The human dose of this was 1 c. c. 
Park, Schroder and Zingher 28 have recently changed this mix- 
28 Park, Schroder and Zingher. Jour. Amer. Pub. Health Assoc., No 
13, 1923, p. 23. THERAPEUTIC IMMUNIZATION 
535 
ture in the direction of less toxicity and, therefore, less danger of 
severe reaction. They experimented with various mixtures, in all 
of which there were different amounts of toxin and antitoxin, but 
all of them striking such a balance between the two, that 1 c. c. caused 
paralysis in 300 gram guinea pigs, and 5 c. c. caused death in guinea 
pigs within 10 days. These mixtures were made on the basis of 
Park’s statement that the best mixture is one that is underneu- 
tralized, and yet perfectly safe. As the mixture stands the toxin 
deteriorates faster than the antitoxin, so that it gradually deteriorates 
in usefulness. Apparently annoying reactions are obtained because 
of protein reactions due to diphtheria bacillus cell substances found 
in the toxin. In using mixtures such as those described, in which 
the balance between toxin and antitoxin was the same, but the actual 
total amount of toxin was reduced, they found that the different 
preparations gave the same immunizing results, but those having the 
least amount of toxin and, therefore, the least amount of accompany- 
ing bacillus substance, showed the least amount of reaction. The 
following table published by Park and his associates gives results 
obtained with mixtures made up in this way. 
TABLE II29 
The antitoxin development produced by three injections of mixtures hav- 
ing different amounts of toxin and antitoxin but all causing severe paralysis 
in guinea pigs receiving doses of 1 c. c. and death within ten days in those 
receiving: 5 c. e. 
Amount of Original 
Per cent, of Nonimmunes Shown 
No. of School Children To Be Immune on Schick 
Toxin in 1 c.c. of Mixture 
Receiving 3 Injections 
Retest 4 Months Later 
*1/10 L + ( 4 lethal doses).. 
490 
90% 
x/z L + ( 20 lethal doses).. 
304 
95% 
3L + (120 lethal doses).. 
318 
92% 
5 L + (200 lethal doses).. 
487 
85% 
* The mixture is made by adding three-fourths of a unit of antitoxin to one L + dose 
of toxin. The toxin and antitoxin should be diluted in cold water and the two solutions 
mixed immediately. If the toxin is diluted in water at room temperature it deteriorates 
rapidly. 
TABLE III 
Comparison as to the amount of local and constitutional reaction caused 
by the new and old preparation. 
No Local Reaction 
* New Preparation 
1/10 L + 
25% 
Old Preparation 
3 to 5 L + 
0% 
Slight Local Reaction 
64% 
41% 
Moderate Local Reaction 
11% 
37% 
Marked Local Reaction 
0% 
22% 
Of those showing marked reactions there 
was a rise of 1° to 3° F., and other con- 
stitutional symptoms in 
0% 
6% 
* If the 1/10 L + preparation is underneutralized there will be a local reaction from 
the excess of toxin. 
29 Taken directly from Park, Schroder and Zingher, loc. cit. INFECTION AND RESISTANCE 
536 
The Intracutaneous Method of Determining Toxin and 
Antitoxin Values 30 
Marks 31 was the first to utilize the prevention of local edema or 
injury for the determination of antitoxin values. He mixed diph- 
theria antitoxin and toxin and injected them subcutaneously into 
guinea pigs, claiming that this method was considerably more deli- 
cate than the Ehrlich method, since the amount of toxin capable of 
causing localized edema amounted to as little as one-twentieth of a 
minimal lethal dose. This method has many points in its favor, and 
has been recently utilized and improved upon by Ilomer. 
Romer 32 has developed a method of diphtheria antitoxin stand- 
ardization which depends upon intracutaneous injections into guinea 
pigs. The principle of this test consists in the observation that, when 
very slight amounts of diphtheria toxin are injected intracutaneouslv 
into the abdominal skin of guinea pigs, small areas of local necrosis 
result within about 48 hours. When such injections are made with 
mixtures of toxin and antitoxin the presence of free toxin is indi- 
cated by the appearance of such necrosis. 
Before proceeding to the standardization by this method it is 
necessary to determine the “limes-necrosis” (just as Ehrlich deter- 
mines his L+ dose), that is, the amount of toxin which, together 
with a given amount of toxin (1/50, 1/200, or 1/2,000), will still 
produce a minimal amount of necrosis after intracutaneous injection 
into guinea pigs. It is necessary, therefore, arbitrarily to choose a 
certain definite fraction of an antitoxin unit and mix this with vary- 
ing amounts of toxin and inject the mixtures into guinea pigs intra- 
cutaneously. Those mixtures in which the toxin is fully neutralized 
will give rise to absolutely no lesion further than, possibly, a slight 
local edema. Those in which there is a large excess of toxin will 
cause extensive necrosis. Between the two, in the series, there will 
be a mixture in which slight local necrosis results from the injection. 
In this mixture the amount of toxin, just sufficient to cause notice- 
able necrosis in spite of admixture with the antitoxin, contains the 
L-n (limes necrosis) dose. 
When this has been determined, then unknown antitoxin can be 
similarly measured against this L-n dose of the standard toxin. The 
method has the advantage of permitting one to work with very small 
quantities, since only a small fraction of a cubic centimeter need be 
used for intracutaneous injections; also it permits great economy of 
30 See also Kellogg’s method described on p. 522, under the discussion of 
methods for the determination of antitoxin in blood serum of human beings. 
31 Marks. Centralbl. f. Bakt., Orig. Yol. 36. 
32 Romer. Zeitschr. f. 1mm., Vol. 3, 1909, p. 208. Romer and Sanies. 
Ibid., p. 344. Romer and Somogyi. Ibid., p. 433. THERAPEUTIC IMMUNIZATION 
537 
animal material, since four or five tests can be simultaneously car- 
ried out upon the abdominal wall of the same guinea pig. 
The technique is not easy. We have found in studying this 
method in connection with some work carried on in our laboratory by 
Dr. M. C. Terry, that a considerable amount of practice and experi- 
ence is necessary, both in carrying out the procedure accurately and 
in judging the lesions. However, when carefully and consistently 
done by an experienced worker, this method gives results which cor- 
respond with fair accuracy to measurements made of the same 
antitoxin by the Ehrlich method. This has been the expe- 
rience of Lewin,33 and also of Terry in the few experiments carried 
out by him. 
The Rbmer method has been recently used by clinicians for the 
determination of the presence of free toxin or antitoxin in the circu- 
lating blood of patients suffering or convalescent from diphtheria. 
Rbmer himself suggested this, since his method is adapted to the 
determination of extremely slight amounts of either substance. A 
recent study by Harriehausen and Wirth 34 illustrates the results 
obtained in such tests. Normal human serum injected intracuta- 
neously into guinea pigs never caused necrosis. Neither did the 
similar injection of the sera of children suffering from varicella and 
other diseases. Of twelve children suffering from diphtheria, how- 
ever, serum taken before the administration of antitoxin caused 
necrosis upon intracutaneous injection into guinea pigs, in every 
case. In spite of the administration of antitoxin, toxin was demon- 
strable in the blood in five cases as long as the 35th day. Of ten 
cases of post-diphtheritic paralysis, toxin was demonstrated in the 
blood of five.35 
Schick Reaction.—In 1912 Michaelis and Schick 36 carried out 
intracutaneous reactions with diphtheria toxin directly upon human 
beings to determine whether or not diphtheria immunity was present. 
Their early experiments were carried out by injecting 0.1 c. c. of a 
1-1,000 dilution of toxin, and their results indicated that a positive 
intracutaneous reaction with this amount indicated an absence of 
antitoxin from the blood, or at any rate, an insufficient protection. 
The reaction was called positive when within 24 to 36 hours there 
appeared a slight infiltration of the skin surrounded by a red areola 
1 to 2 cm. in diameter. The reaction may continue to increase, or 
in weak positives may not come out completely for 48 hours, there- 
after fading gradually. This reaction has attained great importance 
in public health work since its first introduction. At the present 
time it is being carried out on many thousands of children and 
33 Lewin. Centralbl. f. Bakt., Orig., Yol. 67, 1913. 
34 Harriehausen and Wirth. Zeitschr. f. Kinderheilkunde, Vol. 7, 1913. 
35 See Kellogg’s modification, p. 522. 
36 Michiels and Schick. Zeitschr. f. Kinderheilkunde, Yol. 5, 1912. 538 
INFECTION AND RESISTANCE 
adults in the United States and Europe. In America it has been 
studied chiefly by Park and Zingher.36a 
The material used for the reaction at the present time consists 
of a standardized diphtheria toxin, diluted in such a way that 
0.1 c. c. contained 1/50 M L D for a guinea pig. A very large 
number of tests has indicated that this amount will produce no 
reaction on intracutaneous injection into a human being if the blood 
serum of the subject contains more than 1/30 unit of antitoxin 
per cubic centimeter. Park and Zingher have found that negative 
reactions were present in about 93 per cent, of the newborn, and 
became less frequent up to the fifth year at which time 37 per cent, 
were negative. In our own experience with the Army in an age 
group ranging from 20 to 30 years, an average of about 10 per cent, 
positives was obtained. 
In performing the reaction a certain number of precautions must 
be carefully observed. A pseudo reaction may appear a little earlier 
than the true reaction, and disappear within 24 to 48 hours. This 
is occasionally seen in individuals who may have sufficient antitoxin 
but who are sensitive to either the bacterial protein or the cultural 
material present in the injected toxin. It may be obtained in the 
same individual by the injection of autolyzed diphtheria bacilli, 
and sometimes by the injection of uninoculated broth. To avoid 
error, therefore, it is necessary, when doing Schick reactions, to 
carry out a control injection consisting of 1/50 M. L. D. of the toxin 
heated to 80° for 5 minutes. This destroys the toxin, but leaves 
uninjured the substance which produces the pseudo reaction. 
Tetanus Antitoxin and Its Standardization 
The methods employed in the production and standardization of 
tetanus toxin are in every way analogous to those used in the case 
of diphtheria antitoxin. A strong toxin is obtained by growing the 
organisms under anaerobic conditions on suitable media. Accord- 
ing to Vaillard and Vincent 37 it is essential that the media upon 
which the tetanus bacilli are grown should be freshly made and 
sterilized. Apparently this precaution, which has been similarly 
recommended by Wladimiroff, Novy, and others, is made necessary 
by the gradual absorption of oxygen which takes place if the media 
are allowed to stand for a long time without heating. Tt is further 
necessary in preparing tetanus toxin that the culture medium should 
not be acid, and a weakly alkaline initial titre is advised. For the 
same reason, also, most workers have advised against the use of 
glucose or other carbohydrates in the media, since the acid formed 
by the fermentation of these substances inhibits growth and toxin 
36aPark and Zingher. Jour. A. M. A., 65, 1915, p. 2216. 
37 Vaillard and Vincent. Ann. de I’Inst. Past., 1891. THERAPEUTIC IMMUNIZATION 
539 
production. Recently Hall 38 has advised the use of a simple meat 
extract broth to which have been added 1 per cent, of dextrose and 
0.5 per cent, of finely powdered magnesium carbonate. The last- 
named substance, by neutralizing any acid that is formed from the 
glucose, prevents the harmful acidity. Anaerobic conditions are ob- 
tained by growing the organisms under a layer of oil in tightly stop- 
pered flasks. 
Although mice were formerly used in the standardization of 
tetanus toxin and antitoxin, the more recent usage has been to sub- 
stitute guinea pigs as in diphtheria standardization. According to 
the recent directions of Rosenau and Anderson 39 the purposes of the 
standardization are carried out as follows: 
The unit of antitoxin is arbitrarily designated as 10 times the 
smallest amount of serum necessary to preserve the life of a guinea 
pig weighing 350 grams for 96 hours, when given together with an 
official test dose of toxin. The test dose of toxin contains 100 min- 
imal lethal doses. And the minimal lethal dose is measured against 
a 350-gram guinea pig. 
In carrying out the standardization the L-f- dose of the toxin is 
used, but, unlike diphtheria standardization, in this case the L-f- 
dose means an amount of toxin which will kill a guinea pig of 350 
grams in four days, although united with 0.1 unit of antitoxin (it 
must be noted that the L-f- dose in this case is measured against one- 
tenth unit of antitoxin rather than against 1 unit, as in the case of 
diphtheria). 
In determining the value of an unknown antitoxin, mixtures are 
made, each containing the L-f- dose of the toxin and varying quan- 
ties of antitoxin. As in diphtheria measurements, the various in- 
jection volumes are brought to 4 c. c. with salt solution, and are then 
injected subcutaneously into guinea pigs of about 350 grams. The 
table given below is taken from the Bulletin of Rosenau and Ander- 
son. 
No. of 
guinea 
pig 
Weight of 
guinea pig 
’ (grams) 
Subcutaneous 
a mixt 
Toxin 
(Test dose) 
(gram) 
injection of 
ure of 
Antitoxin 
(c. c.) 
Time of death 
1 
360 
0.0006 
0.001 
2 days 4 hours 
2 
350 
0.0006 
0.0015 
4 days 1 hour 
3 
350 
0.0006 
0.002 
Symptoms 
4 
360 
0.0006 
0.0025 
Slight symptoms 
5 
350 
0.0006 
0.003 
No symptoms 
38 Hall. “Univ. of Cal., Publ. in Path.,” Yol. 2, No. 11, 1913. 
39 Rosenau and Anderson. U. S. P. H. Service Hyg. Lab. Pull. 43, 1908. 540 
INFECTION AND RESISTANCE 
In this experiment 0.0015 equals 0.10 antitoxin unit. 
Therapeutic Use of Tetanus Antitoxin.—Prophylactically, te- 
tanus antitoxin has always been used with a gu-eat deal of success, 
and its efficiency was established beyond question of doubt during 
the recent war. Its curative value, however, has always been con- 
siderably limited, probably owing to the early union of tetanus toxin 
with cells of the central nervous system. But recent improvements 
in the methods of administration have considerably increased its 
efficiency. 
As a general rule, for prophylactic use, subject, of course, to in- 
dividual modifications in various types of cases, the following mode 
of procedure is advisable. About 1,000 to 1,500 units are adminis- 
tered subcutaneously, and this dose repeated if secondary injury or 
operation on the wound. It may be said with reasonable certainty 
that tetanus antitoxin after injection disappears in the human being 
almost completely in the course of 2 or 3 weeks, and in cases, espe- 
cially of compound fracture in which operative interference is called 
for within this time, one cannot rely upon a holding over of the pro- 
tective effect of the first injection. In children, of course, propor- 
tionately smaller doses should be given. A similar repetition of the 
dose may be advisable when inflammatory conditions in the wound, 
with discharge, etc., continue over a prolonged period. All wounds 
made by implements soiled with dirt, especially garden earth in 
which horse or cow manure or other faecal material may be suspected, 
or in which any severe contamination of the wound is likely to have 
occurred, and especially those in which there has been extensive tissue 
destruction, laceration, are to be suspected of tetanus. Especially 
is this the case in gun-shot and shell wounds, in which particles of 
soiled clothing are carried into the wound with the projectile. In 
the last war it was a custom to inject every wounded man with tetanus 
antitoxin as soon as he came under medical observation, and there 
is little question that this procedure saved many lives. 
For curative purposes we follow the advice given by Matthias 
Niehol, Jr.,40 who has worked on this problem, both experimentally 
and practically, at the New York Department of Health. In the 
curative use of the antitoxin, the speed of treatment is one of the 
most important factors for success. He states that every hour lost, 
after the development of symptoms makes the prospect of success less 
bright. Every case coming under observation with symptoms sus- 
picious of tetanus, according to Niehol, should receive an intraspinous 
dose of tetanus antitoxin of not less than 10,000 units, and this dose 
should be repeated in 12 hours, and again in 24 hours if the symptoms 
continue. At the same time the intraspinous dose is given, Niehol 
advises an intravenous dose of 5,000 to 10,000 units. 
We have seen one or two cases, also, in which there seems to have 
40 Niehol, M. Jour. A. M. A., Yol. 76, 1921, p. 112. THERAPEUTIC IMMUNIZATION 
541 
been some benefit by adding to this an injection of small amounts of 
antitoxin in the neighborhood of the wound, in the region of the 
afferent nerve trunks. Whether or not this is of value, however, is 
questionable. Nichol believes that the intraspinous use is the most 
important application, and advises the use of a general anesthesia 
when opisthotonos is present, or when any condition exists which 
makes a very careful and deliberate operation difficult. The anti- 
toxin should never be forced into the canal, and a certain amount of 
fluid can be withdrawn through the same needle before the antitoxin 
is injected, if the spinal fluid is under pressure. Nichol also advises 
diluting the antitoxin with sterile salt solution to a total of about 
15 c. c. in the case of children, and 30 c. c. in the case of adults. The 
intraspinous administration of the serum is subject to no greater 
danger than the intravenous administration, according to this worker. 
ANTITOXINS AGAINST SNAKE POISONS 
(Antivenin) 
Antitoxins against snake poisons have been produced by a num- 
ber of different workers, but the subject has been most extensively 
studied by Calmette. As early as 1887 Sewall 41 succeeded in in- 
creasing the resistance of pigeons to snake poison. Later Calmette 
and Physalix and Bertrand independently succeeded in producing 
immunity in rabbits and guinea pigs with the poison of the cobra. 
The serum of animals treated with snake poisons gradually acquires 
antitoxin properties, but the process of immunization is not a simple 
one, and considerable time is needed for the immunizations. 
Snake poisons, as we have seen, have attracted considerable atten- 
tion because of their peculiarities in being antigenic and yet differ- 
ing in heat resistance and a number of other properties from the 
bacterial toxins. It was with snake poisons that Calmette definitely 
showed that the union of toxin and antitoxin is a true neutralization 
and is not accompanied by the destruction of the toxin. These ex- 
periments, as we have seen, were elaborated later by Morgenroth, 
who succeeded in producing the snake poison HC1 combination. It 
is these poisons also that have been the subject of extensive study by 
Flexner and Noguchi, by Kyes, and later by von Dungern and Coca. 
This work has been sufficiently discussed in other places and need 
not occupy us here. The important poisonous snakes may be divided 
into the colubridse, to which class the cobra belongs, and the viper- 
idse, which includes the ordinary European vipers, the rattlesnake, 
and most of the poisonous snakes of North and South America.4la Ac- 
41 Sewall. Cited from Calmette. 
41a There are at least two other varieties of poisonous snakes in India 
against the poisons of which no antitoxic sera have as yet been successfully 
produced. 542 
INFECTION AND RESISTANCE 
cording to Calmette the poison of the cobra is much more heat-stable 
than that of the rattlesnake. Pharmacologically the poisons of these 
two main classes of snakes show considerable difference. In the case 
of the cobra there is very little local disturbance and the systemic 
symptoms dominate the clinical picture. Calmette describes the 
cobra bite as being followed only by a feeling of stiffness at the site 
of the bite, followed very soon by great general weakness, difficulty 
in respiration, slow heart action, and finally death with unconscious- 
ness. In the case of the vipers the local symptoms are very much 
more marked, there being great pain and swelling and apparent clot- 
ting of the blood about the point of the bite, with a rather slower 
onset of systemic symptoms. In a description by Sparr 42 of a case 
of bite by Russell’s viper there was almost immediate swelling of the 
limb with a faint bluish tint around the pin-point puncture, and 
within 15 minutes great weakness, restlessness, and retching. In 
spite of very active local treatment, within a short time after the 
bite, the patient died within less than 24 hours of asphyxia and heart 
failure. 
According to Calmette 0.0002 gm. of cobra poison will kill a 
guinea pig; Noguchi states that 0.0005 gm. of rattlesnake venom will 
kill a guinea pig of 250 gr. within 24-30 hours when injected intra- 
peritoneally. The snake poisons apparently contain substances which 
are especially active upon nerve cells (neurotoxins), and hemolysins 
which act particularly upon the red blood cells. Flexner and 
Noguchi 43 also speak of another poison which acts particularly upon 
the endothelium of the blood vessels producing hemorrhages. 
According to Calmette the antisera which are produced by im- 
munization with cobra poison are most strongly potent against neuro- 
toxic poisons of the colubridse and, to a certain extent, against some 
of the poisons of the vipers. However, the action of the cobra anti- 
toxin against viper poison seems at best to be weak. On the other 
hand, antitoxins produced with rattlesnake poison are not potent 
against the cobra venom since, as Calmette states, the rattlesnake 
poison contains hardly any neurotoxin. Antitoxins may be produced 
by the gradual immunization of horses, and have been produced in 
this way by Calmette in the Pasteur Institute of Lille for some years. 
Calmette standardizes his antitoxin by determining the amount of 
serum which completely neutralizes in vitro 0.0001 gm. of the poison 
as tested upon white light. He also determines the protective power 
by injecting a rabbit with 2 c. c. of the serum and two hours later 
gives 1 gm. of the poison. 
Noguchi has studied rattlesnake poison particularly and suc- 
ceeded in preparing a strong antitoxin by the gradual immunization 
of a goat. Great difficulty has always been experienced in attempts 
42 Sparr. Biochem. Bull., Dec., 1911, No. 2. 
43 Flexner and Noguchi. Univ. of Pa. Bull., Yol. 15, 1902. THERAPEUTIC IMMUNIZATION 
543 
at immunization with rattlesnake poison because of the very violent 
local injury produced by injections of the venom. The potency of 
the serum produced by him was such that 21/2 c. c. of goat seruni 
protected guinea pigs against 12 times the fatal dose of rattlesnake 
poison if given at the same time. If the antivenin was given one 
hour later, 5 times the amount of serum had to be given. 
Serum Treatment in Botulismus.—Antitoxic serum has been 
produced by the injection of botulinus toxin by a number of ob- 
servers. Kempner 44 started research in this direction in 1897. He 
was followed by Jorssman and Lundstrum,45 and Leuchs.46 Hickson 
and Howitt 47 have recently produced potent antitoxins by the im- 
munization of goats. These observers made the important observa- 
tion that various strains of B. botulinus produces at least two 
different toxins which are not serologically homologous. Representa- 
tives of these two types must be used in the production of the anti- 
toxin. 
There is not yet sufficient statistical evidence to permit us to draw 
conclusions concerning its therapeutic value. Dickson advises its 
use, however, by the ordinary methods of intravenous injection. 
Antitoxin Treatment in the Case of Anaerobic Wound Infection. 
B. Welchii Toxin.—Since, in 1917, Bull and Pritchett 48 were able 
definitely to determine the presence of a toxin in cultures of the 
Welch bacillus, the possibility of antitoxin treatment of infections 
with this organism was established. Before this time, a suspicion 
that the Welch bacillus produced a toxin had often been suggested 
but never definitely proven. At about the same time that Bull and 
Pritchett made their observations, similar observations were made 
by Weinberg and Seguin.49 Klose 50 had reported similar observa- 
tions in 1916, but the toxin produced by him was not very powerful, 
and the antitoxin resulting from its injection into animals gave 
limited protection only. Apparently, the production of the toxin 
in the cultures of Bull and Pritchett depended upon the presence of 
fresh muscle tissue and glucose in the broth in which the organism 
was grown. They found no variations in the different strains of B. 
Welchii, irrespective of the source, to produce the toxin, but toxin 
production seemed to be directly proportionate to virulence. Most 
potent toxin production seems to be obtained by inoculating the in- 
fected muscle of a pigeon dying of B. Welchii infection directly into 
the medium to be used for toxin production. Caulfield 61 who has 
44 Kempner. Zeit. f. Hyg., Yol. 26, 1897, p. 481. 
45 Jorssman and Lundstrum. Ann. de Vlnst. Past., Yol. 16, 1902, p. 294. 
46 Leuchs, Kolle and Wassermann. Handbook, 2nd Edit., Vol. 4, p. 939. 
47 Dickson and Howitt. Jour. A. M. A., Yol. 74, 1919, p. 718. 
48 Bull and Pritchett. Jour. Exp. Med., Yol. 26, 1917, p. 67. 
49 Weinberg and Seguin. “La Gangrene gazeuse,” Masson, Paris, 1918. 
50 Klose. Miineh. med. Woch., Vol. 63, 1916, p. 723. 
51 Caulfield. Jour. Inf. Dis., Vol. 27, 1920. 544 
INFECTION AND RESISTANCE 
studied this, claims that good toxin production docs not result unless 
the virulence of the strain is such that 0.02 c. c. of the supernatant 
fluid of a brother culture will kill a 300 gram pigeon. Work by 
Bengston of the Hygienic Laboratory in Washington, however, and 
similar work by DeKruif seems to indicate that the substitution of 
chopped veal in broth, even after autoclaving, furnishes a favorable 
medium for toxin production. The toxin differs from the classical 
toxins in that it varies in potency with virulence of the strain, and 
has no definite incubation period. 
Antitoxin is produced by the injection of horses, and an attempt 
to standardize the antitoxin has been made by Bengston, who suc- 
ceeded in producing an antitoxin of which 1 c. c. of serum contains 
one unit, one unit neutralizing 1/1,000 of an M. L. U. of the B. 
Welchii toxin. 
In laboratory animals injected with pure cultures, the antitoxin 
gives complete protection. Similar protection, however, seems to be 
provided, according to Weinberg and Seguin, by antibacterial sera 
prepared by the injection of whole broth cultures of the bacillus, 
although the antitoxin contents of these sera has not been determined. 
The usefulness of the serum in human beings has not been suffi- 
ciently tested for final judgment, but there is no theoretical reason 
why it should not be very valuable in infections by pure cultures of 
the Welch bacillus. 
Unfortunately, most wounds that contain Welch bacillus are apt 
to contain other anaerobic organisms, particularly the Vibrion 
septique and the B. cedematiens. Efforts, therefore, to prov!..ce 
mixed antitoxins are being made. 
Antitoxin Against Vibrion Septique.—The Vibrion septique, 
according to Weinberg and Seguin,52 occurred in about 12 per cent, 
of the wounds examined by them. It is rarely alone, but usually 
associated with other anaerobes, particularly the Welch bacillus. 
All strains of the Vibrion septique produce a powerful toxin. 
This substance, too, is peculiar in that it kills without an incubation 
period comparable to that noticed in the case of most other toxins. 
It is peculiar, also, in that death occurs with regularity only when 
the substance is injected intravenously, and often when injected sub- 
cutaneously or intramuscularly produces a local necrosis only. 
A powerful toxin may be obtained by growing the organism 
anaerobically in an 0.2 per cent, dextrose broth to which 10 per cent, 
of horse serum has been added. Robertson 53 has recommended the 
addition of liver tissue from a guinea pig dead of Vibrion septique, 
for inoculation into broth. Great care has to be exerted in controlling 
filtration, since filters which are too tight hold back a considerable 
percentage of the toxin. 
52 Weinberg and Seguin. Loc. cit. 
53 Robertson. Jour. Path, and Bad., 1920. THERAPEUTIC IMMUNIZATION 
545 
Antitoxin has been prepared by Weinberg and Seguin, and by 
Robertson, by injecting the toxin into horses and sheep. The de- 
velopment of the antitoxin has been undertaken chiefly in France 
where the standard requires that 0.001 c. c. of the antitoxin should 
neutralize two fatal doses of the toxin after 30 minutes’ incubation 
of the mixture at room temperature and injection into guinea pigs. 
Bacillus (Edematiens.—Weinberg and Seguin isolated this or- 
ganism in 34 per cent, of the war wounds examined. The percentage 
of most other workers has been lower. The organism and the toxin 
described are probably identical with what Sacquepee 54 has spoken 
of as the B. Bellonensis. This organism forms a soluble toxin which 
is of unusual strength. It is best produced, according to Weinberg 
and Seguin, by growing the organism in broth, containing chopped 
veal, for about a week. It never kills acutely on intravenous injec- 
tion, in this respect differing from that previously described, but with 
a potent strain, 0.01 c. c. of the broth filtrate, intravenously injected, 
will kill a 300 gram guinea pig in 48 hours. 
Rabbits, sheep and horses, carefully injected with broth filtrates 
from these strains, will eventually yield an antitoxin of considerable 
strength. Weinberg and Seguin, using horses, prepared an antitoxin 
of which 0.0001 c. c. neutralized two lethal doses for a guinea pig. 
Passive Immunization in Bacillary Dysentery.—In view of the 
fact that a soluble toxin has been claimed by a number of workers 
for the Shiga dysentery, and that animals immunized with organisms 
of the Flexner group yield a serum to some extent neutralizing the 
toxic effects of these organisms, sera have been produced by the im- 
munization of horses with organisms of both the Shiga and the Flex- 
ner type. Flexner advises the separate immunization of horses with 
these organisms, and the use of the sera either mixed or separate as 
indicated by the bacteriological examination of the cases. 
When serum has been employed, it has usually been used sub- 
cutaneously, but it may be used intravenously in quantities ranging 
from 20 to 100 c. c., according to the age of the patient and the 
severity of the case. According to Flexner,55 a single injection has 
often led to marked alleviation of symptoms, the most striking results 
having been obtained in acute cases; but favorable results have also 
been obtained in the second and third weeks. The same writer states 
that the mortality in certain outbreaks in which comparisons between 
serum treated and untreated cases have been made, was reduced from 
10 to 15 per cent, to 1 to 2 per cent, under serum treatment. And 
in epidemics in which the mortality reached 50 per cent., serum treat- 
ment has reduced the mortality to about 10 per cent. 
In institutional outbreaks of the disease he advises prophylactic 
doses of 5 c. c., subcutaneously injected. 
54 Sacquepee. Ann. de I’Inst. Past., Yol. 30, 1916, p. 76, 
55 Flexner. Jour. A. M. A., Yol. 76, 1921, p. 108. 546 
INFECTION AND RESISTANCE 
PASSIVE IMMUNIZATION IN DISEASES CAUSED BY BAC- 
TERIA WHICH DO NOT FORM SOLUBLE TOXINS 
General Principles.—As we have stated, the greatest therapeutic 
successes with passive immunization have been achieved in bacterial 
diseases in which the malady is essentially a toxemia due to a 
toxin. In such cases the serum of actively immunized animals con- 
tains specific antitoxins by virtue of which the toxins circulating in 
the blood of the patient are directly neutralized, quantity for quan- 
tity, with consequent therapeutic benefit. In the case of bacteria in 
which no toxins are formed, the immunization of an animal is not 
followed by the formation of any poison-neutralizing principle. 
Here the injection of bacteria, dead or alive, or the invasion of the 
bacteria in the course of spontaneous disease, is followed by the 
formation of specific antibacterial substances, lytic, opsonic, agglu- 
tinating, or precipitating bodies, the nature of which we have dis- 
cussed in other chapters. The toxemia which occurs in such cases 
is due as we have seen to derivatives of the bacterial protein which 
by some observers are regarded as preformed endocellular poisons 
liberated by the lytic action of the serum, and by others as split 
products of the bacterial protein, non-existent until the bacterial cell 
has been acted upon by the serum components and destroyed. How- 
ever this may be, the recovery from diseases of this nature is accom- 
plished by bacterial destruction; this may be rarely, by the bacteri- 
cidal action of the serum, but more often by opsonic powers which 
induce phagocytosis. The poisons which are liberated from the bac- 
terial bodies, if free, can do their injury, and no neutralizing sub- 
stance is formed in the body fluids to prevent their action as far as 
we know. Immunity in such cases, then, is not an antitoxic immu- 
nity in any sense of the word; it is rather an antibacterial immunity 
in which the disease is prevented or cured only when complete de- 
struction of the bacteria has taken place. If an animal or a human 
being is prophylactically immunized against diseases of this kind 
(typhoid, cholera, etc.), it is easy to see that an increased presence 
of antibacterial substances, bactericidal or opsonic, in the circulation 
would serve efficiently and rapidly in disposing of the small numbers 
of invading micro-organisms which ordinarily enter the body in spon- 
taneous infections. And, indeed, experience has shown that prophy- 
lactic immunization can be successfully carried out in the case of 
cholera, typhoid fever, plague, and other diseases which are suffi- 
ciently prevalent endemically or epidemically to justify prophylaxis 
on an extensive scale. 
However, when in diseases of this kind the body is already exten- 
sively infected and has begun, as is usually the case, to respond spon- 
taneously with the formation of specific antibodies, it has been a 
matter of doubt whether or not passive immunization, that is, the THERAPEUTIC IMMUNIZATION 
547 
introduction of specific antibodies in the form of the serum of a 
highly immunized animal, is therapeutically of value. Indeed, it 
has been feared that the use of such sera may even be harmful in 
that the sudden introduction of large amounts of bactericidal sub- 
stances might lead to a sudden liberation of large quantities of poi- 
sonous products and consequent rapid toxemia. 
The conditions in such cases are exceedingly complex and many 
gaps exist in our knowledge concerning them. The bacteria when 
invading the body, immediately enter into conflict with the protective 
forces, as we have stated in the chapter on Infection. If a consider- 
able degree of resistance exists, let us say as the result of preceding 
immunization or a recent attack of the disease, there is a rapid 
destruction of the bacteria, probably by active phagocytosis. It has 
been shown by Bordet in the case of cholera and more recently by 
Gay with typhoid, that injection of the organisms into immunized 
animals is followed by prompt and high leukocytosis, whereas simi- 
lar injections into normal animals usually induce a temporary leuko- 
penia. When the invaded animal is not particularly resistant the 
bacteria may accumulate and, as in the case of pneumococci and 
streptococci, develop phagocytosis-resisting properties (capsule for- 
mation, etc.) ; or, as in the case of typhoid bacilli, there may be an 
immediate liberation of toxic substances (endotoxins, etc.) by reac- 
tion between bacterial cell and blood plasma, which can induce leuko- 
penia, and by this means the organisms may be protected from 
phagocytic destruction. Experience with curative sera in all of the 
conditions of this class has yielded promising results only when the 
cases have been treated with the sera at early stages of the disease, 
either when the invading germ was still localized or, at least, when 
the septicemic condition was not yet thoroughly established. It may 
be that the doses heretofore given have been insufficient, and indeed 
recent experiences with pneumonia seem to indicate that this may 
have been, in part, the cause of earlier failures. Yet in pneumonia 
the septicemia probably does not represent the firm establishment of a 
foothold by the pneumococcus in the circulation but rather a con- 
tinuous discharge of new organisms into the blood from the localized 
lesion in the lung. 
It is our own opinion, moreover, that septicemia as usually ob- 
served clinically represents in most cases exactly this condition, that 
is, a more or less continuous discharge of the bacteria into the blood 
from some active focus with a continuous destruction of the organ- 
isms after they have entered the blood stream. It is only when the 
resistance of the body is overwhelmed, in the later stages of the 
disease, that the bacteria can continue to grow and develop in the 
circulation, and this stage probably does not occur until death is 
imminent. In such septicemic diseases as streptococcus infection, 
typhoid fever, plague, anthrax, and many others the presence of the 548 
INFECTION AND RESISTANCE 
bacteria in the blood at the time when the patient is still in a condi- 
tion of powerful resistance probably means that the bacteria are 
being supplied to the blood from the local lesions. There is prob- 
ably just such a continuous discharge of bacteria from the focus into 
the blood with active destruction after the bacteria have entered the 
circulation. This seems especially probable from the fact that in 
many of these diseases the protective antibodies, bactericidal and 
opsonic, can often be demonstrated in the blood serum in quantities 
higher than normal at the very time when blood culture yields posi- 
tive results. In typhoid fever, of course, it is well known that bac- 
tericidal titres of over 1-50,000 are often present while the patient 
may still be very sick, and in the more chronic streptococcus condi- 
tions with malignant endocarditis we have often seen that opsonic 
properties on the part of the patient’s serum against the very organ- 
ism invading him are considerably higher than normal. We take 
this to mean that the injection of immune sera would simply aid in 
more rapidly freeing the blood stream of the bacteria, the cure of 
the disease, however, involving a destruction of the focus. This, of 
course, is not possible merely by the injection of the serum. When, 
as in some cases of streptococcus infection, the focus can be surgically 
reached, the septicemia will often disappear and cure result, as we 
have ourselves had the opportunity to observe. When the focus 
cannot be reached surgically, it may nevertheless be a wise procedure 
to inject considerable amounts of immune serum, for, by keeping the 
blood stream free of bacteria, the case may be influenced favorably. 
Pneumonia is an example of this. Former failures have recently 
been turned into partial success by the work of Heufeld and of Cole 
merely by the use of larger quantities of immune sera essentially 
similar to sera used at previous times, and Cole attributes the appar- 
ently favorable results to the fact that the blood stream can be cleared 
of bacteria although the focus cannot itself be affected. 
Rapid and complete cure of such diseases, therefore, can hardly 
be expected. Favorable influence of the disease by energetic serum 
treatment may, however, be hoped for. As to the exact manner of 
action of such sera there is much room for discussion. It is likely 
that a number of factors are involved. Chief among them is prol> 
ably the increased opsonic power of the patient brought about by the 
specific sensitization of the bacteria. In addition to this there is, in 
many cases an active intravascular agglutination of the bacteria, as 
shown by Bull for animals infected with pneumococci and typhoid 
bacilli. By this the bacteria are clumped in the capillaries of some 
of the viscera where active phagocytosis by fixed tissue cells can then 
take place. A certain amount of actual neutralization of so-called 
“endotoxic” substances by the antibacterial sera has also been claimed 
in the cases of sera prepared with Gram-negative organisms like the 
typhoid and Dysentery bacilli and the meningococcus, THERAPEUTIC IMMUNIZATION 
549 
In discussing this subject it must not be forgotten, however, that 
in most of the diseases which we have classified, on the basis of pre- 
vailing opinions, as caused by bacteria that do not form true toxins, 
the formation of such poisons has been claimed by a number of care- 
ful and eminent observers. In the case of the typhoid bacillus, espe- 
cially, Chantemesse, Kraus and Stenitzer, and others have claimed 
the existence of a true toxin and a consequent antitoxin in immune 
sera. Similar claims have been made for the cholera spirillum by 
Kraus and Doerr, for the streptococcus by Marmorek, and for the 
plague bacillus by Markl and Rowland. Since these claims have 
been made on the basis of extensive experimentation by competent 
men the question must be left open, and the possibility of antitoxic 
properties on the part of the sera cannot be completely ignored. 
Since in most however, the poison-neutralizing properties of 
the immune sera in this disease have not exceeded more than 1 to 2 
multiples of the M L D of the bacterial poisons, it does not seem 
impossible that the apparent antitoxic properties may have repre- 
sented merely an acquired tolerance to anaphylatoxic poisons of 
which we have spoken in another place. 
Serum Treatment in Epidemic Cerebrospinal Meningitis 
Serious attempts to produce curative sera against the epidemic 
form of cerebrospinal meningitis were not made until 1906 and 
1907, when this disease appeared epidemically chiefly in Europe, 
where it appeared most severely in Eastern Germany, and in the 
Eastern United States. 
In 1906 Kolle and Wassermann immunized three horses with 
meningococci, using for immunization purposes the dead organisms 
followed by living cultures and cultures taken up in distilled water, 
the so-called artificial aggressins of Wassermann and Citron. They 
obtained sera of considerable potency when measured against menin- 
gococcus cultures, and suggested standardizing the sera by comple- 
ment fixation. They did not at this time treat human beings, but sug- 
gested the, use of the serum subcutaneously and intravenously in 
meningitis cases. Very soon after the publication of the work of 
Kolle and Wassermann Jochmann 56 also produced an antimeningo- 
coccus serum by immunizing horses with proved meningococcus cul- 
tures, in his cases making a polyvalent serum by the use of many 
different strains of fhe organism. The sera which he obtained were 
highly agglutinating, somewhat bactericidal, and, according to him, 
not antitoxic. He first succeeded in immunizing guinea pigs against 
meningococci by injecting the serum 20 hours before infecting the 
animals. He also treated 40 cases of meningitis in man and ob- 
56 Jochmann, Deutsche med. Woch., VqI. 32, 1906, p. 788. 550 
INFECTION AND RESISTANCE 
tained encouraging results in cases treated before the development of 
hydrocephalus. Believing that possibly intraspinous injection of the 
serum might offer advantages, he first determined by experiments 
upon the dead body that the injection of methylene-blue intra- 
spinously passed from the point of injection in the lumbar regions as 
far up as the olfactory nerves. After having determined this he 
treated 17 cases by tapping the spinal canal, taking out 30 to 50 e. c. 
of spinal fluid and then injecting about 20 c. c. of the serum. Of 
these 17 cases only 5 died, and Jochmann expresses himself opti- 
mistically in consequence. 
Meanwhile Flexner 57 had been working upon the same subject, 
laying a rather more thorough basis for therapy in careful animal 
experimentation. He produced the typical disease in monkeys by 
intraspinous inoculation of the meningococci and then saved the 
animals from death by following the infection with the injection of 
serum intraspinously six hours later. In his earlier articles he ex- 
presses himself with much conservatism, but his studies were con- 
tinued and extensive opportunity for testing the serum which he then 
produced, together with Jobling,58 was offered by the continuance of 
the epidemic throughout the United States. 
In 1908 Flexner and Jobling reported upon 47 cases treated with 
the antiserum of which 34 recovered. Of 12 additional cases re- 
ported in an addendum only 4 died. Flexner 59 records of 1,294 
cases that have been treated with the serum prepared at the Rocke- 
feller Institute. Of this number, unselected and treated in many 
different parts of the world, 69.1 per cent, recovered. It is of course 
very difficult to obtain exact comparative data on the efficiency of 
any method of treatment in a disease as irregular in its clinical mani- 
festations as epidemic meningitis, especially since the mortality at- 
tending upon different epidemics is subject to great variations. For 
this reason we can draw conclusions only from a large statistical 
material. However, we know that the average mortality of epidemic 
meningitis before the introduction of specific therapy ranged cer- 
tainly higher than 65 per cent., and in carefully studied epidemics 
usually between 70 and 80 per cent. The statistics of Flexner show 
a mortality hardly exceeding 30 per cent, in unselected cases. It 
must be remembered in considering the benefits of this serum that 
in unselected cases there must be many in which the disease has pro- 
duced marked anatomical changes in the central nervous system 
before the serum is used. It is well known, of course, that the later 
manifestations of this disease, which often lead to death with hydro- 
cephalus, asthenia, and malnutrition, are the remote results of the 
57 Flexner. Jour. Exp. Med., Yol. 9, 1907, and J. A. M. A., Vol. 47, 1906, 
p. 560. 
58 Flexner and Jobling. Jour. Exp. Med., Vol. 10, 1908. 
59 Flexner. Jour, Exp. Med., Yol. 17, 1913. THERAPEUTIC IMMUNIZATION 
551 
anatomical injuries produced by the inflammatory reactions accom- 
panying the earlier manifestations of the acute infection. These 
conditions of course cannot be expected to yield to serum treatment. 
It must be assumed, therefore, that were we able to obtain statistics 
of cases diagnosed and treated soon after the onset the figures would 
be even more favorable than those stated above. 
The following mortality table is taken from a bulletin of the 
Rockefeller Institute published in 1917: 
COMPARATIVE MORTALITY REPORTED BY VARIOUS OBSERVERS 60 
Treatment Begun 
Flexner, 
Per cent. 
Netter, 
Per cent. 
Dopter, 
Per cent. 
Christo- 
manos 
Per cent. 
Levy, 
Per cent. 
Flack, 
Per Cent. 
Before third day 
... 18.1 
7.1 
8.2 
13.0 
13.2 
9.09 
From fourth to seventh 
day 
... 27.2 
11.1 
14.4 
25.9 
20.4 
• • • • 
After seventh day... 
... 36.5 
23.5 
24.1 
47.0 
28.6 
50.00 
From the above table, there seems to be very little doubt that the 
use of the serum has very materially modified the course of the dis- 
ease, and that serum treatment in meningitis is one of our most im- 
portant specific therapeutic agents. 
In view of the fact that we now know that many different types 
of meningococci exist, serological procedures must be governed by 
this knowledge. Typing of meningococci in different countries has 
led to various classifications. The first knowledge concerning the 
existence of various types resulted from the investigations of 
Dopter.61 Dopter found that some of the meningococci isolated from 
cases in and about Paris did not agglutinate in the ordinary so-called 
normal type serum. This organism he calls the parameningococcus, 
and its discovery marks the beginning of our serological classification 
of these organisms. Wollstein62 confirmed Dopter’s work, and 
showed that in addition to the parameningococcus and the so-called 
normal strains, there were a considerable number of other variants 
that could not be classified definitely with either of the other two. 
During the war, Gordon,63 in England, studied many strains of 
meningococci obtained from the epidemic which occurred among 
Canadian and British troops, and divided meningococci into four 
types, now commonly spoken of as the “Gordon types.” Hot all 
organisms could be classified definitely into one or the other type, but 
Tulloch 64 found that of 356 organisms he investigated, 234 agglu- 
tinated or gave specific agglutinin absorption in one of the four type 
sera produced by Gordon. Gordon’s Type I corresponds to Dopter’s 
parameningococcus. His Type II corresponds to the normal or origi- 
80 Flexner. Bull. Rock. Inst, for Med. Res., 1917. 
81 Dopter. Compt. Rend, de l. Soc. Biol., Yol. 67, 1909, p. 74. 
82 Wollstein. Jour. Exp. Med., Yol. 20, 1914. 
83 Gordon. Brit. Med. Res. Com. Rep., London, 1915-1917. 
84 Tulloch. Jour. Roy. Med. Coll., February, 1918, p. 9. 552 
INFECTION AND RESISTANCE 
nal meningococcus, and his Types III and IV represent intermediate 
or irregular strains of which there are a large number that shade 
into one another. In this country we speak of the normal, the para, 
and intermediate strains. In France three types have been particu- 
larly worked with, corresponding to the normal, para, and inter- 
mediate groups, and spoken of as Types A, B, and C. 
While it is, of course, possible to treat cases with type serum when 
cultures can be obtained and type determined with sufficient speed, 
yet the necessity for rapid treatment makes it imperative that we 
work chiefly with polyvalent serum, giving the first few injections 
with such serum, and subsequently, as cultures are deter- 
mining whether the polyvalent serum that is being used includes the 
particular meningococcus from which the patient is suffering . 
In America and most countries, at the present time, polyvalent 
serum is almost entirely used. Horses are immunized with cultures 
of as many different types as possible. Both the normal and the 
paratypes, together with as many intermediates as can be collected 
from different epidemics throughout the country, must be used, and 
it is this factor, namely, the use of many different races of menin- 
gococci, which is perhaps the most important one in successful serum 
production. In laboratories in which serum production is being 
carried on, a constant survey of cases of meningitis throughout the 
country should be made, organisms obtained from new cases and 
tested against the serum. If cultures are received which do not 
agglutinate in the polyvalent serum, they should be added to the 
immunizing collection. Much judgment is necessary in this par- 
ticular respect because the use of too large a number of strains may 
keep down the potency of the serum for any particular one. 
Cultures for injection are grown upon agar and are best washed 
once in salt solution before injection. Usually two or three injec- 
tions are made on consecutive days, and a rest of 7 to 8 days between 
treatments is given. 
The standardization of the serum was first attempted by Flexner 
and Jobling on the basis of opsonin contents. In some laboratories, 
complement fixation has been recommended, but the usual method 
employed at the present time is agglutination. The serum should 
agglutinate normal and para types in from 1-1,000 or 1-2,000 dilu- 
tions, and should react with the intermediate strains in dilutions of 
1-200 or more. 
Administration of the Serum.—As we have stated above, the 
most important factor for success is early diagnosis and immediate 
treatment with serum. For this reason, responsibility of the physi- 
cian in properly appraising the case, having his suspicions of epi- 
demic meningitis aroused early, and doing a lumbar puncture, is a 
grave one. Delaying several days may materially influence the out- 
come, as may be seen from the table given above. Fluid from the THERAPEUTIC IMMUNIZATION 
553 
lumbar puncture should be taken into a sterile centrifuge tube and 
immediately sent to the laboratory for diagnosis. If the fluid which 
runs from the needle is turbid, and an immediate Gram stain shows 
Gram negative micrococci, serum should be administered at once. 
Moreover, we believe that if such a fluid shows a large number of 
polymorphonuclear leukocytes and no Gram positive organisms can 
be found on staining, even in the absence of Gram negative micro- 
cocci, the chances are in favor of epidemic meningitis since the 
meningococci undergo rapid autolysis in the spinal fluid, and in some 
cases it may take a long search before intact micrococci can be found. 
We would, therefore, be in favor of injecting serum even in such 
cases if an “acute” exudate is found as shown by the polymorphonu- 
clear leukocytes and no Gram positive organisms can be seen in 
smears. 
If serum is to be injected, spinal fluid should first be taken until 
the flow slows down to a drop every ten or twenty seconds. The 
serum is then injected through the same needle, slowly, by gravity, 
or if this is impossible, very slowly with a syringe. The serum 
should be at body temperature, and should be injected so slowly that 
the entire amount shall enter the spinal canal in the course of about 
ten minutes. 
It is of the greatest importance that both in withdrawing fluid 
and in injecting serum, the patient should be very carefully watched 
for symptoms resulting from the rapid changes in intraspinous 
pressure. 
The amount injected must to some extent depend upon the amount 
of fluid withdrawn, and should be a few cubic centimeters less than 
the amount taken out. In general, from 25 to 35 c. c. can be injected 
into an adult, but more may be given if a great deal of fluid has been 
withdrawn. Sophian recommends a simultaneous taking of the blood 
pressure and interruption of the procedure if the blood pressure 
drops suddenly. 
Serum injections should be repeated as often as indicated until 
the fluid becomes clear and this repetition must be controlled by 
clinical observation, and the appearance of the spinal fluid, the num- 
ber of pus cells, the number of organisms, etc. 
Since it has recently been shown by Herrick 65 and a number of 
British observers, that in many cases of meningitis the organisms are 
in the blood stream even before the meningeal symptoms appear, it 
is advisable to inject 30 c. c. of the serum intravenously at the time 
at which the first intraspinous injection is made. 
Moreover, it is important to remember that during epidemics, 
there are many cases of meningococcus septicaemia, usually accom- 
panied by septic temperature, and in which there is often a profuse 
rash consisting both of red blotches on the skin and petechial spots 
65 Herrick. Arch. Inter. Med., Vol. 21, 1918, p. 541. INFECTION AND RESISTANCE 
554 
not unlike those in typhus fever. Such cases can be diagnosed by 
blood culture and should be treated by energetic intravenous serum 
injections. 
Serum Treatment in Streptococcus Infections 
The attempts to produce powerful immune sera against strepto- 
cocci date back to the earliest days of immunology. That the sub- 
ject is a particularly difficult one follows from the great confusion 
which has prevailed, and, to a great extent, still prevails, regarding 
the classification of the streptococci and their interrelationship. 
There are apparently a large number of different strains of strepto- 
cocci which vary from each other, not only culturally, but also in 
regard to agglutination and bactericidal reactions. For this reason 
it is not at all a foregone conclusion that a serum prepared by the 
immunization of an animal with a streptococcus of one type will 
have any protective action against other strains. In all cases in 
which streptococcus immune serum is at all used it must be remem- 
bered that the disease produced in human beings by organisms classi- 
fied among the streptococci are by no means necessarily closely 
related in biological reactions, and the same immune serum may be 
extremely potent in one case and entirely useless in another. 
It has long been known that the hemolytic streptococci which can 
infect man are of many different antigenic varieties, and the recent 
work of Dochez, Avery and Lancefield 66 has indicated that many of 
these organisms can be divided into at least four groups. According 
to the earlier work of von Pirquet, Dochez and his associates, and 
Tunnicliff,67 the streptococci which are associated with cases of scar- 
let fever seem to constitute a separate type. However, in recent 
cultural studies, Rice has found marked cultural differences between 
organisms from this source. This question is still an open one, 
though the mass of evidence seems to point to a general serological 
similarity between the scarlatinal hemolytic streptococci. There are, 
nevertheless, a large number of hemolytic organisms which cannot be 
grouped together, and in which serological reactions seem to shade 
one into the other. In the viridans groups, this heterologous con- 
dition is still more marked, and studies such as those of Kinsella and 
Swift 68 show that these organisms are as varied as are the Type IV 
pneumococci. Thus, the production of therapeutic streptococcus 
serum is rendered extremely difficult, and even the hope of obtaining 
a widely useful polyvalent serum must await a more exact definition 
of the grouping of these organisms. 
That animals could be successfully immunized against strepto- 
66 Dochez, Avery and Lancefield. Jour. Exp. Med., Yol. 30, 1919, p. 179. 
67 Tunnicliff. Jour. A. M. A., Vol. 75, 1920, p. 1339. 
68 Kinsella and Swift. Jour. Exp. Med., Yol. 28, 1918, p. 877. THERAPEUTIC IMMUNIZATION 
555 
cocci was shown early in the history of investigations in immunity 
by a number of workers, notably Roger, Behring, von Lingelsheim, 
and Mironoff. The first extensive attempts to produce a curative 
serum for use in passively immunizing human beings were made by 
Marmorek 69 at the Pasteur Institute in 1895. The basic idea from 
which Marmorek worked was the similarity of all the streptococci 
producing disease in human beings. He also believed that the most 
powerful serum could be produced with cultures whose virulence 
had been greatly enhanced by animal passages. When such cultures 
were grown on mixtures of human serum and broth he asserted 
furthermore that soluble poisons were produced which could be ob- 
tained by filtration of the culture fluids. For these reasons he im- 
munized horses with cultures rendered highly virulent by very 
gradual injections first of dead then of living organisms, finally in- 
jecting also considerable quantities of culture filtrates. 
Testing these sera upon animals, he was successful in protecting 
against streptococcus infection when the serum was administered 12 
to 18 hours before the bacteria were injected. He expressed the 
opinion that the serum was antitoxic as wTell as antibacterial. In 
his earliest reports the results of the treatment of 413 cases of ery- 
sipelas leave one very much in doubt as to the value of the serum since 
the difference in mortality between the treated and the untreated 
cases is less than 2 per cent. However, an analysis of the individual 
cases makes the serum treatment appear more favorable. He re- 
ported good results also in 7 cases of puerperal septicemia and in 
scarlatinal angina. Later observers, notably Lenhartz,70 Baginsky,71 
and many others, have not been able to confirm the favorable results 
reported by Marmorek, and it may be stated that at the present day 
the value of Marmorek’s serum is very much in question. Anti- 
streptococcus sera have also been produced by Aronson 72 and Tavel, 
Van de Velde, Menzer,73 Moser,74 and some others. Aronson at 
first worked from the idea which Marmorek also had used that there 
was a close relationship between the various streptococci pathogenic 
for man. He adopted the opinion first developed bv Denys 75 and 
Van de Velde that many different strains should be used for im- 
munization in order to allow for possible difference in the character- 
istics of the pathogenic streptococci. This principle of the necessity 
for the production of polyvalent sera was also emphasized strongly by 
69 Marmorek. Ann. de Vlnst. Past., Yol. 9, 1895. 
70 Lenhartz. “Die Septischen Erkrankungen Holder,” Wien, 1903. 
71 Baginsky. Berl. klin. Woch., 1896, p. 340. 
72 Aronson. Berl. klin. Woch., Vol. 39, 1902; Deutsche med. Woch., Yol. 
29, 1903. 
73 Menzer. Berl. klin. Woch., 1902, and Munch, med. Woch., 1903. 
74 Moser. Wien. klin. Woch., 1902. 
75 Denys. Bull, de VAcad. Beige, 1896, cited from Schwoner K. and 
L. H., Vol. 2. 556 
INFECTION AND RESISTANCE 
Tavel, who based his opinion on careful agglutination tests, and by 
Menzer and Moser. 
That the action of the antistreptococcus sera, however produced, 
is very largely due to its opsonic properties has been shown by 
Bordet,76 by Meier and Michaelis, and a number of other workers. 
If there is any bactericidal power it is probably relatively slight. 
It would be quite impossible to attempt in this place to analyze 
the large number of streptococcus infections of man which have been 
treated with one or the other antistreptococcus sera. Those men- 
tioned, moreover, do not by any means include all the sera which 
have been produced and marketed for this purpose. In general we 
may say that beneficial results have been obtained chiefly in cases 
in which the streptococcus infection has been localized and treated 
early after its inception. In generalized or advanced cases it can- 
not be said that the results are encouraging. Even in animals, in 
which experimental conditions can be so much more thoroughly con- 
trolled, the protective action of even the strongest sera is evident only 
if the serum is administered either before infection or within a very 
definite period after inoculation. The standardization of strepto- 
coccus sera may be accomplished by determining its protective value 
for animals when injected 18 to 20 hours before infection. 
Serum Treatment in Pneumonia 
Attempts to work out a therapeutically valid method of passive 
immunization in pneumonia have been many and date from the very 
beginning of the discovery that pneumonia was a bacterial infection. 
Sera have even been marketed and used, but until recently no very 
encouraging results were obtained. Recent studies have revealed 
that in pneumonia the serum of convalescents contains practically no 
bactericidal properties for the pneumococcus, and that the protective 
powers of such serum depend upon the presence of immune opsonins 
or bacteriotropins, by means of which the pneumococci are ren- 
dered amenable to phagocytosis. Virulent pneumococci are not as a 
rule phagocytable in the presence of normal serum. However, in 
the presence of immune serum powerful phagocytic action can be 
observed. That the agglutinating action of such sera may also play 
an important role in their protective action has recently been shown 
by Bull (vide infra). 
Heufeld has studied the conditions of pneumococcus immunity 
most thoroughly. The most important advance from a practical 
point of view was a discovery made by him, with Handel,77 in 1909. 
They determined that there was a definite difference between various 
76 Bordet. Ann. de VInst. Past., 1897. 
77 Neufeld and Handel. Zeitschr. f. Imm., Yol. 3, 1909, and Arb. a. d. 
kais. Gesundh. Amt., Yol. 34, 1910. THERAPEUTIC IMMUNIZATION 
557 
pneumococci in their reactions to immune serum; in other words, 
pneumococci could be grouped into various serological types. The 
serum produced with organisms of one type did not protect against 
infection with other strains. In consequence they called attention 
to the importance of determining the type of pneumococcus which 
causes the individual pneumonia so that the corresponding immune 
serum might be used. They produced a highly potent anti-pneu- 
mococcus serum by the injection of horses and donkeys with virulent 
pneumococi grown on fluid cultures, then determined the high pro- 
tective power of this serum upon animals and used it in the treat- 
ment of patients by intravenous injection. Their results were ex- 
ceedingly encouraging. In reporting their results Ueufeld and 
Handel stated that considerable doses must be given. They called 
attention to the fact, revealed by their animal experiments, that 
moderate amounts do not, as in the case of diphtheria serum, exert 
a correspondingly slight amount of beneficial action, but that in the 
case of the pneumonia serum amounts smaller than a certain active 
minimum seem to exert absolutely no beneficial action. This is a 
fact which later was also determined by Dochez. 
The entire subject of pneumococcus serum treatment has been 
further cleared by the work of Dochez and Gillespie 78 and Dochez 
and Avery 79 in their studies on pneumococcus types. Taking their 
departure from the observations of Ueufeld and his associates, they 
have determined for the eastern United States the prevalence of four 
pneumococcus types which are too familiar to bacteriologists at the 
present time to necessitate reiteration in this place. We may, how- 
ever, remind the reader that their classification divides pneumococci 
into two fixed types, I and II, a third group, Pneumococcus Mucosus, 
and a fourth heterologous group generally spoken of as Type IY. 
These types have also been found in Europe, and Lister has found 
similar typing among cases in South Africa with, however, the dif- 
ference that his Type C and B corresponded to Type I and II of the 
American classification, but his homologous Type A has so far not 
been found in other countries. 
The discovery of individual serologically homologous types has 
made possible the logical development of serum therapy. Cole, with 
his associates at the Rockefeller Hospital, has given the matter serious 
study, and has carried on extensive therapeutic investigations on a 
considerable number of cases. Sera produced according to his 
method by various health departments and commercial laboratories, 
have further permitted an extensive plan of investigation. 
The production of the sera at the present time consists in the 
injection of horses with, at first, killed cultures and then living pneu- 
78 Dochez and Gillespie. Jour. A. M. A., Yol. 61, 1913, p. 727. 
79 Dochez and Avery. Jour. Exp. Med., Yol. 26, 1917, p. 477. Avery, 
Chickering, Cole and Dochez, Monograph No. 7, Rock. Inst., October, 1917. 558 
INFECTION AND RESISTANCE 
mococci. The organisms are grown on broth and young cultures are 
used for injection. Cole advises injecting daily for 6 days, using the 
sediment thrown down from 50 c. c. of a 12 hour broth culture killed 
at 56°. After six injections, a rest of a week is given, and the serum 
of the horse tested for agglutinins. This is followed by a second 
series of dead cultures, and again an interval is allowed. When the 
agglutination test works out specifically in dilutions of 1-200, and 
0.2 c. c. protects against 0.1 c. c. of a highly virulent culture, the 
serum can be used, but usually immunization is continued with living- 
cultures to a higher point. 
The standardization of the serum is carried out by protection 
experiments in mice, with the determination of the amount of serum 
necessary to protect a 20 gram mouse against a standard virulent 
culture. The most important part of this technique is to possess a 
culture of high and constant virulence. Such cultures are produced 
by passage through mice. The virulence should be so great that a 
millionth of an 18 hour broth culture will kill a mouse in 48 hours. 
Dilutions are then prepared in such a way that 0.5 c. c. of a dilution 
contains various amounts of the pneumococcus culture, ranging from 
0.2 c. c. to 0.000001 c. c.. The dilutions must be freshly made in 
order that the number of organisms in the tubes should not materially 
change by growth before the test is made. With each of these dilu- 
tions, then 0.2 c. c. of the serum to be tested is mixed in a syringe 
and the mixture immediately injected intraperitoneally. The rule 
laid down by Cole is that only such sera should be employed in which 
0.2 c. c. protects against 0.1 c. c. of the culture described above. 
Since type serum is used, it goes without saying that the serum 
should be used only in cases in which the type infecting the patient 
is known. This must be done by obtaining sputum from a patient, 
taking particular care to make sure that the sputum is collected 
after thoroughly rinsing the mouth with bicarbonate of soda or salt 
solution. This is collected in a sterile Petri dish and is immediately 
sent to the laboratory. It is stained by Gram, and a capsule stain 
is made. It is then washed, dipping into successive Petri dishes 
containing salt solution to remove the bacteria sticking to the out- 
side, and a small bit of the sputum is emulsified in salt solution and 
injected intraperitoneally into a mouse. When the mouse is either 
dying or dead, it should be immediately autopsied, cultures taken 
from the heart’s blood and the peritoneal exudate, washed out with 
salt solution into a centrifuge tube by means of a nipple pipette. 
It is centrifuged for a few moments at low speed to throw down 
leukocytes and the turbid supernatant fluid, which should have the 
maximum turbidity of a well grown broth culture, is transferred to 
another centrifuge tube and centrifuged at high speed to throw down 
the organisms. This sediment is resuspended and tested with type 
sera for agglutination according to the following scheme taken THERAPEUTIC IMMUNIZATION 
559 
directly from the Monograph by Avery, Chickering, Cole and 
Dochez.80 
Other methods of typing have been devised for exceptional cases 
which can be found described in any of the more recent textbooks 
of bacteriology.81 
Determination of Pneumococcus Types by Agglutination 
Serum I 
Serum II 
Serum II 
Serum III 
Pneumococcus 
(1:20) 
(undiluted) 
(1:20) 
(1:5) 
Suspension, 0.5 c. c. 
0.5 c. c. 
0.5 c. c. 
0.5 c. c. 
0.5 c.c. 
Type I 
Type II 
++ 
+ + 
++ 
Sub-groups 11a, b, x 
Type III 
Type IV 
. 
+ 
•• 
+ + 
Incubation for 1 hour at 37° C. 
Although much has been written about the results of serum treat- 
ment in pneumonia, we can for the present do no better than accept 
the judgment of Cole and his co-workers who have studied the prob- 
lem extensively and have, in various places, published an unpreju- 
diced opinion. For some unknown reason, no success has been 
obtained with anything but Type I serum. For the present, there- 
fore, there is no experimental basis for serum treatment in cases 
other than those caused by the Type I pneumococcus. According 
to Cole, there seems to be very little doubt that in Type I cases the 
use of serum has reduced the mortality by about 50 per cent., and 
while there are many dissenting opinions, this, as far as we can 
ascertain, is the general judgment of most clinicians with whom we 
have discussed the matter, who have given the method serious test. 
We can assume, therefore, that in Type I pneumonias, the use 
of serum is indicated. It is, of course, necessary to use it as early 
as possible in the disease. Skin reactions for horse serum hyper- 
sensitiveness should be done on the patient and, if necessary, desen- 
sitization should be undertaken, according to the principles described 
in another section of this book. 
For first injection, Cole 82 advises 90 to 100 c. c. of the serum, 
diluted 50 per cent, in salt solution and injected intravenously by 
gravity with such slowness that the first 10 c. c. take about 10 to 15 
minutes to run in. Cole establishes the effectiveness of the serum 
by the fact that the patient’s blood becomes sterile, that often the pro- 
gression of local lesions in the lung is arrested, and that there is a 
80 Avery, Chickering, Cole and Dochez. Loc. cit. 
81 See Hiss, Zinsser and Russell. “Textbook of Bacteriology,” 5th Edi- 
tion, I). Appleton, New York, 1922. 
82 Cole. Jour, 4, M, A., Vol, 76, 1921, p. 111. 560 
INFECTION AND RESISTANCE 
great amelioration in objective and subjective symptoms. He states 
that the mortality rate in untreated lobar pneumonias of Type I has 
been shown to be from 25 to 30 per cent. Of 495 cases collected, the 
mortality was 10.5 per cent, after serum treatment. Further statis- 
tical studies will settle this question. 
In addition to the reactions generally classified as anaphylactic 
following the injections of the serum, serum sickness may occur, and 
there may be a marked thermal reaction within 20 minutes or one 
hour, following administration. This is probably similar to that 
described in another section of this book in connection with the in- 
jection of non-specific protein and protein cleavage products in 
various infections. 
The exact manner of action of the serum is, of course, unclear. 
Dochez has shown that in untreated pneumonia cases, antibodies do 
not appear in the serum until about the time of the crisis. In the 
treated cases, of course, antibodies are supplied immediately after 
the injection of the serum and, according to the studies o'f Bull 
detailed in another part of this book, an almost immediate agglutina- 
tion of the pneumococci in the blood stream takes place. If nothing 
else is accomplished by the serum injection, the reduction of the 
septicemia may be of enormous benefit to the patient. 
More recently the treatment of pneumonias with the so-called 
purified antibody extracts of Huntoon has been extensively attempted 
by Cecil and his co-workers in Hew York. These antibody prepara- 
tions have been described in detail in our section on the dissociation 
of antigen from antibody. 
The Huntoon pneumococcus antibody extracts have been used 
particularly by Cecil and Larsen 82a at Bellevue Hospital in Hew 
York, and by Dr. Lewis A. Conner.8215 Both observers carefully typed 
their cases in every instance. The antibody extracts were made with 
horse serum potent against Types I, II and III, as a point of depar- 
ture. This serum was absorbed with heavy emulsions of the or- 
ganisms of these types until the suspensions were heavily agglu- 
tinated, the sediment washed with salt solution, and finally extracted 
in 0.25 per cent, sodium bicarbonate at 55° C. for from three-quarters 
to one hour. The material was prepared by Huntoon, contained 
protective substances against Types I, II and III, equal in amount 
to potent polyvalent antipneumococcus serum. The conclusions of 
Cecil and Larsen from a considerable number of cases are as follows: 
In 424 cases of pneumococcus pneumonia treated with the antibody 
solutions, the death rate was 21.4 per cent., while in a control series 
of 428 cases in the same institution, the death rate was 28.3 per cent. 
The most striking results of the pneumococcus antibody were ob- 
served with Type I. In 156 treated Type I cases, the death rate was 
82a Cecil and Larsen. Jour. A. M. A., 79, 1922, p. 343. 
82b Conner, Lewis A. Amer. Jour. Med. Scien., 164, 1922, p. 832. THERAPEUTIC IMMUNIZATION 
561 
13.3 per cent., while a control series of 162 showed a death rate of 
22.2 per cent. A definite but less marked effect was seen in the Type 
II and Type IV cases. There was no effect whatever in Type III 
cases. Streptococcus pneumonias were not favorably influenced. 
There seemed to be a shortening of the course of the disease also, in 
that 28.8 per cent, of the treated cases recovered on or before the fifth 
day, while only 7.9 per cent, of the untreated cases recovered as early 
as this. There was a slight difference in the severe complications 
occurring in treated and untreated cases, in favor of the treated cases. 
These statistics, gathered from a very carefully observed material, 
seem definitely to show that a certain amount of therapeutic value 
may be attributed to the antibody extracts, but it may still be ques- 
tioned whether with such relatively small differences, this should be 
interpreted as a specific immunological protection, or perhaps as the 
non-specific effect of the bacterial materials injected with these 
extracts. Moreover, it does not permit us to compare the relative 
values of antibody extracts and pneumococcus serum. 
Similar studies by Lewis A. Conner, again in very carefully typed 
cases, are summarized by him as follows: In Conner’s observations 
there were no untreated controls used, partly because of the relatively 
small number of cases of this kind coming to the hospital in the 
course of a year, and probably also because of a conscientious feeling 
that if the treatment proved effective, it would be unfair not to let 
every patient have the benefit of it. His observations were carried 
over two winter seasons in order to lessen the possibility of seasonal 
variations in the severity of pneumonias and consequent mortality; 
116 cases of lobar pneumonias in adults were treated during this 
period, with a death rate of 14.6 per cent. Among these were 13 
Friedlander and streptococcus pneumonias, which showed a death 
rate of 46.1 per cent. He calls attention to the fact that 54.4 per 
cent, of the cases happened to be Type IV, with the very low mor- 
tality of 4.1 per cent., and he justly concludes that since the effect 
on the Type IV cases seemed to be as potent as the effect on Types 
I and II, one might naturally question the specific nature of the 
benefit derived from it. That the non-specific element is a strong one, 
would also follow from the relatively severe immediate reactions fol- 
lowing the injections. In one patient, Conner says that he believes 
death may have been directly traceable to the severity of this imme- 
diate reaction. He, nevertheless, believes the method to represent a 
forward step in rational treatment of pneumonia. 
The Serum Treatment of Typhoid Fever 
The first extensive attempts to treat typhoid fever by passive im- 
munization with the serum of treated animals were made by Chante- 
jnesse; who immunized horses with filtrates of typhoid cultures sub- 562 
INFECTION AND RESISTANCE 
cutaneously and with emulsions of virulent bacilli intravenously. 
Chantemesse believed that the serum of horses which had been 
treated in this way for very long periods possessed, not only bacteri- 
cidal action, but stimulated phagocytosis, and possessed a certain 
limited amount of neutralizing power against the toxic properties of 
the typhoid filtrates. At the International Congress of Hygiene in 
Berlin in 1907 Chantemesse83 reported upon a thousand cases 
treated with his serum. Of this number 43 only died, whereas the 
average mortality during the same six years at the Paris hospitals 
was 17 per cent. The injection of the serum he claimed very mark- 
edly improved the condition of patients in that, after a preliminary 
period of no apparent change lasting from several hours to 5 or 6 
days, the temperature goes down and the general condition of the 
patient changes considerably for the better. He noticed very few 
complications in these cases, and intestinal hemorrhage occurred four 
times only. 
A remarkable feature of Chantemesse’s treatment is that he in- 
jected into the patients a few drops only of the serum, and rarely 
made a second injection, facts which alone tend to persuade one that 
his apparent therapeutic success was a fortunate accident. 
The opinion originally expressed by Chantemesse that the serum 
of horses vigorously treated with typhoid bacilli possesses in addition 
to its bactericidal and opsonic powers definite antitoxic properties 
recurs again in the work of a number of investigators. Besredka 84 
prepared a serum by the intravenous injection of typhoid cultures 
heated to 60° C., continuing the immunization for 6 months. He 
claims that this serum possesses what he designates as “anti-endo- 
toxic” properties. A dry extract of typhoid bacilli which in dose of 
0.01 gram killed a guinea pig of 300 grams regularly became innocu- 
ous when mixed with small quantities of this horse serum. One c. c. 
of the horse serum neutralized often as much as two fatal doses of the 
serum, but it is important theoretically to recognize that Besredka 
states particularly that even an increase of the quantity of serum 
never neutralized more than two fatal doses. This is particularly 
important in connection with the more recent studies on toxic split 
proteins by Vaughan, and on anaphylatoxins by Bessau and by Zins- 
ser and Dwyer, in which it has been shown that an animal acquires 
a tolerance against the toxic substances produced from bacterial and 
other proteins which, however, never exceeds one or two multiples 
of the minimum lethal dose. This fact alone would militate against 
considering the serum of Besredka in any way antitoxic in the sense 
in which the word is used concerning diphtheria and tetanus anti- 
toxins where neutralization of poison follows roughly the law of 
83 Chantemesse. International Congress of Hygiene, Berlin, September, 
1907; Ref. Bull, de Vlnst. Past., Yol. 5, 1907, p. 931. 
B<jsr§dka. Ann. de Vlnst, Past., Yol. 19, 19Q5, and Yol. 20, 1906A THERAPEUTIC IMMUNIZATION' 
563 
multiples. Besredka’s anti-endotoxic sera has recently been very 
thoroughly investigated by Pfeiffer and Bessau.85 These investi- 
gators have found that Besredka’s serum exerted a very definite 
beneficial influence upon typhoid infection in guinea pigs if injected 
at the same time with the bacilli. In their experiments it also pro- 
tected somewhat against the toxic properties of substances derived 
from the typhoid bacillus, and Pfeiffer and Bessau did not believe 
that this was due to a true antitoxic action, nor that the serum was 
superior in this respect to the ordinary bactericidal sera prepared by 
inoculating animals with typhoid bacilli. Kraus and Stenitzer 86 
have also taken up the study of typhoid immunization from the 
point of view that the typhoid bacillus produces a true toxin, and that 
therefore a true antitoxic action could be expected from the sera pro- 
duced by immunization with typhoid filtrates. It should be noted 
that, in spite of the most common opinions against this at present, 
a similar point of view was advanced by MacFadyen,87 and more 
recently by Arima.88 Kraus and Stenitzer 89 immunized horses and 
goats very highly with extracts of agar cultures and with broth fil- 
trates by intravenous injection. The serum which they produced in 
this way not only possessed the ordinary bactericidal action, but, 
they claimed, neutralized also toxic broth filtrates, not only of the 
typhoid, but of the paratyphoid bacilli. The serum of Kraus and 
Stenitzer has been used by a number of observers, among whom are 
Ilerz,90 Forssmann, Unger, Russ, and others, and the results are said 
to be encouraging in early cases. 
Rodet and Lagrifoul 91 immunized horses with living typhoid 
cultures, and also claim favorable results. 
Mathes,92 continuing the work of Gottstein after the death of the 
latter, employed the method of immunizing with the product ob- 
tained by digesting typhoid bacilli with trypsin. The poison so pro- 
duced he speaks of as “fermotoxin.” Liidke 93 slightly modified the 
Gottstein-Mathes method by digesting the typhoid bacilli with pepsin 
and hydrochloric acid, and with the poison so produced immunized 8 
goats, reenforcing the immunization by the subsequent injection of 
the bacilli themselves. He claims that 0.05 to 0.1 c. c. of the serum 
so produced protected animals against five times the lethal dose 
85 Pfeiffer and Bessau. Centralbl. f. Bakt., Vol. 56, 1910. 
86 Kraus and Stenitzer. Wien. klin. Woch., Yol. 20, 1907, pp. 344, 753, 
and Vol. 21, 1908, p. 645. 
87 MacFadyen. Cited from Stenitzer in “Kraus und Levaditi Handbuch,” 
Vol. 2. 
88 Arima. Centralbl. f. Bakt., Yol. 65, 1912, p. 183. Orig. 
89 Kraus and Stenitzer. Wein. klin. Woch., Vol. 22, 1909, p. 1395; 
Deutsche med. Woch., March, 1911. 
90 Herz. Wien. klin. Woch., Vol. 22, 1909, p. 1746. 
91 Rodet and Lagrifoul. C. R. de la Soc. de Biol., April, 1910. 
92 Mathes. D. Archiv f. klin. Med., Vol. 95, 1909. 
93 Liidke. D. Archiv. f. klin. Med., Vol. 98, 1910. 564 
INFECTION AND RESISTANCE 
of the poison. In a small series of human cases treated by this 
method he reports good results. 
Garbat and Meyer 94 immunized animals with sensitized typhoid 
bacilli, and claim that the most potent sera for typhoid immunization 
can be obtained by the combination of sera produced by the injection 
of sensitized and of unsensitized bacteria. They assert that the 
typhoid bacillus contains two definite antigens, one particularly as- 
sociated with the bacterial ectoplasm, which becomes active when the 
bacteria enter the animal body, and a truly endocellular poison which 
does not become active until the surrounding ectoplasm is dissolved. 
They believe that sensitizing bacteria is a method for the production 
of endotoxin, and think that therefore the ideal serum for the treat- 
ment of typhoid consists of a mixture of two sera produced each 
with one of the antigens, that is, with sensitized and unsensitized 
bacteria. Rommel and Herman 95 did not obtain encouraging results 
with this serum. 
From a study of the literature it seems to us that in spite of the 
many different methods of production employed by various observers 
in their studies on typhoid sera it is quite likely that all these sera 
are essentially alike, containing, quantitatively, according to the de- 
gree of immunization, bactericidal, agglutinating, and opsonic prop- 
erties, with possibly a limited amount of neutralizing power for the 
poisons liberated from the typhoid bacilli in the body. As far as we 
can judge from clinical reports the therapeutic value of the sera 
so far produced is not very great. It seems that cases treated early 
in the disease may be benefited, and possibly an early cessation of 
the bacteriemia can in this way be attained. However, it does not 
seem either theoretically or from the study of clinical publications 
that any very marked effects have followed the use of any of the sera 
in advanced cases. 
The Serum Treatment of Plague 
That the serum of animals immunized with killed plague cultures 
may actively protect normal animals from experimental infection 
was first shown by Yersin, Calmette, and Borrel.96 The serum which 
they produced possessed apparently powerful bactericidal action, 
but no antitoxic properties were demonstrated. They determined 
its protective powers by injecting measured quantities into mice and 
infecting them with fatal doses of virulent plague bacilli 24 hours 
later. The Yersin serum which was produced for the treatment of 
plague as a result of these experiments was made, then, by the 
gradual immunization of horses with first dead plague bacilli, finally 
with virulent living organisms. The serum has been extensively 
04 Garbat and Meyer. Zeitschr. f. exp. Path. u. Ther., Yol. 8, 1911. 
95 Rommel and Herman. Centralbl. f. Bakt. Ref. Yol. 53, 1912. 
06 Yersin, Calmette, and Borrel. Ann. de VInst. Past., 1895. THERAPEUTIC IMMUNIZATION 
565 
used by many observers with results that leave one much in doubt 
as to its efficacy. Yersin 97 himself, reporting on an epidemic in 
Nhatrang, reports a general mortality of 73 per cent, for the whole 
epidemic, a mortality of 100 per cent, in untreated cases, and of 42 
per cent, among those treated with his serum. Good results were 
also reported from the epidemics in Amoy and Canton in 1896. 
However, these results apparently were not accepted by all observers 
as proving the efficiency of the serum, since the number of cases 
observed were few, and the irregularity in the gravity of the disease 
in different individuals makes statistical evidence unreliable unless 
large material can be studied. Kolle and Martini 98 announce that 
Dr. Choksy reported very poor success with the Yersin serum, and 
cite a number of later writers whose results with this serum were 
also unsatisfactory when used on human beings. That the serum 
unquestionably contains antibodies against the plague bacillus is 
testified to, not only by the French observers themselves, but also by 
the German Plague Commission of 1899, and by Kolle and Mar- 
tini 99 themselves. The Commission experimented with this serum 
upon monkeys, and showed that it possessed unquestionable protec- 
tive powers in rodents and in monkeys when given 24 hours before 
the plague infection, and in monkeys possessed fair curative prop- 
erties when injected 24 hours later than inoculation with the plague 
bacilli. Because of the doubtful success in the treatment of human 
beings with this serum Yersin and Roux at the Pasteur Institute 
later altered their methods of serum production by injecting, not 
only dead and living plague cultures, but considerable quantities of 
culture filtrates after the horses had attained a high degree of im- 
munity. Later observations on the Yersin 100 serum have been pub- 
lished by the British Plague Commission in 1908 and 1911. In this 
investigation the cases were controlled as to their severity by blood 
culture, since it had been claimed by a number of earlier investi- 
gators that the Yersin serum was efficient in mild cases, but failed 
entirely in the severe ones. It seems from the report of this Commis- 
sion that ordinarily 70 per cent, of cases of plague without bacilli in 
the blood survive while three-quarters of those with mild septicemia 
die, and all of those with a marked septicemia succumb. In the 
summary given of 146 cases treated with Yersin’s serum by the 
British Commission 65.1 per cent, died, whereas of 146 untreated 
controls 71.90 per cent. died. These fixures, together with an 
analysis of the percentages, classified according to the severity of the 
infections, do not show a very marked curative action on the part 
of the serum. 
97 Yersin. Ann. de VInst. Past., 1899. 
98 Kolle and Martini. Deutsche med. Woch., 1902, p. 29. 
99 German Plague Commission. Arb. a. d. kais. Amt., Yol. 16, 1899. 
100 British Plague Commission. Jour. Hyg., Yol. 12, Sup., 1912, p. 326. 566 
INFECTION AND RESISTANCE 
Mark!,101 who claims that the plague bacillus produces a soluble 
toxin, has produced a plague serum by immunization of animals by 
filtrates of broth cultures. He claims that 0.1 c. c. of his serum, as 
produced at Vienna, will protect various animals against lethal doses 
of plague bacilli if given at the same time. He attributes much of 
its curative action to the fact that in the presence of this serum active 
phagocytosis takes place. 
Dean,102 in 1906, also claimed to have produced strongly anti- 
toxic plague sera by treating horses with filtrates from 8 to 10 weeks’ 
old bouillon cultures. He claims that 1 c. c. of his serum will neu- 
tralize 150 or 450 minimal lethal doses of the plague poison according 
to whether one measures the M L D by death in 48 hours or in 4 days. 
Rowland 103 also has produced a serum by the immunization of ani- 
mals with the “toxins” produced by his sulphate process. Rowland 
has apparently utilized the idea previously advanced by Lustig of 
immunizing with “nucleoproteins” derived from the plague bacillus 
instead of with the whole bacteria. Lustig5s 104 method consisted of 
washing up agar cultures of plague bacilli in 1 per cent, sodium 
hydrate solution, precipitating with ascitic acid, taking up the pre- 
cipitate in an indifferent fluid and injecting it into horses. The 
serum produced by Lustig’s method was used in Bombay, and is re- 
ported by Hahn as effective in milder cases, but without action in 
the more severe ones. There was but slight difference in the latter 
type between the treated and the untreated cases. 
Rowland’s 105 method consisted in the treatment of the moist 
bacteria with enough anhydrous sulphate of soda to combine with 
all the water present, freezing and thawing the mixture and filtering 
off the bacterial deposit at 37° C. Subsequently he extracted this 
bacterial mass with water. The extract so obtained was fatal to 
rats in quantities of 0.05 to 0.1 mg., killing them in 18 hours. In 
his experiments doses of 0.001 to 0.01 afforded protection, the last- 
named quantity reducing the mortality after inoculation of fatal 
doses from 80 per cent, to 10 per cent. 
The sera produced by the immunization of horses with these 
supposed nucleoproteids are taken to be antitoxic in nature by Row- 
land himself and by MacConky. They were used in the treatment of 
plague cases in the epidemics of 1908 and 1911 by the Maratha Hos- 
pital in Bombay, and reported upon by the Indian Plague Commis- 
sion on the basis of observations made by Dr. Choksy. The cases in 
101 Markl. Centralbl. f. Bakt., Yol. 24, 1898; Zeitschr. f. Hyg., Yol. 37, 
1901; Zeitschr. f. Hyg., Vol. 42, 1903. 
102 Dean. Cited from MacConkv, Jour. Hyg., Vol. 12, Plague Suppl. 
II, 1912, p. 402. 
103 Rowland. Jour. Hyg., Vol. 11, Plague Suppl. I, pp. 11-19. 
104 Lustig. “Monograph Sierotrapia e Vaecin Prev. Control la Peste,” 
Turin, 1899; cited from Kolle and Martini, loc. cit. 
105 Rowland. Jour. Hyg., Vol. 10, p. 536. THERAPEUTIC IMMUNIZATION 
567 
this series were controlled, as were those treated by the Yersin serum, 
by blood culture. Here the results were not striking—68.40 per 
cent, of the serum-treated cases died, while 77.60 per cent, of the 
controls died. 
Altogether we cannot draw any definite conclusions as to the 
value of the serum treatment in plague. On the whole it does appear 
that the milder cases are materially benefited by the treatment, and it 
is not at all impossible that in such cases aggravation of a milder 
case into fatal septicemia may be prevented by the timely adminis- 
tration of the plague serum. Animal experimentation also seems to 
indicate that the administration of the serum may be of great value 
as a prophylactic measure. It seems, on the other hand, as far as we 
can judge from the evidence of statistics, that when a case of plague 
has developed into the condition of active septicemia the administra- 
tion of even the strongest plague sera at present available is of little 
use. And this is indeed unfortunately true of all passive immuniza- 
tion where the activity of the serum seems to depend chiefly upon 
bactericidal and opsonic properties. For we cannot definitely accept 
at the present day the claims that a true antitoxic serum, in the sense 
of those produced against diphtheria and tetanus poisons, can be 
really produced in the case of plague. The toxic substances derived 
from plague bacilli by a number of observers do not correspond in 
many particulars to true toxins. 
Passive Immunization with Anthrax Serum—Anthrax serum 
has been prepared by a great many observers, and has been more 
closely studied by Sobernheim.106 The serum may be prepared by 
the active immunization of sheep. Beginning with a preliminary 
treatment with Pasteur’s vaccines and following this with gradually 
increasing doses of eventually virulent anthrax bacilli. Sheep are 
preferred by Sobernheim for serum production. He believes that 
a potent serum has very definite beneficial effects in the passive pro- 
phylactic protection of cattle, sheep and horses, and may even save 
animals that have been infected. Sclavo has also applied passive 
immunization in infected human beings. According to him, the 
best method of treatment is the injection of 30 to 40 c. c. of serum 
subcutaneously, followed in 24 hours by a similar dose. He has 
injected 10 c. c. or more intravenously in severe cases. As far as 
we can tell from the collected cases, serum treatment is advisable 
in all cases in which a severe anthrax infection has been determined. 
INFECTION AND IMMUNITY IN POLIOMYELITIS 
The active experimental investigation of poliomyelitis started 
with the discovery by Landsteiner and Popper,107 that monkeys 
106 Sobernheim. Zeitschr. f. Hyg., Yol. 25, 3897, and Yol. 31, 1899. 
107 Landsteiner and Popper. Zeitschr. f. Immunitdts., Yol. 11, 1909. 568 
INFECTION AND RESISTANCE 
could be inoculated with this disease by the intracerebral or the 
intraperitoneal injection of saline emulsions of brain or spinal cord 
of individuals dead of the disease. The same observations by Flex- 
ner and Lewis 108 a little later, and then by a large number of work- 
ers throughout the world, gave an opportunity for the careful study 
of immunological conditions, and of the nature of the virus. It is 
our own opinion that the disease is caused by the globoid bodies of 
Noguchi and Flexner 109 and not by streptococci. However, the work 
on this controversy may be regarded as unfinished at the present 
writing, and we may abstain from a prolonged discussion of the 
etiological factor. The mere fact that the virus of poliomyelitis has 
been found to keep as long as four to six years in glycerinated nerv- 
ous tissue, and the analogy which this offers to such virus as that 
of rabies, small-pox, etc., make it alone seem likely that true bac- 
teria are not concerned in the cause of the disease. Our own ex- 
periments, and those of Dr. Tsen of this laboratory, on the isolation 
of streptococci from poliomyelitis animals, incline us to think that 
animals afflicted with this disease are readily subject to secondary 
tissue infection with organisms of various kinds, but chiefly with 
streptococci of the viridans type, which are so universally distributed 
throughout the body. 
It has long been suspected that one attack of poliomyelitis pro- 
tected from subsequent infection. It is plain therefore that some 
form of active immunization takes place in the course of the disease. 
Experimentation on monkeys subsequently confirmed this, in that it 
was shown that monkeys that had contracted the disease, and re- 
covered, were thereafter resistant to inoculation. Strangely enough, 
however, monkeys that have been unsuccessfully inoculated are just 
as susceptible as they were before, showing that the immunity in 
this disease is closely analogous to that existing in syphilis and some 
other diseases where inoculation with attenuated virus, dead virus, 
or sub-infectious doses of living virus, is entirely incapable of pro- 
ducing immunity. 
Early in the study of poliomyelitis it was found that one attack 
protected and that the serum of a human being or monkey that had 
recovered from the disease was capable of neutralizing the virus. 
The artificial production of immune sera by the injection of virus 
has been generally unsuccessful. The only hopeful fact we know 
in this regard is the fact that the serum of individuals who have 
recovered from an attack of the disease will neutralize the virus if 
mixed with it before injection. For this reason, Netter,110 Netter 
and Salanier,111 Amoss and Chesney 112 and others have attempted 
108 Flexner and Lewis. Jour. Amer. Med. Assn., Yol. 44, 1910, p. 45. 
109 Flexner and Noguchi. Jour. Exp. Med., Yol. 18, 1913, p. 461. 
110 Netter. Bull, de l’Acad. Med., Series 3, Vol. 74, 1915, p. 403. 
111 Netter and Salanier. Bull. Soc. Med. des Hop. de Paris, Series 3, Vol. 
40, 1916, p. 299. 
112 Amoss and Chesney. Jour. Exp. Med., Vol. 25, 1917, p. 581. THERAPEUTIC IMMUNIZATION 
569 
to use the serum of recovered cases in the treatment of patients with 
the disease. The following summary of the treatment is taken largely 
from Amos.113 Blood is withdrawn from patients that have re- 
covered from the disease in quantities varying according to the size 
and age of the subject. In children of ten, as much as 200 c. c. may 
be taken, and in adults 500 c. c. or even more. The serum, separated 
from the clot, is inactivated at 55° for one hour and tested for 
sterility. Wassermann reactions should be done on the donors. The 
administration of the serum after the febrile period is, according to 
Amoss, of very doubtful value. It must be administered early before 
the onset of paralysis and, for this reason, a microscope carried to 
the bedside may make it possible to examine the fluid for cells and 
globulin while the needle is still in place, and if the diagnosis is a 
positive one, intraspinous injection of the serum may be given imme- 
diately through the same needle without a second puncture. Fluid 
is withdrawn, until the flow is quite slow—drop by drop. Amoss 
advises slow injection, by the gravity method, of an amount of serum 
which equals the amount of fluid withdrawn, less ten in terms of 
cubic centimeters. Care should be taken that the serum is warmed 
to body temperature before injection. Repetition of the treatment 
may be practiced after 24 hours, if the temperature is not normal. 
ISTo other form of serum treatment has had any effect in this disease 
so far. 
113 Amoss. Jour. A. M. A., Yol. 76, 1921, p. 110. CHAPTER XXI 
THERAPEUTIC IMMUNIZATION (Continued) 
Active Immunization in Man. (Prophylactic and Therapeutic.) 
In discussing the value of active immunization in man we must 
distinguish sharply between active immunization which is prophy- 
lactic and that which is carried out after the disease has gained a 
definite foothold in the body. In the former case we are dealing with 
an old method and with one upon which the very foundations of 
our knowledge of immunity have been built. It is the method of 
Jenner in small-pox. It is that of Pasteur in chicken cholera, in 
anthrax, and in many other infections. It has been used as a routine 
in animal experimentation in laboratories since the first days of the 
systematic study of infections. There is no question about its being 
a rational and logical procedure. The immunity which can be easily 
conferred upon a healthy individual in this way need not be exten- 
sively above the normal in order to protect from invasion by the small 
numbers of pathogenic germs which may gain entrance under con- 
ditions of accidental, spontaneous infection. 
The possibilities of the method were recognized by Ferran, a 
pupil of Pasteur, who applied it to cholera, and, since his time, it has 
been extensively attempted in many of the infectious diseases which 
occur epidemically, and therefore justify attempts in this direction. 
In essence also Pasteur’s method of active immunization in rabies 
represents such prophylactic vaccination, since, in this case, although 
treatment is begun after infection has taken place, nevertheless the 
process of immunization is carried out during the incubation period 
before active manifestations of the disease have set in. Prophylactic 
vaccination, therefore, is a valuable procedure which has reaped 
remarkable results of recent years, especially in protection against 
typhoid fever. In a subsequent section this phase of vaccination 
is more extensively discussed, and we may therefore leave it for the 
present. 
An entirely different problem is presented by the conditions pre- 
vailing in cases in which a bacterial infection is actually going on 
at the time that vaccine treatment is proposed. Here the bacteria are 
already present and growing in the body, and a certain amount of 
antibody reaction is being stimulated automatically as a result of 
their presence. In how far the administration of vaccines is justifi- 
570 THERAPEUTIC IMMUNIZATION 
571 
able or even logical in such cases, is a question which cannot be 
briefly answered and which depends upon an analysis of the pre- 
vailing conditions in each individual case. 
We can approach the problem best by roughly classifying the 
various forms in which infection occurs in the human being. 
When bacteria gain entrance into the tissues of the human body, 
granted that the organisms are pathogenic, an immediate struggle 
ensues between the offensive properties of the micro-organisms and 
the defensive powers of the tissues. The factors wTiich determine 
the outcome of such a combat have been more fully considered in 
Chapter I. Briefly, if the defensive powers of the body greatly 
preponderate the result is localization and rapid destruction of the 
micro-organisms—with cure. In such a case any form of treatment 
is unnecessary. On the other hand, the balance of power may be 
turned in the opposite direction, in which case the infectious process 
becomes rapidly generalized, the bacteria enter the blood stream 
and lymphatics, and the defensive powers are overwhelmed. In 
such a case also active immunization with vaccines is entirely use- 
less. 
There are cases, however, in which the struggle is a more equal 
one, and in which the infectious process is held in check by the 
defenses, so that it takes a slow, chronic, localized form, and spreads, 
if at all, very slowly. What is it in such a case that prevents com- 
plete healing of the process ? The answer to this may be found both 
in local and in systemic causes. Locally the lesion, after the pre- 
liminary skirmishes, may become encapsulated either by fibrin 
formation, clot, or other tissue changes so that, as Wright suggests, 
the fluid constituents of the blood-plasma cannot easily approach the 
organisms in the lesion. The same effect may result from internal 
pressure by fluid and possibly by the presence of considerable quan- 
tities of tissue detritus, by which protective serum constituents are 
fixed and thus diverted from the bacteria. Against these factors, of 
course, no form of immunization can be of value. Wright recog- 
nizes this, and suggests the use of surgical evacuation, Bier’s method, 
X-rays, Finsen light, heat, and a number of other localized methods 
of increasing the blood supply. This, too, may be the reason for 
the benefits derived from wet dressings, in that they keep the tissues 
macerated, soft, and moist. At any rate, it is a matter of local 
surgical treatment. At the same time, however, there may be sys- 
temic causes which prevent the complete healing of such lesions, 
namely, an insufficient supply of circulating antibodies, opsonic or 
bactericidal substances. These may be sufficient to hold the lesion 
in check, but since small quantities of bacteria only are in contact 
with the blood stream, relatively small amounts of antigen are ab- 
sorbed and antibody formation is consequently deficient. Here we 
have an ideal condition for vaccine therapy. By isolation of the 572 
INFECTION AND RESISTANCE 
organisms from the patient’s lesion, for which, in this case, there 
is time, and the careful immunization of the patient with these 
organisms, the immunity may be considerably increased and cure 
effected. 
Closely related to this type of lesion are those conditions in 
which there are localized infections which heal rapidly but recur in 
quick succession again and again. Such are the common cases of con- 
secutive crops of boils; and not dissimilar are the manifestations of 
erysipelas where the lesion extends along the edges while it heals in 
the center. There is in this type, probably, a very close balance 
between protection and offense; the defensive reaction is sufficient to 
overcome the localized lesion, but insufficient to set up a permanent 
systemic protection. A certain amount of local immunity acquired 
by the tissues of the affected areas may suffice to throw slight weight 
into the balance on the side of protection, enough at least to decide 
the struggle; and this element of locally acquired tissue resistance is 
in all probability also the cause for the failure of these lesions to 
recur immediately in the same area. Here, too, treatment with vac- 
cines is not illogical and may yield good results if properly carried 
out. 
In generalized systemic infections we must sharply distinguish 
between cases of acute sepsis in which the bacteria are actively grow- 
ing and multiplying in the circulation and other cases in which blood 
cultures are positive only because the bacteria are being constantly 
discharged into the circulation from a focus in the tissues. In the 
former the defenses of the body are overwhelmed by an extensive 
flooding with the bacteria, and vaccines, if not harmful, are, at any 
rate, utterly useless since the antigen is already so extensively dis- 
tributed throughout the tissues that if the body were capable of 
responding with sufficient antibody formation this would unques- 
tionably occur without the small additional amount furnished in the 
bacterial emulsion. Vaccination in such cases is entirely analogous 
to an attempt to stimulate a degenerated heart muscle with strychnin. 
Such cases of septicemia, however, are not in our opinion the 
most common ones in the human being. It is probable that all 
localized infections of more than a very trifling nature discharge 
living bacteria into the circulation from the very beginning. How- 
ever, in most cases the bacteria, though able to hold their own in 
their entrenched position at the focus where accumulated offensive 
factors and local injury reenforce them, are yet rapidly destroyed 
when, in small detachments, they get into the open circulation where 
the plasma antibodies and phagocytes are freely active. There are 
cases which take a middle course between such purely localized 
lesions and the acute septicemia, conditions in which a well-estab- 
lished focus continues to furnish bacteria to the blood stream as fast 
as they are destroyed. An example which illustrates our meaning THERAPEUTIC IMMUNIZATION 
573 
well is that of the so-called subacute endocarditis caused by the 
Streptococcus viridans and its close biological kin, where blood cul- 
tures are often consistently positive for a long period or may show 
occasional intervals in which the blood is bacteria-free. The focus 
on the heart valves apparently can continue uncured in spite of a 
relatively high or at least normal systemic resistance to the micro- 
organisms. If, as we ourselves have done, we isolate the organisms 
by blood culture from such cases, and then measure the opsonic prop- 
erties of the patient’s own serum against them, using the patient’s 
own leukocytes, we may often find that active phagocytosis takes 
place, in a degree equal or even superior to that taking place in the 
serum of normal individuals. Neither does there seem to be a dimin- 
ished phagocytic power of the patient’s own leukocytes. For a long 
time these conditions may continue, with a constant destruction of 
bacteria in the blood and a corresponding renewal of the supply from 
the lesion. The same condition can be observed in rabbits in which 
chronic endocarditis with persistently positive blood culture has been 
produced by injections of these bacteria. In such animals measure- 
ments similar to those described above have been made by Miss Gil- 
bert in our laboratory, and it has seemed as though persistently 
positive blood cultures could be obtained only when a localized focus 
was set up in the animals. Unless this is the case the blood cultures 
rapidly become negative. 
Conditions essentially similar may exist in any other form of 
severe localized infection. Positive blood cultures do not necessarily 
mean a multiplication of the bacteria in the blood stream and a rapid 
overwhelming of the body. We have had occasion to see a number of 
cases of bacteriemia in which the focus of infection was surgically 
accessible; and in some of these cases early removal of the focus and 
purely surgical treatment resulted in a clearing up of the infection. 
Similar experiences have been reported by Libman and a number of 
others, and for this reason general septicemia, if not fulminating, 
may still be less desperate than ordinarily supposed. 
Now, having outlined the conditions obtaining in such cases, let 
us briefly consider whether, under the circumstances, vaccine therapy 
may logically be regarded as a hopeful form of treatment. We may 
assume, on the one hand, that the bacteria, being consistently present 
and destroyed in the blood, should furnish antigen sufficient to 
stimulate the body tissues to their utmost reactive ability. This 
would seem a strong argument against vaccine therapy. On the 
other hand, we must take into consideration another phase of the 
subject, one which has some experimental justification. In discus- 
sing the origin of antibodies in another section it will be remembered 
that we called attention to the fact that many different tissue cells 
probably participate in the production of these protective reaction- 
bodies. We cited an experiment of Wassermann and his pupils in 574 
INFECTION AND RESISTANCE 
which they proved that antibodies were produced most energetically 
in the tissues about the point of injection of the antigen, namely, in 
the place at which it came into most concentrated contact with the 
cells. They injected bacteria into the subcutaneous tissues of the ear 
of a rabbit, measured the progressively increasing appearance of 
antibodies in the blood stream, and then amputated the ear. A 
sudden drop of antibody contents followed, showing that the supply 
of antibodies had largely emanated from the tissues surrounding 
the injection point. Park 1 has pointed out another reason why vac- 
cine treatment might be expected to exert beneficial action in such 
cases. He calls attention to the fact that when very large amounts 
of antitoxin are added to toxin before injection no antibody produc- 
tion results, and assumes that in chronic or subacute general infec- 
tions the circulating bacteria are in contact with specific antibodies, 
partially “sensitized,” and therefore not efficient as antigen. In 
consequence the injection of homologous unsensitized bacteria may 
hasten antibody formation. This assumption of Park is theoretically 
valid, but it is not in accord with the more recent experiments of 
Metchnikoff and Besredka, who claim to have obtained the best 
results in prophylactic typhoid vaccination by the injection of sensi- 
tized bacteria. 
In acute diseases which run a definite course, typhoid fever, 
pneumonia, dysentery, cholera, plague, and a number of other con- 
ditions, vaccine treatment during the course of the disease has not 
much theoretical justification. In typhoid fever, especially, specific 
antibodies appear in the blood in amounts enormously increased 
above the normal at periods when the patient is still actively ill in 
spite of the fact that the blood stream has been freed of the micro- 
organisms. Whatever may be our opinion as to the continuance of 
the disease after bacteria have been driven out of the blood stream, 
the use of vaccines can only tend to further increase of antibodies 
which are already present in amounts far exceeding normal. In 
pneumonia the micro-organisms seem curiously resistant against the 
attack of the serum antibodies, and in spite of the presence of large 
amounts of antigen both in the lungs and, for a time, in the circula- 
tion the development of immunity is delayed until just before or near 
the crisis. Since this, however, is usually only a matter of 7 or 8 
days, it is hardly likely that the injection of vaccines during this 
period could markedly alter the ultimate outcome. In a later sec- 
tion we shall see that vaccine treatment in typhoid fever is neverthe- 
less being extensively tried and gives non-specific reactions of a nature 
which cannot entirely be explained on the basis of the above con- 
siderations. 
Thus, the use of vaccines in subacute or chronic cases of infection 
finds much theoretical justification in all conditions in which a slug- 
1 Park. Tram, of Americ. Phys., Vol. 8, 1910. THERAPEUTIC IMMUNIZATION 
575 
gish, localized process persists without much tendency to spread, but 
also without sufficient progress in the direction of cure. Theo- 
retically, also, we may justify the attempt at vaccine treatment in 
subacute or chronic cases in which bacteria appear in the blood 
stream, constantly fed into it from a localized focus. Like so many 
other phases of this question, the ultimate answer lies with clinical 
experience, for experimentation upon animals cannot imitate en- 
tirely the conditions prevailing in man. Also, in man there are many 
modifying accidental factors which cannot be controlled by animal' 
experiment. Clinical experience has not yet brought the necessary 
reply, and in spite of many attempts to summarize the available 
literature, no generalization is as yet justified. All that we can say 
at the present time is that active immunization with homologous 
vaccines is one of our most useful methods of prophylactic protec- 
tion. It is extremely useful, though not universally successful in 
cases of often repeated localized infections, such as boils and fu- 
runcles in which there are free intervals between attacks during which 
an opportunity is given for artificial enhancement of the resistance. 
It is justified and hopeful in cases of sluggish, chronic infections if 
properly combined wuth rest, nutrition, fresh air and other thera- 
peutic measures to improve general systemic conditions, and in many 
such cases has justified its use if, of course, the vaccines in these 
latter cases were autogenous. In subacute conditions, such as endo- 
carditis in which the focus cannot be surgically reached, and in which 
bacteria appear in the blood, autogenous vaccination has some theo- 
retical justification, but has, in our own experience, and in an un- 
prejudiced study of the literature, not impressed us as having given 
much hopeful therapeutic result. In acute infectious diseases any 
effect that has been claimed for vaccines, we believe, has been due 
to the nonspecific reactions which are discussed in a later paragraph 
and in no sense to a specific immunizing effect of the injected 
bacteria. 
In prophylactic vaccination, such as typhoid, plague, cholera, 
etc., described in the subsequent sections, homologous vaccines made 
from stock cultures are, of course, necessary, and have justified them- 
selves by brilliant results. In the treatment of existing infections, 
whatever their nature, the use of stock commercial cultures, or stock 
cultures kept in laboratories have very little specific value, largely 
because of the many antigenic varieties existing within the same 
species of most of the bacteria that give rise to infections in which 
this question usually arises. Autogenous vaccines, only, can be ex- 
pected to give any sort of result. The random use of stock vaccines 
without laboratory diagnosis and without control is, of course, an 
entirely unjustifiable procedure in cases of this kind. 
Moreover, the control of vaccine treatment by opsonic index de- 
terminations, as suggested by Wright and his coworkers, is not a 576 
INFECTION AND RESISTANCE 
practicable method for clinical application. The measurements, as 
we have pointed out, are entirely too laborious and, therefore, ex- 
pensive. Their accuracy is subject to so many uncontrollable vari- 
ables that there is nothing to be gained for the patient, and equally 
nothing to be gained for the physician in his judgment of the case. 
Accurate and skilled clinical supervision is of far greater value. 
As we have said before, the opinions expressed above are given 
with the purpose of stating as clearly as we can the logic of vaccine 
therapy as we see it at present. The next ten years of clinical ex- 
perience may largely modify these views. One thing is certain, how- 
ever, and that is that the problem can only be settled if treatment by 
this method is undertaken with the guidance of an accurate bacteri- 
ological diagnosis, and with bacteriological control of the individual 
case, so that, when occasion arises, estimations of antibodies can be 
made. 
THE PRODUCTION AND STANDARDIZATION OF VACCINES 
Vaccines in the sense of Wright consist merely of killed cultures 
of the bacteria with which the patient is infected. In all cases it is 
extremely desirable to make such vaccines “autogenous,” by which 
we mean that the organism used is one which has been isolated from 
the case. The difference between various strains of the same species 
of bacteria seems to make this imperative whenever it is at all pos- 
sible. The recent investigations of Reufeld and Ilaendel in de- 
termining that there are a number of types of pneumococcus which 
are antigenically distinct illustrates this point. The same principle 
is made clear by the recent work of Rosenau on the streptococcus- 
pneumococcus group. Especially important is Rosenau’s observa- 
tion that a pneumococcus which he had been able to transform cul- 
turally by special methods was found to be altered also in its reaction 
to agglutinins. 
In the development of prophylactic methods of vaccination 
against epidemic disease like typhoid, cholera, plague, etc., many 
different methods of antigen preparation have been developed. In 
typhoid prophylaxis the bacteria have been used dead, living, and 
sensitized, and even extracts have been employed. In cholera the 
early use of living cultures by Ferran has given way, in the hands 
of Kolle and others, to that of dead bacterial emulsions. In plague 
and a number of other conditions the impression seems to be general 
that the bacteria should be used in the living, but attenuated, state. 
Special methods which have been developed in these cases are dis- 
cussed in another section. 
In treatment of developed diseases with vaccines the method most 
commonly used is that which has been introduced by Wright, namely, 
the use of dead cultures. In his earlier experiments Wright culti- THERAPEUTIC IMMUNIZATION 
577 
Fated the bacteria on agar slants for about 24 hours, then washed off 
"he growth wTith 10 c. c. of sterile salt solution. It will be well to 
iescribe in detail the preparation of such a vaccine. 
The bacteria must be isolated from the patient by the usual 
method of plate cultivation and colony fishing on suitable media. 
We do not think that any satisfactory substitute for careful isolation 
by plating has been devised. After a pure culture of the organism 
bas been obtained this is grown on relatively large surfaces of agar, 
glucose-agar, or ascitic agar, as the case may require. 
These cultivations may be made in Kolle flasks or, as 
Wright and others have suggested, on large agar sur- 
faces obtained when the culture medium is allowed to 
harden in a square 3-oz. medicine bottle laid on its 
side. Any device of this kind in which a large surface 
of agar is exposed may be used. 
After suitable growth of the micro-organisms has 
taken place, 24-48 hours, the growth is gently washed 
off with 10 c. c. or more of sterilized salt solution. 
Care must be taken to do this in such a way that no 
agar is drawn away with the emulsion. The thick 
emulsion so obtained is removed from the culture bottle 
with sterile nipple pipettes or Pasteur pipettes and 
transferred to a sterile thick-walled test tube into 
which glass beads have been placed. By drawing out 
the neck of this test tube in the flame a glass capsule 
is formed, in which we now have our so-called stock 
emulsion. (See figure.) 
The next thing to be done is to standardize this 
stock emulsion, or, in other words, determine approxi- 
mately the number of bacteria to the cubic centimeter. 
There are a number of methods by which this can be 
done. 
The method most extensively used by Wright and 
his followers was that in which the bacteria are counted 
against red blood cells. The bacteria in the capsule 
are shaken thoroughly with glass beads so that clumps 
may be broken up and even distribution obtained. A 
little of the emulsion is then put into a clean watch 
glass, a step which can be accomplished most easily by 
breaking off the tip of the drawn-out part of the cap- 
sule, tilting it very gently and heating the closed end 
over a small flame, so that some of the emulsion will be 
driven out by the expanding air. With a nipple pipette marked 
about an inch from the tip, as in the taking of an opsonic index, a 
little of the emulsion is drawn up. This is placed into another clean 
watch glass and is mixed with about 2 volumes of salt solution and 
Capsule Made 
op Test Tube 
to Hold 
Stock Vac- 
cine Emul- 
sion from 
Which Dilu- 
tions are 
Made. 
(It is well to 
keep capsule 
open to capil- 
lary tip until 
it has cooled 
off, otherwise 
it may crack 
when quickly 
cooled.) 578 
INFECTION AND RESISTANCE 
one volume of blood from the finger, these quantities being measured 
with the same nipple pipette. We then have a mixture in which, 
in a total of 4 volumes, there are equal parts of blood and of bac- 
terial emulsion. After this emulsion has been thoroughly mixed by 
drawing in and out through the nipple pipette smears are made on 
slides and stained with Jenner or any other suitable blood and bac- 
terial stain. Under the field of the microscope the ratio between 
the bacteria and blood cells is then determined, and from our 
knowledge of the number of the red blood cells in this blood to each 
c. mm. we can easily calculate the number of bacteria to the c. mm. 
or c. d.2 
A more accurate method of enumerating the bacteria in a sus- 
pension to be used for vaccine is by direct count of an accurately 
made dilution in a hemocytometer chamber, as was first suggested 
by Malory and Wright in 1908.3 The bacterial suspension is diluted 
in blood-counting pipettes, 
1-20 to 1-100 dilutions of 
thick bacterial suspensions 
being as a rule satisfactory. 
As a diluent one may use 
either salt solution or some 
dilute anilin dye, such as 
one made by mixing one 
part alcoholic methylene 
blue with 40 parts of 1 per 
cent, carbolic acid. The 
dilute suspension is then 
placed in an ordinary 
Thoma-Zeiss chamber, 
which was designed for 
counting blood platelets and 
has a depth of 0.02 mm. 
This enables one to use an 
oil immersion lens or high 
power dry system with a 
short working distance. From such a count one may readily estimate 
the number of bacteria in the original suspension; for example, if 20 
squares in the Helber-Zeiss chamber are counted the result gives the 
number of bacteria in 0.001 c. mm.4 
Another method of standardization of vaccines which is suffi- 
Microscopic Field as Seen in Standardi- 
zation op Vaccines by Wright’s Method. 
2 For such counts it is convenient to contract the field of the microscope 
by using a diaphragm or simply marking a circle on the eyepiece with a 
grease pencil. 
3 Malory and Wright. “Pathological Technique,” 4th Ed., New York, 
1908. 
4 Glynn, Powell, Rees, and Cox. Jour. Path, and Bad., Vol. 18, 1914, 
p. 379. THERAPEUTIC IMMUNIZATION 
579 
ciently accurate for clinical purposes is that of Hopkins, which con- 
sists in measuring the volume of the sediment 5 after centrifugaliz- 
ing the preparation under standard conditions in a graduated tube. 
The tubes may be made with a capacity of 10 to 15 c. c. with a capil- 
lary tip about one inch in length, having a capacity of about 0.05 
c. c. graduated in 0.01 c. c. The bacterial suspension, after being fil- 
tered through sterile cotton to remove fragments of the agar or other 
foreign bodies, is centrifugalized in such a tube for half an hour at 
about 2,800 revolutions a minute. The supernatant fluid and bac- 
teria are removed down to the 0.5 c. c. mark and the sediment resus- 
pended in 5 c. c. sterile salt solution by means of a capillary pipette 
which gives a 1 per cent, suspension. 0.05 c. c. of streptococci sedi- 
mented in this way represent quite constantly 16 mm. of dried bac- 
terial substance. The number of organisms per cubic centimeter 
contained in 1 per cent, suspension in this way are as follows: 
Streptoceocus aureus and albus 10 billion 
Streptoccocus 8 “ 
Gonococcus 8 “ 
Pneumococcus (capsulated) 2.5 “ 
Bacillus typhosus 8 “ 
Bacillus coli 4 “ 
After the vaccines have been standardized suitable dilutions can 
be made in salt solution to which 0.5 per cent, carbolic acid or some 
other antiseptic has been added. The dilutions are usually so made 
that from 100 to 500 million bacteria are contained in the cubic cen- 
timeter, this being a suitable initial dose of most organisms. The 
dilutions are placed in sterile bottles containing beads and fitted 
with rubber caps. These bottles can be shaken 
before use, the emulsion thoroughly distributed, 
and the desired quantity can be taken out with a 
sterile hypodermic syringe thrust through the 
rubber cap after th}£ has been covered with a 
small amount of lysol or strong carbolic (see fig- 
ure). After the dilutions have been made both 
these and the stock vaccines should be sterilized. 
Some workers sterilize always the stock vaccines 
and make the dilutions with asceptic proportions 
in such a way that no further sterilization is 
necessary. This is preferable because the less 
heat that is applied the better it is for the 
preservation of their antigenic properties—sterilization is usually 
accomplished by heat in the water bath. Wassermann’s earlier 
technique called for heating to 60° C. for one hour for a 
number of consecutive days. It is generally considered at the 
Vaccine Stock 
Emulsion in 
Rubber Top 
Bottle. 
5 Hopkins, Jour, A, M, A.f Yol, 60, 1913, p. 1615. 580 
INFECTION AND RESISTANCE 
present time that it is better not to heat above 55° C. After the 
vaccine has been heated its sterility mnst be controlled to aerobic 
and anaerobic cultivation, and possibly by animal inoculation, al- 
though, except in special cases, this is unnecessary. Some workers, 
especially when the vaccine is to be extensively used, as in typhoid 
immunization, inject some of the vaccine into white mice to exclude 
the possibility of contamination with tetanus. In such cases also it 
is not inadvisable to test out the antigenic value of the vaccine upon 
animals, measuring the agglutinins, etc., which result from a number 
of inoculations. In the preparation of a therapeutic vaccine where 
speed is required this of course is not feasible. Moreover, it is un- 
necessary in view of the fact that we wish to inject that particular 
organism into the patient from whom it has been cultivated. What- 
ever its antigenic value may be from animal experiments, it is pre- 
ferable for the given purpose to any other strain. 
Sensitized vaccines are easily made by exposing emulsions of the 
bacteria to moderate amounts of a strong immune serum which has 
been heated to 56° C. to destroy the complement. Bacteria will 
usually agglutinate under these circumstances and can easily be 
centrifugalized to the bottom. The excess serum is then washed olf 
and the bacteria emulsified as in the case of the preparation of vac- 
cines with dead organisms. 
Prophylactic Immunization in Typhoid and Paratyphoid Fever. 
—The first attempt to inoculate human beings with typhoid bacilli 
with the intention of producing prophylactic active immunization 
was probably that made by Pfeiffer and Kolle 6 in 1896. During 
the same year also Wright 7 made similar studies in England, and 
soon after this he reported upon the development of antibodies in the 
blood of 17 people inoculated with typhoid. By these studies it was 
shown that human beings could be inoculated with dead typhoid 
bacilli without danger, and this logically led to the attempt to vac- 
cinate human beings on a large scale. 
It is hardly necessary to dwell upon the desirability of such a 
procedure. From tables recently published by Russell 8 we take the 
information that, in our own Spanish American war, 20,738 cases 
of typhoid with 1,580 deaths occurred in a total enlistment of 107,- 
973. In this entire war 243 men were killed in action or died of 
their wounds, while almost 7 times as many died of typhoid fever. 
In the British army during the Boer war there were over 75,000 
cases of typhoid in 380,000 men, and in the Russian army during the 
Russo-Japanese war over 17,000 cases of typhoid occurred, over half 
as many as the number of men killed in action. Such appalling fig- 
ures leave no possible doubt as to the desirability of prophylactic im- 
6 Pfeiffer and Kolle. Deutsche med. Woch., Yol. 22, 1896, p. 735. 
7 Wright. Brit. Med. Jour., January, 1897, p. 256. 
8 Russell. Amer. Jour, of Med, Sc., Yol, 146, December, 1913, THERAPEUTIC IMMUNIZATION 
581 
munization in armies, and there can be little question that typhoid 
fever is sufficiently prevalent in many parts of the civilized world to 
encourage prophylactic immunization of individuals, even when not 
living under the especially dangerous conditions of camps. 
Following the preliminary studies of Pfeitfer and Kolle and of 
Wright extensive practical studies of vaccination were made in the 
German colonial army during the Herrero war, and by British 
bacteriologists during the Boer war. Leishmann 9 also studied care- 
fully the results of vaccination among regiments of the British army 
in India. 
The vaccine employed by Wright and his associates in England 
consisted of broth cultures of a typhoid bacillus killed by exposure to 
53° C., and by the further addition of 0.4 per cent, of lysol. The 
German vaccine consisted of emulsified agar cultures similarly killed. 
The results obtained with these vaccines, although encouraging, 
were not as striking as had been hoped. Russell10 summarizes the 
earlier attempts by stating that the morbidity was reduced to about 
one half among vaccinated persons with a slightly greater reduction 
of mortality. The last-named writer also attributes the early fail- 
ures to the overheating of the vaccines with a consequent reduction 
of their antigenic properties, and to timidity in their administration 
resulting from Wright’s fear of a negative phase. Russell, of the 
United States Army Medical Corps, made a most extensive study of 
typhoid vaccination in this country. After careful consideration of 
the methods of others he produces his vaccines as follows: A single 
strain of typhoid bacilli was used (culture Rawlings obtained from 
England), and this grown on agar in Kolle flasks for 18 hours. The 
purity of the culture was tested out both morphologically and by 
transplantation upon the double sugar media devised by Russell. 
Agglutination tests are also made. After 18 hours the growth was 
washed off in small quantities of salt solution and the emulsion heated 
in a water bath for one hour at 53° C.; it was then diluted with sterile 
salt solution to a concentration of one billion bacteria to a cubic 
centimeter. Then 0.25 per cent, of tricresol was added. Before use 
the safety of the vaccine was ascertained both by aerobic and anaero- 
bic cultivation and by the injection into mice and guinea pigs of 
considerable quantities for the exclusion of possible tetanus contami- 
nation. The efficiency of the vaccine was then tested by injecting 
rabbits with three doses at 10 day intervals, and determining the 
resulting agglutinating power. 
With these vaccines under the direction of the United States 
Army Medical Corps the troops mobilized in Texas, California, and 
along the Mexican border were treated. Compulsory vaccination was 
9 Leishmann. Glasgow Med. Jour., Vol. 77, 1912, p. 408, cited from 
Russell. 
10 Rugsell, Amer. dour. Med, Sc., Vol, 146, December, 1913, 582 
INFECTION AND RESISTANCE 
established in March, 1911, and the results as reported by Russell 
have fully justified the measure. The following table taken from 
Russell’s paper will illustrate the results obtained: 
Typhoid Fever. Officers and Enlisted Men, United States Army 
Yr. 
Jan. 
Feb. 
Mar. 
Apr. 
May 
June 
July 
Aug. 
Sept. 
Oct. 
Nov. 
Dec. 
Totals 
for 
9 months 
Volun- 
1908 
5 
6 
4 
2 
3 
11 
14 
31 
25 
26 
12 
8 
101 
tary 
1909 
4 
10 
6 
4 
11 
15 
26 
14 
16 
45 
20 
6 
106 
1910 
8 
11 
1 
4 
2 
6 
12 
27 
21 
16 
20 
11 
92 
Com- 
1911 
3 
3 
3 
7 
4 
4 
4 
7 
4 
4 
1 
0 
39 
pulsory 
1912 
1 
2 
2 
0 
0 
3 
1 
3 
1 
4 
0 
1 
13 
. 1913 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 
0 
Paratyphoid fever included in figures for 1908, but excluded in other years. Cases paratyphoid 
1909, 3; 1910, 3; 1911, 2; 1912, 3; 1913, 0. 
We have mentioned in another place that Metchnikoff and Bes- 
redka in their studies on typhoid vaccination in the chimpanzee 
have concluded that very little protective value resided in vaccina- 
tion with dead typhoid vaccines, whereas animals vaccinated wTith 
small amounts of living cultures were very efficiently protected. 
Metchnikoff and Besredka adopted finally the method of immunizing 
with living sensitized vaccines. By this is meant typhoid bacilli that 
have been exposed to the action of heated immune serum, or, in 
other words, typhoid bacilli that have absorbed specific antibodies. 
There is no question as to the efficiency of this form of vaccination. 
The method of employing sensitized bacteria for these purposes 
utilized by Besredka in the case of plague has unquestionably won 
an important place in active immunization. However, the results 
of Bussell and others seems to indicate that in human beings the use 
of dead vaccines is certainly of considerable value, and there are 
certain practical objections to the use of living vaccines in immuniza- 
tion of large numbers of people as in armies to which Russell calls 
attention. In the first place, living vaccines cannot be stored for 
any considerable period, and may become a source of possible infec- 
tion by mouth if carelessly handled; furthermore, contamination is 
not so easily ruled out in the case of living vaccines when used on 
a large scale, and it is not possible at present to require compulsory 
vaccination with living bacteria. The choice of strains is important. 
For vaccines in the British and American armies the Rawlings strain 
originally used by Wright is employed at present. The work of Weiss, 
Hooker and others recently has shown that perhaps wre may have to 
use for best results a polyvalent typhoid injection, and that in im- 
munizing with paratyphoid B, a polyvalent vaccine will be necessary THERAPEUTIC IMMUNIZATION 
583 
eventually. The paratyphoid A strains are so uniform antigenically 
that this will be unnecessary in their case. 
Gay has recommended the use of sensitized killed vaccines. He 
controls the efficiency of his vaccines by testing them out on rabbits 
in which typhoid septicemia has been produced by inoculation with 
cultures grown on rabbit blood agar. These vaccines have not yet 
been used upon sufficient numbers to justify conclusions. It would 
seem, however, that any one of the methods mentioned must possess 
considerable value, since they all represent merely slight variations 
of the same procedure. The method at present used in the German, 
British, and American armies, namely, vaccination with dead cul- 
tures, seems certainly, according to Bussell’s statistical studies, to 
have yielded excellent results and recommends itself by its extreme 
simplicity and safety. 
The proof of this was brought conclusively during the last war 
by the freedom of armies from typhoid fever, in spite of the most 
unfavorable and unpreventable sanitary difficulties existing during 
prolonged battles in areas of concentration. 
Since, therefore, the methods first recommended by Bussell and 
others have stood the actual test on many millions of men, it seems 
unadvisable to experiment on a large scale for the present with other 
methods. 
The vaccine as prepared at the present time, is made up of 
typhoid, paratyphoid A and B, so distributed that each cubic centi- 
meter contains one million typhoid bacilli and 750 million of each 
of the two paratyphoids. Three doses are given at 7 to 10 day in- 
tervals, the first one of 1/2 c. c. of this vaccine, and the second and 
third of 1 c. c. each. In cases of children or adults in which care is 
indicated, the dosage can be reduced and then the entire amount dis- 
tributed in a larger number of injections. Indeed, we have, on a 
number of occasions, given the vaccine in 4 and 5 injections, 6 to 7 
days apart, in amounts ranging from 0.25 to 0.75 c. c., in the hope of 
getting more thorough immunity in cases in which there was time 
for this procedure. 
In vaccinating it is extremely important not to put the vaccine 
into a small vein and not to inject it intramuscularly. It should be 
administered subcutaneously so that the finger feels a slight bulge 
under the needle. Intracutaneous injection should also be avoided. 
There is always a local reaction which may attain an area of 5 or 
6 c. c. in diameter over which there is redness, swelling and some 
pain. There may be swelling of regional lymph nodes. Suppuration 
should not occur in vaccination properly performed. General reac- 
tions vary in different individuals. In most healthy adults the first 
injection gives little trouble. But after the second or third, there 
may be headache, nausea and temperature, coming on several hours 
after the injection. In some cases these symptoms may persist 584 
INFECTION AND RESISTANCE 
for several days, and the patient feel very miserable, but no severe 
harm is done. If, as occasionally may happen, the injection passes 
into a vein, there may be very violent symptoms within 20 minutes 
after injection, consisting of a chill, followed by temperature, vomit- 
ing and nausea, and a condition almost bordering on collapse. We 
have seen two such cases in both of which the symptoms were alarm- 
ing. Great attention should, therefore, be paid to this point. 
Typhoid vaccine is contraindicated in chronic nephritis, in any 
acute illness in which there is fever, and during convalescence. There 
was at one time very much literature on the dangers of typhoid vac- 
cination in cases of tuberculosis. The judgment of Russell, Nichols 11 
and others who have had most experience with vaccination, is that 
tuberculosis is not a contraindication if there is sufficient reason for 
vaccination, that is, danger of infection. They state that neither 
typhoid fever nor vaccination have any harmful effects on the prog- 
ress of tuberculosis. However, should a tuberculous case be pre- 
sented for vaccination, it is well to withhold vaccination if there are 
acute symptoms and temperature, and to wait until these have 
subsided. 
Therapeutic Vaccine Treatment in Typhoid Fever.—A number 
of workers have attempted to treat typhoid fever after it occurred 
with typhoid vaccine sensitized and unsensitized. Frankel did this 
as early as 1893. He was followed by Petruschky 12 and Netter and 
in 1912 Ichikawa began the intravenous injection of typhoid vaccine 
in patients with the rather astonishing results of obtaining in many 
cases rapid falling of temperature and often apparent shortening of 
the disease. Boinet in 1912 obtained similar results and recently 
the same thing has been done by Gay.13 
Of these writers the only one who recognized immediately that 
perhaps the intravenous injection of such vaccines was not due to 
purely specific activity was Ichikawa, who thought the paratyphoid 
bacilli injected into typhoid cases often produced similar results, 
and soon after this Kraus obtained similar degrees of temperature 
and favorable results by injecting colon bacilli into typhoid patients, 
into a few cases of pyocyaneus infection, and into streptococcus septi- 
cemias. 
With Jobling and others we have believed from the beginning 
that at least an important factor in the activity of these vaccines 
was non-specific, due perhaps to a rapid mobilization of leukocytes 
and of ferments. This is discussed in another section. 
At any rate the therapeutic results obtained justify the cautious 
continuance of such intensive treatment in various diseases. 
Active Prophylactic Immunization in Cholera.—Attempts to 
11 Russell and Nichols. Jour. A. M. A., Yol. 76, 1921, p. 177. 
12 Petruschky. Centralbl. f. Bakt., 1, Vol. 19, 1896. 
13 Gay. Arch, of Int. Med., 1914. THERAPEUTIC IMMUNIZATION 
585 
protect human beings against cholera by prophylactic vaccination 
were made as early as 1885 by Ferran,14 a pupil of Pasteur. 
At the time at which Ferran’s experiments were done little was 
known regarding the production of immunity with killed cultures 
or with bacterial extracts, and Ferran, under the influence of 
the French school and its endeavors to immunize with living attenu- 
ated organisms, applied similar methods to cholera. First experi- 
menting with guinea pigs, he soon applied his method to human 
beings, inoculating them with small quantities of living broth cul- 
tures of cholera spirilla. In many of his experiments he gave, at the 
first injection, 8 drops of a fresh broth culture, following this after 
8 days with 0.5 c. c. of a similar culture. There is no reason why 
Ferran’s method should not have yielded excellent results. How- 
ever, it is stated that he worked with impure cultures, and other 
observers, notably Nikati and Rietsch, van Ermengen, da Lara, and 
others, failed to obtain encouragement in their subsequent investiga- 
tion of this method of vaccination. 
The method which Haffkine 15 worked out some years after Fer- 
ran’s experiments also depended upon the injection of living cul- 
tures, but Haffkine attempted, by a rather elaborate technique, to 
produce two separate vaccines, one attenuated, the other enhanced in 
virulence. Attenuation was accomplished by growing the cholera 
spirilla at a temperature of 39° C. in broth over the surface of which 
a constant stream of sterile air was passed. Under these conditions 
the first crop of cholera organisms died rapidly, but Haffkine prac- 
ticed reinoculation into new broth flasks before complete death of the 
original culture had taken place; after a series of generations of 
cultivation in this way he obtained cultures which produced merely 
temporary and slight edema when injected under the skin of guinea 
pigs. This weakened virus was used for the first inoculation. 
He enhanced the virulence of cholera cultures with the purpose 
of producing a strain of maximum potency comparable to virus fixe. 
His procedure was as follows: 
a. Giving an animal an intraperitoneal injection of cholera 
spirilla larger than the fatal dose. 
b. Taking out the peritoneal exudate and exposing it for a few 
hours to the air. 
c. Injecting this exudate into another animal and treating the 
resulting peritoneal exudate in the same way. 
After a series of such animal passages he claims to have obtained 
a virus of great virulence, and this is his second and stronger vaccine. 
In applying the method to human beings Haffkine planted the 
cholera spirilla upon agar slants of the standard size, emulsified the 
growths in sterile water, and injected 1/5 to 1/20 c. c. of such a 
14 Ferrari. Compt. Rendu de I’Acad. des Sc., 1885. 
15 Haffldne. Lancet, February, 1893; Brit. Med. Jour., December, 1895. 586 
INFECTION AND RESISTANCE 
culture, using first the weak vaccine and five days later a more 
virulent culture. 
Beginning his work as early as 1893, Haffkine and others vac- 
cinated as many as 40,000 people in India. On the whole, the results 
obtained were very encouraging. It is a question, however, whether 
or not his method is unnecessarily complicated. In the light of our 
more recent knowledge concerning cholera immunity it seems likely 
that the importance which Haffkine attached to the virulence of the 
cholera culture used for injection was exaggerated, and we have 
reason to believe that simple immunization with killed cultures may 
produce results fully as efficacious. After all, we could not expect, 
at least at present, to produce by active artificial immunization an 
immunity as permanent as that which results from an attack of the 
disease. Concerning the reasons for the acquisition of such perma- 
nent immunity we have as yet little knowledge. Even Haffkine’s 
method of inoculation with living virus does not, by his own estima- 
tion, last longer than possibly two years. It is therefore likely that 
prophylactic immunization in cholera is efficacious by reason of the 
appearance in the blood serum of the specific bactericidal and opsonic 
substances by which the small numbers of cholera organisms enter- 
ing during spontaneous infection can be disposed of before a foot- 
hold in the body is gained. 
Tamancheff later used Ilaffkine’s method, but killed the cultures 
by the addition of a 0.5 per cent, solution of carbolic acid. 
Kolle 16 later recommended the injection of dead cholera or- 
ganisms, maintaining that a single injection of about 2 milligrams 
of a culture killed by exposure to 50° C. for a few minutes, and 
by the addition of 0.5 per cent, of phenol, is sufficient to immunize 
successfully. Good results with Kolle’s method have been reported 
from Japan. 
Strong,17 also proceeding from the idea that the immunizing 
antigen is present, as such, within the cell body of the cholera 
spirilla, recommends the injection of autolytic products obtained by 
digesting cholera spirilla in aqueous suspension and filtering. He 
prepared his “prophylactic” by growing the organisms upon agar, 
then suspending the growth in sterile water and keeping it at 60° C. 
for from one to twenty-four hours. The mixture was then exposed 
to 37° C. for from two to five days and filtered through Reiehel 
filters. One to 5 c. c. of this was used in his experiments upon 
human beings. 
According to Teague 18 and others, none of the more complicated 
modifications of cholera vaccine have any particular advantage over 
simple salt solution suspensions of young agar cultures killed by heat- 
16 Kolle. Deutsche med. Woch., 1897, No. 1. 
17 Strong. Jour. Inf. Dis., Yol. 2, 1905. 
18 Teague. Jour. A. M. A., Yol. 76, 1921, p. 243. THERAPEUTIC IMMUNIZATION 
587 
ing one hour at 53° C. The vaccine made in the Government Labora- 
tories in Bombay contains, according to Teague, eight thousand mil- 
lion organisms per cubic centimeter, and 0.5 c. c. administered sub- 
cutaneously at the first dose, 1 c. c. at the second. It has been used 
extensively in India, and reports from there seem to indicate favor- 
able results. During the last war, Teague states that several million 
men in the Austrian and German armies were vaccinated, and many 
thousands of them spent months in districts where cholera was pres- 
ent. A relatively small amount of cholera occurred. 
Contraindications in cholera are similar to those stated for 
typhoid fever. Vaccine treatment in acute attacks of cholera is use- 
less, as far as we know. 
Prophylactic Immunization Against Plague.—The first attempts 
to immunize human beings prophylactically against plague were 
those of Haffkine.19 The first vaccinations were carried out with 
broth cultures killed at 65° C. He tested out his vaccines on a 
large scale in Bombay, and obtained apparently promising results. 
In a plague epidemic occurring in a Bombay prison only 2 of 151 
vaccinated persons became ill, and neither of these died; whereas, 
of 177 unvaccinated persons 12 became ill and 6 died. In large 
series of vaccinated people only 1.8 per cent, were infected with 
plague, with a mortality of 0.4 per cent, for the total, whereas of 
unvaccinated individuals in the same epidemic 7.7 per cent, fell 
victim to the disease, with a mortality of 4.7 per cent. 
The German Plague Commission, consisting of Gaffky, Pfeiffer, 
and Dieudonne, recommend a vaccine of killed agar cultures. 
Kolle and Otto,20 basing their earlier results upon experiments car- 
ried out with monkeys, mice, guinea pigs, and rats, have come to the 
conclusion that vaccination with dead plague cultures is much in- 
ferior to that obtained when attenuated living cultures are used. 
The same conclusion has been reached by Kolle and Strong.21 Kolle 
and Otto found that the immunization of animals with large doses 
of killed agar cultures of plague bacilli and with Haffkine’s prophy- 
lactic did not protect them against subsequent inoculation with viru- 
lent cultures. 
Strong 22 subsequently made a very careful comparative study of 
the various methods of plague vaccination, and concluded that the 
most efficient method is immunization with attenuated living cul- 
tures. He showed that when carefully done this method can be 
safely employed in human beings, but admits that his work must be 
as yet considered as experimental and further studied before it can 
be universally employed. 
19 Haffkine. Bull, de I’Inst. Past., Yol. 4, 1906, No. 20, p. 825. 
20 Kolle and Otto. Deutsche med. Woch., 1903, p. 493, and Zeitschr. f. 
Hyg., Vol. 45, 1903. 
21 Kolle and Strong. Deutsche med. Woch., Vol. 32, 1906, p. 413. 
22 Strong. Jour. Med. Res., N. S., 13, 1908. 588 
INFECTION AND RESISTANCE 
Besredka 23 lias advised the use of sensitized dead plague cultures, 
claiming, from animal experimentation, that such vaccines produce 
an efficient and relatively durable immunity. 
Rowland 24 confirms the immunizing properties of Besredka’s 
vaccines in plague, and believes that the antigenic properties of the 
plague bacillus are attached to the bacterial nucleoproteins, and can 
be extracted with these. Rowland prepares a vaccine by the treat- 
ment of the moist bacteria with enough anhydrous sodium sulphate 
to combine with all the water present, freezing and thawing the 
mixtures, then filtering off the bacterial deposits at 37° C., and ex- 
tracting them with water. The solution so obtained was fatal to 
rats in small quantities and afforded substantial protection, reducing 
the mortality on subsequent inoculation of a standard culture from 
80 to 10 per cent. 
Haffkine’s virus which consists of broth cultures grown for six 
weeks at room temperature and then killed by heating to 65° C., for 
one half hour and the addition of 0.5 per cent, carbolic acid, is still 
extensively used in India and other British possessions, but is 
gradually being replaced by simple saline suspensions of 24 hour agar 
cultures heated for one hour at 65° C. The latter type of vaccine is 
stated by Teague as being more effective in producing immunity in 
laboratory animals. 
The efficiency of the vaccine can best be judged by reports from 
the British Indian Plague Commission. In the report of 1911, pre- 
pared by Major Glen Liston, it is stated that 1,211,170 doses of vac- 
cine were distributed. His figures of comparison between the in- 
oculated and uninoculated, as far as they could be gathered reliably, 
are very favorable to the efficiency of the vaccine. A comparative 
table indicates the degree to which this is true. 
Comparative Morbidity and Mortality from Plague Among Inoculated and 
Uninoculated25 
Inoculated 
Population 118,148 
Attacks 941 
Deaths 371 
Incidence, per 1,000 7.96 
Case mortality, per cent. 39.5 
Not Inoculated 
Population 321,621 
Attacks 11,041 
Deaths 8,695 
Incidence, per 1,000 34.4 
Case mortality, per cent. 78.6 
Since a population like that of India offers so many difficulties 
to statistical study, some of the most important contributions to our 
knowledge of plague vaccination are obtained from small outbreaks 
in towns where the population has been under the observation of 
23 Besredka. Bull, de I’Inst. Past., Yol. 8, 1910. 
24 Rowland. Jour. Hyg., Vol. 12, 1912, p. 344. 
25 Teague. Cited from Jour. -4. M. A., Yol. 76, 1921, p. 244. THERAPEUTIC IMMUNIZATION 
589 
individual physicians for very long periods. Thus, Beals 26 inocu- 
lated in 1916-1917 4,378 people in the town of Wai, Bombay. In 
that year, of the 6,000 uninoculated people in Wai, 270 got the dis- 
ease, and 167 died. Among 4,378 inoculated persons, there were 
only 39 attacks and 10 deaths. In 1917-1918, among the uninocu- 
lated, less than 6,000 people, there were 76 attacks of plague and 47 
deaths, while with the 4,081 inoculated there were 17 attacks and 
only two deaths. 
Prophylactic Vaccination in Pneumonia.—Owing to the high 
death rate from pneumonia in mining camps, such as those in South 
Africa, in Panama, and among soldiers during the war, it has seemed 
desirable to work out a method of vaccination against pneumonia 
on a large scale. The most extensive experiments on record in this 
direction are those of Cecil and Austin in this country, and the pre- 
ceding ones of Lister in South Africa. Lister prepared a vaccine 
by using his own types A, B and C (B and C corresponding, re- 
spectively, to II and I of our classification), and injected these in 
salt solution suspensions in amounts of from six to seven billion at a 
dose. His first injections were done intravenously. Later, he used 
subcutaneous inoculations, giving only three doses of two billion each 
at weekly intervals. His results were inconclusive but encouraging. 
Cecil and Austin,27 and Cecil and Vaughan 28 have continued this 
work in America. 
At Camp Upton during the war, they vaccinated over 12,000 men 
against Types I, II and III, administering three or four doses of a 
saline vaccine at intervals of 5 to 7 days, with a total dosage of six 
to nine billion of Types I and II, and four to six billion of Type 
ILL Forty per cent, of the Camp were vaccinated. Among the un- 
vaccinated, there were 173 cases of pneumonia of all types, and 
among the 40 per cent, vaccinated, there were 17 cases. A similar 
experiment was carried out at Camp Wheeler, the only difference 
being that a lipo-vaccine, that is, a suspension of pneumococci in 
neutral fat, was used. Over 13,000, or 80 per cent., of the men were 
vaccinated. Here the period of post-vaccination observation was 
longer, that is, two to three months. During this period there were 
32 cases, among the 80 per cent, vaccinated, and 42 among one-fifth 
of the men who were unvaccinated. These figures, while encourag- 
ing, are not as convincing as they should be in order that we can 
pass definitely favorable judgment. 
Of value to such judgment, however, are the monkey experiments 
of Cecil and Blake.29 They found that the ordinary small sub- 
cutaneous doses of vaccine are not sufficient to protect monkeys 
26 Beals. Jour. A. M. A., Vol. 75, 1920, p. 955. 
27 Cecil and Austin. Jour. Exp. Med., Yol. 28, 1918, p. 19. 
28 Cecil and Vaughan. Jour. Exp. Med., Yol. 29, 1919, p. 457. 
29 Cecil and Blake. Jour. Exp. Med., Vol. 31, 1920, p. 519. 590 
INFECTION AND RESISTANCE 
against a subsequent attack of pneumonia, but that they would 
modify the course of the disease favorably. Subsequently, Cecil and 
Stephen 30 found that if large subcutaneous or intravenous doses of 
pneumococcus vaccine were used, prophylactic immunization of the 
monkeys were successful. 
These workers believe that there is a large field for pneumonia 
vaccination under circumstances where epidemics are expected, such 
as the conditions of concentration camps during cold weather men- 
tioned above, and perhaps in the course of the high pneumonia mor- 
tality encountered in the second and tertiary waves of influenza 
epidemics. 
The vaccine is contraindicated in acute infections, in pulmonary 
tuberculosis and nephritis. Also, it should not be given to patients 
with chronic heart disease, or pregnant women after the fifth month. 
The active treatment of pneumonia as it has appeared with vac- 
cines, is, so far, without experimental or therapeutic foundation. 
Prophylactic Vaccination in Influenza.—Since we are not sure 
about the relationship of the Pfeiffer bacillus to epidemic influenza, 
prophylactic immunization with this organism is not entirely logical. 
.Nevertheless, since the most severe complications in secondary waves 
of such epidemics are unquestionably due to influenza bacilli, strepto- 
cocci and pneumococci, many attempts have been made to at least 
protect against serious complications by prophylactic vaccination. 
Extensive experiments were made on this during the last great 
epidemic. Eyre and Lowe 31 inoculated 16,000 men, leaving 5,700 
controls which were either uninoculated or received only one dose. 
These experiments were done with a mixed vaccine containing 
pneumococci, streptococci, influenza bacilli, staphylococcus aureus, 
micrococcus catarrhalis, Eriedlander bacillus, and a bacillus which 
they called B. septus. They obtained a slightly greater incidence 
among the uninoculated, that is, 4.1 per cent., as against an in- 
cidence of 1.3 among the inoculated. Cadham 32 in 1919 reported 
upon the use of mixed vaccines of streptococcus, pneumococcus and 
influenza bacillus upon soldiers. He claimed that the incidence of 
pneumonia was one-half and the mortality less than one-third among 
the inoculated as compared with the uninoculated. Wirgman33 
also in 1919, reported observations on 11,000 people of whom 800 
were inoculated in November and December, 1918, and January, 
1919. He, too, reported favorable results in that the incidence 
among the inoculated was one-half of that occurring among the un- 
inoculated, and that there was no death rate among the inoculated 
as against those who had not been treated. In spite of these ap- 
30 Cecil and Stephen. Jour. A. M. A., Yol. 76, 1921, p. 178. 
31 Eyre and Lowe. Lancet, Oct. 12, 1918. 
32 Cadham. Lancet, Vol. 1, 1919, p. 885. 
33 Wirgman. Lancet, Yol. 1,1919, p. 357, and Lancet, Yol. 2,1918, p. 324. THERAPEUTIC IMMUNIZATION 
591 
parently favorable results, a critical survey of the situation by 
McCoy 34 points out many significant sources of error. The chief 
one of these is the fact that the inoculations had usually been made 
during the progress of an epidemic, while the cases among the gen- 
eral population or control groups, are calculated from the be- 
ginnings of the epidemic. Also, the vaccine cannot be expected to 
have any appreciable prophylactic effects in less than 7 to 10 days. 
He has been able to collect, furthermore, a considerable number of 
observations such as the one done on the Naval personnel at the 
Pelham Bay Training Station (U. 8. Naval Bulletin Ho. 50 and 
Ho. 51), and the experiments of Ilinton and Kane 35 in an epileptic 
colony in which there were absolutely no appreciable results. 
McCoy’s opinion is that in spite of a generally favorable impression, 
the evidence so far, whenever experiments have been properly con- 
trolled in every particular, does not demonstrate any effects on either 
incidence or mortality. 
The Use of Vaccines in Streptococcus Infections.—In strepto- 
coccus infections the problem of vaccine treatment is very similar 
to that encountered in staphylococcus lesions. It is less frequently 
applicable, largely because the conditions caused by the strepto- 
coccus are usually much more acute and severe, and under such 
circumstances vaccine treatment has very little purpose. When, 
however, streptococcus infections are sluggish, connected with old 
sinuses or sequestra, wounds that will not heal, or in the form of 
furuncles, vaccine treatment may be applied with some hope of 
raising the general resistance of the patient. 
Here, again, it is necessary to use a strain obtained from the 
patient, since not only the streptococcus viridans, but the hemolytic 
streptococci, as well, are antigenically heterologous. In the case of 
the viridans, it is rare to find two entirely homologous strains. They 
vary in many subgroups like the Type IV pneumococci. The recent 
work of Dochez, Avery and Lancefield,36 moreover, has shown that 
the hemolytic streptococci can be divided into groups. Stock or 
commercial vaccines are, therefore, useless, except for a certain 
amount of non-specific effect, such as that described in another sec- 
tion. The vaccines are made in the usual manner, sterilized, tested, 
and standardized to about twenty million per cubic centimeter. 
Vaccination in Infections with the Staphylococcus Pyogenes 
Aureus.—In no infection has vaccination been so extensively at- 
tempted as in the various staphylococcus lesions of the skin and 
subcutaneous tissues commonly spoken of as boils and furuncles. 
This subject has been discussed to a considerable extent in the chapter 
34 McCoy. Jour. A. M. A., Yol. 73, 1919, p. 401. 
35 Hinton and Kane. Bimonthly Bulletin, State Bd. Health, Mass., 28, 
1919, p. 6. 
36 Pochez, Avery and Lancefield. Jour, Exp. Med,, Yol. 30, 1919, p. 179. 592 
INFECTION AND RESISTANCE 
on opsonic index in which Wright’s experiments on therapeutic 
vaccination in these conditions have been detailed. The immunologi- 
cal circumstances in these infections are peculiar in that a consider- 
able degree of local immunity is probably developed by the tissues 
immediately surrounding the staphylococcus focus; for, as is well 
known, a patient suffering from boils may be completely getting 
the better of an infection at one site, while a few inches away 
another focus is starting and progressing vigorously. There are a 
number of possible explanations for this, in addition to local re- 
sistance. It may be that the leukocytic concentration at the focus, 
due to purely chemotactic influences, accumulates at the point of 
attack a defensive mechanism which is not dependent upon acquired 
powers of resistance of the fixed tissues at this point; from this it 
follows that if we could hasten the accumulation of leukocytes at 
the focus, by artificial means, the lesion might be arrested without 
any fundamental increase of immunity. At any rate, the fact re- 
mains that boils may follow each other, one upon the heels of the 
other, or may come out in small simultaneous crops with intervals 
of a week or more between their appearance. Since staphylococci 
of pathogenic varieties are probably in constant contact with the skin 
of human beings, getting into small abrasions, hair folicles, etc., 
it is likely that the incidence of boils is not merely due to accidental 
contact or uncleanliness in a person of normal resistance. It is 
more than likely that, in addition to such contact, a lowered re- 
sistance against staphylococci must exist, at least in the usual case. 
It is noticeable that boils occur particularly in individuals suffering 
from some systemic disease such as nephritis or diabetes, in persons 
with chronic gastro-intestinal disturbance, or in perfectly healthy 
individuals who are temporarily below physical par. We have also 
noticed their very frequent appearance in men and women who have 
subjected themselves to a sudden change of physical daily habits, 
such as athletes going too rapidly into severe training and individuals 
having led a sedentary existence, living a vacation life of strenuous 
physical activity. Whether or not the indirect influences which 
determine an increased coagulability of the blood and a lessened 
water content in the circulation have any direct bearing upon this 
is doubtful, though it has been suggested by Wright. However 
this may be, all these factors indicate a lowered resistance in these 
cases which it is logical to assume could be improved by specific 
immunization. 
It is impossible to lay down general rules for the treatment of 
such cases. It may be said, however, that treatment of staphylo- 
coccus skin infections should never be undertaken with stock vac- 
cines. There is always time for the very simple procedure of 
isolation of the staphylococci which are directly responsible for the 
lesion in the individual, and the production of a vaccine. In treat- THERAPEUTIC IMMUNIZATION 
593 
ment with a properly prepared vaccine, the chief rule we would 
like to emphasize is the avoidance of any kind of a reaction, keeping 
the dosage below that which gives either local or systemic effects. We 
prefer a more frequent injection, but of smaller amounts, rarely 
giving more than a billion organisms in a dose. If the boils are of 
sufficient severity to cause severe temporary illness in the patient, a 
condition which is rare, treatment should be omitted, but pushed in 
the intervals between crops of boils and at times when the patient is 
in good condition. Injections of small doses can be made every third 
or fourth day, of the larger ones once a week. We have seen no advan- 
tage from any considerable gradual increasing of the dosage. It may 
be necessary to continue the treatment for a considerable period, and 
even when there has been no new boil for ten days or two weeks, it 
is advisable to add one or two further injections. Results of treat- 
ment can only be judged by clinical observation and the opsonic index 
is not of practical value in our opinion. 
Vaccine Treatment in Gonorrheal Infections.—Here, again, 
autogenous vaccines should be used because, as Torrey has shown, 
the gonococcus group is composed of many antigenically varying 
members. The organisms should be cultivated from the patient 
whenever possible, therefore, and a homologous vaccine made. 
Although the literature on the beneficial effects of gonococcus vac- 
cine is very confusing, a few general conclusions can be drawn which, 
though not absolutely definite, will indicate the trend of opinion. 
Geraghty,37 in summarizing the results obtained, states that he be- 
lieves that there is no evidence indicating successful treatment of 
acute or subacute urethral infections. He states, however, that he 
has seen occasional excellent results in acute complications in arthritis 
and epididymitis. He adds that in most instances in which he has 
seen favorable results, he has observed acute reactions with tempera- 
ture and leukocytosis, occasionally of sufficient severity to be alarm- 
ing. He concludes therefore that since the vaccine is of absolutely 
no use in acute or chronic urethritis, and only occasionally gives 
symptomatic relief in the complications mentioned, the benefit de- 
rived is not proportionate to the risk incurred. 
Without having any practical experience with this form of treat- 
ment, we gather from Geraghty’s article, in the cases in which the 
severe reactions occurred, there might have been accidental intra- 
venous or intramuscular introduction of the vaccine with the severe 
non-specific effects noted in such cases. 
Active Immunization against Anthrax.—This method, exten- 
sively used for the protection of cattle, sheep and other domestic ani- 
mals subject to anthrax, was first developed by Pasteur. The prin- 
ciple is the administration of anthrax bacilli which have been at- 
37 Geraghty. Jour. A. M, A.f Vol. 76, 1921, p, 35, 594 
INFECTION AND RESISTANCE 
tenuated by cultivation at 42° to 43° C. in broth. The resultant 
strain, cultivated under these conditions for a sufficiently long time, 
is not only less virulent, but loses its power of spore formation. The 
reduction of virulence is in direct relationship to the length of time 
for which the cultivation is continued. The relative virulence of a 
given culture can be estimated by a method described by Koch, Gaffky 
and Loeffler.38 Rabbits are less susceptible than guinea pigs, and 
virulent anthrax cultures grown for 2 to 3 days under the conditions 
of increased temperature, lose their power to kill rabbits in certain 
doses, but are still virulent for guinea pigs in these amounts. After 
10 to 20 days of further cultivation at 42°, the virulence for guinea 
pigs disappears, but the culture is still potent for mice. Even mouse 
virulence may eventually be eliminated by further cultivation at this 
temperature. 
Pasteur’s procedure was to grow anthrax cultures for different 
periods, under these conditions. A culture which has lost its viru- 
lence for guinea pigs and rabbits, and is still potent against mice was 
his “premier vaccin.” A culture which was still definitely virulent 
for mice and guinea pigs, but no longer for rabbits, was his “deuxieme 
vaccin.” Such strains in which the attenuation has been definitely 
determined are grown in broth for 48 hours, and are then injected 
into cattle in doses of 0.25 c. c., and into sheep in about one-half of 
this dose. The “premier vaccin” is first injected, and 12 days later 
the “deuxieme.” The immunity may last one or even two years. 
A modification of Pasteur’s method was that of Chauveaux 39 
who grew the bacilli in bouillon at 38° to 39° C., under a pressure 
of eight atmospheres. The Pasteur method has been introduced all 
over the world, and, according to statistics gathered by Sobernheim, 
has had excellent results. For statistical facts on the success of the 
treatment, we refer the reader to Sobernlieim’s article in the Third 
Volume of Kolle and Wassermann’s Handbook. 
Prophylaxis against Small-pox.—In the case of small-pox the 
general method of active prophylactic immunization is in principle 
identical with that devised by Jenner in the 18th century. The 
original observation from which Jenner worked -was that dairy maids 
and other individuals who had been infected with cow pox were 
thereafter spared when a small-pox epidemic appeared in the region 
in which they lived. It is now agreed by most observers who have 
studied the problem that the virus of cow pox and that of small-pox 
are identical in nature; the former representing a strain attenuated 
by passage through the animal body. This is based chiefly upon the 
observation that true variola can be transmitted to cattle, and that 
it can be thus carried from animal to animal, during this process 
38 Koch, Gaffky and Loeffler. Mitt. a. d. Kais. Gasundheitsamt, 1884, 
39 Chauveaux, Compt. Bend, de VAcad. Sc,} 1884, THERAPEUTIC IMMUNIZATION 
595 
becoming attenuated for human beings to such a degree that reinocu- 
lated into man a simple vaccinia is produced.40 
Small-pox, therefore, represents in principle active immunization 
by means of attenuated virus. When vaccination was first introduced 
the virus was taken from preceding pustules produced in other human 
beings. This has been given up in most countries to-day largely be- 
cause of the dangers of transferring syphilis and other diseases in 
this way. At present the method of obtaining virus for vaccination 
purposes is carried out as follows: The initial material consists of 
what is known as “seed” virus, which can be obtained from spon- 
taneous cow pox or from vaccination pustules in children, or again 
from pustules obtained in calves after several passages of small-pox 
virus through these calves. From such seed virus calves may be in- 
oculated for vaccine production or else the calves may be inoculated 
from the material obtained from other calves in the usual way. 
Healthy young animals are used; they are washed along the 
abdomen, strapped down upon specially prepared tables, and the 
abdominal skin thoroughly cleansed with soap and water. The 
exact procedure varies in different places; often the skin is thor- 
oughly cleansed with carbolic solution, and this is thoroughly re- 
moved with sterile water before inoculation, or else cleansing is re- 
lied upon without the use of germicides. Over the clean area 
longitudinal scratches 1 to 2 c. c. apart are made, and into these 
the seed virus is rubbed. The animals are then kept in a clean 
stall, preferably over asphalt floors, and rigid cleanliness is observed 
during the period of development of the pustules. After the 6tli or 
7th day, when the vesicles are beginning to appear, the abdomen is 
well washed and cleansed of superficial dirt without the use of an 
antiseptic, and the pulp removed from the lesions with a curette. 
The pulp so removed is placed into 60 per cent, glycerin and thor- 
oughly ground up in a specially constructed mill. According to 
Rosenau, the animal should always be killed before the vesicles are 
removed, not only for humane reasons, since the same object might 
be attained with anesthesia, but because a thorough autopsy can then 
be performed to determine the health of the calf. 
Vaccines so obtained always contain bacteria, the glycerin there- 
fore serving a double purpose: one, the preservation of the virus, the 
other a gradual destruction of the bacteria. Rosenau has shown that 
the addition of 2 to 4 parts of 60 per cent, glycerin to one part 
weight of the pulp prevents the growth of bacteria and probably 
destroys them by dehydration. Most of the bacteria are destroyed 
within one month at 20° C. During this period, then, from 4 to 6 
weeks, the glycerinated virus should not be used, and should from 
time to time be controlled by cultivation. At the end of this time 
the lymph is ready for use. 
40 Haecius. Cited from Paul, Kraus and Levaditi, Yol. 1, p. 593. 596 
INFECTION AND RESISTANCE 
Formerly the material for the vaccination of human beings was 
obtained very simply by dipping ivory splinters into the fluid of pus- 
tules, allowing this to dry, and rubbing these ivory or bone points 
into the exudate obtained by scratching the skin of the individual to 
be vaccinated. This method has practically gone out of use, and 
to-day the ripened glycerin pulp prepared as above is taken up in 
small capillary glass tubes and from these blown upon the vaccina- 
tion scratch. The efficiency of vaccine virus can be tested for po- 
tency by the inoculation of the ears of rabbits before use. 
That the ordinary vaccine virus ripened in glycerin is often in- 
effective in cases that might take if vaccinated with a more potent 
virus, such as the “green” or unripened vaccine, is a fact which has 
been forced upon our notice a number of times when soldiers of the 
United States Army, twice negatively vaccinated with glycerinated 
virus, developed takes after being vaccinated with unripened or 
“green” French virus. This fact makes it very desirable that further 
research should be carried on in the production of more potent 
vaccines. 
Noguchi has attempted to produce a vaccine without bacterial 
contamination by the intratesticular inoculation of male rabbits, the 
pulp from the infected testicles being employed. This method has 
not been widespread, partly because of the small yield, and partly, 
perhaps, because of the claim that such virus is not as potent for 
human beings, as a general rule, as vaccine derived from calves. As 
to this point, however, we have not sufficient evidence. 
The method of vaccination is an important factor in success. 
Wright 41 has attempted the intracutaneous injection of 0.1 c. c. of 
a 50 per cent, diluted ordinary glycerinated virus. He obtained 70 
per cent, successful takes in 227 negro soldiers by this method, when 
only slightly over 8 per cent, were obtained on the same cases by the 
incision method. 
It is our opinion that one of the most important factors in suc- 
cessful vaccination is to completely avoid the drawing of blood. The 
testing of the potency of vaccine virus can be carried out on the 
shaven skins of rabbits, and in performing such tests in our labora- 
tory, Frederic Parker, Jr., found that a very marked take, amount- 
ing almost to an eruption, could be obtained on one side of the abdo- 
men of a rabbit in which the virus had merely been rubbed into the 
shaven skin, while a very much more moderate take with a larger 
separation of papules resulted on the other side where scarification 
with slight bleeding had been practiced. 
Within slightly over two weeks after vaccination, it has been 
found by Kinyoun that the blood serum of the vaccinated individuals 
will neutralize vaccine virus if allowed to stand with it in a test tube 
overnight. 
41 Wright. Jour. A. M. A., Yol. 71, 1918, p. 654. THERAPEUTIC IMMUNIZATION 
597 
In judging of the efficiency of vaccination, we may draw upon 
a very convincing and voluminous literature. The statistical studies 
of Jurgensen in Sweden show that in the pre-vaccinal period from 
1774 to 1801, there were 2,050 deaths of small-pox per 1,000,000 
inhabitants. 
In the transitional period of nine years from 1801 to 1810, there 
were 680 per million. In 1810, the practice became generally em- 
ployed, although it was not enforced until 1816, but in the 35 years 
between 1810 and 1855, there were only 169 deaths. 
These statistical studies brought together, show that in the pre- 
vaccinal period there was a death rate of 20 per thousand, whereas, 
in the vaccinal period the death rate was 0.17 per thousand. 
There is no question about the fact that the simple procedure of 
systematic vaccination will control small-pox, but it must be remem- 
bered that this implies re-vaccination at least every 7 to 10 years, 
and that a negative vaccination should be followed in the same indi- 
vidual by a second attempt about two weeks later. 
Active Prophylactic Immunization in Rabies (Hydrophobia.)— 
Although many modifications have been suggested and actually 
used in different parts of the world, the most common method of 
immunizing against rabies still remains that originally devised by 
Pasteur. The Pasteur treatment takes advantage of the prolonged 
incubation period of rabies and is planned to confer immunity be- 
tween the time of inoculation and the time at which the disease 
would naturally appear. Since this period in ordinary street infec- 
tion by dog bite is usually 40 days or more, a considerable interval 
for active immunization is available. Formerly much of this time 
was lost in that the diagnosis of hydrophobia in the dog or other 
animal that had caused the injury could not be made with certainty 
until the results of rabbit inoculations had been obtained. Nowadays 
the ease with which a diagnosis of hydrophobia can be made within 
a few minutes by finding negri bodies in the hippocampal and cere- 
bellar cells has added considerably to safety in that it has made pos- 
sible a gain of almost two weeks in determining whether treatment 
should be instituted or not. 
Here again, although the infectious agent of rabies is not known 
with certainty even at the present day, the method of Pasteur de- 
pends upon active immunization by means of an attenuated virus. 
In standardizing the virus for the purpose of treatment Pasteur 
first produced what he calls the “virus fixe.” This consists of the 
ordinary street virus as obtained from rabid animals passed through 
a considerable series of rabbits (20-30) until its virulence for these 
animals has reached a maximum. After a sufficient number of such 
rabbit passages the incubation time after intracerebral inoculation is 
reduced to 7 or 8 days, but can no longer be shortened by further 
passage. The brain and cord material of rabbits dead of rabies after 598 
INFECTION AND RESISTANCE 
such repeated passages constitutes virus fixe. This can be preserved 
for considerable periods in 60 per cent, glycerin, and this is the 
initial material from which the attenuated preparations for treat- 
ment are produced. 
In preparing the material for treatment a small amount of virus 
fixe is injected subdurally into rabbits, about 0.2 c. c. of a salt solu- 
tion emulsion being given. The inoculation is very easily made 
through a small trephine opening in the skull, and contamination 
is very easily avoided. Just before the rabbit dies when completely 
paralyzed it is killed by chloroform and the cord is removed best by 
the method of Oschida.42 The rabbit is nailed to a board, back up- 
permost, and washed with a weak antiseptic; a longitudinal incision 
is then made along the backbone from the occiput to the lumbar 
region, and the vertebral column laid bare. 
After searing the tissues around the back of the head the spine is 
cut across just behind the occiput, and again in the same way just 
above the sacrum. The neck and lumbar regions are dissected loose 
from the skin and gauze is inserted under it to avoid contamination. 
The assistant grasps the end of the spinal cord as it appears in the 
cervical region and pulls on it very slightly while the operator with 
a glass rod or a piece of wire pushes against it from below. If this 
is carefully done the spinal nerves are torn and the cord can be 
gradually pulled out of place. This procedure is by far the best, 
although it requires a certain amount of practice. 
The cords so removed are hung up by a thread in bottles contain- 
ing sticks of caustic potash and exposed in a dark place to 22° to 23° 
C. Under these conditions of drying and temperature the virus is 
gradually attenuated until at the end of 13 days or more the viru- 
lence is practically nil. If removed from the drying bottles at any 
time during the process and kept in a refrigerator in sterile glycerin 
the virulence, whatever it may be at the time of placing into the 
glycerin, remains fairly constant for long periods. When any of 
this material is used for treatment little pieces of the cord % cm. in 
length are cut off and emulsified in 2.5 c. c. of salt solution, and this 
emulsion is used for injection.43 
When patients are to be treated the principle of the treatment is 
to inoculate them first with cords that have been dried for consider- 
able periods, gradually proceeding toward those that have been dried 
for less prolonged times and are therefore more virulent. The treat- 
ment is varied in the individual case according to the severity of the 
injury. Formerly treatment was begun with cords dried as long as 
16 days. More recently it has been found that cords dried for longer 
than 8 days are practically non-virulent and correspondingly lack 
42 Oschida. Centralbl. f. Bakt., Yol. 29, 1901. 
43 In oui’ description of the methods of drying rabies, for the sake of 
adhering to a standard, we follow closely the directions laid down by A. M. THERAPEUTIC IMMUNIZATION 
599 
in antigenic value. They are no longer employed, therefore, since 
their use is regarded as a waste of time. The following tables, on 
Scheme for Mild Treatment 
Day 
Cord 
(injections) 
Amount injected 
Day 
Cord 
(injections) 
Amount injected 
Adult 
(c. c.) 
5-10 
yrs. 
(c. c.) 
1-5 
yrs. 
(c. c.) 
Adult 
(c. c.) 
5-10 
yrs. 
(c. c.) 
1-5 
yrs. 
(c. c.) 
1 
8-7-6 = 3 
2.5 
2.5 
2.0 
12 
4=1 
2.5 
2.5 
2.5 
2 
5-4 = 2 
2.5 
2.5 
1.5 
13 
4 = 1 
2.5 
2.5 
2.5 
3 
4-3 = 2 
2.5 
2.5 
2.0 
14 
3 = 1 
2.5 
2.5 
2.0 
4 
5=1 
2.5 
2.5 
2.5 
15 
3 = 1 
2.5 
2.5 
2.0 
5 
4=1 
2.5 
2.5 
2.5 
16 
2=1 
2.5 
2.0 
1.5 
6 
3 = 1 
2.5 
2.5 
2.0 
17 
2 = 1 
2.5 
2.0 
1.5 
7 
3 = 1 
2.5 
2.5 
2.0 
18 
4 = 1 
2.5 
2.5 
2.5 
8 
2=1 
2.5 
1.5 
1.0 
19 
3 = 1 
2.5 
2.5 
2.5 
9 
2 = 1 
2.5 
2.0 
1.5 
20 
2 = 1 
2.5 
2.5 
2.0 
10 
5 = 1 
2.5 
2.5 
2.5 
21 
2 = 1 
2.5 
2.5 
2.0 
11 
5 = 1 
2.5 
2.5 
2.5 
Scheme for Intensive Treatment 
Day 
Cord 
(injections) 
Amount injected 
Day 
Cord 
(injections) 
Amount injected 
Adult 
(c. c.) 
5-10 
yrs. 
(c. c.) 
1-5 
yrs. 
(c. c.) 
Adult 
(c. c.) 
5-10 
yrs. 
(c. c.) 
1-5 
yrs. 
(c.c.) 
1 
8-7-6=3 
2.5 
2.5 
2.5 
12 
3 = 1 
2.5 
2.5 
2.0 
2 
4-3 = 2 
2.5 
2.5 
2.0 
13 
3=1 
2.5 
2.5 
2.0 
3 
5-4 = 2 
2.5 
2.5 
2.5 
14 
2=1 
2.5 
2.5 
2.0 
4 
3 = 1 
2.5 
2.5 
2.0 
15 
2 = 1 
2.5 
2.5 
2.0 
5 
3 = 1 
2.5 
2.5 
2.0 
16 
4 = 1 
2.5 
2.5 
2.5 
6 
2=1 
2.5 
2.0 
1.5 
17 
3 = 1 
2.5 
2.5 
2.5 
7 
2 = 1 
2.5 
2.5 
2.0 
18 
2 = 1 
2.5 
2.5 
2.0 
8 
1 = 1 
2.5 
1.5 
1.0 
19 
3 = 1 
2.5 
2.5 
2.0 
9 
5 = 1 
2.5 
2.5 
2.5 
20 
2 = 1 
2.5 
2.5 
2.5 
10 
4 = 1 
2.5 
2.5 
2.5 
21 
1 = 1 
2.5 
2.5 
2.0 
11 
4 = 1 
2.5 
2.5 
2.5 
Stimson, in the U. S. P. H. S. Bull. 65, 1910. There are various modifica- 
tions used in different countries, in many cases unimportant, and it seems 
well to adhere to the U. S. regulations as a standard for this country. 
p. 600, taken from Stimson’s article in Bulletin 65 of the Hygienic 
Laboratory of the IT. S. Public Health Service, give the standard 
methods of treatment as recommended by the United States Public 
Health Service. 
This is the standard treatment used almost everywhere in the 
world at present. 
Recently, there has been a tendency to begin with cords that have 600 
INFECTION AND RESISTANCE 
been dried for a less prolonged period. The following scheme of 
treatment is taken from an article by Stimson of the United States 
Public Health Service.44 
Treatment for Adult * 
Day of Treatment 
Cord, 
Dried, 
Days 
Amount 
of 
Cord Cm. 
First 
6 
1 
Second 
5 
1 
Third 
4 
1 
Fourth 
3 
0.5 
Fifth 
3 
0.5 
Sixth 
2 
0.5 
Seventh 
2 
0.5 
Eighth 
1 
0.5 
Ninth 
5 
0.5 
Tenth 
4 
0.5 
Eleventh 
4 
0.5 
Day of Treatment 
Cord, 
Dried, 
Days 
Amount 
of 
Cord Cm. 
Twelfth 
3 
0.5 
Thirteenth 
3 
0.5 
Fourteenth 
2 
0.5 
Fifteenth 
2 
0.5 
Sixteenth 
4 
0.5 
Seventeenth .... 
3 
0.5 
Eighteenth 
2 
0.5 
Nineteenth 
3 
0.5 
Twentieth 
2 
0.5 
Twenty-first .... 
1 
0.5 
* Cited from Stimson. Jour. A M. A,. Yol. 76, 1921, p. 241. 
Other methods have been recommended. One of these is that of 
Ilogyes, in which virus fixe unattenuated is used in dilution. 
Hogyes begins by injecting 3 c. c. of a 1 to 10,000 dilution of virus 
fixe, gradually proceeding within 14 days to 1 c. c. of a 1 to 100 
dilution. 
Fixed virus attenuated by the addition of antirabic serum and 
chemical disinfectants (carbolic acid) and by partial digestion in 
gastric juice has also been used, but none of these methods has at- 
tained widespread application. 
A further method used in laboratories is that of Harris and 
Shackell.45 Harris 46 believes that the attenuation of rabic virus 
by the Pasteur method depends primarily upon the method of ex- 
tracting the water. The slow desiccation of the Pasteur procedure, 
he thinks, acts by a gradual concentration of the salts in the brain, 
the action being essentially a chemical one. He has altered the 
method by rapidly drying the cord at 0° in a vacuum jar over sul- 
phuric acid. Virus so dried, will retain its virulence for as long as 
four months if guarded against moisture. Of this he emulsifies 10 
mg. in 10 c. c. of salt solution, that is one milligram to each cubic 
centimeter. He then tests virulence by injecting 0.1 c. c. of the dilu- 
tion into the brain of a rabbit, passing his needle into the lateral 
ventricle so that none of the material may escape. By this method 
he finds that the material after three weeks’ preservation may be 
equivalent in virulence to fresh cord. After 50 days it is 25 per cent, 
more infective than by the old method. After 200 days its infec- 
44 Stimson. Cited from the Jour. A. M. A., Yol. 76, 1921, p. 241. 
45 Harris and Shackell. Jour. Inf. Dis., Yol. 8, 1911, p. 47. 
46 Harris. Jour. Inf. Dis., Yol. 13, 1913, p. 155. THERAPEUTIC IMMUNIZATION 
601 
tivity is equal to that of the two day cord of the old method, and after 
500 days it is two and one-half times as infective as the three day 
cord. Harris begins his treatment with material which is about six 
months old. As the treatment continues, he gradually increases until 
he uses material which contains 100 minimal infectious doses per 
milligram. He controls his treatment by the preparation of charts 
on which he tabulates the proportion of infectious material to noil- 
infectious material in a brain after preservation for various periods. 
For further particulars we refer the reader to Harris’s publications. 
It should be remembered that rabies treatment is a serious under- 
taking which should be carried out with a complete knowledge of the 
procedure and with the advice of laboratory workers trained in the 
preparation of the material. 
The treatment is severe and in its course, after the first few in- 
jections, large blotches of local hypersensitiveness may appear at the 
sites of injection, and the patient’s general health may suffer. The 
nervous strain, as well as the physical discomfort, may bring further 
severe nervous reactions. 
In the extermination of rabies as a menace to human beings, 
recent attempts have been made to develop a method of prophylactic 
immunization of dogs. Rabies it must be remembered is in many 
parts of the world sufficiently common to make it a public health 
problem, and in southern countries, particularly, seems to be rather 
increasing than diminishing. During the war there was a consider- 
able increase of the disease in Japan, and Umeno and Doi 47 at- 
tempted to introduce the prophylactic treatment of dogs by methods 
in principle the same as those used in human beings. They showed 
that with a single injection of virus, it is possible to immunize a 
dog sufficiently to protect him for one year against infection by the 
bites of other dogs. According to Eichhorn and Lyon 48 from whom 
we take this account, their method consisted of the injection of a 
single dose of mixed virus treated with phenol. The brain of a rab- 
bit which has died from fixed virus was ground up in four times its 
volume of a mixture of 60 parts of glycerin and 40 parts of water 
containing 1.25 per cent, phenol. This was stored at 22° C. for two 
weeks, or in the icebox for 30 days to reduce its virulence. At first 
it was given in dilution, but later in the original condition, that is, 
dilution of the brain 1 in 5. After considerable preliminary work, 
they determined that one injection of 5 c. c. per 15 kilogram weight, 
could be used, and in puppies of 4 and 1/2 kilograms or less, one- 
half the dose should be given. Up to 1921, over 31,000 dogs were 
vaccinated, and in only one case did the immunity fail. Eichhorn 
and Lyon in 1922 reinvestigated this problem, and confirmed the 
findings of the Japanese investigators. In a rather small but fairly 
47 Umeno and Doi. Kitasato. Archiv. Exp. Med., Yol. 4, No. 2, p. 89. 
48 Eichhorn and Lyon. Jour. Amer. Vet. Med. Assn., April, 1922, p. 1. 602 
INFECTION AND RESISTANCE 
conclusive series, they injected one dose of 5 c. c. of the virus, as 
recommended by the Japanese investigators, subcutaneously, and 
subsequently injected street virus into the anterior chamber into 
these dogs and into three controls. None of the vaccinated dogs died 
of rabies. 
The method seems to us of great importance if consistently fol- 
lowed among pet dogs and reenforced by the proper control of stray 
animals. By these methods, there seems no reason to doubt that 
rabies can be eventually completely exterminated. 
Immunity in Syphilis.—Until relatively recent years there seems 
to have been little question in the mind of clinicians regarding the 
existence of true immunity in syphilis. It was stated uncom- 
promisingly by Kicord 49 that “an individual who had once acquired 
syphilis was thereafter protected against reinfection.” This opinion 
acquired wide acceptance and was shared by most of his contempo- 
raries. Baumler 50 in 1875 summarizing the authoritative opinions 
of that period stated that “One who has once had small-pox, scarlet 
fever, typhus, etc., is, as a rule, not liable to these diseases again for 
the rest of his life. The same is true of syphilis.” However, even 
at the time Baumler wrote this, exceptions to the supposed rule were 
accumulating and he appends, to the positive statement given above, 
references to observed instances of second infection reported by 
Bidenkap,51 H. Lee,52 Diday,53 Kober,54 Zeissl,55 and others. 
The conception of the existence of a true acquired immunity was 
also expressed and widely accepted in the so-called “laws” of Colies 
and Profeta. The former, first enunciated by Beaumes and later, 
in 1837, stated by Abraham Colles,56 of Dublin, is the well-known 
generalization based on the observation that mothers who have borne 
syphilitic infants were not infected by their children while suckling 
them, although such children might often infect wet nurses. Pro- 
feta’s 57 observation was the converse of this, namely, that children 
born of mothers who suffered from active syphilis during the period 
of conception did not acquire the disease from their mothers. 
Both of these phenomena appear too well founded in clinical 
observation to be questioned, at least as frequent occurrences. More- 
49 Ricord. Becherches sur le Chancre, Paris, 1858, cited from Baumler. 
60 Baumler. Ziemssen’s Cyclop, of Bract. Med., Amer. Ed., Wm. Wood 
and Co., N. Y., 1875, iii. 
51 Bidenkap. Wien. med. Woch., 1865, cited from Baumler. 
52 Lee, H. Lecture on Syphilis, London, 1863. 
53 Diday. Arch, gener., 1862, ii, cited from Baumler. 
54 Kober. Berl. klin. Woch., 1872, No. 46. 
55 Zeissl. Zeitschr. d. K. K. Gesell. d. Aertze in Wien, 1858. 
56 Colies. Cited from Osier and Churchman, “Syphilis,” Osier’s System of 
Medicine. 
57 Profeta. Cited from Osier and Churchmann, “Syphilis.” Trattato 
Practico delle Malattie Veneree, Palermo, 1888. THERAPEUTIC IMMUNIZATION 
603 
over, considering the intimate contact between mother and infant 
during the first months after birth, they acquire unusual importance. 
However, as we shall see, they have been deprived of much of their 
bearing as proofs of true acquired or inherited immunity by serologi- 
cal investigations such as those of Bauer,58 Knopfelmacher,59 and 
others, which have shown that mothers of syphilitic children usually 
give positive Wassermann reactions, a fact which makes it seem 
likely that such women are suffering from syphilis in a latent form 
and are not immune in the ordinary sense. This indeed was the 
interpretation given to the ‘‘laws” of Colles and Profeta by 
Fournier ,i0 and by Matzenauer 61 at a time prior to that at which 
serological data were available. It is still unclear why such mothers 
should so frequently exhibit the disease in a latent form. However, 
this is a matter for the intelligent discussion of which we are not 
at the present time in the possession of sufficient information. 
In the years just preceding the period of experimental investi- 
gation of syphilis upon animals much purely clinical research was 
carried out on the problem of reinoculation and what is commonly 
known as “superinfection” of syphilitic human beings. Much of this 
work is unavailable as scientific evidence owing to the difficulties of 
distinguishing, at that time, between the true chancre and the 
chancroid, but a considerable number of the observations then made 
have been of much value in pointing out directions of later research 
upon animals. The work has been so thoroughly analyzed both by 
Neisser and by Levaditi that it would be needless repetition to do so 
again. The records include both accidentally occurring superinfec- 
tions, and purposeful experimental reinoculations. Without, there- 
fore, going into details concerning the individual cases we may sum- 
marize the conclusions justified from a study of these observations. 
1. The reports of Lasch,62 Jadassohn,63 Sabeareanu,64 
Queyrat,65 Taylor,66 H. Lee, Knowles,67 and many others, have 
shown that patients are susceptible to a second inoculation during 
the first incubation time, that is, during the period elapsing between 
the first infection with syphilis and the appearance of the chancre. 
Second positive inoculations have also been successful at periods 
shortly subsequent to the appearance of the primary sore. 
58 Bauer. Wien. klin. Woch., 1908, No. 28. 
59 Knopfelmacher and Lehndorff. Med. klin., 1909, No. 40: Wien. med. 
Woch., 1909, No. 38. 
60 Fournier. L’Heredite Syphilitique, Paris, 1890. 
01 Matzenauer. Vererbung d. Syphilis, Wien, 1905, cited from Brack. 
62 Lasch. Arch. f. Dermat. u. Syph., 1891, p. 61. 
63 Jadassohn. Arch. f. Dermat. u. Syph., 1907, p. 86; Festschr. f. Neisser, 
64 Sabeareanu. These de Paris, 1905, Chancres Syph. Successif. 
65 Queyrat. Bull, de la Soc. Med. des Hop., 1904, No. 28, p. 905. 
66 Taylor. Jour. Cutan. Dis., December, 1890. 
67 Knowles. N. Y. Med. Jour., December, 1906. 604 
INFECTION AND RESISTANCE 
Autoinoculations and reinoculations undertaken after the chancre 
has become well developed have been, in the main, negative, though 
Queyrat reports a case successfully inoculated daily up to the 
eleventh, and Taylor one inoculated on the fourteenth day after the 
appearance of the primary induration. In contrast to these and other 
successful attempts, many observers record failure. However, the 
point is not one of particular importance, since, after all, the ap- 
pearance of the chancre does not mark off any fundamental change 
in the progressive pathological development of the disease, and indi- 
cates only the completed reaction at the point of entrance. The fact 
remains that analysis has revealed that reinoculation, with the ap- 
pearance of the second initial lesion, is possible up to about the twen- 
tieth to the thirtieth day after the first infection with the trepone- 
mata (Mauriac 68 and Neisser state the twenty-second day, Queyrat 
cites a case eleven days, Linderman one twenty-four days after the 
appearance of the chancre). It is claimed by some of the observers, 
however, that even when reinoculation during this period is success- 
ful, the incubation of the second and subsequent lesions is shorter, 
that the induration itself is less severe, may not ulcerate, and heals 
more readily than the first. 
In judging of the success or failure of re inoculation practised 
during the later days of the period above referred to, the possibility 
must be borne in mind that the trauma produced at the inoculation 
might have served to favor the development of a localized focus, the 
treponemata at this time being very probably well distributed 
throughout the body. Unsuccessful control inoculations with non- 
syphilitic materials in Queyrat’s cases would tend to eliminate this 
possibility, whereas lesions resulting at the sites of such control 
abrasions in the experiments of Neuman and Celiak,69 and of 
Levaditi,70 appear to support it. However, this is not of great im- 
portance inasmuch as it determines merely a relative lengthening 
or shortening of the period during which reinoculation or superin- 
fection is possible. 
2. After the disease is well established as a systemic infection, 
that is, from the time of development of the chancre throughout the 
so-called active “secondary” period, reinoculation is either impos- 
sible or, at any rate, extremely difficult. Neisser cites Rollet as 
follows: 
“Although I and my predecessors have a thousand times at- 
tempted to reinoculate luetic subjects, we have never observed a 
successful case. I know no single fact more thoroughly proven than 
the insusceptibility of a syphilitic to the action of a new virus, and, 
68 Mauriac. Cited from Neisser, loc. cit. 
69 Neuman and Cehac. Wien. med. Bl., 1890, cited from Neisser. 
70 Levaditi. De Vlmmunite acquise dans la Syphilis. Zeitschr. f. Imm. 
Ref., pp. 191, 277-318. THERAPEUTIC IMMUNIZATION 
605 
moreover, these experiments are so harmless that they may be per- 
formed without scruple.” 
The same opinion was held by Mauriac and is in a general way 
assented to by Eeisser in his summary of this place of his studies. 
(Neisser, loc. cit., pp. 180-181.) 
That the patient with well-developed lues has acquired a con- 
siderable degree of resistance to fresh inoculations is pretty gener- 
ally accepted, therefore, by all who have studied the question—but 
there are investigators, notably Finger and Landsteiner, who believe 
this resistance to be less absolute than stated by earlier writers. 
Their observations on patients with active syphilis seem to indicate 
that superinfection is possible “under certain circumstances in all 
stages of the disease,” but “the positive effect can be obtained only 
with considerable quantities of the virus.” Furthermore, Land- 
steiner 71 states that, in general, lesions so obtained are relatively 
slight in severity, do not appear as primary indurations, but have 
the tendency to simulate the particular variety of lesion spontane- 
ously manifest in the individual at the time. Were it not for Finger 
and Landsteiner’s monkey experiments, to which we will refer di- 
rectly, and which seem to bear them out in their interpretation of 
these inoculations as “positive” results, the simulation of the spon- 
taneously occurring lesions by the inoculation-products would again 
justify suspicion that these experimental results represented merely 
traumata in which, as points of less resistance, the patient’s pre- 
existent disease had found a favorable spot for localization. 
Observations in this respect, also, are corroborated by the older 
literature. A report which has direct bearing on this point, is one 
of Queyrat and Pinard who inoculated a tertiary patient with chan- 
cre material, obtaining not a primary sore but an ulcerated lesion 
having the clinical characteristics typical of the late skin manifesta- 
tions of the disease. The autoinoculation experiments of Ehr- 
mann 72 also would tend to show that the resistance of the luetic 
subject is a relative one only. Ehrmann succeeded in producing 
positive autoinoculation-products in 45 syphilitics with papular erup- 
tions. Control inoculations with sterile water were negative. Al- 
though his observations and those of Finger and Landsteiner, with 
other similar ones, teach us that superinfection during the disease 
is possible the nature of the lesions, their short incubation time, and 
their exceptional character when averaged with the total of such at- 
tempts, prevent them from invalidating the conclusion that there is 
a resistance at this stage higher than that of the normal subject. 
We may conclude, therefore, concerning the secondary period of 
the disease, that the luetic individual has acquired a resistance which 
while not absolute is at least very high, and protects him from fresh 
71 Landsteiner. Centralbl. f. Bakt., Ref. 1908, Yol. 41, p. 785. 
72 Ehrmann. Cited from Neisser, loc. cit. 606 
INFECTION AND RESISTANCE 
external inoculation, although at the time his disease may still be 
progressing and in no sense overcome. 
3. During the late stage of syphilis, the stage at which, accord- 
ing to the more or less arbitrary divisions of Ricord, we are accus- 
tomed to speak of it as “tertiary,” the resistance is still manifest, 
though apparently not so regularly potent as during the preceding 
“secondary” period. Neisser expresses himself with great caution 
and accepts few, if any, of the observed reinoculations of tertiary 
cases as surely representing unquestionable freshly acquired infec- 
tions. Nevertheless, taking into consideration a detailed study of 
individual reports, he concludes that resistance during the late stages 
is pronounced but already beginning to wane. He himself in the 
Festschrift to F. Joseph Pick in 1898 cites a case of the development 
of a chancre in a tertiary case which, however, was not followed by 
constitutional symptoms. 
4. In the preceding paragraphs we have concerned ourselves 
entirely with questions of “superinfection,” that is, the implantation 
of the syphilitic virus into subjects still suffering from manifesta- 
tions of the disease. A problem of equal theoretical, and of much 
greater practical importance, is that dealing with true “reinfection.” 
By “immunity” in the ordinary sense, we mean an increased resist- 
ance to specific infection which persists for a more or less prolonged 
period after the active disease has been overcome and the causative 
agents removed from the body. It is not easy to draw conclusions 
concerning this point from observations on human beings, since it 
is most difficult to decide, even with the aid of serological methods, 
whether or not a given case is cured in the bacteriological sense; 
for syphilis is preeminently the disease in which there occur fre- 
quent and prolonged latent periods terminated, often after lapses of 
years, by the reappearance of foci of a grave nature. 
Moreover, our own experiments both with syphilis in rabbits and 
the treponemata in culture have convinced us that these micro- 
organisms may assume for long periods a condition of metabolic 
latency, a sort of resting period, during which they incite no reactions 
of any kind on the part of the tissues, apparently do not multiply 
to any great extent and yet remain alive and capable of development 
when conditions favor this. 
In spite of these difficulties, however, careful clinical studies by 
Jonathan Hutchinson, Taylor, Hudelo,73 Neisser, and many others, 
have furnished data which warrant an opinion upon this problem. 
Of especial value is the painstaking analysis of reported cases of re- 
infection in syphilis made in 1909 by Felix John.74 
In contrast with some others who have attempted similar analyses 
73 Hudelo. Ann. de dermat. et de syph., 1890, p. 353. 
74 John, Felix. Reinfectio Syphilitica, Samml. klin. Vortr., Volkmann, 
1907-1909, p. 559 et seq. THERAPEUTIC IMMUNIZATION 
607 
John takes the utmost care to separate these cases into the ones in 
which the evidence of true reinfection is absolute, and those in which 
the reported details are insufficient to exclude the possibility of 
recrudescence. In agreement with Taylor he insists upon a symptom- 
free interval of five years between the last manifestations of the first 
attack and the appearance of the second. As an example of what he 
calls an “Ideal Fall” we may cite the following: 
X. J., April 1, 1872, primary sore. 
Roseola, polyadenitis, mucous patches. September, 1872, papular rash. 
March, 1873, palmar and plantar spots, iritis. 1875, gumma of tibia and 
serpiginous syphilide of right thigh. 
Four courses of inunctions and KI. 
1876, married—two healthy children. 
No symptoms till 1887 when he acquired a second chancre followed by 
typical roseola. In 1888 wife had still-birth. 
John has analyzed in this way 356 cases of supposed reinfection, 
in 34 of which the first attack was of congenital origin. Of the 
remaining 322, fourteen were cases which seemed unquestionable 
instances of reinfection and in 16 more there was practically no 
doubt of this. Of the 34 hereditary cases there were three, one of 
Emery, one of Taylor, and one of Hochsinger, in which there was 
practically no question of their nature as valid reinfections. In all 
of the others there was one point or another which rendered them 
doubtful as evidence. 
John concludes that true reinfection can unquestionably occur 
in syphilis but that it is relatively rare. 
To John’s cases Ffeisser has added others reported between 1909 
and 1911. Yet even with these, the total number is not a large one 
Nevertheless, we should not be tempted to conclude from this that 
the relative infrequency of such cases is evidence in favor of the 
existence of a true immunity analogous to that following typhoid, 
plague, etc. For we have seen that a very definite insusceptibility 
is coincident with the persistence of actual disease in the subject 
and, as Neisser points out, the more recent investigations carried out 
with the aid of serological tests have shown that the number of cases 
uncured though long without symptoms, is much larger than for- 
merly supposed. The scarcity of true reinfection, therefore, may 
well be due to the relative scarcity of completely cured cases. More- 
over, it must be remembered that, in this disease, even when final 
recovery results, it is usually achieved only at an age when the in- 
dividual is less exposed to reinfection because of changed economic 
and domestic conditions, or by reason of the virtue which comes with 
arteriosclerosis and the belated wisdom of the burnt child that fears 
the fire. 
Granted, then, that true reinfection is possible, is there any evi- 
dence that when it does occur, the second attack is less severe and 608 
INFECTION AND RESISTANCE 
more easily cured than the first, a fact which would also tend to 
support the opinion that a certain degree of immunity persists. 
Jonathan Hutchinson 75 expressed this view in a clinical lecture in 
which he states that “second chancres are far more common than 
second attacks of constitutional syphilis.” However, John in his 
summary, in which Hutchinson’s cases are included, finds that in 
general the disease has run a second course very similar as to severity 
to that incident to the first infection. 
If we gather together, then, the facts revealed by clinical study we 
may conclude with Levaditi, Neisser, and most others, that: 
1. The syphilitic subject acquires definite resistance to rein- 
oculation which becomes manifest soon after the appearance of the 
primary sore, at a time when the virus may be regarded as having 
gained universal systemic distribution. 
2. This resistance, high though not absolute, persists through- 
out the secondary or most active period of the disease and into the 
tertiary stage. During the latter, however, it appears somewhat tc 
decline, reinoculation or superinfection being more frequently pos- 
sible at this period. 
3. When syphilis is entirely cured, susceptibility may in all 
probability be regarded as returning, possibly, though not certainly, 
to the same degree as it exists in the normal subject. The reasons 
for this last belief will become more clear when we study the evidence 
contributed by animal experimentation. 
Notwithstanding the admirable thoroughness with which clinical 
data had been collected and analyzed in the study of syphilis, prog- 
ress beyond the points indicated in the preceding sections was quite 
impossible without the aid of animal experimentation and a knowl- 
edge of the causative agent. Fortunately these two deficiencies in 
our methods were removed when in 1903 Metchnikoff and Roux 
succeeded in transmitting the disease to a chimpanzee, and in March, 
1905, Schaudinn 76 discovered the treponema pallidum. 
As a matter of fact, probable transmission of syphilis to lower 
monkeys had been accomplished as early as 1879 by Klebs 77 and 
subsequently by Neumann,78 Martineau,79 and Charles Nicolle,80 
but in none of these experiments had it been possible to prove beyond 
question the syphilitic nature of the inoculation-products. In the 
chimpanzees inoculated by Metchnikoff and Koux, the animals de- 
veloped not only primary sores but also secondary eruptions, poly- 
adenitis, and enlarged spleens in such characteristic manner that the 
75 Hutchinson, J. Brit. Med. Jour., Yol. I, 1882, p. 699. 
76 Schaudinn. Deutsch. med. Woch., 1905, No. 42; Arb. a. d. Kais. 
Gsndlitsamte, Yol. 26, 1907. 
77 Klebs. Arch. f. exper. Path. u. Pharmalcol., 1879. 
78 Neumann. Cited from Muhlens, loc. cit. 
79 Martineau. Arch. f. Dermat. u. Syph., 1884, No. 16. 
80 Nicolle, Ch. Cited from Muhlens, loc. cit. THERAPEUTIC IMMUNIZATION 
609 
identity of the inoculation disease with human syphilis could no 
longer be doubted. 
Successful transmission to other anthropoids and to lower mon- 
keys were then announced, in rapid succession, by Metchnikoff and 
Roux, Ch. Nicolle, Neisser, Baermann and Halberstaedter, Finger 
and Landsteiner, Hoffman, and others. The susceptibility of mon- 
keys was tabulated by ISTeisser in the following series: Chimpanzee, 
Gibbon, Orang-Outang, Cynocephalus babuin, Cynocephalus sphinx, 
Cynocephalus hamadryas, Cercopithecus fulginosus, Macacus niger, 
M. nemestrinus, M. cynomolgus, M. sinicus, M. speciosus, M. rhesus. 
In 1906 Bertarelli 81 produced syphilitic keratitis in rabbits, dem- 
onstrating the treponema pallidum in sections of the cornea and in 
1907 Parodi 82 first produced syphilitic orchitis in the same animals. 
Apart from monkeys and rabbits, no animal species have so far 
been shown sufficiently susceptible to be available for systematic 
study. It is true that the production of keratitis is claimed in dogs 
and sheep (Bertarelli, Hoffman and Briinig), in guinea pigs (Ber- 
tarelli), in cats (Levaditi and Yamanouchi), and in goats (Ber- 
tarelli). However, these experiments have been isolated and too 
uncertain with present methods to offer material for experimentation. 
Our own attempts on cats, pigs, guinea pigs, rats, mice, and a few 
birds, have yielded negative results only. There are many observa- 
tions of great scientific interest which might be discussed in con- 
nection with the problem of susceptibility of various animal species; 
however, we will confine ourselves at present to those phases of the 
work only which have bearing on the questions of immunity. 
In the fundamental premises, the work on monkeys has pretty 
accurately confirmed the observations concerning reinoculation and 
superinfection previously made on human beings. 
The most extensive studies in this respect are those of Heisser 83 
and his associates in the Javan expedition. Briefly stated, ISTeisser 
reinoculated 135 animals 165 times with negative result on the second 
and subsequent inoculations. The second inoculations were made at 
periods ranging from 21 days to two years after the first. In but 27 
animals did the reinoculation show positive results, and in ten of 
these cases only does ISTeisser recognize the experiments as valid. 
Although these ten positive reinoculations, it is true, add an element 
of irregularity to the series, they constitute but 6.8 per cent, of the 
entire number, a proportion which in no way invalidates the experi- 
ments when we consider that the work was done entirely on the 
lower monkeys, animals that are far less susceptible to syphilis than 
are human beings and in many of which, therefore, systemic dis- 
tribution of the virus (a generalization apparently necessary for the 
81 Bertarelli. Centralbl. f. Bakt., Orig. Yol. 41, 1906, and Yol. 43, 1907. 
82 Parodi. Centralbl. f. Bakt., Orig. Vol. 44, 1907. 
83 Neisser. Beitr. z. Pathol, u. Ther. d. Syphilis, Springer, Berlin, 1911. 610 
INFECTION AND RESISTANCE 
development of resistance) may not have taken place. Neisser’s 
conclusions, therefore, that monkeys, like human beings, are not 
reinoculable while suffering from systemic syphilis, seem entirely 
justified. 
His work, as well as that of Finger and Landsteiner, and of 
Kraus and Volk 84 on lower monkeys, has shown that resistance 
does not develop until the twelfth to the twentieth or twenty-first day 
after the first inoculation; that is, again as in man, when the virus 
has become generally distributed. Finger and Landsteiner, further- 
more, noticed that reinoculation-products, obtained by reinfection 
during the first incubation period, that is, before the development of 
the primary lesion, were less severe and developed in a shorter time 
than did the first lesion. This phenomenon which would tend to 
mark another analogy to the conditions prevailing in human beings, 
was not observed in the experiments of Neisser and of Kraus and 
Volk. However, like the similar observation in human beings, it 
seems to indicate the gradual acquisition of resistance as the virus 
begins to exert its influence upon the tissues. 
Again, with monkeys as in man, the question arises whether the 
resistance so unquestionably proven is a condition merely coexistent 
with active disease, or whether it may be interpreted as a true im- 
munity which persists after the micro-organisms have been com- 
pletely removed. The most directly pertinent experiments are those 
of Heisser. Neisser reinoculated monkeys at periods ranging from 
27 to 645 days after the first infection. After waiting a time suffi- 
ciently long to insure the negative result of the reinoculation, he used 
organ-substance from these animals to inoculate other monkeys. In 
22 experiments of this kind he obtained positive results—showing 
that the organs of the apparently immune animals still harbored 
virulent treponemata. In seven animals only did the inoculations 
with organ-substance fail to produce lesions, but of these all but one 
died before the 30th day after inoculation. 
In contrast to these results Neisser found that animals which had 
been “cured” by various treponemacidal agents such as atoxyl, 
arsacetin, etc., were almost regularly reinoculable. In fact, these 
experiments were so uniform that Neisser later utilized the reinocu- 
lation method as an index of cure or persistence of the disease. 
The results obtained in monkeys, therefore, are very similar to 
those determined by clinical observations in man, and the following 
statements may be taken as summarizing the conditions revealed by 
monkey experiment: 
1. That the body develops resistance, progressively increasing 
as the virus becomes generally distributed. 
84 Kraus and Yolk. Wien. klin. Woch., 1906, No. 21, Ref. IX. Kongress 
d. Dermatol. Gsell. in Bern. THERAPEUTIC IMMUNIZATION 
611 
2. That the resistance probably reaches its highest development 
during the early tertiary period. 
3. That complete cure is probably synonymous with gradual 
return of susceptibility. 
The only other animals on which systematic experimentation has 
been possible up to the present time have been rabbits. Since 
Bertarelli’s successful production of keratitis and Parodi’s inocula- 
tion of the testes in these animals, they have been studied carefully 
by a number of workers, chiefly Uhlenhuth and Mulzer,85 and 
Noguchi; 8U and in our own laboratory, with Hopkins and Mc- 
Burney, the writer has observed rabbit syphilis for a number of 
years. From a large mass of observations it appears that the condi- 
tions in rabbits are not identical with those observed in man and 
monkeys. So far, the testes and the eyes are the only organs in 
these animals in which syphilitic lesions can regularly be produced, 
and although treponema pallidum may apparently be distributed 
generally to the organs after inoculation, it does not easily arouse 
pathological reactions except in the organs named, and the lesions 
produced are not accompanied by a generalized resistance compa- 
rable to that discussed in connection with the higher animals in pre- 
ceding sections of this paper. 
Bertarelli found that he could reinoculate the cornea in a rabbit 
that had previously been inoculated with syphilis. Uhlenhuth and 
Mulzer, and Neisser and Purekhauer, showed that infections of the 
eye could be produced while the opposite eye was still syphilitic. 
Tomazewski found that scrotal infection did not protect against in- 
fection of the cornea, and vice versa. Furthermore, Uhlenhuth and 
Mulzer, on the basis of a very thorough study, believe that such rein- 
fections are neither less extensive nor more rapid in healing than 
was the first lesion. The same authors noticed resistance to reinocu- 
lation in two animals only, and these were young rabbits in the first 
weeks of life, which, they believed, had been generally syphilized by 
intracardial injections. Interesting in this connection are claims by 
Ossola and Truth who believe that successful skin inoculation in 
rabbits confers a certain amount of skin resistance and this is in 
harmony with the belief of Kraus and Volk, that a specific skin 
immunity in syphilis is possible. Of great interest to us and of a 
possible theoretical bearing which we will discuss below, are observa- 
tions made by the writer with Hopkins and McBurney 87 on 20 
rabbits which were reinoculated into the testes after apparent healing 
of lesions in these organs. It appeared that in rabbits the opposite 
85 Uhlenhuth and Mulzer. Arb. a. d. k. Gsndhtsamte, Vol. 44, 1913. 
86 Noguchi, H. Jour. A. M. A., Yol. 58, 1912, p. 1163. 
87 Zinsser, H., Hopkins, J. G., and McBurney, M. Jour. Exp. Med., Yol, 
23, 1916, pp. 329, 341, 612 
INFECTION AND RESISTANCE 
testis can be successfully inoculated before, during, or after, the 
existence of a testicular lesion on one side, but that reinoculation of 
the same testis which had apparently returned to normal, at periods 
ranging from 6 weeks to one year, was not often successful. There 
is a certain amount of evidence here of a purely local immunity, a 
matter which we will discuss at greater length presently. Perhaps 
the difference between rabbits and the higher animals lies chiefly in 
the fact that syphilis is not generalized in the same sense that it is 
in man and monkeys, and even though inoculations from the organs 
of syphilitic rabbits may often result positively, this might signify 
only that the treponemata have been generally distributed by the 
blood stream and latently lodged in the organs, without, however, 
arousing in the tissues any sort of pathological response. This idea 
would seem to be borne out by the two experiments of Uhlenhuth 
and Mulzer cited above in very young rabbits since in such animals 
true generalization seems to be more common. The study of very 
young rabbits will be continued with this point in view. 
It is apparent from the preceding considerations that resistance 
to syphilis differs from that acquired in many bacterial infections 
chiefly in the fact that it does not persist after the disease is over, 
bnt probably coexists only with the presence of the living incitants 
in the body. In this respect it seems to be similar to the conditions 
prevailing in many protozoan diseases. Thus Schilling88 cites 
a case of Trypanosoma Brucei in a steer, experimentally infected 
by Koch, in which reinoculation was repeatedly negative, this result 
being at first falsely interpreted as immunity; but six years later 
Kleine found that the same steer still harbored the trypanosomes 
in his blood—showing that the resistance to reinocnlation was not, 
in this case, an evidence of true immunity, but rather represented a 
condition of insusceptibility to “superinfection” analogous in every 
respect to that existing in syphilis. Similar conditions have been 
shown to prevail in Texas fever and Schilling believes that they may 
be regarded as also existing to a certain extent in Malaria and Sleep- 
ing Sickness. In both of these conditions complete spontaneous 
sterilization of the body (without medicinal aid) is probably rare, 
possibly does not occur at all, and the apparent immunity to rein- 
fection is, as in syphilis, an evidence of persistence of the disease in a 
latent form. 
In weighing the analogy of syphilis to such protozoan diseases, 
one is inclined to wonder whether syphilis in man may be regarded 
as at all spontaneously curable. So few are the cases left untreated 
and so rarely does the reinfection occur even in the face of specific 
remedies, that it seems to us more than likely that in syphilis, as in 
Sleeping Sickness, a spontaneous “sterilizing” immunity does not 
88 Schilling. In Kolle u. Wassermann Ilandb. d. Path. Mikroorg.f Vol. 17; 
p. 600, THERAPEUTIC IMMUNIZATION 
613 
occur. This is a point, however, regarding which it is impossible 
to gather data. 
In order to distinguish the conditions outlined above tersely from 
the ordinary conception of immunity, Neisser 89 speaks of the altera- 
tions which govern the reactions of the luetic body to freshly intro- 
duced virus as “Anergie;” and “Umstimmung” or “Allergie.” By 
“Anergie” (a term first used by v. Pirquet and subsequently intro- 
duced by Siebert in working out the analogy between pigeon-epitheli- 
oma and syphilis), Neisser designates a condition of inability to 
react by cellular change to contact with the virus. As he uses it, it 
implies a passivity on the part of the invaded tissues (in which “die 
Zellen auf die Spirochaeten schwer oder garnicht reagiren), by 
which there is not necessarily a destruction of the invading trepone- 
mata, and which, therefore, cannot in any sense be interpreted as a 
“Schutz wirkung.” By the “Umstimmung” of Neisser, the “change- 
ment dans la mode de reaction” of Levaditi, is meant the changed 
reaction capacity of the syphilized tissues which determines the char- 
acters of the lesions at various stages of the disease. Thus it is 
obvious that cellular reactions which result in the primary induration 
are quite different from those which produce the tertiary gumma, 
and that the histological changes of the roseola are distinct from 
those of the serpiginous syphilide of the late stages. And since, as 
we shall see, there is no valid reason to assume that the incitant has 
been modified in virulence or vitality, we are forced to believe that 
the reaction capacity of the body cells has been altered. 
Since it is a fact, then, that syphilitic infection so changes the 
body tissues of man and monkeys that, during its course, resistance 
to reinfection is produced, it should be possible to analyze this re- 
sistance into its responsible factors and perhaps utilize the knowledge 
so gained for practical therapeutic purposes. Before we proceed 
to do this, it will be of advantage to review briefly the attempts at1 
active and passive immunization which have been made in animals 
and man. 
Metchnikoff and Roux 90 followed their first animal inoculation 
studies by extensive vaccination experiments. Some of their first 
work along these lines is perhaps marred to some extent by an in- 
sufficient recognition of the resistance which depends upon per- 
sistence of the disease rather than upon true immunity, but a good 
many of their observations are of fundamental importance. It will 
be convenient to classify their work and that of others into experi- 
ments dealing with “active” and those dealing with “passive” im- 
munization. 
®9 Neisser. Baermann u. Halberstaedfer d. med. Woch., 1906, Nos. 1 
and 3. 
90 Metchnikoff and Roux. Ann. de I’Inst. Past., Vol. 17, 1903, p. 809; 
Vol. 18, 1904, pp. 1, 605; Vol. 19, 1905, p. 673; Vol. 20, 1906, p. 785. 614 
INFECTION AND RESISTANCE 
Metchnikoff and Roux first worked with filtered virus and virus 
killed at 51° C. All their attempts with such material were negative 
in that the monkeys treated with it could not be regarded in any 
sense as immunized. In their reports of these experiments they 
remark that they believed this to be due to an absolute loss of power 
to incite reaction on the part of the vaccine-material. We emphasize 
this point here because our own subsequent work inclines us to be- 
lieve, with them, that the production of a reaction is necessary for 
the development of any considerable degree of resistance. 
Keisser carried out a large number of attempts at vaccination in 
which he used extracts of syphilitic primary lesions and of the organs 
of congenitally syphilitic children, killed by the addition of carbolic 
acid. Unfortunately he assumed that his extract contained syphilitic 
antigen because it gave positive reactions by the complement fixation 
technique of Wassermann, an assumption which we of course know 
now to be unfounded as far as any relation to the body substance of 
the treponemata is concerned. This, to our mind, deprives these 
particular experiments entirely of their negative importance. 
In rabbits a large amount of work has been done by Uhlenhuth 
and Mulzer. They injected living material from rabbit lesions in- 
travenously and subcutaneously, without ever observing any evidence 
of protection against subsequent inoculations. 
Of perhaps the greatest importance in connection with active im- 
munization are the attempts made upon human beings by different 
investigators. 
Casagrandi and De Luca 91 tried prophylactic immunization on 
six human beings by injection of filtrates obtained from primary 
lesions. Two of these people later contracted syphilis in the ordinary 
way. 
Possibly the most hopeful results are those obtained by Spitzer 92 
by a method suggested by Kraus. Kraus,93 reasoning from the fact 
that syphilis like hydrophobia was a disease with long incubation 
time, expressed the hope that perhaps the method of Pasteur in 
hydrophobia, that is, active immunization during the period of in- 
cubation, might, in syphilis also, tend to abort the disease. Accord- 
ingly, Spitzer treated 15 cases of early syphilis immediately after 
the appearance of the chancre by subcutaneous injections of emul- 
sions made from human chancre material in dilutions of 1:200 to 
1:20. The cases received from 11 to 20 injections and in seven of 
them the disease was uninfluenced. In the others, however, subse- 
quent symptoms were delayed, and in four, no generalized symptoms 
occurred. In a later communication, Spitzer reported 23 further 
cases similarly treated, 10 of which failed to develop generalized 
91 Casagrandi and de Luca. Gior. ital. de mol. ven., 1905. 
92 Spitzer. Wien. klin. Woch., 1905, 45, and 1906, 38. 
93 Kraus. Wien. klin. Woch.} 1905, No. 41, and 1906, No. 21. THERAPEUTIC IMMUNIZATION 
615 
symptoms and in 9 of these the Wassermann test remained negative. 
One of them, a fact which is of great importance, was spontaneously 
reinfected with syphilis two and one-half years later. These results, 
if accurate in every way, are of the greatest importance, but are 
diametrically opposed to the experience of all other investigators. 
Monkey experiments along the same lines by Neisser gave entirely 
negative results and Brandweiner94 as well as Kreibich95 were 
unable to confirm Spitzer’s results in man. Further comment on 
the Kraus and Spitzer method is valueless without more experi- 
mental data. It is one of the few rays of hope, but so isolated that 
one is forced to skepticism. Metchnikoff and Roux did almost the 
identical thing in an orang-outang and obtained lesions both at the 
point of the original inoculation as well as that at which the subse- 
quent “protective” injection was made. 
The only experiments in which an attempt at vaccination with 
attenuated virus was made with some indication of efficacy, is one of 
Metchnikoff and Roux, the outcome of an accidental laboratory in- 
fection. It appears that a laboratory assistant who had been attend- 
ing to the animals, noticed a small lesion on his lip which did not 
look like a typical syphilitic chancre. In order to allay the patient’s 
fears, however, Metchnikoff and Roux did inoculations from the 
patient to monkeys and these were positive. Nevertheless, Fournier 
after examining the original lesion declared it so unlike the ordinary 
primary sore that he did not advise treatment. No secondaries de- 
veloped in the patient nor in the three chimpanzees inoculated with 
the material. From this occurrence Metchnikoff and Roux con- 
cluded that the patient had probably been infected in handling the 
monkeys, and that the virus had become attenuated by passage 
through these animals. On the basis of this observation they later 
inoculated a willing subject 79 years old with virus carried for five 
generations in lower monkeys. The lesion which developed was very 
slight, consisting only of a local induration, and no generalized symp- 
toms developed. A previous attack of syphilis Metchnikoff believes 
could be reliably excluded in the subject. The experimenters suggest 
that the passage through lower monkeys may attenuate the virus for 
man, this leading to relative immunity to subsequent inoculations, 
and furnish a possible means of protection. Their experiments are 
too few to permit conclusions as yet, but even should they hold good, 
the method would seejn to imply a considerable degree of risk and 
our experience with superinfections and reinfections in syphilis does 
not encourage the hope that a method so drastic would be justified 
when the benefit to the patient is apt to be of such duration only. 
The history of passive immunization is an extensive one and 
hardly worth going into with any degree of detail since many times 
94 Brandweiner. Wien. med. Woch., 1905, No. 45. 
95 Kreibich. Wien. klin. Woch., 1906, No. 8. 616 
INFECTION AND RESISTANCE 
extravagant claims have been made only to be refuted by accurate 
study. There is a great similarity in respect to this between syphilis 
studies and those in tuberculosis and cancer. A brief examination of 
the bibliography in Neisser’s book is sufficient to convince one of the 
many attempts that have been made in this direction and often by 
methods as ludicrous as the claims of success for which they formed 
the basis. The most careful and skillful workers have uniformly 
reported failure. Metchnikoff and Roux treated various monkeys 
with blood from syphilitic patients and used the serum of these ani- 
mals for protective experiments. There are a few instances in which 
mixtures of such serum with syphilitic virus rendered this inactive 
on inoculation. A powder made of this serum was supposed to have 
some protective effect when applied to fresh inoculation spots within 
the first hour after inoculation. However, injection of the serum 
had no effect whatever. Casagrandi and De Luca using serum of a 
dog treated with syphilitic virus, obtained entirely negative results, 
and Finger and Landsteiner 96 report negative results with monkey 
blood in man. The most extensive experiments were again those 
carried out by Neisser and his associates. They were done by the 
treatment of animals with dead and living syphilis virus, organ 
extracts, with the blood of syphilitic man and monkey, and horses, 
sheep, and monkeys were used for the production of “immune” 
serum. In no case was there the slightest protective effect on the 
part of the serum either in vitro or in vivo, and the results of the 
experiment were unqualifiedly negative. 
When the method of complement fixation was successfully ap- 
plied to the diagnosis of syphilis first by Detre, and then by Wasser- 
mann, Neisser and Bruck,97 it was generally assumed that this 
reaction incidentally demonstrated that specific antibodies were 
formed in syphilis. As matters have developed, however, this point 
of view can no longer be maintained. It was soon discovered that 
the antigen used for these reactions by Wassermann and his associates 
derived its “fixing” constituent not from the body substances of the 
treponemata contained in the syphilitic organs, but from certain 
tissue extractives chiefly of lipoidal nature which could be obtained 
readily from normal as well as syphilitic tissues. Although Bruck 98 
and others who have occupied themselves with the theoretical basis 
of the Wassermann reaction, still maintain that a specific antibody 
may be incidentally involved, they admit that this is the less im- 
portant factor in the reaction which depends chiefly upon the ex- 
istence of lipotropic substances which appear in the course of the 
96 Finger and Landsteiner. Centralbl. f. Bakt., Ref. Yol. 38, 1906. 
97 Wassermann, A., Neisser, A., and Bruck, C. Deutsch. med. Woch., 
Yol. 32, 1906, p. 755. 
98 Bruck. Imm. bei Syph. in Kolle u. Wassermann Handb. d. Pathog_ 
Mikroorg., Yol. 7, p. 1045 et seq. THERAPEUTIC IMMUNIZATION 
617 
disease as metabolic products either of the body or possibly of the 
treponemata. This is a subject which we will deal with in extenso in 
another paper. It is sufficient for our present purposes to point 
out that, although to a slight extent specific antibodies may play a 
part in the Wassermann reaction, this is certainly not the chief 
element or even a very important factor involved. The truth of this 
conception has been further confirmed by the work of Noguchi," 
of Craig and Nichols,100 of Kolmer,101 and ourselves, in which it 
has been found that antigens made with pure cultures of treponemata 
produced a complement fixation with syphilitic sera to a very limited 
extent only, and in our work in which we have been able to duplicate 
the Wassermann reaction in a large series of antigens made from the 
treponema cultures, we have found that similar results could be 
obtained with cultures of colon and typhoid bacilli identically pre- 
pared, the fixing power to a large degree depending upon the lipoidal 
constituents of the bacteria. Whatever the ultimate explanation of 
the Wassermann reaction may turn out to be (and this is a subject 
which it would not be profitable to discuss at length at this place), it 
cannot be maintained that as it exists at present, it can be interpreted 
as demonstrating the existence of true circulating antibodies, analo- 
gous to those found in bacterial diseases in the blood of syphilitic 
patients. 
Early in the history of such investigations, Fornet,102 with his 
collaborators, Schereschewsky, Eisenzimmer, and Rosenfeld,103 found 
that when the sera of syphilitics were mixed with clear extracts of 
syphilitic livers similar to those used in the Wassermann reactions, 
precipitates formed which were not seen in similar experiments done 
with normal sera. This Fornet104 in a number of communications 
interpreted as showing the formation of precipitins in the course of 
syphilis. At almost the same time, L. Michaelis 105 made similar 
observations giving them the same interpretations. The experiments 
of Fornet have not found universal confirmation, but even if such 
precipitin reactions can be occasionally observed, they do not indicate 
true precipitins for the same reasons that the Wassermann reaction 
does not demonstrate true complement fixation of antibodies. The 
antigens did not represent strong treponema antigens and Jacobs- 
thal106 and others have showm that with the ordinary Wassermann 
99 Noguchi, H. JourSExp. Med., Yol. 16, 1911, p. 99. 
100 Craig, C. F., and Nichols, H. J. Jour. Exp. Med., Yol. 16, 1912, p. 336. 
101 Kolmer, J. A. Jour. Exp. Med., Vol. 18, 1913, p. 18. 
102 Fornet. XIV Internat. Kongress f. Hyg. und Berm., Sept. 1, 1907. 
103 Fornet, Schereschewsky, Eisenzimmer and Rosenfeld. Deutsch. med. 
Woch., 1901, No. 41. 
104 Fornet, Schereschewsky. Munch, med. Woch., 1907, No. 30; Berl. 
klin. Woch., 1908, p. 85. 
105 Michaelis. Berl. klin. Woch., 1907, No. 46. 
106 Jacobsthal. Munch, med. Woch., Yol. 57, 1910, p. 214. 618 
INFECTION AND RESISTANCE 
antigens syphilitic sera can produce precipitations visible under the 
dark field, the basis of the complement fixation being, therefore, one 
of probable colloidal precipitation. The formation of true precipi- 
tins, therefore, has not been shown for syphilis. 
Investigations of the effect produced upon virulent treponemata 
by the sera of syphilitic individuals have likewise been unsatisfac- 
tory. Hoffmann and Prowazek 107 reported in 1906 that the serum 
of syphilitics in the later stages produced immobilization of the 
virulent treponema pallidum, an observation which was confirmed 
by Zabolotny.108 Landsteiner and Mucha 100 were not able to observe 
such immobilization, nor did they see any evidence of agglutination 
in such experiments. In a limited observation of our own, we have 
also failed to see any regular or distinct influences of this kind. It is 
not impossible that a slight difference may exist in this respect 
between syphilitic and normal sera, but even if it occurs, the action 
is feeble, irregular, and entirely insufficient to be interpreted as 
having much practical importance in the acquisition of syphilis 
immunity. 
Treponemacidal experiments have been done by some workers 
also with negative results. As we have stated in another part of 
this paper, the attempt to treat syphilitic animals and man with the 
serum of syphilitics has led to no reliable results, and no evidence 
whatever that can be accepted has been adduced to show that virulent 
treponemata may be killed by active luetic serum. 
As far as any opsonic action is concerned, Levaditi 110 in histo- 
logical studies in congenitally syphilitic children has seen phagocy- 
tosis or at least intracellular localization of treponemata in the alveo- 
lar cells of the lung and in the parenchyma cells of the liver and 
kidneys. For reasons not entirely clear to us, he interprets the 
former as true phagocytosis and the latter as a penetration of the 
treponemata into the liver cells to the detriment of the latter. We 
ourselves have occasionally seen phagocytosis in sections of syphilitic 
rabbit testes, but in all cases the process was not a very active one 
and not much can be said about its importance at present. As a 
matter of fact, Hopkins and the writer,111 in studying the mechanism 
of the natural resistance of mice against syphilis, injected virulent 
organisms into the peritoneal cavities and observed the treponemata 
107Hoffmann, E., and Prowazek, S. Centralbl. f. Bakt., I. 0., Vol. 41, 
1906, pp. 741, 817. (Balanitis and montu spirocliaetae.) 
108 Zabolotny, I)., and Maslakowitz. Centralbl. f. Bakt., I. 0., Yol. 44, p. 
532. 
109 Landsteiner and Mucha. Wien. klin. Woch., Vol. 19, 1906, p. 1349; 
Centralbl. f. Bakt., Vol. 39, 1907, I. R., p. 540. 
110 Levaditi. Compt. rend. Soc. de biol., Vol. 60, 1906, p. 134; Ann. de 
VInst. Past., Vol. 20, 1906, p. 41. 
111 Zinsser, H., and Hopkins, J. G. Jour. A. M. A., Vol. 62, 1914, p. 1802; 
Jour. Exp. Med., Vol. 21, 1915, p. 576. THERAPEUTIC IMMUNIZATION 
619 
in peritoneal puncture fluid, alive, actively motile, and unphagocyted, 
though surrounded by masses of leukocytes, as long as three days 
after their injection. It seemed almost as though the natural im- 
munity of such animals might be similar to the “atreptic” cancer 
resistance spoken of by Ehrlich. The treponemata did not multiply 
in the mice but though, naturally, diminishing in number, were ap- 
parently neither killed nor even inhibited in motility by the perito- 
neal exudates and, for several days, swam in and out among the 
accumulating leukocytes, often adhering to them peripherally but 
not taken by them. Lack of entirely satisfactory methods of 
staining cells containing treponemata make it difficult to speak 
with certainty of the actual occurrences. But we gained the distinct 
impression that the treponemata were not actively injured or de- 
stroyed until they had spontaneously died out owing to lack of suit- 
able environment, i. e., nutrition. 
In the case of natural immunity at any rate, we do not think that 
phagocytosis by the mobile leukocytes plays a primarily important 
role. However, these experiments will need further elaboration. 
The search for antibodies gained new vigor when the efforts of 
Sehereschewsky,112 Miihlens,113 Hoffmann,114 and especially Nogu- 
chi, had resulted in successful cultivation of treponemata from syphil- 
itic lesions in man and animals. It was hoped that with the causative 
organisms isolated, immunization and a clear understanding of the 
antibodies in syphilis might yield practical results. Kolmer 115 ob- 
served that cultivated treponemata were agglutinated in the sera of 
rabbits treated with culture material, and such agglutinins were pro- 
duced in high potency by the writer with Hopkins. Subsequently 
in our own laboratory with Hopkins and McBurney, extensive ex- 
perimentation on the production of antibodies with culture pallida 
was carried out. 
We were able to show that not only were agglutinins formed by 
the immunization of rabbits with such cultures, but also that trep- 
onemacidal antibodies which were analogous to the ordinary bac- 
tericidal substances in such sera were present. Also, it was shown 
by cross agglutination and absorption experiments, that treponemata 
cultivated from various sources were related in group reactions. 
Indeed, experiments done with cultivated treponemata (much 
facilitated by the discovery of a simple method of obtaining mass 
cultures of old strains) seemed at first very encouraging in that 
animals immunized with the cultures responded by powerful anti- 
112 Sehereschewsky, J. Deutsch. med. Woch., 1911, Yol. 37, Nos. 20 and 
39. 
113 Miihlens. Treponema Pallickim in v. Prowazek Handbuch der Pathog. 
Protozoen, i, Barth, Leipzig, 1912; Klin. Jahrb., Vol. 23, 1910, p. 339. 
114 Hoffmann. Berl. Jclin. Woch., 1905, No. 46. 
115 Kolmer, J. A., Williams, W. W., and Laughbaugh, E. E. Jour. Med. 
Res. Yol. 28, 1913, p. 345. 620 
INFECTION AND RESISTANCE 
body formation. It was perfectly justified hope, therefore, that 
antibodies produced with these “attenuated” or rather “avirulent” 
strains might have some action on the virulent treponemata in luetic 
lesions. However, subsequent work in this direction disappointed 
such expectations. We may briefly review this work as follows: 
The serum of rabbits immunized with “culture” pallida although 
potent against “culture” pallida, had no effect either in agglutinating 
the virulent organism from rabbit lesions, nor did it exert any pro- 
tective influence when the virulent organisms were subjected to its 
action before injection. 
Conversely, the serum of syphilitic rabbits showed but a very 
slightly increased agglutinating power for the “culture” pallida. 
This increase of potency in a few experiments was definite but very 
slight, a few of the syphilitic animals agglutinating as highly as 1:25 
and 1:50, whereas most of the normal rabbits agglutinated in 1:10 
and some of them 1:25. 
Although Kissmeyer has recently reported that diagnostic use 
might be made of the fact that sera of syphilitic individuals ag- 
glutinated the “culture” pallida, we have tried this with a consider- 
able number of cases and found that although the sera of tertiary 
syphilitics will sometimes agglutinate a little more highly than will 
the sera of normal individuals, yet many patients suffering from 
non-syphilitic diseases agglutinated as highly and almost as regularly 
as did the syphilitics. 
We have come to the conclusion, therefore, as far as our work 
has gone, that in the syphilitic human being there is as little ag- 
glutinin formation against the “culture” treponema as there seems 
to be against the virulent organisms. If the slightly greater ag- 
glutinating power found in some of the tertiary syphilitics can be 
considered at all, the reaction is so feeble that it is negligible from 
the points of view either of diagnostic value or protective importance. 
Furthermore, vaccination either intravenously or locally into the 
testis with cultures has thus far failed to protect rabbits against 
subsequent inoculation with virulent material, and passive immuniza- 
tion with sera produced with “culture” pallida has been without 
effect. 
From all this it appears that the “culture” treponema has im- 
munologically no relation to the virulent organism. It has lost its 
virulence completely, as six and more successive inoculations into 
rabbit testes have sufficiently demonstrated to us. 
The sera produced by many injections of dead and living culture 
organisms have no effect whatever on the virulent organisms in vitro, 
and vaccination with it does not protect against subsequent infection 
with living treponemata. 
The luetin reaction is the only method by which the relationship 
between the two is demonstrable at all, and there, too, we have to THERAPEUTIC IMMUNIZATION 
621 
reckon with what is generally spoken of as the non-specific increased 
sensitiveness of the syphilitic skin. 
Were it not for the production of lesions with cultures in their 
early test tube generations by Hoffmann 116 and by Noguchi in a few 
experiments, one would be almost in doubt as to the identity of the 
virulent with the culture organisms. 
The reactions in syphilis between the invading micro-organism 
and the invaded subject thus differ in certain fundamental premises 
from those prevailing in diseases caused by most bacteria. The 
treponema pallidum is an organism which, unlike many bacteria, is 
rarely subjected to the necessity of adapting itself to extra-corporeal 
existence during the interval between its passage from one host to 
another. It practically always infects directly, being inoculated from 
one human being to the next and has in consequence developed a very 
delicate parasitism not unlike that seen in certain trypanosome 
diseases of rats and that which we ourselves have observed in the 
well-known spirochete infections of white mice. A considerable 
percentage of laboratory white mice have been found to harbor 
actively motile spirochetes, often in considerable numbers in the 
blood and peritoneal fluid without there being any objective signs of 
illness in the animals. It is an instance of what has been spoken of 
by Bail as “infection without disease,” and approaches what biolo- 
gists speak of as symbiosis, except that the host in this case does not 
benefit in any way by the invasion. Even in the case of the mice, 
a certain amount of gradual injury, perhaps only metabolic, by the 
slow removal of nutritive material, is taking place. In syphilis the 
mutual adaptation may perhaps be less complete, a sufficient accumu- 
lation of the invaders and especially a mechanical injury of tissue 
cells, of closing of tissue spaces together with a certain amount of 
toxic action, leading eventually to pathological changes. 
The virulent treponemata apparently do not arouse true antibody 
formation in any marked degree. When they have been cultivated 
and have become accustomed to the test tube conditions they entirely 
lose their virulence, are easily attacked by the active constituents of 
animal serum, and are probably amenable to phagocytosis. When 
such cultures, living or dead, are injected into animals they act like 
other specific protein antigens and incite the formation of antibodies. 
However, these antibodies have no effect whatever upon the virulent 
■organisms. 
In consequence, it cannot astonish us that all efforts at passive 
immunization with the sera of syphilitic man and animals, or with 
those of animals systematically treated with dead virulent materials, 
have been unqualified failures. 
However, this does not preclude the theoretical possibility of 
immunization or vaccination with such materials since, in this 
116 Hoffmann, W. H. Deutsch. med. Woch., Yol. 37, 1911, p. 1546. 622 
INFECTION AND RESISTANCE 
case, the antigen distributed from the points of injection might act 
upon tissue cells throughout the body. However, with the exception 
of the unconfirmed reports of Spitzer, all attempts to vaccinate either 
with dead virulent material, or with living and dead culture mate- 
rial, have been disappointing. The few experiments of Metch- 
nikoff with virus “attenuated” by passage through monkeys have 
indeed seemed to indicate some possibility of approaching the subject 
from this direction, but these isolated observations have been very 
logically criticized by Heisser and should not bear too much weight. 
The observations were made on two cases only, both of them well 
along in life, and the validity of the important conclusions drawn 
rests entirely on the always problematical fulcrum of complete ex- 
clusion of previous syphilitic infection in the two subjects. More- 
over, attempts in this direction would be fraught with a considerable 
amount of danger and it is therefore questionable whether experi- 
mentation along these lines is sufficiently promising to be justified. 
There is certainly no attenuation for man by passage through rabbits 
as has been sufficiently proven by a number of accidental infections, 
an instance of which in a laboratory has been reported by 
Graetz and Delbanco.117 
We may state, therefore, as safely summarizing our knowledge 
of the conditions in syphilis, that the resistance which undoubtedly 
develops during the course of the disease is one which depends upon 
reaction to the living virus only, cannot so far be produced in animals 
by systemic treatment with dead treponemata, and does not express 
itself in the formation of significant amounts of circulating anti- 
bodies analogous to those observed in bacterial diseases. Moreover, 
it is a well-known fact that the treponemata can continue to do injury 
to many organs and tissues at a time when reinfection by the paths 
of skin and mucous membranes is no longer possible. 
How, then, are we to explain this peculiar state of affairs ? A 
clue to the problem we think is found in the 20 rabbits which Hop- 
kins, McBurney and the writer 118 inoculated into the testes after 
apparent recovery from a previous lesion. Ordinarily in rabbits, 
as we have stated before, no generalized resistance is developed dur- 
ing the disease, and the opposite testis can be successfully inoculated 
before, during, and after the existence of a lesion on the other side. 
In these rabbits it was found that testes that had apparently recov- 
ered from a previous lesion, were not subsequently as easily infected 
as were normal testes. It has seemed to us from this as well as from 
a careful study of the observations of other investigators, that re- 
sistance in syphilis was probably a matter of localized reaction. 
Tissues which have sustained active invasion with the living virus 
117 Graetz and Delbanco. Med. klin., 1914, pp. 375 and 420. 
118 Zinsser, H., Hopkins, J. G., and Gilbert, R. Jour. Exp. Med., Yol. 21, 
1915, p. 213. THERAPEUTIC IMMUNIZATION 
623 
react and gain thereby a certain degree of resistance which expresses 
itself in a failure to react to subsequent inoculation. This would 
explain why in syphilis of the human being reinoculation is unsuc- 
cessful and reinfection of the skin and mucosse does not occur spon- 
taneously at a period later than the early secondary stages when the 
virus has become systematically distributed. It would furthermore 
explain why in this disease organ after organ may be pathologically 
involved when skin infection is no longer possible. This point of 
view is entirely in harmony, though perhaps from a slightly different 
aspect, with the skin immunity suggested by Kraus and Volk, where 
the resistance is attributed to the tissue as a whole rather than to a 
local cell group. The cells which have once reacted to the living 
virus no longer respond, i. e., can no longer be injured for an in- 
definite period after recovery. However, the factors which lend 
them this resistance, whatever they may be, are not distributed to 
the blood stream in a way analogous to that in which antibodies are 
mobilized in bacterial diseases, and the effect of the resistance of the 
local area is not distributed to remote parts of the body. It is of 
course likely that a certain amount of phagocytosis of treponemata 
by the now resistant fixed cells may account for the absence of local 
injury. This, however, we have not yet been able to prove suffi- 
ciently, and further studies on this point are necessary. As far as 
any positive evidence can be adduced at the present time, the newly 
entering treponemata may not be entirely destroyed. It may well 
be that the tissues do not react and are in the state which Heisser 
calls “Anergie.” The treponemata that enter during this period 
may nevertheless remain uninjured and be as capable of subsequently 
causing lesions in other locations as are those already present in thr 
patient. This phase is being studied by comparative histological 
observation on the fate of treponemata which have been injected into 
normal tissues and into tissues rendered resistant bv previous in- 
fection. 
In animals like monkeys and man where generalization is rapid 
and apparently complete, the xesistance becomes a general one. In 
animals like the rabbit in which the lesion—or in other words 
pathological response—occurs in a few organs only, the resistance is 
limited to the particular organ or organs that have previously de- 
veloped a lesion. 
It must not be forgotten that such a resistance probably persists 
for a limited time only, and does not imply the sterilization of the 
body and the complete destruction of the micro-organisms. These 
may, and probably do, remain alive and potent in various parts of the 
body, capable of again setting up new lesions in parts hitherto unin- 
volved or again susceptible after a diminution of their local, acquired 
resistance. That the virulent treponemata may remain thus latent, 
alive and virulent has been sufficiently shown in animal experimen- 624 
INFECTION AND RESISTANCE 
tation by successful inoculation with tertiary lesions and by the 
frequent late accidents, especially of the nervous system, in indi- 
viduals apparently cured or for a long time without symptoms. 
It would seem when we analyze the conditions in syphilis, that 
complete sterilizing immunity or, in other wTords, complete cure, 
occurs but rarely without specific medicinal aid, and that the un- 
treated syphilitic (if such an unfortunate individual exists in a 
civilized country) might go on to apparent cure, in that a general 
syphilization of his body would bring about a general resistance, 
but would always harbor virulent treponemata which could cause 
recrudescences in parts in which resistance was diminished, and 
eventually kill by degenerative processes in the central nervous sys- 
tem where many injuries cannot be compensated for as is possible in 
other organs. 
The resistance which develops is apparently a new attribute only 
of the cell groups which have undergone direct reaction with the 
treponemata. This resistance may consist merely in the complete 
failure of the tissue cells to react to the virus, a sort of “tissue indif- 
ference” or “Anergie.” It may be, however, and probably is, ac- 
companied by a certain amount of active defense in the form of local 
phagocytosis of the treponemata by the fixed tissue cells. CHAPTER XXII 
NON-SPECIFIC EFFECTS OF PROTEIN AND PROTEIN 
DERIVATIVES. SERUM ENZYMES. LEUCOCYTIC 
ENZYMES. 
THE INFLUENCE OF INJECTIONS OF NON-SPECIFIC SUB- 
STANCES UPON INFECTIOUS DISEASES 
The therapy of infectious diseases has been very logically domi- 
nated in the past by attempts to increase resistance, either passively, 
by the injection of specific anti-sera, or actively, by treatment with 
bacterial antigens or their derivatives. The idea of specificity, in 
other words, has dominated such attempts almost exclusively. In 
spite of this, however, there has gradually grown in the minds of bac- 
teriologists an impression that not all the effects of the injection of 
bacterial protein were purely specific. Jobling, who has recently 
summarized most of this work, calls attention to the fact that 
Matthes 1 as early as 1895 showed that the effects of tuberculin injec- 
tions could be obtained equally well with deuteroproteose. From that 
time on, many observations have been made which show that profound 
physiological effects can often be produced in human beings and 
animals suffering from infectious diseases, if they are treated not 
with specific antigen but with proteins and protein derivatives of 
many sources. 
The work that the writer did with Hiss on the injection of 
leucocytic extracts is a case in point. Similar in significance, prob- 
ably, are the results obtained by the injection of substances such as 
the filtrate of bacterial culturev commercially sold as phylacogens, 
the intramuscular milk injection practiced by Schmidt 2 and the 
injection of various ferments, reported by a number of writers 
throughout the literature of the past ten or fifteen years. Into the 
same category fall the favorable reports supposed to have been ob- 
tained in syphilitics with tuberculin reported by Blach.3 Perhaps 
the clearest example of this principle has been obtained in connection 
with attempts at vaccine therapy in typhoid fever. Attempts to 
cure typhoid fever by the injection of typhoid bacilli date back to 
Fraenkel, who treated it in this way as early as 1893. Since then, 
1 Matthes, M. Deutsch. Arch. f. klin. Med., Yol. 45, 1895, 
2 Schmidt, R. Med. klin., 1910, No. 43. 
3 Plach, M. Wien. klin. Woch., 1915, No. 49, 
625 626 
INFECTION AND RESISTANCE 
extensive studies have been made on this subject by Petruschky, 
Netter, and others. Ichikawa in 1912 and Boinet in 1914 obtained 
astonishing results by the use of sensitized vaccine, the former being 
the first to introduce the intravenous method of injection. The re- 
sults of intravenous injection into a patient suffering from typhoid 
fever consisted in rapid falling of temperature, often followed 
by a chill, with occasionally rapid general improvement of the 
patient. 
Gay, too, has made similar such observations by the methods 
of Ichikawa, and most of these writers were inclined to believe that 
the treatment was effective by some sort of specific reaction. Doubt 
has been cast upon this point of view, however, by observations such 
as those of Kraus. Ichikawa alone had some doubt of this, as is 
indicated by the fact that he injected typhoid bacilli into some of his 
paratyphoid patients with like results. Kraus subsequently ob- 
tained similar effects by injecting colon bacilli into typhoid patients 
and used non-specific bacterial virus with good results on cases of 
pyocyaneus infection and upon a single streptococcus puerperal septi- 
cemia. Leudke injected typhoid patients with non-bacterial proteose 
and Jobling and Petersen obtained very severe reactions in typhoid 
patients by injecting them intravenously with quantities of 1 to 2 c. c. 
of 2 per cent, solution of proteoses. 
There seems to be little doubt of the fact that the injection of 
bacterial proteins intravenously produces a profound therapeutic ef- 
fect without being in any but a minor degree specific. Boinet is still 
an adherent of specific reaction, believing that his typhoid patient was 
benefited by a rapid mobilization of antibodies and Gay had the idea 
that a specific hyperleucocytosis was the basis of improvement. This 
last phenomenon has been discussed in another place. We, ourselves, 
are inclined to believe that the specific reactions have little to do 
with the improvement in such cases and that the results are due tc 
the injection of foreign protein and perhaps proteoses; the specific 
nature of the sources of these have little significance. 
The entire problem of non-specific resistance and what is spoken 
of as protein therapy has recently been made the subject of a mono- 
graph by William F. Petersen.4 Among the non-specific agents that 
have been used for various purposes in medicine, in addition to bac- 
terial substances, Petersen lists normal serum, various proteins, egi: 
albumen, milk, gelatin; protein-split products, proteoses, globin, pep- 
ton; enzymes, and cell such as trypsin, leucocyte extracts, 
autolysates of tumor tissue, colloidal metals, gold, silver, magnesium, 
platinum, mercury, zinc; hypertonic and hypotonic salt solution, 
sugar solution, distilled water, etc. 
Whatever our opinion of the wisdom or therapeutic possibilities 
4 Petersen. “Protein Thei’apy and Non-Specific Resistance,” Macmillan 
Co., New York, 1922. SERUM ENZYMES 
627 
of the administration of non-specific therapy of this kind may be, 
the fact remains that many of the agents used exert a very profound 
physiological action upon the body. The symptoms following injec- 
tions will vary, both in degree and kind, with the nature of the mate- 
rial used, and the quantities injected. For details, we must refer 
the reader to Petersen’s monograph. In most cases, however, we may 
state that the symptoms include a chill which may set in as early as 
15 minutes after the injection of such materials as bacterial extracts 
or proteose solutions, but may be delayed for several hours. The 
same has been noticed after intramuscular milk injections. The 
temperature curve varies tremendously, but will usually rise, reach- 
ing its maximum within 4 or 5 hours. The pulse rapidly increases 
and the leucocytes go up. After the subsidence of the chill, there is a 
progressive decline of the blood pressure, and at the same time pro- 
fuse sweating may be noticed. Hausea and gastrointestinal symp- 
toms occasionally supervene. There may be great nervous irritability, 
and sometimes enlargement of the spleen. Albuminuria is irregular. 
There may be an increased lymph flow, and, as Jobling and Petersen 
ascertained, there may be a considerable change in the concentration 
of the serum enzymes. According to Petersen, the serum protease in 
patients almost invariably decreases after the shock, but later in- 
creases for a period of from 3 to 4 days. The serum peptidase usually 
increases in the cases that respond with improvement. Lipolytic 
enzymes showed no change, and the diastatic activity was usually 
diminished. 
Many theories have been brought forward to explain the mech- 
anism of the reaction. Attempts to explain it all by a mobilization 
of antibodies and eventually bringing it back into specific channels 
of reasoning, have not been adequate. It is not impossible, of course, 
that the entire process is related to a preliminary intoxication on the 
basis of anaphylatoxin formation, such as that more thoroughly dealt 
with in another chapter. Petersen seems to lay a considerable amount 
of stress upon the detoxicating effects of increased enzyme activity. 
Erepetase, that is, the ferment winch is able to digest partially hydro- 
lyzed proteins and is active at neutral reactions, is partially a de- 
toxicating agent, and he considers a mobilization of this enzyme as 
being of distinctly beneficial significance, and never a factor in the 
production of an intoxication. As a further favorable effect we may 
consider the increased leucocytosis. Petersen also suggests a change 
in the permeability of the cell membranes in the direction of in- 
creasing the resistance of the cells to the toxic effect. The entire 
problem is still clothed in considerable mystery, and entirely too 
complex to be dealt with in this book. Petersen, himself, has ap- 
praised all the evidence on probable mechanism and states that the 
non-specific agents may all of them produce reactions that are funda- 628 
INFECTION AND RESISTANCE 
mentally alike, but that the determining effects in different disease 
processes may vary considerably. 
Various forms of this non-specific treatment have been utilized 
in arthritis, in acute fevers, like typhoid, paratyphoid, etc.; in gonor- 
rhea and its complications; in anthrax, bacillary dysentery, erysipe- 
las, influenza, scarlet fever, septicemias; in asthma, various diseases 
of the nervous system, and in skin diseases. 
Before leaving this subject, it seems to us important to express 
some opinion as to the possibility of the treatment and the judgment 
to be employed in its use. Again quoting Petersen, we may say that 
“protein therapy offers a possibility, perhaps the most powerful 
method at our command, for the alteration of the current of cellular 
activity either in the direction of acceleration of function, or depres- 
sion of function.” Proper dosage is, therefore, of the greatest im- 
portance since over-dosage with a toxic agent may lead to a depres- 
sion of function and great injury. His idea is that in order to in- 
telligently apply the method in infectious diseases it is necessary to 
thoroughly realize that non-specific therapy is merely a method of 
stimulation of the cellular and humoral forces of the body for short 
periods, and is, therefore, useless if overdone or applied when the 
body is in a state of intense depression. 
In the choice of substances for use, he states that for intravenous 
injection, protein-split products are more satisfactory than vaccines; 
if relatively mild reactions are desired, the various serums are use- 
ful ; and where moderate general reactions are desired, intramuscular 
injections of boiled market milk are to be considered. 
Even with the greatest precautions as to dosage and choice of 
material, there are definite contra-indications of which we know at 
the present time, and there may be a great many contra-indications 
that have not yet been revealed. The most serious contra-indication 
is the intense physical depression of a patient already severely ill 
from any disease. Furthermore, the history of serum sickness, 
asthma, urticaria, or other signs of hypersensitiveness should be re- 
garded as contra-indications, as well as severe nervous diseases. 
Alcoholism, Petersen states, is an absolute contra-indication, preg- 
nancy as well. In severe cardiac conditions, the greatest care should 
be observed. Diabetes is considered a contra-indication. In infec- 
tious diseases, like typhoid fever, the older cases towards the latter 
part of the normal clinical course, and the very septic cases should 
be excluded from the treatment. 
Finally, we would like to state that while including this problem 
of non-specific therapy in this book because of its obvious bearing 
upon general problems of resistance, we do not think that the treat- 
ment, at the present time, has reached the stage at which it can be 
considered as anything but experimental, and should not be attempted 
by physicians who have not very thoroughly familiarized themselves SERUM ENZYMES 
629 
with clinical records, and who do not approach it verv distinctly with 
the precautionary accuracy of a worker conscious that he is carrying 
out an experimental measure. 
Serum and Leucocytic Enzymes.—In our section upon the nature 
of the precipitins, especially in the discussion of Gengou’s conception 
of “albuminolysins,” we have called attention to the probable signifi- 
cance of protein antibodies as a mechanism for the disposal of such 
foreign substances. In the bodies of the higher animals in which a 
special alimentary system, with its many digestive ferments, is well 
developed, it is most probable that the normal condition of digestion 
is one in which the foreign substances utilized for nutrition are com- 
pletely split into their simpler components before they gain entrance 
to the circulation. Nevertheless, abnormal conditions or accidents, 
such as gastro-enteric diseases, digestive disturbances, and bacterial 
infections, may lead to a condition, probably frequent enough in ordi- 
nary life, during which such foreign substances may get into the 
blood stream without previous cleavage. The problem is to determine 
where and how such substances, protein or otherwise, are broken up 
so that they may be either assimilated or eliminated. We have re- 
ferred in another place to the fact that foreign proteins may occa- 
sionally pass through the kidneys and be eliminated unchanged. This 
has been shown actually to occur by Oppenheimer, Ascoli, and others, 
but probably represents a very unusual state of affairs produced by 
special experimental conditions. As a rule these substances are dis- 
posed of within the body by chemical cleavage or by assimilation. 
It is more than likely, therefore, as has been emphasized by such 
workers as Jacoby, Salkowski, Wells, and others, that every fluid 
and cell in the body contains enzymes which, by their action upon 
proteins, fats, and carbohydrates, play an important part in the 
metabolism of the body. It has been known for a long time that the 
organs of animals removed from the body will undergo self-digestion 
by a process spoken of as autolysis. The action of bacteria as caus- 
ing such autolysis can be excluded by covering organ emulsions with 
toluol, chloroform, and some other substances which have the peculiar 
property of preventing the growth of bacteria without in any way 
interfering with the action of enzymes. Thus, any organ of the body 
so treated will show rapid changes consisting in the splitting of its 
own proteins, fats, and carbohydrates. That such processes also go 
on in the living body is probable, although, as we have stated in an 
earlier chapter, it is a well-known fact that living cells oppose a more 
or less mysterious resistance to enzymic digestion, just as they oppose 
a similar resistance to bacterial invasion. The phenomenon is more 
fully discussed by Wells in his “Chemical Pathology,” to which the 
reader is referred. It is not improbable that this resistance to in- 
vasion and enzyme attacks, spoken of vaguely as “vital resistance,” 
can be analyzed into more exact factors, one of which seems to be 630 
INFECTION AND RESISTANCE 
the question of reaction, digestion depending to some extent upon 
the reduction of alkalinity by the formation of acid in the tissue cells. 
However, the physiological importance of tissue enzymes is a 
subject which we cannot go into in this connection, since the rela- 
tionship of these enzymes to normal body metabolism is a subject 
more fully dealt with in text-books of physiology and is too far 
removed from the subject of our present discussion to warrant ex- 
tensive analysis. It must, of course, be self-evident that the existence 
of cellular ferments of this nature must play an important role 
wherever and whenever cell death occurs, and since such cell death 
is an accompaniment of many phases of bacterial infection it may 
well be that the enzymes which destroy dead cells, whether they be 
bacterial or those of the body itself, contribute to the general patho- 
logical picture characterizing such diseases. Moreover, the impor- 
tance of the enzymes is particularly enhanced by the knowledge we 
have gained from the work of Vaughan and others, who have shown 
that in the course of proteolytic cleavage toxic substances are lib- 
erated. It is such proteolysis which is the probable basis of the 
formation of the “anaphylatoxin” of Friedberger, the “proteotoxin” 
about which we ourselves have written, and the “sereotoxin” of 
Jobling and Petersen; and similar processes are involved in the toxic 
substances, studied by Whipple, which form in the intestine as a 
consequence of ligature of the gut. It is conceivable that wherever 
and whenever a proteolytic ferment reacts in the body with its 
substrate, toxic cleavage products result, perhaps in the form of 
albumoses, peptones, etc., which are rapidly absorbed and cause 
symptoms of varying intensity. In the chapter on anaphylaxis we 
have seen that many writers have attributed the injury occurring in 
this phenomenon to the formation of such split products and there is 
much evidence in favor of such a view, although the rapidity and 
vehemence of anaphylactic shock indicate that probably there are 
purely physical elements also involved which we do not as yet com- 
prehend. 
Of the cells in the body with which enzyme study has been most 
assiduously followed, the most important are the leucocytes. We 
have already had much to say in the chapters on phagocytosis of the 
digestive functions of the white blood cells. However, we dealt with 
them there chiefly from the point of view of their phagocytic and 
antibacterial functions. It will be useful to consider at* a little 
greater length the importance of these cells from the point of view 
of purely enzymotic activity.5 
As early as 1885 Hammarsten called attention to the fact that 
leucocytes aid in dissolving fibrin. Leber 6 subsequently studied pus 
5 See Wiens. Ergebnisse d. Allg. Pathol. Lubarsch & Ostertag, Vol. 15, 
1911. 
6 Leber. Entstehung der Entzundung. Leipzig, 1891. SERUM ENZYMES 
631 
and found that it possessed a powerful digestive action for gelatin, 
fibrin, and other protein substances. He correlated his studies par- 
ticularly with the processes of inflammation, and his work, as well 
as that of later investigators, brought into the foreground the fact 
that in the resolution of abscesses, especially of the staphylococcus 
variety, the leucocytes might play an important role in liquefying 
the necrotic tissue cells and bringing about the breaking down of 
the center of the abscess. Any one who has carefully studied the 
histology of a staphylococcus abscess can easily see the halo of disin- 
tegrating tissue lying just inside the ring of aggregated polymorpho- 
nuclear leucocytes. That these cells carry on an important part in 
the liquefaction is certain, although as yet we are not sure whether 
or not the proteolytic bacterial enzymes participate. Friedrich Mul- 
ler,7 in 1902, also studied these processes and attributed clinical 
significance to the proteolytic activities of the white blood cells in the 
resolution of the pneumonic lung. The study of inflammatory exu- 
dates by Opie 8 confirmed much of the preceding work and revealed 
that the ferments of all white blood cells were not the same. He 
found a leucoprotease which was contained in the polynuclear leuco- 
cytes which was active chiefly in a slightly alkaline medium, whereas 
the large mononuclear cells of uncertain origin which appear toward 
the end of an inflammatory process contained a protease which was 
active in weak acid. This differentiation of a leucoprotease and a 
lymphoprotease was more or less confirmatory of assumptions made 
earlier by Metchnikoff. It seemed probable that the acid protease 
noticed by Opie is specific to the large mononuclear cells of such 
exudates and is not common to lymphocytes in general, for the con- 
sensus of opinion of other workers seems to be that the small lympho- 
cytes contain no proteolytic enzyme whatever. The study of tubercu- 
lous exudates particularly has failed to reveal proteolytic properties 
when only the characteristic small lymphocytes were present, al- 
though some writers, Bergell 9 especially, have shown that these cells 
contain a fat-splitting enzyme. Spch lipase could also be determined 
in the press-juice of lymphoid organs, such as the spleen and lymph 
nodes. It is not improbable that the characteristic dry abscess or 
“caseous” abscess accompanying typical tuberculous lesions owes its 
peculiar histological characteristics to the lack of proteolytic 
enzymes. 
Jochmann 10 and his collaborators have extensively studied leuco- 
cytic extracts in this regard and, curiously enough, found that the 
proteolytic enzymes of leucocytes of which we have been speaking 
7 Fr. Muller. Verh. d. 20 Kongress f. innere Mediz., Wiesbaden, 1902. 
8 Opie. Jour, of Exp. Med., Yols. 7 and 8, 1905 and 1906. 
9 Bergell. Munch, med. Woch., 1909. 
10 Jochmann. For summary of his work and literature see Kolle u. Was- 
sermann, 2nd ed., Yol. II, 2. 632 
INFECTION AND RESISTANCE 
can be found only in the cells of man and monkeys and, to a slight 
degree, in dogs. In these species only, according to Pappenheim,11 
do the leucocytes contain true neutrophile granules and it is there- 
fore a possibility that the neutrophile granules and the enzyme action 
are related to each other. The leucocytes of rabbits do not ap- 
parently contain protease, but recently in our laboratory Mrs. Parker 
and Miss Francke have shown that rabbit leucocytes contain erepsin. 
The proteolytic enzymes of leucocytes curiously enough continue 
their activity at temperatures as high as 55° C., a fact which makes 
it possible to investigate their activity at temperatures at which most 
bacteria will no longer grow and functionate, and this incidentally 
facilitates sterile experimentation. Their activity is astonishing in 
that Jochmann found instances where dilution even as high as 500- 
fold with salt solution did not completely eliminate the proteolytic 
activity of pus. The simplest method of demonstrating such action 
is to place pus or washed leucocytes, in droplets, upon the surface of 
plates of Loeffier’s coagulated blood serum, such as that used in the 
cultivation of diphtheria bacilli. On such plates, small indentations 
rapidly give evidence of the liquefaction of the coagulated protein. 
Casein also can be used as an indicator of proteolytic digestion, 
a casein solution being made by dissolving a gram of casein in 100 
c. c. of N/10 NaOH solution and neutralizing this to litmus with 
deci-normal 1IC1. A very curious property of the leucocytic ferment 
has been described by Jochmann and Ziegler, who reported that pres- 
ervation in 10 per cent, formalin solution will long preserve the fer- 
mentative activities of the cells. 
It was formerly supposed that the proteolytic properties of 
leucocytes were more or less pathological. But we have since learned 
that these powers are common to leucocytes both in health and in 
disease. The older idea was due to the fact that the earliest investi- 
gations of the ferments were made in connection with myelogenous 
leukemia. Jochmann and Ziegler studied the organs of patients 
dead of this disease and found that the bone-marrow, spleen, and 
lymph nodes of such cases had powerful proteolytic properties in 
contrast to the very weak activities in this regard of normal organs. 
The proteolytic activity was more or less in direct proportion to the 
degree of myelogenous infiltration. In lymphatic leukemia no such 
activity was discovered. In consequence there has been a tendency 
to draw a fundamental physiological distinction between these types 
of cells, a distinction which would have considerable importance in 
discussions of their origin and significance. 
Jochmann has an idea that the activities of these ferments in 
connection with tissue destruction and other pathological conditions 
where they become active, may be an important phase in the produc- 
tion of fever. 
11 Pappenheim. Cited from Wiens, loc. cit. SERUM ENZYMES 
633 
An important question which immediately arises is whether these 
ferments can be identified with the bactericidal substances described 
in a preceding chapter as existing within leucocytes. According to 
Jochmann the enzyme extracts have no bactericidal properties. In- 
deed, it is a curious fact that living bacteria oppose a very powerful 
resistance to digestion by these and other ferments. Kantorowicz 13 
who has studied this particularly, has shown that living bacteria 
contain a very powerful antiferment which is similar to the anti- 
ferment presently to be described for blood serum. He has shown 
that living bacteria cannot be digested by trypsin, a resistance which 
is lost when the bacteria are heated to 80° C., and an extract of 
living bacteria will prevent the tryptic digestion of the heated bac- 
teria. It will appear, therefore, that in the process of phagocytosis 
the bacteria are first killed by the bactericidal substances contained 
in the cells and are later digested by the leucoprotease. 
That the lymphocytes contain lipase has been mentioned above 
and since these cells are specifically accumulated about tuberculous 
foci it has been many times suggested that their function is particu- 
larly directed against these acid-fast bacteria in whose constitution 
the presence of waxes and fats plays such an important role. It is 
a fascinating thought, though entirely conjectural, that perhaps the 
specific benefit of feeding fats in tuberculosis may have some basis in 
the possible increase of lipolytic ferments which appear in response 
to the stimulation of introducing larger quantities of fats. 
It is hardly necessary to reiterate here the possible importance 
of these cellular enzymes in the many different phases of the ab- 
sorption of larval organs which occurs in the lower animals in the 
course of normal development, or of dead tissue in mammalia in 
disease and in the processes of senescence. 
In addition to proteolytic enzymes, it has been long known that 
the leucocytes also contain an oxidase. This enzyme can be demon- 
strated by the well-known guaiac test, in which tincture of guaiac is 
added to leucocytic exudates or .pus, and a blue color results as a 
consequence of oxidation of the guaiac. The oxidizing ferment is 
apparently limited to the polynuclear leucocytes. Exactly what its 
significance is we do not know but it is highly probable that it plays 
an important part in the metabolism of the leucocyte. 
It has only recently become clearly apparent that the function of 
intracellular digestion of foreign substances, such as invading bac- 
teria, is not limited to the circulating leucocytes but is to a very 
large extent also carried on by the fixed cells of organs. Studies bv 
Wyssakowitz,13 by Kyes,14 by Bartlett,15 and, recently, by Hopkins 
12 Kantorowicz. Miinch. med. Woch., Yol. 56, 1909, p. 897. 
13 Wyssakowitz. Zeitschr. f. Hyg., Vol. 1, p. 1. 
14 Kyes. Jour, of Inf. Dis., Vol. 18, 1916, p. 277. 
15 Bartlett. Jour. Med. Res., Nov. 5, 1916, p. 465. 634 
INFECTION AND RESISTANCE 
and Parker,16 have shown that in the course of many bacterial infec- 
tions the greater proportion of invading bacteria is taken care of 
by cells of the liver, spleen, and especially the lung. Streptococci 
injected into rabbits rapidly disappear from the circulation, largely 
because they are taken up very actively by the (probably) endothelial 
cells of the lung where apparently they are killed, their digestion 
taking place within the cells very slowly. The endocellular enzymes 
bringing this about can, of course, not be separately studied as can 
those of leucocytes, for obvious technical reasons. 
While the cellular enzymes have been very carefully studied, it 
is only of relatively recent years that much attention has been paid 
to the enzymes present in the circulating blood. This is to some ex- 
tent due to the fact that German writers especially have taken for 
granted that enzymes in the blood were nothing more or less than 
liberated leucocytic enzymes, and also because the activity of the 
circulating enzymes has been largely masked by the existence of a 
powerful antiferment which must needs be there for physiological 
reasons to prevent injurious autodigestion. We have already men- 
tioned the fact that the study of pathological conditions, such as 
myelogenous leukemia, was the point of departure for work upon the 
leucocytic ferments, and this, to a certain extent, was also the be- 
ginning of the studies on serum ferments. When blood serum is 
incubated with various substrates, it is not unlikely, as we shall see 
in our subsequent discussion of the Abderhalden reaction, that the 
antiferment is absorbed and thereby the proteolytic and other fer- 
ments of the blood are liberated. There is normally an excess of 
antiferment in the blood, and normal human serum usually does not 
contain strong proteolytic enzyme. These two factors together there- 
fore prevent powerful proteolysis by normal serum. There are con- 
ditions, however, under which the enzyme contents of the blood are 
increased. Thus, the sera of pregnant women have been shown 
to be relatively rich in proteolytic properties, and such an increase 
is also present during starvation, as Schultz,17 and Heilner and 
Poensgen 18 have shown. The same thing has been noted in cases 
of pneumonia, in which condition Palls 19 has shown a sudden di- 
minution of this enzyme at the time of crisis, and the same author 
has made analogous observations in the fluctuation of serum protease 
in malaria, typhoid fever, and some other diseases. 
The lipolytic activity of the serum has been found increased in 
syphilis, in diseases involving liver function, such as chloroform 
poisoning, and in a number of other conditions. In syphilis, indeed, 
16 Hopkins and Parker. In press at present writing. 
17 Schultz. Deutsch. med. Woch., No. 30, 1908; Munch, med. Woch., Yol. 
60, 1913. 
18 Heilner and Poensgen. Munch, med. Woch., Yol. 61, 1914. 
19 Falls. Jour, of Inf. Dis., Yol. 16, p. 466. SERUM ENZYMES 
635 
the increased lipase contents have been held responsible for the Was- 
sermann reaction by a number of writers. They believed that the 
lipolytic ferment acting upon the lipoid antigen produced fatty acid 
which by increasing the H-ion contents rendered the complement 
inactive. This to a certain extent was strengthened by the knowledge 
that in the Wassermann reaction the end-piece of complement re- 
mained free. However, we ourselves in hitherto unpublished experi- 
ments have been able to convince ourselves that there is no relation- 
ship between lipase contents and Wassermann positiveness. 
Activity of intravascular enzymes of any kind may perhaps be 
going on constantly in the normal course of life as a part of general 
body metabolism; yet it is probable that any great extent, especially 
of proteolytic action, would be incompatible with health or even 
life. The work of Vaughan, of Friedberger, and others has shown 
us that in the course of proteolysis toxic cleavage products, probably 
of the albumose variety, are liberated. This we have seen has been 
made the basis of some theories of anaphylaxis, and the “split prod- 
ucts” of Vaughan, the “anaphylatoxin” of Friedberger, the “proteo- 
toxins,” as we ourselves have chosen to call them, and the “sero- 
toxins” of Jobling and Petersen, all are probably the results of such 
proteolysis. Indeed, as we have seen, the incubation of almost any 
tissue substrate with fresh serum will result in this proteolytic 
change and even in the formation of toxic substances. It was at first 
thought that the toxic split products were derived from the substrate, 
but we have seen in another chapter that the mechanism is probably 
one in which antiferments are removed by the substrate, and the 
serum-protease is liberated to act upon the serum protein itself. 
This was undoubtedly the case in such instances as those in which 
anaphylatoxin was obtained by digestion of serum with kaolin, 
barium sulphate, tubercle bacilli, etc. How it appears that the 
injection of almost any substance, and especially of bacterial pro- 
teins, produces a sharp rise of enzymes in the blood. This seems to 
be true not only of proteolytic enzymes but of lipolytic and amylolytic 
ferments as well. Ferments so stimulated are entirely non-specific. 
This greater serum activity after protein injection may be due both 
to a direct increase of production and a simultaneous diminution of 
antiferments, by means of which activity hitherto inhibited is al- 
lowed to run riot. The result may be in some respects beneficial 
and in other respects injurious. The marked increase of the enzymes 
may aid in rapidly disposing of the unchanged foreign substance, 
and perhaps the beneficial effects following the injection of typhoid 
vaccines into patients suffering from this disease may be due to the 
mobilization of ferments as suggested by Jobling and Petersen.20 
On the other hand, the proteolytic activity may result in protein 
cleavage, by means of which albumoses, etc., are liberated in greater 
20 Jobling and Petersen. Jour. A. M. A., Vol. 65, 1915, p. 515. 636 
INFECTION AND RESISTANCE 
or smaller quantities which act injuriously and lead to clinical 
symptoms of various kinds. That fever may perhaps be explained 
by poisoning with albumoses so produced has already been suggested 
by Jochmann, and that albumoses are present in various products 
of suppurative inflammation has also been shown. It has been shown 
by Pfeiffer and Jarisch 21 and others, that fluctuations in proteolytic 
enzymes accompany anaphylaxis. Jobling and Petersen22 have 
formulated a theory of anaphylaxis based upon studies of serum 
protease. They observed that during the course of sensitization 
there occurred a gradual mobilization of non-specific protease which 
increased in intensity up to the time of maximum sensitization. 
They attributed acute shock to the fact that on the second injection 
there occurs an instantaneous mobilization of large amounts of non- 
specific protease together with a decrease in antiferment and an 
increase in non-coagulable nitrogen and amino-acids, and they be- 
lieve that the cause of the acute intoxication is the rapid cleavage of 
the serum protease. The specific element they believe consists in the 
rapid mobilization of the ferments and the colloidal serum changes 
which bring about the change in antiferment titre. Fascinating as 
this theory is, and although it has a number of things in common with 
our own ideas concerning the colloidal balance in the plasma which 
ordinarily prevents the rapid union of antigen with antibody, we feel 
that recent knowledge concerning the essentially cellular nature of 
anaphylaxis in the guinea pig prevents its adoption. Furthermore, 
we believe that the specific element of anaphylaxis is insufficiently 
explained by Jobling and Petersen’s conclusions. 
Again the question arises: Are these serum proteases in any 
way to be identified with bacteriolytic antibodies or with alexin or 
complement? It is more than likely that no such relationship exists. 
In the first place, the seroproteases are non-specific, and it has been 
shown that in the course of bacteriolysis no increase in non-coagulable 
nitrogen occurs. Moreover, Jobling and Petersen,23 to whom we owe 
much of our recent knowledge on this subject, have shown that the 
treatment of bacteria with complement alone or with complement 
together with human serum renders them more resistant to proteoly- 
sis, probably owing to the absorption of antiferments from the blood 
serum. The identification of complement with lipase has been sug- 
gested. The idea gains likelihood from the fact that both complement 
and lipase are destroyed by lipoid solvents and that many of the 
substances upon which complement acts, such as red blood cells, con- 
tain lipoid constituents. On the other hand, no definite knowledge 
is available in this regard, and it has been so far impossible to prove 
even the enzymotic nature of complement, though this, at least, seems 
21 Pfeiffer and Jarisch. Zeitschr. f. 1mm., Yol. 16, 1912. 
22 Jobling and Petersen. Jour. Exp. Med., Yol. 20, 1914. 
23 Jobling and Petersen. Jour. Exp. Med., Yol. 20, 1914, p. 321. SERUM ENZYMES 
637 
likely. The peculiar resistance exhibited by bacteria to digestion by 
ferments lias already been alluded to. 
It develops therefore that the antiferments in the blood are ex- 
tremely important factors in maintaining the balance of enzyme 
activity, and in consequence the antienzymatic properties of serum 
have been investigated by many workers. For, depending upon their 
fluctuation, cleavage processes are permitted or prevented from tak- 
ing place in the circulating blood. The earliest workers concerned 
themselves largely with the phenomenon that blood serum would 
prevent the proteolytic activity of leucocytes. They utilized the ob- 
servation therapeutically by injecting serum into suppurating ab- 
scesses for the purpose of preventing tissue destruction by leuco- 
protease and increasing the ability of the body to limit the processes. 
Indeed it may be that this form of therapy in chosen cases may still 
have possibilities. 
Later investigators studied antiferment fluctuations in disease. 
In septic conditions, Wiens 24 claims to have noticed a diminution 
of serum antiferments. Similar observations have been made after 
shock in anaphylaxis by Pfeiffer and Jarisch. In patients with can- 
cer, Landois 25 claimed to have noticed a marked increase in anti- 
ferments, which for a time was believed to be sufficiently regular 
to have diagnostic significance. Brenner 26 noticed the same phe- 
nomenon in severe anemias. In cachexia the same phenomenon has 
been noticed and fluctuations in this constituent have been noticed 
in a great many different diseases, such as diabetes, pneumonia, etc., 
without our being at present in possession of any definite or corre- 
lated knowledge of the principles upon which these fluctuations 
depend. 
Earlier observers believed that the anti-ferment in the blood was 
a function of the albumin fraction of the serum. More recent 
workers have attributed it to lipoid constituents. Schwarz,27 Sugi- 
moto,28 and Kirchheim suggested the lipoid idea because of the fact 
that the antiferment properties were destroyed by lipoid solvents. 
They did not settle the question satisfactorily because their results 
were inconsistent, but Jobling and Petersen29 found that anti- 
proteolytic activity of blood serum was almost wholly inhibited when 
sera were extracted for several days with chloroform or ether at room 
temperature or with chloroform in the incubator for an hour. The 
antiferment could be recovered from the ether and chloroform solu- 
tion, and its action could be destroyed by incubation with iodin or 
24 Wiens. Munch, med. Woch., No. 53, 1907. 
25 Landois. Berl. klin. Woch., No. 10, 1909. 
26 Brenner. Deutsche med. Woch., No. 9, 1909, and Mediz. Klinik, No. 28, 
1909. 
27 Schwarz. Wien. klin. Woch., 22, 1909. 
28 Sugimoto. Arch. f. Exp. Path. u. Pharmakol., Yol. 74, 1913. 
29 Jobling and Petersen. Jour. Exp. Med., Yol. 19, 1914, p. 480. 638 
INFECTION AND RESISTANCE 
potassium iodid. They believed this to be due to the saturation of 
the free carbon atoms. Furthermore, the same authors showed that 
toxic substances analogous to anaphylatoxins could be produced in 
serum when incubated with chloroform. The lipoidal nature of the 
antiferments is a likely one although it does not account for the fact 
noticed uniformly by all workers that the antitryptic activity can 
be entirely eliminated by heating the serum to 70° C. for a half hour. 
This was the observation which at first gave rise to the idea that the 
antiferment was in itself an enzyme. For this phenomenon of heat 
liability no satisfactory explanation has as yet been advanced. 
The Abderhalden Reaction.—The recent researches of Abder- 
halden 30 upon the intravascular digestion of foreign substances in- 
troduced into animal bodies promised to have considerable bearing 
upon problems of immunity. Abderhalden, whose work we cite 
chiefly from his monograph, “Die Schiitzfermente des tierischen Or- 
ganismus,” took as his point of departure the conception that the ani- 
mal body must necessarily preside over a mechanism whereby it can 
assimilate foreign substances which obtain entrance unchanged into 
the circulation. Abderhalden believes that this process depends upon 
the mobilization of “protective ferments,” a term which he borrows 
from Heilner,31 and suggests the possibility that these ferments may 
originate in the leucocytes. 
Experimentally Abderhalden approaches his problem by deter- 
mining the presence of specific ferments in the blood of animals into 
which various foreign substances have been introduced by paths other 
than the alimentary canal. For this purpose he has developed a 
number of methods, the most important of which are his optical 
method and his dialysis method. The optical method used for the 
determination of the proteolytic properties of the serum depends upon 
the fact that many of the amino-acids are optically active. More- 
over, most of these substances are chemically known and their optical 
activity determined, so that it is possible to take blood serum 
which is to be examined for its contents of particular ferments, mix 
them with a suitable protein, or preferably a polypeptid, and de- 
termine with a polariscope the rotation which takes place. We will 
not go into the technique of this method more extensively because 
we have no personal experience with it, and the method is one of 
such delicacy that it is best obtained directly from Abderhalden’s 
original publications.32 His dialysis methods depend upon placing 
the blood serum and fermentable substance into dialyzing bags, sus- 
80 Abderlialden. “Schiitzfermente des tierischen Organismus,” Springer, 
Berlin, 1912. 
31 Heilner. Cited from Abderlialden Zeitschr. f. Biol., Yol. 50, 1907. 
32 See especially Abderlialden, Hoppe-Seyler, Zeitschr. f. physiol. Chemie, 
Vols. 60, 65, and 66; also “Handbuch der Biochem. Arbeitsmethoden,” Yol. 
5, 1911, p. 575. SERUM ENZYMES 
639 
pending these into distilled water, and determining the presence of 
peptone, amino-acids, or total nitrogen in the liquid outside of the 
bag after definite intervals of time. 
By these and other methods Abderhalden 33 has carried out tests 
with a large number of different substances. Experimenting first 
with proteins, he injected egg albumen, horse serum, silk peptone, 
gelatin, edestin, casein, etc., into dogs and rabbits, then, several days 
later, bled the animals and mixed 0.5 c. c. of the serum with 0.5 c. c. 
of a solution of the respective substances which had been injected. 
He found in such cases that definite proteolytic action was exerted 
upon the injected substances by the active serum of a treated animal, 
whereas, in the case of most of the substances used, the normal serum 
possessed no proteolytic action whatever. These results were con- 
sistently obtained both by the dialysis and by the optical methods. 
It should he especially noted that the ferments studied by Abder- 
halden were not as specific as are the antibodies which we have dis- 
cussed in another place. For Abderhalden found that the serum of 
an animal treated with proteins developed enzymes which were active, 
not only against the particular protein used for injection, but rather 
against proteins in general. They were specific only in that, when 
produced with proteins, they were not active against fats or carbo- 
hydrates. This is especially important in connection with the recent 
discussion concerning the identity of Abderhalden’s protective fer- 
ments and the specific protein antibodies. 
In later experiments Abderhalden claimed to show further that 
similar ferments could be induced in animals by treatment with car- 
bohydrates and with fats. The serum of normal dogs is not capable 
of splitting cane sugar. However, the blood serum or plasma of a 
dog that has been treated with cane sugar develops the property of 
inverting the cane sugar into dextrose apd fructose within fifteen 
minutes after injection. This could easily be determined both by 
putting together the serum with cane sugar and determining the 
increase of reducing powers, and by means of subjecting such active 
plasma or serum, together with saccharose, to polariscope examina- 
tion. 
The earlier experiments with fats were negative because the 
simple method of titration for fatty acids proved insufficient as an 
indicator of activity. However, Abderhalden succeeded in determin- 
ing fat splitting properties in the blood of treated dogs by using the 
method of Michaelis and Bona.34 The presence of fats largely in- 
creases the surface tension of mixtures, and their cleavage in such 
mixtures consequently leads to reduction of this tension. Utilizing 
this principle, Abderhalden claims to have determined that the paren- 
33 Abderhalden. “Schutzfermente,” p. 49. 
34 Michaelis and Rona. Cited from Abderhalden, loc. cit. 640 
INFECTION AND RESISTANCE 
teral introduction of fats into dogs is followed by a reactionary in- 
crease of lipases. 
The general conclusion drawn from Abderhalden’s researches is 
this: When any foreign substances, protein, carbohydrate, or fats, 
gain entrance to the circulation of an animal, the animal body reacts 
by the mobilization of ferments or enzymes specifically capable of re- 
ducing these substances to assimilable form. It is likely that these 
ferments represent a mobilization of substances normally present but 
not concentrated in the blood stream under ordinary conditions, since 
they appear with a speed out of all proportion to that obtaining in 
the case of the antibodies discussed in another place. In one case 
cited by him a dog injected on November 25th, 29th, and December 
4th showed powerful peptolytic serum properties on December 6th. 
Apparently the injection of homologous proteins into animals (i. e., 
rabbit serum into rabbits, etc.) does not incite reaction. 
These enzymes seemed to differ from specific antibodies in that 
they did not react solely with the substance injected, but also with 
other substances belonging to the same chemical group. Other dif- 
ferences from antibodies are the rapid appearance of the ferments 
after treatment and their rapid disappearance after the inciting 
stimulus is removed. Thus Abderlialden reports that the enzymes 
found in a case of pregnancy disappeared within eight days after 
abortion or child birth. 
It is plain that these researches of Abderlialden offer many op- 
portunities for diagnostic utilization, and he has applied them to the 
diagnosis of pregnancy. In this condition substances from the 
chorionic villi get into the blood. These, according to Abderlialden, 
may be looked upon as in a certain sense foreign in nature, and must 
be chemically disintegrated by the body. In consequence it is likely 
that the ferments which accomplish this would appear in the sera of 
pregnant individuals and could be determined by his methods. 
When he prepared peptone from the placental substances of human 
beings and allowed the blood plasma of normal individuals to act 
upon it, observing it both by the dialysis and the optical method, no 
peptolytic action could be observed. However, when the plasma of 
pregnant women was used proteolytic action was determined. In 
these cases the ferment seemed to be specific for peptones produced 
from placental tissue both in animals and human beings, but did not 
act upon casein, gelatin, or other proteins. There are certain techni- 
cal difficulties connected with the production of a test material from 
the placental tissue which render this method difficult. For their 
more detailed description we refer the reader to the original articles. 
The Abderlialden reaction and its various logical ramifications 
promised for some time to bring an entirely new and important factor 
into the field of immunology, namely, that specific ferment mobiliza- 
tion in definite response to the introduction of various substances. SERUM ENZYMES 
641 
and yet distinct from the specific antibody heretofore known. The 
reaction rapidly became the subject of active research mostly of a 
clinical nature, and carried out with insufficient critical judgment. 
This literature is extensive and, because almost entirely refuted at 
the present day, is hardly worth citing. In this country the re- 
action has been carefully studied by a great many workers, more 
especially by Bronfenbrenner, and Jobling, Petersen and Eggstein. 
From all these studies, we may say for the sake of brevity, it has 
become clear that the enzymes involved in the Abderhalden reac- 
tion are probably not specific as supposed by this writer, that during 
the reaction the tissues employed (that is placenta, etc.), act by 
absorbing antienzymes from the serum. In consequence the fer- 
ments in the serum act upon the serum protein itself which there- 
fore becomes the substrate of the cleavage product determined as a 
result of the reaction. This robs the reaction of any claim to the 
specific and diagnostic value attributed to it by Abderhalden and his 
collaborators, and as a final result of these extensive and intricate 
researches we have merely a better understanding of the non-specific 
enzyme activity of normal serum. INDEX 
Abderhalden, protective fer- 
ments of, 638, 639 
Abderhalden’s “building 
stones,” 111, 114 
reaction, 638, 641 
Abortion reaction, 497 
Abscesses, secondary, 
caused by bac- 
teria, 27, 28 
Acne, opsonic index in 
vaccine treatment 
of, 399 
Adrenal cytotoxin, 102 
Adsorption theory in toxin- 
antitoxin reaction, 
139 
Agglutination, absorption 
experiments of 
Castellani on, 254, 
255 
acid agglutination, 269 
action of salts in, 240, 
260, 261 
agglutinability of bac- 
teria in, in agglu- 
tinoids, 251 
normal differences in, 
between strains of 
same species, 250, 
251 
agglutinins in, 245 
absorption of, 248, 254 
heating of, 247 
explanation of, 259 
major, 252, 255 
normal, 102, 255 
explanation of, 256 
qualitative identity 
with “immune” 
agglutinins, 256 
para- or minor, 252, 
255 
agglutinogen, 245 
effects of heating of, 
247 
agglutinoids, 257 
biological relationship 
and, 252, 253 
Bordet’s explanation of, 
260 
by means of cell body 
proper, 247 
by means of ectoplasmic 
substances of bac- 
teria, 246, 257 
by means of flagella, 247 
cataphoresis of bacteria 
in, 262, 264 
cohesive force affecting, 
267, 269 
description of, 240 
effect of gelatin addition 
in mastic solution 
on, 265 
effect of immediate shak- 
ing on, 140 
Ehrlich’s interpretation 
of process of, 256 
diagrammatic repre- 
sentation of, 257 
Agglutination, Eisenberg 
and Volk’s inter- 
pretation of, 257 
Ficker’s reaction in, 245 
flocculation of colloids 
and, 247, 261 
mechanism of, 262 
mutual, 263 
functional importance of, 
in infectious dis- 
ease, 270, 271 
group, 251 
cause of, 252, 253 
diagnostic value of, 
253 
hemagglutination analo- 
gous to, 259 
history of, 240 
importance of electro- 
lytes in, 260 
in colloidal solutions, in- 
hibition zones in, 
266 
in immune serum, 160 
in motile and non-motile 
bacteria, 242, 244, 
247 
in salt-free environment, 
by addition of or- 
ganic substances, 
266 
influence of immuniza- 
tion with different 
animal species on, 
254 
with different species 
of bacteria on, 
253, 254 
influence of salts on sen- 
sitized and unsen- 
sitized bacteria in, 
264, 265 
experiments of Neisser 
and Friedemann, 
in, 265 
inhibition zones in. 177 
iso-agglutinins in, 299 
grouping of, 299 
value of presence of, 
303 
methods of, 240, et seq. 
Bordet’s, 245 
Ficker’s reaction in, 
245 
Gruber-Widal, 242 
macroscopic, 241 
microscopic, 241, 242 
Proescher’s, 242 
thread reaction of 
Pfaundler, 245 
nature of, 245 
Northrop and deKruif’s 
work on, 267, 
269 
not associated with life 
of bacteria, 244 
not dependent upon 
alexin, 244 
of “agglutinin” bacteria, 
264 
Agglutination, of bacteria 
in active immun- 
ization, 99 
of capsulated bacteria, 
14, 251, 264 
phenomenon of, 240 
power of, alterations in, 
by cultivation in 
immune serum, 13, 
249 
effect of heating on, 
247, 248 
spontaneous alteration 
in, 248 
pro-agglutinoid phenome- 
non in, 259 
pro-agglutinoid zone in, 
177, 243, 258 
specificity of, 240, 241, 
251 
diagnostic value of, In 
group reaction, 
253 
limitations to, 251, 252 
thread reaction of 
Pfaundler in, 245 
“two phase” theory of, 
261 
Agglutination reaction, 
diagnostic use of, 
241, 244 
in diagnosis of dysentery, 
241, 243 
of glanders in horses, 
244 
of paratyphoid fever, 
243 
of pneumonia, 243 
of streptococcus in- 
fections, 244 
of typhoid fever, 242, 
243 
nature of, 245 
precipitin reaction analo- 
gous to, 288 
with dead bacteria, 245 
Agglutinins, 245 
absorption of, in agglu- 
tination, 248, 254 
complete, — impossibil- 
ity of, 255 
bacteriotropins and. 101, 
359, 379, 382 
definition of, 99, 100 
group, 251 
chief or major, 252, 
253 
para- or minor, 252, 
253 
heating of, effects of, 
247, 257 
explanation of, 259 
“immune,” 256 
in hemolytic serum, 
103 
nature of, 246 
normal, 102, 255 
explanation of, 256 
production of, 149, 245, 
246 
643 644 
INDEX 
Agglutinogen, 245 
effects of heating on, 247 
localization of, in cell 
body proper, 247 
nature of, 246 
Agglutinoid Phenomenon, 
so-called, 257, 259 
explanation of, 320 
by Ehrlich school, 258 
Agglutinoids, 251, 257 
Aggressins, 387 
action of, 17, 18 
obtaining of, 17 
secretion of, by bacteria 
in body, and viru- 
lence, 17, 19 
Albuminolysins, 210, 211, 
231, 232, 317, 629 
Alexin, 155 
a combination of soaps 
and proteins, 191 
absence of, from aqueous 
humor of the eye, 
186 
action of acid and alkali 
on, 195 
analogy between fer- 
ments or enzymes 
and. 155, 163, 164, 
192 
bactericidal powers of, 
155 
deterioration of, by 
heat, 92 
definition of, 90 
dependence of, on con- 
centration, 192 
effect of hydrogen ion 
concentrations on, 
194, 195 
extraction of, from leuko- 
cytes and lympha- 
tic organs, 365, 
et seq. 
filtration of, 193 
in hemolysis, 162 
inactivation of, by salt, 
193, 194, 201 
by heat, 194, 195 
by salt-removal, 195 
by shaking, 193, 194, 
203 
reversibility of, 201 
Gramenitski’s exper- 
iments on, 201, 
202 
increase of. on clot, 188 
influence of salts on ac- 
tion of, 188, 194 
leukocytic origin of, 90. 
156, 184, et seq. 
multiplicity of. 169, et 
seq., 193 
Bordet’s views on, 
171, 172 
Ehrlich’s views on, 170 
nature of, 155. 169, 170 
chemical, 190. et seq. 
physical. 193 
photoinactivation of, 203 
presence of, in blood 
plasma, 155, 186 
in blood stream. 186 
in normal blood, Gen- 
gou’s view of, 187, 
364, 365 
Metchnikoff’s view 
of, 185, 186. 363 
presence of, in normal 
blood, other exper- 
imentors on, 187 
364 
preservation of, (foot- 
note), 194 
Alexin, production of, in 
liver, 189 
varieties of : 
macrocytase, 185, 362 
microcytase. 185, 362 
Alexin fixation, 203 
albuminolysin in, 210, 
211 
identity of, with pre- 
cipitins, 211 
writer’s opinion on, 
211 
analogy of, with colloidal 
phenomenon, 211, 
214 
Bordet and Gengou’s ex- 
periment on, 204, 
205 
Bordet-Gengou phenome- 
non in. 204 
Gay’s experiments sup- 
porting, 207, 208 
in diagnosis of infec- 
tious diseases, 205, 
206 
in diagnosis of syph- 
ilis. 206 
Moreschi’s experi- 
ments, 207 
“Bordetscher Antikor- 
per” in. 212 
by immune animal sera 
with their specific 
antigens, 206 
by protein and antipro- 
tein sensitizers, 
206, 210 
distinguished from com- 
plement deviation, 
204 
during hemolysis, na- 
ture of. 192, 193 
Ehrlich’s (schematic) 
conception o f, 
205 
excess of antigen or anti- 
body in. 178 
experiments of, on syphi- 
litic monkeys, 216, 
217 
forensic tests in, 206, 
209. 214 
in anaphylaxis. 462. 463 
in determination of na- 
ture of unknown 
protein. 231 
delicacy of, 209, 232, 
233 
technique of. 232 
in diagnosis of glanders, 
235, 236 
in diagnosis* of gonor- 
rheal infections, 
236 
in diagnosis of malignant 
neoplasma. 233 
von Dungern’s method 
of, 234. 235 
antigen production 
for. 234 
results of, 235 
technique of. 234 
in diagnosis of syphilis 
in human beings, 
217 
in diagnosis of tubercu- 
losis, 216, 236, 
237 
in Wassermann reaction, 
206, 216 
nature of. 190 
non-specific, 212 
by heated normal 
serum, 214 
Alexin fixation, by lipoidal 
substances in tis- 
sues, 213 
by preserved normal 
serum, 214, 215 
by protein emulsion 
and other ex- 
tracts, 214 
by unsensitized bac- 
teria, 213 
of specific precipitates, 
208, et seq. 
albuminolysin identical 
with, 211 
writer’s opinion on, 
211 
Gay on, 207, 208 
Pfeiffer and Fried- 
berger on, 208 
Sachs on, 208 
writer on, 211 
practical applications of, 
206, 216 
precipitin reaction and, 
208, 210 
Dean on, 211 
Gengou on, 210 
Neufeld and Ilaendel 
on, 212 
writer on, 211 
preparation of bacterial 
antigens for, 237 
specific antiprotein sen- 
sitizers in, 206, 
210 
with syphilitic serum in 
antigens from nor- 
mal organs, 218 
Alexin splitting, 195 
Brand on, 11)6, 197 
by method of Ferrata, 
195, 197 
by method of Liefmann, 
201 
by method of Sachs and 
Altmann, 201 
effect of acid reaction on 
fractions of. 198 
end-piece in, 197, et 
seq. 
mid-piece in, 197 et seq. 
heat sensitiveness of 
fractions of, 196, 
198 
interchange of fractions 
of. in different 
animals, 199 
is it the inactivation of 
the mid-piece? 200 
mid-piece only bound In 
Wassermann reac- 
tion, 198 
persensitized cell in, 198 
physical occurrence of 
fractions of. 196 
properties of fractions of, 
196 
quantitative relations be- 
tween fractions 
of, 199, 200 
Alexocytes, 184 
Alkali-albuminate precipi- 
tin, 284 
“Alkalinity theory’’ of im- 
munity, 94 
“Allergie,” 30, 106, 416 
of smooth muscle in 
anaphylactic ani- 
mals, 453 
in syphilis, 613 
Allergy, 408, et seq. 
Amanita phalloides, specific 
antitoxin from, 
U2 INDEX 
645 
Amboceptor, 164 
Bordet’s definition of, 
174, 175 
complementophile groups 
or polyceptors of 
(Ehrlich), 164, 
172, 174 
cytophile group of, 164, 
et seq., 174 
multiplicity of, 165, 166, 
169 
Ehrlich and Morgen- 
roth on, 165, et 
seq. 
quantitative determina- 
tion of, in immune 
serum, 175, 176 
specificity of, 165, 205 
Amboceptor and comple- 
ment, Ehrlich and 
Sachs’s views on 
union of, 180, 181 
Noguchi’s measurement 
of quantitative re- 
lations of, 179, 180 
quantitative ratio be- 
tween, 179 
Amebae, artificial, 355 
Amoss’s work on polio- 
myelitis, 568, 569 
Anaphylactic reaction, in 
determination of 
antigenic proper- 
ties, 412 
Anaphylactic reaction, site 
of, 450 
cellular theory of, 428, 
451, 467 
disappearance of anti- 
body from circula- 
tion, 452 
Doerr and Russ’s ex- 
periments on, 452 
isolated muscle tech- 
nique, 453 
Dale’s isolated uter- 
us method, 454 
Schultz’s physiologi- 
cal method, 453 
Weil’s experiments, 
454. 455 
time intervals neces- 
sary between in- 
jection of anti- 
serum and devel- 
opment of hyper- 
susceptible state, 
451. 452. 460 
transfusion experi- 
ments in support 
of. 452 
Coca’s work with 
guinea pigs, 452 
Manwaring’s work 
with dogs, 452 
Pearce and Eisen- 
brey’s work with 
dogs, 452 
Weil’s work with 
guinea pigs, 453 
humoral theory of, 428, 
451, 459 
Anaphylactic shock, 430 
See also Anaphy- 
laxis, clinical 
manifestations of, 
430, et seq. 
absence of antibody in 
circulation at time 
of, 452 
diminution of, by injec- 
tions of hyper- 
tonic salt solu- 
tion, 445, 463 
Anaphylactic shock, effect 
of atropin and 
other drugs on, 
444, 445 
histamin causing symp- 
toms in guinea 
pigs analogous to, 
458 
peptone shock, analo- 
gous to, 458 
prevention of, 442 
quantity of antigen to 
produce, in com- 
parison with that 
necessary for sen- 
sitization, 423 
reinjection for produc- 
tion of, 422 et seq. 
Anaphylactin, 446, 449 
Anaphylatoxin, 18, 443, 
445, 459, 465, 
506, et seq. 
action of alexin in, 507 
aggressive action of, 250 
coagulation and toxifica- 
tion of blood, in 
relation to, 466, 
467 
conception in anaphy- 
laxis, 460 
importance of, 467 
Jobling and Petersen’s 
experiments on, 
466, 509, 510 
Keysser and Wasser- 
mann’s work on, 
509 
Novy and deKruif’s 
studies on forma- 
tion of, 466, 467 
proteolysis in formation 
of, 630, 635 
source of, 507, 509, 510 
summary, 510 
Zinsser and Dwyer’s 
studies on, 444 
Anaphylatoxin theory ap- 
plied to bacterial 
infection, 506, et 
seq. 
Anaphylaxis. 404 
alexin fixation in, 462 
“allergy” and, 408, /et 
seq. 
altered proteins in, 413, 
414 
analogy of immediate re- 
action in serum 
sickness to, 471, 
472 
analogy of tuberculin re- 
action to, 500 
anaphylactic poisoning, 
from precipitates, 
464 
protein-split products 
of Vaughan and 
Wheeler in, 461, 
462 
anaphylatoxin conception 
of, 460 
importance of, 467 
Anderson and Schultz’s 
work on, 432 
anti-anaphylactic state 
in, See Antiana- 
ph.vlaxis, 407 
antibody involved in, 428 
antigen-antibody reaction 
as basis of, 442 
antigen in, 411, et seq. 
identity of sensitizing 
and toxic sub- 
stance, 448 
Anaphylaxis, antigen in, 
identity of sensi- 
tizing and toxic 
Substance, Doerr 
and Russ’s work 
on, 448, 449 
Wells’s work on, 449 
intervals between ad- 
ministration of, 
419 
nature of, 412 
path of introduction 
of, 418 
intracerebrally, 422 
intra - intestinally, 
418 
by feeding, 418 
by rectum, 419 
into large intes- 
tine, 419 
intravenously, 418 
subcutaneously, 418 
quantity of, adminis- 
tered, 419, 420 
specificity of, 414 
degree of, 414, 415 
individual, of vari- 
ous fractions of 
antigen, 415 
organ, 416 
autosensitization 
in, 412 
species, 416 
two separate sub- 
stances in, 429 
Doerr and Russ’s ex- 
periments on, 448, 
449 
Gay and Adler’s ex- 
periments on, 448 
Pick and Yama- 
nouchi’s experi- 
ments in, 448 
Arthus phenomenon in, 
456 
work on, 406 
asthma and, 438, 485 
attempted explanation of 
differences in 
symptoms in, 438 
Auer and Lewis’s work 
on. 432 
autosensitization in. 416 
Besredka and Stein- 
hardt’s work on, 
423 
Besredka’s theory of, 
447, 448 
Besredka’s work on, 419 
Biedl and Kraus’s work 
on, 432. 436, 458 
Bogomolez’s work on, 414 
Boughton’s work on. 439 
Calvary’s work on. 436 
cell participation in, 442, 
467 
cellular site of reaction 
in, 451, 467 
clinical manifestations 
of, 430, et seq., 
455, 456 
attempted explanation 
of differences in, 
438 
in dogs, 436 
fall in blood pres- 
sure in, 436, 437 
increase of lymph 
flow in, 436 
intestinal reaction 
in, 437 
lowered coagulabili- 
ty of blood in, 436 
in goats, 438 646 
INDEX 
Anaphylaxis, clinical mani- 
festations of, in 
guinea pigs, 430 
alexin reduction in, 
434 
circulation symp- 
toms in, 433 
effect of atropin and 
other drugs on, 
432 
fall of temperature 
in, 433 
fever in, 434 
lowered coagulability 
of blood in, 435 
pulmonary emphy- 
sema in, 431 
respiratory symp- 
toms in, 431 
temporary diminu- 
tion of polynu- 
clear leucocytes, 
435 
in mice, 439 
in monkeys, 438 
in rabbits, 435 
Coca’s experiments 
on, 435, 436 
reaction of heart in, 
435 
clinical significance of, 
469 
Dale’s isolated uterus 
method used in, 
454 
degenerations and 
chronic disease as 
a result of, 439 
dependence of, on pre- 
ceding inoculation, 
406 
differentiation between 
true, and idiosyn- 
crasies, 477 
diminution of alexin 
after anaphy- 
lactic shock in, 
462. 463 
significance of, 463 
drug idiosyncrasies and, 
489 
early work on, 405 
Fink’s work on, and pro- 
tein cleavage prod- 
ucts, 412, 413 
Flexner’s work on, 405 
food idiosyncrasies, 469, 
478, et seq. 
Friedberger’s work on, 
419. 433, 434, 464 
Friedemann’s experi- 
ments in, in vitro, 
464 
fundamental principles 
of, 404 
Gay and Southard’s ar- 
guments against 
antigen - antibody 
reaction theory of, 
447, 449 
hay fever and, 485, 489. 
See also Hay 
Fever, 479 
humoral theory of, 428, 
451, 459 
incubation time in, 406, 
420 
injury a factor in, 457 
in serum therapy. See 
Serum Sickness, 
469 
Lesne and Dreyfus’ work 
on. 419 
local, 456 
Anaphylaxis, local, Arthus 
phenomenon as, 456 
Auer’s experiments on 
dogs and rabbits 
producing, 456, 
457 
Dongcope’s work on, 439 
Magendie on, 405 
Manwaring’s work on, 
437 
mechanism of antiana- 
phylaxis or sensi- 
tization in, 439, 
442 
desensitization in, 439 
difference in method 
of, 420, 422 
duration of hypersen- 
sitive state, 423 
specificity in, 441 
tolerance to anaphy- 
lactic poison in, 
443. 444 
Nicolle’s theory of, 462 
organ specificity in, 416 
Otto’s work on, 420 
Pearce and Eisenbrey’s 
work on, 436, 437 
Pfeiffer’s work on, 433 
phenomena related to, 
493, et seq. 
toxic action of normal 
sera, 513 
toxin hypersuscepti- 
bility. 510 
Pick and Yamanouchi’s 
work on, 414 
quantitative methods ap- 
plied to study of, 
428, 429 
Ranzi’s work on, 434 
relation of alexin to, 462 
relation of antibodies of, 
to other anti- 
bodies, 429, 430 
Richet and Hericourt’s 
work on, 405, 406 
Richet and Portier’s 
work on, 405, 406 
Rosenau and Anderson’s 
work on 407, 413, 
415, 418, 420, 433 
“sessile receptors” in, 
450 
specificity of, 407, 414 
sympathetic ophthalmia 
and. 517 
Theobald Smith phenom- 
enon in, 407 
toxic action of normal 
sera and, 513 
toxin hypersusceptibility 
and, 510 
transference of, 407, 424, 
et seq. 
true antigen-antibody re- 
action. 442. 444 
tuberculin ophthalmo-re- 
action and, 497, 
499 
tuberculin reaction and, 
497. See also Tu- 
berculin reaction, 
500 
tuberculin skin reaction 
and, 497. 499 
Vaughan and Wheeler’s 
theory of mechan- 
ism of. 461, 462 
Vaughan and Wheeler’s 
work on toxic 
fraction of protein 
molecule in, 461, 
462 
Anaphylixis, Vaughan’s 
work, 434 
Wells’ review of, 438 
Weichardt and Schitten- 
helra’s work on, 
437 
Weil’s work on, 437, 442 
with bacterial extracts, 
408 
with normal serum, 407 
with proteins, 407, 408 
Zunz’s work on, 412, 413 
Anaphylaxis, bacterial, 493 
anaphylatoxin formation 
in, 506. See also 
under Anaphyla- 
toxin. 
cellular nature of, 496 
endotoxin theory of pro- 
duction of, 506 
Friedberger’s experiments 
in, 496, 507, 509 
general correlation, 515, 
518 
nature of bacterial infec- 
tions and. 508 
serum anaphylaxis and, 
494 
specificity of, 493, 494 
technique of sensitiza- 
tion with bacteria, 
494, 495 
tuberculin and similar 
reactions as. 497 
Friedberger’s anaphy- 
latoxin theory as 
explanation of, 500 
Vaughan’s theory of bac- 
terial spitting as 
cause of. 506 
Anaphylaxis, passive, 423, 
et seq. 
Doerr and Russ’s work 
on, 425 
duration of. 426 
Friedemann’s work on, 
424. 425, 459 
Gay and Southard’s work 
on, 424, 425 
interval between injec- 
tion of sensitized 
serum and injec- 
tion of antigen in, 
426, 427 
methods of production 
of, 424 
nature of reaction of, 
424 
Nicolle’s work on, 424 
Otto’s work on, 424. 425 
production of, in differ- 
ent species, 425, 
426 
Richet’s work on, 424, 
459 
summary. 427, 428 
Weill-Halle and Le- 
maire’s work on, 
459 
Anderson and Schultz’s 
work on anaphy- 
laxis. 432 
“Anergie,” 613, 623, 624 
Anthrax bacilli, attenua- 
tion of virulence 
of, 14. 24 
as pure parasites, 11 
capsule formation, 15 
immunization against 
arsenic of, 14 
virulence of. 11, 15, 25 
Anthrax, relative suscepti- 
bility of man and 
animals to, 62 INDEX 
647 
Anthrax, serum therapy of, 
567 
toxin of, heat stability 
of, 43 
vaccination against, his- 
tory of, 74 
method of, 74 
Anti-agglutinin, 256 
Anti-alexin. 172 
action ox, 172, 173 
Anti-amboceptor, 167, 168 
Anti-anaphylaxis, or desen- 
sitization, 407, 
439 
duration of, 443 
Besredka’s work on, 419, 
440, 442 
Friedberger and Mita’s 
work on, 442 
mechanism of, 441, 442 
tolerance to anaphy- 
lactic poison in, 
443, 444 
methods of producing, 
439, et seq. 
Besredka and Stein- 
hardt’s methods, 
440 
Rosenau and Ander- 
son’s methods, 
440, 441 
quantitative relations in, 
441 
specificity of, 441 
"Anti-antibodies,” 298 
Antibodies, concentration 
of, in lymphatic 
organs, 116 
in active immunity, 117 
in active immunization, 
95, 96, 101 
in other organs in active 
immunization, 117 
locality of production 
of, dependent on 
locality of antigen 
concentration, 117 
normal, explanation of, 
256 
resistance to, 16 
41, 96, 100, 
Antibodies, further consid- 
eration of the 
nature of, 306, 
et seq. 
cell permeability in for- 
mation of, 309, 
310 
chemical structure of 
tissue cell in for- 
mation of, 306, 
307 
classification of these 
phenomena. 308 
Ehrlich’s side-chain 
theory concerning, 
309 
hypersusceptibility and 
tolerance as alter- 
ations in, 307, 308 
dissociation, of antigen 
and, 315 
early attempts to pro- 
duce protein-free 
antibodies, 315, 
316 
Huntoon’s method of, 
316, 321, 323 
influence of hydrogen 
ion concentration 
on, 316 
Landsteiner’s early 
work on, 315 
Antibodies, further consid- 
eration of the na- 
ture of, essential 
identity of, 317 
Bail and Hoke’s study 
on, 318, 321 
identity of agglutinins 
and precipitins, 
317, 321 
identity of hemolytic 
sensitizer and ag- 
glutinin, 324 
Coulter’s work, bear- 
ing of on, 324-5 
identity of precipitat- 
ing and anaphy- 
lactic, 323 
Doerr and Russ’s ex- 
periments on, 323, 
428, 429 
Friedberger on, 323, 
429 
identity of precipitins, 
and bactericidal 
and protective, 
321 
identity of precipitins 
with complement 
fixing, 322 
Dean’s work on, 322 
writer on, 322, 323 
opposition to view of, 
319, 321 
Gay and Chickering’s 
experiments on, 
321 
nondiffusibility of anti- 
gen, import of in- 
formation of, 310 
physical measurements 
applied to reac- 
tions of, 325, et 
seq. 
production of, by bac- 
terial antigens, 
306 
as protective process, 
306 
as a sensitizing proc- 
ess, 306 
purification of anti- 
bodies, 320, 321 
union of antigen with, 
311 / 
chemical nature of, 311 
colloidal nature of, 
311, 312 
Coulter’s work on nor- 
mal and sensitized 
cells in electric 
field, changing 
conception of, 313, 
314 
Loeb’s work on col- 
loids, changing 
conception of, 
312, 313 
Antibody effects, resistance 
to, 16 
Antibody formation, bodv 
cell in, 103, 110, 
116, 144 
against bacteria, 111 
chemical action of anti- 
gens in, 148 
chemical nature of, 145 
in active immunity, re- 
moval of spleen 
and, 116, 117 
inheritance of a capacity 
for, 484 
mechanism of, 110, 113 
mechanism of (side chain 
theory), 143 et seq. 
Antibody formation, mech- 
anism of, by in- 
ternal secretion of 
body cell, 144 
processes of metabo- 
lism and, 145 
overproduction of recep- 
tors in, 148 
principles of, 104, 106 
Antibody involved in ana- 
phylaxis, 428 
Doerr and Russ’s quan- 
titative methods 
• of studying, 428 
429 
identity of precipitin s 
and anaphylactic 
antibody, 429, 430 
Weil’s experiments on, 
430 
Anticomplement, 169, 172 
action of, 172 
existence of, questioned, 
207 
Anticomplementary action, 
207, 213, 214 
of bacterial antigens, 
237 
Anticomplementophile, 169, 
173 
Anticytophile interpreta- 
tion of anti-sensi- 
tization (Ehrlich 
and Morgenroth), 
168 
controversy on, 168 
“Antiformin,” 80 
Antigen-antibody reactions, 
129. See also 
Toxin - antitoxin 
reaction, analogy 
of enzyme reaction 
with, 141, et seq. 
enzyme reaction with, 
141, et seq. 
adsorption theory and, 
139, 141 
antibody production in 
body cells in, 110, 
149 
in relation to formation 
of receptors, 173, 
174 
possible dependence of, 
on molecular size, 
110 
relationship between sus- 
ceptibility of tis- 
sue and toxin 
binding properties 
in, 150 
side-chain theory in, 148 
similarity of, to colloidal 
reactions, 177, 178 
specificity of, 96, 113, 
148 
varieties of antibodies in, 
101 
agglutinins, 100, 106, 
149 
antitoxins, 96, 149 
cytotoxins, 106, 129 
precipitins, 100, 106, 
140, 149 
Antigen-antibody theory of 
anaphylaxis, 442, 
444 
Antigenic properties of 
cells, relation of, 
to lipoidal constit- 
uents, 113 
demonstration of, by 
anaphylactic ex- 
periment, 106 648 
INDEX 
Antigenic properties of 
cells, of protein 
cleavage products, 
107 
Antigenic properties of pro- 
teins, demonstra- 
tion of, by an- 
aphylactic experi- 
ment, 106 
of protein cleavage prod- 
ucts, 107 
specificity of, altered, 108 
Antigen in anaphylaxis, 
411 
altered proteins as, 413 
drug idiosyncrasies ex- 
plained by, 414 
protein cleavage prod- 
ucts as, 412 
racemization, effects of, 
414 
Antigens, action of, 105 
active immunization 
with,analogous to 
drug tolerance, 
115 
bacterial, 237, 239 
characteristics of, varia- 
tions in, 105, 109 
chemical nature of, 106, 
166 
classification of, 105 
complex structure of, 
Ehrlich and Mor- 
genroth’s concep- 
tion of, 166 
definition of, 41, 105 
“local” immunity in or- 
gans directly in 
contact with, 117 
locality of production of 
antibodies depend- 
ent on locality of 
doncentration of, 
117 
organ specificity of, 114 
protein nature Oil, 41, 
105, 113 
residue, 15, 178, 237 
specificity of, 113, 114 
Antigens, Wassermann re- 
action, 216, 223 
Antihemolysin a,, induction 
of, 167 
Anti-isolysins, 298 
“Antiricin,” 95 
Antisensibilisin. 447, 419 
Antisensitization, anticyto- 
phile interpreta- 
tion of (Ehrlich 
and Morgenroth), 
168 
controversy on, 168 
Antisensitizers, 167, et scq. 
non-specificity of, 169 
Antitoxic serum, direct 
effect of, on toxin, 
120 
indirect protective action 
of, against toxin, 
120 
“normal” antitoxin of 
Behring, 123 
Antitoxin, B. Welchii, 543, 
544 
Bacillus Oedematiens, 
production of, 
544, 545 
chemical relations of, 
with toxin, 96, 
130, et seq. 
definition of, 95, 96 
diphtheria. See Diph- 
theria antitoxin. 
Antitoxin, immunity trans- 
ferred by, 66 
production of, 149 
by true toxins, 41, 96 
snake venom, 121, 541 
effect of heat on, 121, 
541 
specific, substances incit- 
ing, 41, 96, 97 
standardization of, 123, 
525 
by means of toxin, 123 
guinea pigs used in, 
123, 124 
tetanus, production of, 
538 
use of, in passive immu- 
nization, 96 
Vibrion septique, prepa- 
ration of, 544 
Antitoxin unit, dipth- 
theria, 123, 126 
“Antitoxinogen,” differen- 
tiated from other 
antigens, 41, 105 
Antivenin, 541 
Arrhenius and Madsen on 
neutralization in 
toxin-antitoxin re- 
action, 136, 137 
Arthus, phenomenon of, 
422, 456 
work of, on anaphy- 
laxis, 406 
Coca’s suggested expla- 
nation of, 422, 436 
local, anaphylaxis and, 
456 
Nicolle, work of, on, 
424 
relationship between im- 
mediate local re- 
action in humans, 
472 
Ascoli and Izar’s work on 
meiostagmin reac- 
tion, 328, 329 
Asiatic cholera, relative 
susceptibility of 
man and animals 
to, 62 
Asthma, anaphylaxis and, 
438, 485, 489 
Atrepsie, 65 
similarity of Pasteur’s 
“Exhaustion The- 
ory” to, 93 
Attenuation of bacteria 
by chemicals, 76 
by cultivation under 
pressure, 76 
by drying, 76 
by heating, 75 
by passage through ani- 
mals, 75 
by prolonged cultivation 
above optimum 
temperature, 75 
by prolonged growth on 
artificial media in 
presence of own 
metabolic prod- 
ucts, 73, 75 
loss of capsule forma- 
tion in, 14 
Auer and Lewis’s work on 
anaphylaxis, 432 
Autocytotoxins, 103 
in precipitin formation, 
288 
Autogenous vaccines, 395, 
576 
“A'utohemolysins/’ 297, 
298 
Auto-inoculation, by mas- 
sage in active im- 
munization, 390, 
400 
Autolysins, 297 
Autosensitization in ana- 
phylaxis, 416 
Auxilysin, 183 
Avian tuberculosis, relative 
susceptibility of 
animals to, 20, 60 
Bacillus botulinus, action 
of, 4 
different toxins produced 
by various strains 
of, 543 
Bacteria, adaptation of, in 
body, 5, 7, 10, 11, 
12, 19, 65 
agglutinability of, altera- 
tions in, 13, 247, 
248 
by cultivation in im- 
mune serum, 13, 
249 
caused by heating, 15, 
248 
normal differences in, 
between strains of 
same species, 12, 
250, 251 
spontaneous, 248, 249 
in agglutinoid, 251 
agglutination in. See 
under Agglutina- 
tion. 
“agglutinin” bacteria, 
agglutination of, 
264 
aggressin secretion of, in 
body, and viru- 
lence of, 17, 19 
alexin fixation reaction 
with, 237 
anti-ferment in, 633 
anti-opsonic properties 
of, and anti-chem- 
otactic substances, 
387 
artificial cultivation of, 
11 
attenuation of, 12, 24, 
26, 75 
methods of, 75. See 
also under Atten- 
uation. 
by laboratory man- 
ipulations, 12, 
75 
capsulated, agglutination 
of, 14, 251, 264 
resistance to leuko- 
cytes of, 15, 388 
virulence of, 14, 388 
capsule formation in, 
and virulence, 13, 
14 
destruction of, by cytases 
in leukocytes. 362, 
363 
by exudates, 361 
by phagocytes, 5, 360- 
361 
different types of infec- 
tions caused by 
same, 26 
different strains of, vari- 
ation in infection 
from, 12 
ectoplasmic hypertrophy 
of, in relation to 
virulence, 15, 18 INDEX 
649 
Bacteria, effect of body tem- 
perature on inva- 
sive powers of, 20 
effect of cultural adapta- 
tion of, on viru- 
lence, 11, 13, 17, 
19, 24 
effect of path of intro- 
duction of, on in- 
fection, on viru- 
lence of, 3, 21, 22 
effect of quantity of, in- 
troduced, on in- 
fection, 3, 23 
effects of, on immunized 
and normal ani- 
mals, 45 
entrance of, into body 
tissues, 7, 26, 29 
enzyme-like substances 
formed, 17 
generalized action of, 27 
growth of, within leuko- 
cytes, 359 
immunization of, against 
host, 13 
in blood stream, 5, 8, 9, 
27 
in latent condition, 31 
in localized infection, re- 
action to, 28 
through accidental con- 
ditions, 2, 24, 28 
in normal serum, resist- 
ance to phago- 
cytosis of, 16, 19, 
387 
incubation of, 29, 30 
localized action of, 25, 26 
measurement of relative 
degrees of viru- 
lence of, 24 
negative charge of, in 
suspension, 6, 263 
number of, .introduced, 
and relative viru- 
lence, 23 
occurrence of, 2 
parasitic and saprophy- 
tic, classification 
of, 10 
phagocytosis of. See un- 
der Phagocytosis, 
ptomain production by, 
33, 37 
relative virulence of dif- 
ferent strains of 
same, 12, 23, 24 
resistance of, to antibody 
effects, 16 
to phagqfcytosis, due 
to nonabsorption 
of opsonin, 386 
resistance of living cell 
to, 6 
secondary ab scesses 
caused by, 27, 28 
selective action of, in 
localized infection, 
28 
selective lodgment of, in 
tissues, 28, 48, 49 
sensitized, immunization 
with, 78 
sensitized and unsenti- 
tized, influence of 
salts on agglutina- 
tion of, 264, 265 
similar conditions pro- 
duced by different, 
26 
specificity of, and infec- 
tion, 25, 26 
symbiosis of, 30, 31 
Bacteria, use of, in active 
immunization, 75, 
95, 98, 99 
variation in virulence of, 
when successively 
passed through 
animals, 12, 24 
Bacterial alexin fixations, 
237, 238 
Bacterial anaphylaxis. See 
Anaphylaxis, bac- 
terial. 
Bacterial differentiation, 
99 
Bacterial extracts, active 
imm unization 
with, 78 
extraction of bacteria 
for, by mechani- 
cal methods, 81 
by permitting them to 
remain in fluid 
media, 80 
Bacterial hypersensitive- 
ness and specific 
reactions, 493, et 
seq. 
Bacterial infections, con- 
ceived as reaction 
of body against a 
foreign protein, 
19, 508 
Bacterial precipitins, group 
reactions in, diag- 
nostic value of, 
275 
partial or minor, 276, 
278 
Bacterial products, active 
immunization 
with, 82 
toxins, 82 
Bacterial proteins, 39, 45 
agglutinating sera and, 
100 
immunization and, 79 
injection off, producing 
therapeutic ef- 
fects, 626 
Bacterial toxins, 33. See 
also Toxins. 
action of, after distribu- 
tion in bods, 43, 
48 / 
active immunization 
with, 82 
chemical structure of, in 
relation to toxic- 
ity, 51, 52 
classes of, 39 
endotoxins. See Endo- 
toxins. 
lesions produced by, in 
cofurse of excre- 
tion, 48, 53, 54 
local injury by, 48 
locality of production of, 
47 
nature of, 38 
nature of union of, to 
body cells, 51 
nerves attacked by, 49, 
50, 54 
obtaining of, 39 
from dead cultures, 39, 
41 
from living cultures, 
39, 41 
physical relationship 
with body cells in 
action of, 43, 48, 
51 
production of antitoxin 
by, 41, 96 
Bacterial toxins, ptomaines 
in relation to, 35, 
37 
selective action of, 48, 
51 
principles of, 51 
reasons for, 53 
selective localization of, 
47, 48 
specific distribution after 
introduction of, 
48, 49 
specific susceptibility of 
tissues to, 49, 55 
chemicals inhibiting, 55 
specificity of, 38 
true, 39, et seq. 
analogy of, with en- 
zymes, 17, 40, 42 
bacteria producing, 11, 
40, 44 
characteristics of, 41 
heat sensitiveness of, 
42, 43 
incubation time of, 
30, 43 
Bactericidal powers of 
blood serum, 59, 
89, et seq. 
alexin in, 90, 156 
in vivo, 156 
cholera experiments 
in, 59, 98, 156 
nature of, 156, 157 
Bactericidal properties of 
blood serum, 153 
Bactericidal substances, 
origin of, from 
leukocytes, 155, 
184, 365 
Bacteriemia, 8, 17 
prognosis of, 9 
Bacteriolysins, agglutinins 
and, 383 
in active immunization, 
99, 102 
Bacteriolysis, 156 
extracellular theory of, 
159 
heat a factor in, 156, 157 
mechanism of, 159 
immunity conferred by, 
156, 157 
in immune serum, 156, 
157, 159 
Bordet’s findings in, 
159, 161 
inactivation and reac- 
tivation in, 159 
Pfeiffer’s phenomenon 
in, technique of, 
156, et seq. 
specificity of protec- 
tion of, 156 
transference of power 
of, 156 
leukocyte action in, 184, 
365 
leukocytes in, 159 
side chain theory in, 161 
Bacteriotropins, bacteri- 
cidal sensitizers 
and, 383, et seq. 
normal opsonins and, 
382, et seq. 
presence of, in immune 
sera without 
lysins, 384 
specificity of, 383 
thermostability of, 382 
Bail’s aggressin theory, 17, 
18 
acquired immunity and, 
77 650 
INDEX 
Bail’s aggressin theory, 
classification of para- 
sites, 10 
Bauer’s modification of 
Wassermann test, 
227 
Baumgarten’s osmotic 
theory of bacteri- 
cidal powers of 
blood, 94, 154 
Bayliss, on conception of 
cell, 309 
Behring, Kitasato and 
Wernicke, anti- 
body theory in ac- 
tive immunity of, 
94 
Besredka and Steinhardt’s 
work on anaphy- 
laxis, 423 
work on antianaphylaxis, 
439, 440 
Besredka’s, anti-endotoxic 
serum in treat- 
ment of typhoid 
fever, 562 
method of preparation 
of typhoid (endo- 
toxin), 81 
Besredka’s, theory of ana- 
phylaxis, 447, 448 
vaccines in prophylactic 
immuniza tion 
against plague, 
588 
work in antianaphylaxis, 
419, 442 
work on toxin absorp- 
tion, 55, 151, 152 
Biedl and Kraus’s, work 
on anaphylaxis, 
432, 436 
on peptone shock, 458 
Blood, anti-ferments in, 
637 
coagulation and toxifica- 
tion of, in rela- 
tion to production 
of anaphylatoxin, 
466, 467 
non-putrefaction of, 89, 
153, 280 
phagocytic activities of, 
in immunity, 89 
Blood plasma, cell free, in- 
hibition of bacte- 
rial growth in, in 
immunity, 89 
presence of alexin in, 90, 
186, 188 
Blood serum, agglutination 
in, immune, 160 
alexin in, 155 
nature of, 155, 157 
antibacterial powers of, 
in immunity, 89, 
153 
anti-isolysins in, 298 
autohemolysins in, 297, 
298 
bactericidal action of im- 
mune, 91 
alexin in, 90, 156 
early theories re- 
garding, 153 
in natural immunity, 
89, 90 
in vivo, 156 
cholera experi- 
ments in, 156, 
et seq. 
mechanism of, 157 
assimilation theory of, 
155 
Blood serum, assimilation 
theory of, by 
chemically unfa- 
vorable environ- 
ment, 155 
osmotic theory of, 154 
bacteriqUdal and agglu- 
tinating powers 
of, W r i g h t’s 
studies, 389 
bactericidal powers of, 
153, 155 
bacteriolysis in immune, 
156 
Bordet’s findings in, 
159 
cholera experiments in, 
156, et seq. 
summary of facts in, 
156, 157 
heat a factor in, 156 
mechanism of, 159 
bacteriolysis in immune, 
immunity con 
ferred by, 156, 157 
inactivation and reac- 
tivation in, 159 
intracellular theory of, 
157, 159 
leukocytes in, 159 
Pfeiffer’s phenomenon 
in, technique of, 
156, et seq. 
specificity of protec- 
tion of, 156 
bacteriolytic powers of 
transferable, 156, 
, 157 
cell-free, bacterial 
growth in, 91 
enzymes in, 634 
hemolysis in immune, 
166, 161 
alexin, or complement 
in, 162 
analogy of, to bacter- 
iolysis, 160, 161 
Bordet’s work on, 160 
Ehrlich and Morgen- 
roth on mecha- 
nism of, 161 
haptophore groups in, 
161 
relation of antigen, 
amboceptor and 
complement in, 
161, 165 
work of Ehrlich and 
Morgenroth on, 
162, 163 
work of Liefmann 
and Cohn on, 163 
isohemolysins in, 297, 299 
protective action of, 
against bacteria, 
59 
Blood stream, presence of 
alexin in, 186 
Blood transfusion, 101, 
298, 301, 303, 304 
experiments in proving 
cellular site of 
anaphylactic re- 
action, 452 
isoagglutinins and, 303 
toxic action of normal 
sera in, 513 
Body fluids, bactericidal 
powers of, in nat- 
ural immunity, 90 
Bogomolez’s work on ana- 
phylaxis, 414 
Borden’s method of agglu- 
tination, 245 
Bordet and Gengou's ex- 
periments on alex- 
in fixation, 204, 
205 
Bordet-Danysz phenomenon, 
in connection with 
nature of anti- 
bodies, 327 
in neutralization in 
toxin-antitoxin re- 
action, 141 
Bordet, explanation of, on 
agglutination, 100, 
260 
findings of, on bacterio- 
lytic power of 
immune serum, 
159 
on neutralization in 
toxin-antitoxin re- 
action, 138, 141 
views of, concerning 
relation of anti- 
gen, amboceptor 
and complement, 
173, et seq. 
concerning relation of 
antigen, ambocep- 
tor and comple- 
ment, action of 
complex in, 174 
schematic representa- 
tion, 174 
work of, on antisensi- 
tizers, 169 
on hemolysins, 102 
on precipitins, 104 
Bordet-Gengou phenomenon 
in alexin fixation, 
204, 206, et seq. 
Gay’s experiments sup- 
porting, 207, 208 
Moreschi’s experiments 
supporting, 207 
Botulinus poisoning, action 
of, 4, 36, 48 
Botulinus toxin, 4, 48, 54 
production of antitoxic 
serum by injection 
of, 542, 543 
Bovine tuberculosis, rela- 
tive susceptibility 
of man and ani- 
mals to, 20 
Brownian movement, ag- 
glutination and, 
242 
Buchner on bactericidal 
power of blood in 
natural immunity, 
90, 154, 155 
Bull, studies of, on agglu- 
tination in vivo, 
271 
Calmette’s investigations 
in snake poisons, 
96, 120, 121, 190, 
541 
ophthalmo-reaction, 497, 
499 
Capsule formation in bac- 
teria, destruction 
of, by heat and 
acid, 15 
in relation to infectious- 
ness, 14 
loss of, 14 
virulence and, 14, 15 
Carriers, 2 
Castellan!, absorption ex- 
periments of, in 
agglutination, 254, 
255 INDEX 
651 
Cecil’s work on prophylac- 
tic vaccination in 
pneumonia, 589, 
590 
Cell death, cause of, 6, 7 
part played by cellular 
ferments in, 630 
poisons freed by, 44, 99 
Cell receptors, overproduc- 
tion of, 168 
structure of, 145, 146 
Cellular defenses. See un- 
der Phagocytes. 
Cellular theory, of anaphy- 
laxis, 428, 451 
of immunity, 90, 106 
Cerebrospinal meningitis, 
administration of 
serum in, 552 
death after second injec- 
tions in spinal 
cord in, 472 
Dopter’s investigations 
on typing, 551 
early investigations in, 
549 
epidemic, serum therapy 
in, 549 
Flexner and Jobling’s 
work on, 550 
Jochmann’s investiga- 
tions in, 549 
Kolle and Wassermann’s 
investigations in, 
549 
method used in, 553 
necessity of polyvalent 
serum in, 551, 552 
results of, 550, 551 
standardization of serum 
in, 552 
Chantemesse’s early experi- 
ments in serum 
therapy of typhoid 
fever, 561, 562 
Chemotaxis, 346 
anaphylatoxin and, 352 
Engelmann’s studies in, 
348 
influence of bacteria in, 
349, 350 
influence of bacterial ex- 
tracts in, 39, 349, 
350 
malic acid in, 347 
of slime-molds or myxo- 
mycetes, 347 
of spermatozoa of ferns, 
347 
Pfeffer’s technique in, 
347 
physical explanations of, 
353 
selective, 352, 356 
surface tension in, 353, 
354 
“artificial amebae” 
and. 355 
Chicken cholera, vaccina- 
tion against, his- 
tory of, 73 
Cholera, active prophylac- 
tic immunization 
against, 585 
Ferran’s early investi- 
gations in, 585 
Haffkine’s method in, 
585, 586 
Kolle’s method in, 586 
Strong’s method in, 
586 
Asiatic, relative suscepti- 
bility of man and 
animals to, 62 
Cholera, chicken, vaccina- 
tion against, his- 
tory of, 73 
effect of path of intro- 
duction of bacteria 
of, on infection, 
22 
experiments in, showing 
bactericidal pow- 
ers of blood serum 
in, 98, 156 
hog, immunization with 
bacterial products 
in, 17, 82 
Coagulins, 210 
Cobra, antitoxin, action of, 
542 
standardization of, 542 
Cobra lecithid, 191 
Cobra venom, action of, 
541 
inhibition of action of, 
56 
Coca’s, classification of 
hypersensi- 
tiveness, 409 
objections to analogy of 
anaphylaxis with 
serum sickness, 
474, 475 
suggested explanation of 
Arthus phenome- 
non by, 422, 436, 
456 
transfusion experiments 
on guinea pigs, 
strengthening cel- 
lular theory of 
anaphylaxis, 452 
work of, on anaphylaxis 
in rabbits, 435, 
436 
Coctoprecipitin, 283 
“Coefficient of extinction,” 
393 
Cole’s work on serum 
therapy of pneu- 
monia. 557, 559 
Colles’ law, 602 
Colloidal gold reaction, 
329, 332 
Colloids, flocculation of, 
precipitin reac- 
tion analogous to, 
289, et seq. 
physical properties of, 
surface tension 
in. 202 
reaction in. analogous to 
complement devia- 
tion phenomenon 
of Neisser-Wechs- 
berg. 177 
inhibition zones in, 177 
Complement. See also 
Alexin. 
amboceptor and, No- 
guchi’s measure- 
ment of quantita- 
tive relations of, 
179, 180 
quantitative ratio be- 
tween, 179 
union of, Ehrlich and 
Sachs’s views on, 
180, 181 
definition of, 162, 163 
in hemolysis, 162 
dependence of action of, 
on concentration, 
192 
multiplicity of, 169, et 
seq. 
Bordet’s views on, 171 
Complement, multiplicity 
of, Ehrlich’s views 
on, 170 
nature of, 169, 170 
chemical, 190, et seq. 
colloidal, 202 
Complement deviation, 175, 
et seq. 
argument in favor of 
Bordet’s views, 
178 
colloid reactions, analo- 
gous to, 177 
Gay’s contribution to, 
178 
in hemolytic reactions, 
179 
Morgenroth and Sachs’s 
experiments sup- 
porting, 179 
pro-agglutinoid zone re- 
action analogous 
to, 177 
Complement fixation, 203. 
See also Alexin 
fixation. 
excess of antigen or anti- 
body in, 178 
in determination of na- 
ture of unknown 
protein, 231 
delicacy of, 209, 232, 
233 
technique of, 232 
practical applications of, 
206, 216 
test of, in diagnosis of 
glanders, 235, 236 
in diagnosis of gonor- 
rheal infections, 
236 
in diagnosis of malig- 
nant neoplasms, 
233 
von Dungern’s meth- 
od of, 234, 235 
in diagnosis of syph- 
ilis, 216, 217, 616 
in diagnosis of tuber- 
culosis, 216, 236, 
237 
Complementoid, 173 
Complementopbile group of 
amboceptor, 164, 
172, 175, 203 
Complement splitting, 195. 
See also under 
Alexin, splittingof. 
Bronfenbrenner and No- 
guchi on. 200 
Conglutinin effect, 180. 
183, 321 
Conjunctiva, susceptibility 
of, to infection, 22 
Corpus luteum cytotoxin, 
102 
Coulter’s work on normal 
and sensitized 
cells in electric 
field, 313, 314 
“Cryptogenic tetanus,” 5 
Cultivation of bacteria, 
artificial, 11 
Cytases, in phagocytes, de- 
struction of bac- 
teria by, 362, 363 
Cytolysins, 102 
Cytolysis, 153 
Cytolytic substances, origin 
of, from leuko- 
cytes, 185 
Cytophile group of am- 
boceptor, 164, 167, 
168 652 
INDEX 
Cytotoxins, 102 
ncn-specittcity of organ, 
103 
problems of formation 
of, 298 
production of, 149 
specificity of, 102 
Dale’s isolated uterus meth- 
od, 454 
passive sensitization in 
bacterial anaphy- 
laxis demon- 
strated with, 495, 
496 
used in showing anti- 
genic nature of 
pollen extracts, 
487 
work with histamin, 458 
Dauysz effect in neutrali- 
zation in toxin- 
antitoxin reaction, 
141 
Daphnia, Metchnikoff's 
study of, 89, 334, 
357 
phagocytosis in, 357 
Dean’s antiplague sera, 566 
Denys and Leclef on im- 
portance of phago- 
cytosis in immun- 
ity, 373, 374 
Desensitization, 439. See 
also Anti-anaphy- 
laxis. 
in food idiosyncratic pa- 
tients, 485 
in serum sickness, 476, 
477 
mechanism of, 442 
methods of, 440, 441 
in man, 477, 443 
non-specific, 443 
Bessau’s experiments 
on, 443 
Biedl and Kraus’s 
studies on, 443 
Dale’s observations in, 
443, 482 
specificity of, 441, 444 
Diphtheria, active immuni- 
zation in, with 
t o x i n-antitoxin, 
531 
relative susceptibility of 
man and animals 
to. 62 
Diphtheria antitoxic serum, 
normal, 123 
Diphtheria antitoxin, 519 
antitoxin production in, 
525, 528 
concentration of, 530, 
531 
Kellogg’s modification of 
Roemer’s method 
of titrating, in 
human serum, 
colostrum, milk, 
etc., 522 
Park and Biggs’ experi- 
ments on single 
and divided dos- 
age of, 523, 525 
presence of, in blood of 
normal individ- 
uals, 67, 69, 521 
preservation of, 124 
speed in administration 
of, 520 
speed in absorption of, 
520, 523, 524 
Diphtheria toxin-antitoxin, 
standardization of 
antitoxin, Hom- 
er’s method or 
536, 537 
toxic action of, 532 
Dochez and Gillespie’s 
work on typing of 
pneumococci, 557 
Doerr and ltuss’s experi- 
ments on two 
separate sub- 
stances in anaphy- 
lactic antigen, 
448, 449 
experiments on identity 
of precipitating 
and anaphylactic 
antibodies, 323, 
429 
work on passive anaphy- 
laxis, 425 
work on quantitative 
methods of study- 
ing anaphylactic 
antibody, 428, 429 
Doerr, classification of 
hyp e,.r sensitive- 
ness, 408 
Drug idiosyncrasies, 489 
artificial sensitization 
unsuccessful, 490 
differences in, 490, 492 
experiments on analogy 
of, with anaphy- 
laxis, 490, 491 
Landsteiner’s work on 
alterations of pro- 
teins as a possible 
explanation, 491 
manifestations of condi- 
tions in, 490 
Obermeyer and Pick's 
work on altera- 
tions of proteins 
in explanation of, 
491 
relationship of sensitive- 
ness to chemical 
constitution o f 
drug in, 490 
Swift’s work on, 491 
Drug tolerance, analogy be- 
tween, and active 
immunization with 
antigens, 115 
as function of tissue 
cells, 489 
cellular reactions and, 
307, 308 
von Dungern and Hirsch- 
feld, work on in- 
heritance of iso- 
antibodies accord- 
ing to Mendelian 
laws, 301 
von Dungern’s method of 
alexin fixation in 
diagnosis of malig- 
nant tumors, 234, 
235 
antigen production in, 234 
results of, 235 
technique of, 234 
view of precipitin reac- 
tions, 292, 293 
“Dust cells” of the lungs 
in phagocytosis, 
340 
Dysentery, agglutination 
reaction in diag- 
nosis of, 241, 243 
bacillary, passive immu- 
nization in, 545 
Diphtheria antitoxin, speed 
in absorption of, 
on intravenous in- 
jection, 523, 524 
on subcutaneous injec- 
tion, 523, 524 
speed of diagnosis for, 
necessity of, 525 
stability of, 124 
standardization of, 529 
by means of toxin, 123 
early attempts in, 123 
statistics showing reduc- 
tion of mortality 
with, 519 
transfer of, from mother 
to child, 67, 521, 
522 
toxin production for, 525 
choice of culture in, 
525, 526 
cultivation of strain 
in, 526 
culture medium in, 
526, 527 
“maturing” of toxin 
in, 527 
testing of toxin in, 
528 
Theobald Smit h’s 
method of, 527 
unit of, 123 
Diphtheria bacillus, action 
of, 11 
as a saprophyte, 4 
growth requirements, 10 
in blood, 5 
in normal individuals, 2 
symbiosis and, 30, 31 
Diphtheria toxin, action of, 
49 
construction of, 133 
determination of diph- 
theria immunity 
with, 538 
normal, 123 
preservation of, 124 
specificity of formation 
of, 38 
stability of, 124 
unit of, 123 
Diphtheria toxin-antitoxin, 
neutral mixtures 
of, 531 
Behring’s method of im- 
munization with, 
531, 532 
advantages of. 532 
chief value of, 533 
danger of anaphylaxis, 
in, 533 
determination of presence 
of free toxin or 
antitoxin in con- 
valescents follow- 
i n g treatment 
with, 537 
human susceptibility to, 
532 
Park’s recommendations 
actually used in 
mixtures of, 534, 
535 
production of homologous 
antitoxin in hu- 
man beings with, 
for passive im- 
munization, 534 
results of treatment 
with, 533, 534 
standardization of anti- 
toxin, 536 
limes-necrosis of toxin 
in, 536 INDEX 
653 
Ehrlich and Morgenroth on 
multiplicity of 
amboceptor, ex- 
ample of, 165, 166 
Ehrlich, conception of ac- 
tion of toxin, 126, 
et seq. 
of alexin fixation 
(schematic), 205 
experiments of, on mice 
with ricin and 
abrin, 66, 67 
on relation of antigen, 
amboceptor and 
complement, 164, 
et seq. 
interpretation of agglu- 
tination by, 122, 
256 
diagrammatic repre- 
sentation of, 257 
on action of a poison, 43 
on multiplicity of alexin 
or complement, 
170 
“partial adsorption’’ 
method of, 131 
side chain theory of, in 
identity of anti- 
bodies, 309 
in toxin-antitoxin re- 
action, 144, et seq. 
working out of passive 
immunization on a 
quantitative basis 
by, 85 
Ehrlich-Sachs phenomenon 
in sensitization, 
180, 181 
Bordet and Gay’s inter- 
pretation of, 181, 
182 
Ehrlich’s “antiricin,” 95 
work on toxins and anti- 
toxins, 122, et seq. 
Eisenberg and Volk’s inter- 
pretation of ag- 
glutination, 257 
Eisenberg on residue anti- 
gen and antibody 
in precipitin reac- 
tion, 292 
Encephalitis, filtrable virus 
of, selection of 
tissues, 28 
Endocarditis, malignant, 27 
Endocomplement, 190 
Endolysins, 186, 366 
Endotoxins, 39, 44, 46, 79 
characteristics of, 18, 39, 
40, 44 
production of, in typhoid 
fever, 564 
toxic cleavage products 
of, 45 
Engelmann’s studies in 
chemotaxis, 348 
Enzymes, analogy of, with 
alexin or comple- 
ment, 155, 163, 
164, 192 
analogy of, with true 
bacterial toxins, 
17, 40, 42, 105 
difference between toxins 
and, 42 
in phagocytosis, endo- 
cellular and extra- 
cellular, 362, 366 
intestinal, and destruc- 
tive action on bac- 
terial poisons, 48 
leukocytic, 367, 630 
serum, 629 
Epiphanin reaction, 327 
328 
Epithelioid cells, action of, 
in phagocytosis, 
345 
Epitoxoids, definition of, 
128 
Epitoxonoids, 141 
Erythrocyte laking, 101 
Essential identity of anti- 
bodies, 317 
“Exhaustion theory’’ of 
immunity, 93 
Exotoxins, 39, 96. See also 
Toxins, true, 
bacteria producing, 40, 
44 
characteristics of, 41 
chemically indefinable 
nature of, 42 
immunization with, 83 
Fermentation, infectious 
disease and, 1 
intestinal flora and, 37 
microorganisms causing, 
1 
side-chain theory and ex- 
periments on, 146 
Ferments, protective, Ab- 
derhalden’s experi- 
ments with, 638, 
641 
analogy of, with alexin 
or complement, 
155, 156, 163, 164, 
192 
as antigens, 105 
immunization against, 
97, 167 
in animal body, 629 
Ferran’s investigations in 
active prophylac- 
tic immunization 
against cholera, 
585 
Ferrata, experiments of, in 
complement split- 
ting, with salt 
solution, 195, 197 
Ficker’s reaction in agglu- 
tination, 245 / 
Filtrable virus, cultural 
procedure. 11 
in encephalitis, 8 
in herpes, 8 
in smallpox, 8 
selection of special tis- 
sues by, 28 
“Fixateur” or sensitizer in 
phagocytosis, ac- 
tion of, in im- 
munized animals, 
363 
Fixed cells, intracellular 
digestion of for- 
eign substances 
by, 533, 534 
as phagocytes, 337, 341 
Flexner and Jobling’s work 
on serum therapy 
in cerebrospinal 
meningitis, 550, 
552 
Flexner and Lewis’ work 
on poliomyelitis, 
567, 568 
Flexner’s early observation 
on anaphylaxis, 
405 
on serum therapy in cere- 
brospinal menin- 
gitis, 550 
Fliigge school, 90 
bactericidal powers of 
blood, 94, 98 
Food idiosyncrasies, 469, 
478, et seq. 
analogy of, with protein 
anaphylaxis, 481, 
485 
congenital cases of, so- 
called, 481, 482 
Cook and VanderVeer’s 
work on, 478, et 
seq. 
hay fever and horse 
asthma classified 
with, 479 
passive transfer of hyper- 
susceptibility in 
cases of, 484 
possibility of inheritance 
of, 478, et 8eq. 
Mendelian principles 
in, 483 
Schloss’s studies on, 479, 
et seq. 
summary of, 485 
Forensic alexin fixation 
tests, 231 
Forensic blood examination, 
281 
Forensic complement fixa- 
tion, basic prin- 
ciple of, 322 
Forensic determination of 
unknown proteins, 
231 
delicacy of, 232, 233 
precipitin tests in, 280 
technique of, 232 
Fornet and Muller, ring 
test of, for pre- 
cipitin blood tests, 
281 
Friedberger, experiments of, 
in bacterial ana- 
phylaxis, 507, 509 
method of desensitization 
of, in serum sick- 
ness, 441, 477 
on the nature of bac- 
terial infections, 
508 
work of, on anaphylaxis, 
419, 433, 434, 464 
Friedemann, experiments 
of, on anaphylaxis 
in vitro, 464 
work of, on passive ana- 
phylaxis, 424, 425 
in rabbits, 459 
Garbat and Meyer’s work 
on serum therapy 
of typhoid fever, 
563, 564 
Gastric juice, action of, on 
stomach itself, 6 
Gastrotoxin, 102 
Gay and Adler’s experi- 
ments on two 
separate sub- 
stances in anaphy- 
lactic antigen, 448 
Gay and Southard’s objec- 
tions to antigen- 
antibody theory of 
anaphylaxis, 447 
theory of anaphylaxis, 
423, 446 
work on anaphylaxis, 
424 
on passive anaphy- 
laxis, 424, 425 654 
INDEX 
Gay’s sensitized killed vac- 
cines in prophy- 
lactic typhoid 
fever immuniza- 
tion, 583 
Gay’s work on injection of 
non-specific sub- 
stances in infec- 
tious diseases, 626 
Gengou, studies of, on 
alexin fixation 
with non-cellular 
antigens, 266, 207 
Giant cells in phagocytosis, 
118, 341 
foreign body, 118, 341 
tuberculosis, 341 
Glanders, alexin fixation 
test in diagnosis 
of, 235, 236 
in horses, agglutination 
reaction in diag- 
nosis of, 244 
relative susceptibility of 
man and animals 
to, 62 
uninjured conjunctival 
sac infected with, 
22 
Gold reaction, colloidal, 
329, 332 
Gonococcus, relative sus- 
ceptibility of man 
and animals to, 
62 
Gonorrheal infections, 
alexin fixation test 
in diagnosis of, 
236 
vaccine treatment in, 593 
Gordon types of meningo- 
cocci, 551 
Gottstein-Mathes’ work on 
serum therapy of 
typhoid fever, 563 
Gramenitski’s experiment 
in reversal of 
alexin inactivation 
by heating, 201, 
202 
Grohman on inhibition of 
bacterial growth 
by cell-free blood 
plasma in immun- 
ity, 89 
“Group reactions,” of bac- 
teria, as evidence 
of multiplicity of 
amboceptor, 169 
Gruber-Widal reaction in 
diagnosis, 242 
Haemotoxins, 17 
as true toxins, 40 
selective action of, 48 
Haffkine’s early work on 
prophylactic im- 
m u n ization 
against plague, 
587, 588 
Half parasites, 11, 76 
“Haptenes.” 110, 495, 504, 
517 
Haptines, 149 
of the second order, 257, 
288 
of the third order, 165 
varieties of, 149 
Haptophore group in toxin, 
126 
action of, 126 
Haptophore groups, in he- 
molysis, 161, 173 
Haptophore groups, in ag- 
glutination, 256, 
257 
Harris and Hhackell’s meth- 
od of treatment in 
rabies, 500, 601 
Hay fever and asthma, 
479, 485 
anaphylactic nature of, 
487 
desentitization in, 488-9 
differences between true 
anaphylaxis and, 
489 
due to a hypersensitive- 
ness to some mate- 
rial in pollen, 486 
history of observations 
in, 485, 486 
intracutaneous test in, 
488 
ophthalmic test in, 488 
pollen extracts in, 487 
production of anti- 
bodies in animals 
with, 488 
sensitization of guinea 
pigs with, 487 
treatment of, 488 
Heat alkali-precipitin, 284 
Heat-precipitins, 283 
Hemagglutination, agglu- 
tination by bac- 
teria by serum 
analogous to, 259, 
260 
Hemoglobinuria, paroxys- 
mal, 298 
autohemolysins in, 298 
hemolysis in, 298 
Hemolysins, anti-isolvsins 
in, 298 
as true toxins. 44 
autolysins in, 297 
isohemolysins in, 297 
specific, definition of, 102 
specific inciting of, in 
animal, 101, 102, 
113 
Hemolytic properties of 
normal serum, 
101 
Hemolytic reactions, com- 
plement deviation 
in, 179 
Hemolytic serum, agglu- 
tinins in, 104 
Ehrlich and Morgen- 
roth’s conception 
of neutralization 
of, by antilysin or 
anti- amboceptor 
reacting with 
eytophile group, 
167, 168 
precipitins in, 104 
Hemolytic substances, ori- 
gin of, from leuko- 
cytes, 185 
Hemolysis, alexin or com- 
plement in, 160, 
162 
alexin splitting and, 195, 
et seq. 
amboceptor in, action of, 
163, 164 
anti-amboceptor in, 167, 
168 
anti-cytophile interpreta- 
tion of antisensi- 
tization in (of 
Ehrlich and Mor- 
genroth), 168 
controversy on, 168 
Hemolysis, antisensitizer 
in, 167, et seq. 
anti-isolysins in, 298 
experiments of Leifmann 
and Cohn on, 163, 
164 
in immune serum, 160 
analogy of, to bacter- 
iolysis, 161 
Bordet’s work on, 160, 
161 
Ehrlich and Morgen- 
roth on mecha- 
nism of, 161 
haptophore groups in, 
161 
relation of antigen, 
amborceptor and 
complement in, 
161, 165 
multiplicity of ambocep- 
tor in, 165 
recapitulation of views 
of Ehrlich and 
Morgenroth on, 
168 
side chain theory, 161 
Ilepatotoxin, 102 
Hirschfeld’s studies on 
presence of iso-ag- 
glutinins in differ- 
ent races, 305 
Iliss and Zinsser’s work on 
injection of leuko- 
cytic substances in 
infectious dis- 
eases, 625 
Hiss, investigations of, on 
therapeutic use of 
leukocyte ex- 
tracts, 369, 370 
Hog cholera, immunization 
with bacterial 
products in, 17, 
82 
Ildgyes method of treat- 
ment in rabies, 
600 
Holobut’s work on bacterial 
anaphylaxis, 494 
Hopkins’ method of stand- 
ardization of vac- 
cines, 579 
‘‘Horror autotoxicus,” 
298 
Human isolysins. See Iso- 
lysins, 299 
Humoral theory, of im- 
munity, 90, 95, 
155 
of anaphylaxis, 428, 451, 
459 
Iluntoon's method of dis- 
sociation of anti- 
gen and antibody, 
316, 321, 323 
IIydrophobia>, active pro- 
phylactic immuni- 
zation against, 
597. See also un- 
der Rabies. 
Hyperleukocytosis, specific, 
370 
Hypersensitiveness phenom- 
ena, classification 
of, 409, 410 
Hypersusceptibility. See 
Anaphylaxis 
bacterial, and incubation 
time, 30 
food idiosyncrasies as 
forms of, 478, et 
seq. INDEX 
655 
Hypersusceptibility, inher- 
itance of, 407, 425, 
428, 483, 484 
varieties and classifica- 
tion of phenomena 
of, 408, et seq. 
virus, and incubation 
time, 30 
transference of, 425 
toxin, and anaphylaxis, 
510, 512 
Idiosyncrasies to non-anti- 
genic stubstances, 
477, et seq. 
Immune bodies, structure 
of, 145, 146 
Immune isolysins, 299 
Immune serum. See also 
Serum, immune, 
384 
agglutination in, 160 
bactericidal effect, 175, 
177 
bacteriolytic power of 
transferable, 156 
bacteriolytic properties 
of, Bordet’s find- 
ings in, 159, 161 
direct neutralization of 
toxin-antitoxin re- 
action, the protec- 
tive power of, 143, 
144 
hemolysis in, 160 
alexin or complement 
in, 162 
analogy of, to bac- 
teriolysis, 160, 161 
Bordet’s work on, 160 
Ehrlich and Morgen- 
roth on mecha- 
nism of, 161 
haptophore groups in, 
161 
relation of antigen, 
amboceptor and 
complement in, 
161, 165 
work of Ehrlich and 
Morgenroth on, 
162, et seq. 
work of Liefmann and 
Cohn on, 163, 164 
phagocytosis in, 101 
specific agglutination of 
bacteria in, 99 
precipitin formation 
in, 100 
Immunity. 58 
acquired, 70 
artificially, 73, 95 
definition of, 1, 3, 58, 
73 
history of, 71, 75 
increased phagocytosis 
and, 360 
active, relation to phago- 
cytosis to. 390 
antibacterial, 546 
cellular theory of, 154, 
155 
definition of, 3, 58, 73 
diphtheria, determination 
of, with diph- 
theria toxin, 538 
“high tide” of, 399 
humoral theory of, 155 
lasting, diseases in which 
one attack con- 
veys, 70 
diseases in which one 
attack does not 
convey, 71 
Immunity, local, in organs 
directly in contact 
with antigens, 117 
in skin infections, 118 
“passive” theories of, 
154, 155 
natural, 59 
agglutination in vivo 
in, 270, 271 
cellular theory of, 90 
definition of, 59, 73 
humoral theory of, 90 
inflammation in, 88 
mechanism of, 88 
theories concerning, 
88, 92 
bacterial destruc- 
tion by phago- 
cytic cells, 88, 
89 
bacterial growth 
in cell-free 
blood serum, 91 
bactericidal power 
of blood in nat- 
tural immunity, 
89 
bactericidal prop- 
erties of extra- 
vascular plasma 
or serum, 91 
inhibition of bac- 
terial growth by 
cell-free blood 
plasms, 89 
intracellular de- 
struction of bac- 
teria, 88 
mechanism of, 
theories con- 
cerning, phago- 
cytic activities 
of blood, 89 
phagocytic activi- 
ties of blood in, 
89 
principles of, 59 
body tempera- 
ture in, 59, 
60 
cultural condi- 
tions f<* bac- 
teria in' body 
a factor in, 
65 
increased inva- 
sive powers 
of bacteria in, 
65 
individual dif- 
ferences in, 
68 
inheritance in, 
65, 66 
racial differ- 
ences in, 64 
relative resist- 
ance of ani- 
mals in, 59, 
60 
serum, 9 
species resist- 
ance in, 59, 
60 
Pfeiffer phenomenon 
in. 99 
phagocytosis in, 100 
Immunization, 70 
against snake venoms, 
85. 541 
history of, 71, 72 
Immunization, active. See 
also Vaccine 
therapy, 571 
Immunization, active, 
against anthrax, 
75, 76 
against chicken cholera, 
73, 75 
against rabies, 77 
against small-pox, 72 
agglutination of bacteria 
in, 99 
in vivo, 271 
“alkalinity theory” of. 94 
antibacterial, 95, 98, 99 
antibodies in, 94 
bodies in, fundamental 
principles of 
theory of, 94 
origin of, 116, 117 
antitoxic, 95, et seq. 
as a therapeutic measure, 
action of, in gen- 
eralized systemic 
infections, 572 
in local infections, 571 
in successive local in- 
fections, 572 
value of, 570 
in acute diseases, 
574 
in subacute or 
chronic cases, 574, 
575 
as a prophylactic meas- 
ure (value of), 
95, 546, 570, 571 
autogenous vaccines, 576 
auto-inoculation by mas- 
sage in, 399, 400 
bacteria used in, 95, 98, 
et seq. 
bacteriolysins in, 99 
by means of livng but 
attenuated cul- 
tures, 75 
concentration of anti- 
bodies in lym- 
phatic organs in, 
117 
in other organs in, 117 
definition of, 74 
“exhaustion theory” of, 
93 
“high tide” of immunity 
in, 399 
in diphtheria, with toxin- 
antitoxin mix- 
tures, 138, 351. 
See also Diph- 
theria toxin-anti- 
toxin. 
invasion of bacteria in, 
mechanism of re. 
action in tissue 
cells against, 118 
locality of production of 
antibodies depen- 
dent on locality of 
antigen concentra- 
tion in, 117 
negative phase in, 396, 
397 
second injection in, 
396, 397 
successive inoculations 
in, 398 
summation of, 398 
non-bacterial anti-toxin- 
stimulating sub- 
stances in, 96, 97 
“osmotic theory” of, 94 
phagocytosis in, 101 
phenomena following, 93, 
et seq. 
precipitin formation in, 
100 656 
INDEX 
Immunization, active, re- 
action of tissue 
cells to invasion 
in, 118 
removal of spleen in, and 
antibody - forma - 
tion, 116, 117 
“retention theory” of 
93 
second positive phase in, 
398 
specificity of antibodies 
in, 94, 95 
summation of positive 
phase in, 398 
vaccines in, 576 
production of, 576 
sensitized, 580 
with dead bacteria, 
576 
with living bacteria, 
576 
standardization of, 576 
Hopkins’ method or, 
579 
Wright's method of, 
577 
with antigens, analogy 
between drug tol- 
erance and, 115 
with bacterial extracts, 
78 
extraction of bacteria 
for, by mechanical 
methods, 81 
by permitting them 
to remain in fluid 
media, 80 
with bacterial products 
(toxins), 82, 94 
with dead bacteria, 78 
methods used in killing 
bacteria for, 79 
with fully virulent cul- 
tures in sublethal 
amounts, 76 
with sensitized bacteria, 
78 
Immunization, active pro- 
phylactic, in man, 
570 
against anthrax, 593 
against cholera, 585. See 
also under Choi 
era. 
against influenza, 590. 
See also under In- 
fluenza. 
against paratyphoid 
fever, 580 
against plague, 587. See 
also under Plague, 
against pneumonia, 589. 
See also under 
Pneumonia. 
against rabies. 597. See 
also under Rabies, 
against small-pox, 594. 
See also under 
Small-pox. 
against syphilis, 602, 
613, 614 
against typhoid fever, 
580. See also un- 
der Typhoid fever. 
Immunization, passive, 67, 
74, 84, 95 
antitoxins in. 67, 85 
definition of, 74 
history of, 84 
in bacillary dysentery, 
545 
in diphtheria. See Diph- 
theria antitoxin. 
Immunization, passive, in 
diseases cuused by 
bacteria which do 
not form soluble 
toxins, 85, 545. 
See also under 
Serum therapy, 
quantitative basis of, 85 
therapeutic application 
of, 85, 96 
toxin-antitoxin reaction 
in, 120 
underlying principles of, 
74 
Immunized animals, anti- 
bodies in blood of, 
148 
bacteriolysis in, 156 
summary of facts in 
156, 157 
Inagglutinability, artificial 
production of, 249, 
250 
Incubation of bacteria, 29, 
30 
Infection, accidental fac- 
tors favoring, 24 
acquired rsistance to, 70 
adaptation of bacteria in 
tissues in, 7, 9, 11, 
12, 19 
aggressin secretion of 
bacteria in body 
and, 17 
body temperature and, 
20, 60 
capsule formation of bac- 
teria and, 14, 15 
chronic, adaptation of 
bacteria in, 8, 19 
classification of various 
forms, occurring 
in human body, 
571 
clinical manifestations of, 
19, 20, 25, 26, 33 
latent, 32 
concomitant, 24 
conjunctiva susceptible 
to, 22 
criteria governing, 3, 20, 
25 
“cryptogenic,” 32 
cultural conditions for 
bacteria in body 
and, 10, 65 
defense of intestinal 
tract in, 21, 22 
defense of mucous mem- 
branes in, 21. 22 
defense of skin in, 21, 22 
definition of, 3, 5, 7 
different, produced by 
same bacteria. 26 
effect of body tempera- 
ture on invasive 
powers of bacteria 
in, 2. 20 
effect of cultural adapta- 
bility of bacteria 
on, virulence of, 
9, 10 
effect of path of intro- 
duction of bac- 
teria on, 3, 21, 
22 
effect of symbiosis upon, 
30. 31 
effect of quantity of 
bacteria intro- 
duced on, 3. 23 
entrance of bacteria in 
body tissues in, 
7, 28 
Infection, focus in, 8 
from bacteria In blood 
stream, 8, 9, 27 
generalized, 9, 27 
increased invasive powers 
of bacteria a fac- 
tor in, 65 
incubation of bacteria in, 
29 
individual differences 
and, 65 
inheritance and resist- 
ance to, 65, 66 
localized, 25, 26 
accidental conditions 
governing, 24, 28 
reaction in, 29 
selective action of 
bacteria in, 28 
natural resistance against 
58 
of various diseases rela- 
tive susceptibility 
of man and ani- 
mals to, 20, 61 
protective action of blood 
serum against, 59, 
156 
protective action of 
leukocytes against, 
59 
protective action of tis- 
sues against, 59 
ptomains and, 33, 37 
ptomains as indirect 
cause of, 36 
racial differences and, 
64 
resistance of living cell 
to, 5, 6 
secondary abscesses in, 
27, 28 
secondary modifying fac- 
tors in, 3 
selective lodgment of 
bacteria in body 
and, 28, 48, 49 
similar, produced by dif- 
ferent bacteria, 
25, 26 
species resistance to, 59, 
60 
specificity of bacteria 
and. 25. 26 
“sub-infection,” 27 
susceptibility to, racial 
differences in, 64 
types of, 25 
variation in, of different 
strains of same 
bacteria, 12, 24 
variation in degree of, in 
bacteria succes- 
sively passed 
through animals, 
12, 13, 24 
without infectious dis- 
ease, 7 
Infectious disease, defini- 
tion of, 3, 5, 7, 
111 
functional importance of 
agglutination in, 
270 
influence of injections of 
non-specific sub- 
stances upon, 625 
Inflammation, process of, 
and phagocytosis, 
341, et 8eq. 
with pyogenic staphylo- 
cocci, 342 
with tubercle bacilli, 344, 
346 INDEX 
657 
Influenza, prophylactic vac- 
cination in, 590 
McCoy’s survey of, 591 
mixed vaccine in, 590 
relative susceptibility of 
man and animals 
to, 62 
Inheritance, a factor in re- 
sistance to infec- 
tion, 65, 66 
iso-agglutinins and, 68 
Inhibition zones in colloid 
reactions, 177 
in precipitation and ag- 
glutination, 177 
Inhibitors, 143 
Injured tissue, chemotaxls 
in relation to, 349 
infection in relation to, 
4, 5, 24, 28 
Intestinal tract, defense of, 
in infection, 21 
Iso-agglutinins, 299 
grouping of, 299 
in blood serum, 68 
inheritance of, 68, 301 
medico-legal use of, 
301 
transfusion and, 303, 
304 
value of presence of, 303, 
305 
Iso-antibodies, 297 
“anti-antibodies,” 298 
“anti-isolysins,” 298 
“nutohemolysins,” 297, 
298 
autolysins, 297 
“Horror Autotoxicus,” 
298 
immune isolysins, 299 
iso-agglutinins, 299 
grouping of, 299 
in blood serum, 68 
isohemolysins, 297 
laws of inheritance in 
transmission of, 
in humans, 301 
“Landsteiner” phenome- 
non of autohemol- 
ysis in hemoglo- 
binuria, 298 
Moss’s classification of 
human, 300 
paroxysmal hemoglo- 
binuria, 298 
work of Ehrlich and 
Morganroth on, 
297, et seq. 
Isohemolysins, 297. 299 
Isolysins, human, 299 
importance of presence 
of. 303, 305 
iso-agglutinins and, 299 
Isoprecipitins, 278, 279 
Jacobsthal’s ultramicro- 
scopic method of 
finding precipi- 
tates in syphilitic 
sera, 222, 229, 231 
Jansky’s classification of 
isolysins and iso- 
agglutinins in hu- 
man sera, 300 
Jenner, Edward, experi- 
mentation of, for 
immuniza tion 
against small pox, 
72 
Jobling and Petersen’s, 
work on anaphy- 
laxis, 466 
Job ling and Petersen's, 
experiments on 
anaphylatoxin pro- 
duction, 509, 510 
theory of anaphylaxis 
based on serum 
protease studies, 
636 
work on change in con- 
centration of 
serum enzymes, 
627 
Jochmann’s investigations 
in serum therapy 
of epidemic cere- 
brospinal menin- 
gitis, 549 
studies on proteolytic 
enzymes of leuko- 
cytes, 631, 633 
Kolle and Otto’s investiga- 
tions in prophy- 
lactic immuniza- 
tion against 
plague, 587 
Kolle and Wassermann’s in- 
vestigations of 
epidemic cerebro- 
spinal meningitis, 
549 
Kolle’s method of prophy- 
lactic vaccination 
in cholera, 586 
Kraus and Doerr’s study of 
bacterial anaphy- 
laxis, 493, 494 
Kraus, Rudolf, discovery of 
specific precipi- 
tins, by, 272 
Kraus and Stenitzer’s 
serum in treat- 
ment of typhoid 
fever, 563 
Kupfer cells, phagocytic 
action of, 92, 338 
“Landsteiner” phenomenon 
of autohemcdysis 
in hemoglobinuria, 
298 
“haptenes” of, 495, 504, 
517 
Le, definition of, 125 
method of determination 
of, 125 
toxin determination of, 
528 
Lo, definition of, 125 
constancy. 126 
method of determination 
of, 125, 528 
Ln-Limes necrosis—Roe- 
mer, 536 
Leishmann’s technique for 
determination of 
opsonic index, 390 
“Leistungskern,” definition 
of, 145 
Leprosy, relative suscepti- 
bility of man and 
animals to. 63 
Lesne and Dreyfus’ work 
on anaphylaxis, 
419 
Leukine, 366 
Leukocidins, 17 
as true toxins, 40, 44 
selective action of, 48 
Leukocyte extracts, thera- 
peutic use of, 368, 
370 
Leukocytes, alexin extrac- 
tion from, 365, et 
seq. 
capsulated bacteria and, 
15 
destruction of, 17 
growth of bacteria in, 
359 
in bacteriolysis, 159 
action of, 184 
in leukocytosis, action of, 
336; 337 
in phagocytosis, 386 
origin of bactericidal and 
hemolytic sub- 
stances from, 155, 
184, 365 
oxidase in, 633 
phagocytic powers of, 59 
proteolytic enzymes from, 
in phagocytosis, 
367, 368, 630, 631 
Leukocytic bactericidal sub- 
stances, 365, 633 
Leukocytosis, 351 
fever and, 33 
increased by injection of 
non-specific sub- 
stances, 627 
sources of leukocytes in, 
351 
Leukoproteases, 367, 368, 
631 
Leukotoxin, 102 
Liesenberg and Zopf, work 
on capsulated and 
n o n - c a psulated 
leukonostoc mesi- 
enteroides, 16 
Lipoid constituents of cells, 
relation of, to 
antigenic proper- 
ties, 113 
to toxins, 56 
Lister, on phagocytic ac- 
tivities of blood in 
natural immunity, 
89 
on prophylactic vaccina- 
tion in pneumonia, 
589 
Liver, production of alexin 
in, 189 
Loeb’s work on colloids, 
changing concep- 
tion of union of 
antigen with anti- 
body, 312, 313 
Lubarsch on bactericidal 
properties of ex- 
t r avascular 
plasma or serum 
in immunity, 91 
Liidke’s work on serum 
therapy in typhoid 
fever, 563 
Lustig’s antiplague serum, 
566 
Lysins, production of, 101, 
149 
principles of, 104 
production of antilysins 
by injection with, 
167, 168 
Macrocytase, 185, 362 
Magendie on anaphylaxis, 
405 
Mallein reaction, 497, 506 
Malta fever, relative sus- 
ceptibility of man 
and animals to, 
62 658 
INDEX 
Manwaring’s, work on ana- 
phylaxis, 437 
transfusion experiments 
on dogs, strength- 
ening cellular 
theory of anaphy- 
laxis, 452 
Markl’s serum in treatment 
of plague, 565, 
566 
Marmorek’s work on serum 
therapy of strep- 
tococcus infec- 
tions, 554, 555 
Mass action, law of, appli- 
cation of to neu- 
tralization of tox- 
in and antitoxin, 
135, et seq., 140, 
142, 143 
union of agglutinogen 
with agglutinin 
according to, 255, 
Maternal transmission, of 
antibodies, confer- 
ring a temporary 
passive sensitiza- 
tion, 478, 484 
of diphtheria antitoxin, 
67, 69 
of hypersusceptibility, 
407, 425, 428 
of immunity in abrin, 
anthrax and ricin, 
66 
tuberculosis and, 69, 70 
Measles, relative suscepti- 
bility of man and 
animals to, 63 
Meat poisoning, 4, 36 
Meinecke reaction for syph- 
ilis, 231 
Meiostagmin reaction, 328 
Ascoli and Izar’s experi- 
ments in, 328, 329 
value of, in diagnosis, 
328 
Meningitis, epidemic cer- 
ebrospinal, serum 
therapy of, 549. 
See also un- 
der Cerebrospinal 
meningitis, epi- 
demic. 
Meningococcus typing, in 
cerebrospinal men- 
ingitis, 551 
Metchnikoff and Besredka’s 
living sensitized 
vaccines for 
prophylactic ty- 
phoid immuniza- 
tion, 78, 582 
Metchnikoff, on agglutina- 
tion in acquired 
immunity, 270 
on bacterial growth in 
cell-free blood 
serum, 91 
theory of, on bacterial 
destruction b y 
phagocytic cells 
in natural im- 
munity, 88, 89 
work of, on poisons in 
circulation, 55, 
149, 150 
early investigations on 
phagocytosis, 333, 
et seq. 
soured milk therapy, 36 
Microcytase, 185, 362 
Minimum lethal dose, defi- 
nition of, 123, 124 
Minimum lethal dose, meth- 
od of determina- 
tion of, 123 
M L D, definition, 123, 124 
method of determination 
of, 123 
toxin determinations of, 
528 
Morgan, application of 
genetic reasoning 
to transmission of 
isoantibodies, 302, 
303 
Moigenroth’s toxin-HCl 
modification in 
toxin-antitoxin re- 
action, 122 
Mucous membranes, defence 
of, in infection, 
21, 22 
Mushroom, specific anti- 
toxin from, 112 
Narcotics, reduction of 
phagocytosis by, 
360 
Natural immunity. See 
Immunity, nat- 
ural. 
Nature of antibodies, 306 
Necroparasites, 10 
“Negative” phase in active 
immunization, 397 
second injection in, 398 
successive inoculations 
in, 398 
“summation” of, 398 
Neisser and Friedemann, 
experiments of on 
influence of salts 
on sensitized bac- 
teria in aggluti- 
nation, 264 
Neisser and Sachs, method 
of, for forensic 
determination of 
unknown proteins, 
232 
on basic principle of for- 
ensic complement 
fixation. 322 
Neisser and Wechsberg, 
phenomenon of, 
175, et seq. 
analogous to colloid re- 
actions, 177 
argument in favor of 
Bordet’s views, 
178 
Gay’s explanation of, 178 
Morgenroth and Sachs’s 
experiments sup- 
porting, 179 
pro-agglutinoid zone re- 
action analogous 
to, 177 
Neisser’s studies in syphi- 
lis, 609, 610 
Neoplasms, malignant, 
alexin fixation in 
diagnosis of, 233 
von Dungern’s method 
of, 234, 235 
antigen production for, 
234 
results of, 235 
technique of, 234 
Nephrotoxin, 103 
Neufeld and Haendel’s 
work on serum 
therapy of pneu- 
monia, 556 
Neurotoxin, 102 
in snake venom, 121, 542 
Nicolle’s theory of ana- 
phylaxis, 462 
work on passive anaphy- 
laxis, 424 
Noguchi’s, modification of 
the Wassermann 
test, 226 
acetone-insoluble antigen 
for Wassermann 
test, 218, 219 
schematic presentation 
of, 227 
Non-bacterial, antitoxin- 
stimulating sub- 
stances, 95, 96 
lysin production, 101 
Non-cellular antigens, alex- 
in-fixation with, 
206 
Non-specific therapy, 627 
method of stimulation, 
628 
“Normal” antitoxic serum, 
123 
“Normal” diptheria toxin, 
123 
“Normal” serum. See un- 
der Serum, 
agglutinins in, 102 
hemolytic properties of, 
101 
opsonins in, 102 
toxic action of, and ana- 
phylaxis, 101 
Northrop and deKruif, 
work of. on “co- 
hesive force” in 
a g g 1 u t i nation, 
267, 268 
Novy and deKruif’s studies 
on anaphylatoxin 
formation, 466, 
467 
Nuttall on bactericidal 
power of normal 
Mood in natural 
immunity, 89, 90 
experiments on deter- 
mining zoological 
classifications by 
means of precipi- 
tin reaction, 279 
experiments on effect of 
temperature on in- 
fection, 60 
Ophthalmia. sympathetic, 
Elschnig's expla- 
nation of, as ana- 
phvlactic reaction, 
517 
Opium, reduction of pha- 
gocytosis by. 360 
Opsonic action, phagocyto- 
sis due to, 375 
with patients’ own serum 
and leukocytes, 8 
Opsonic index, 389 
determination of, Leish- 
mann’s technique 
for, 390 
Simon. Lamar and 
Bispham’s tech- 
nique of, 393, 394 
Wright’s technique for, 
391, et seq. 
difficulties in, 393, 
394 
value of, 394 
fluctuation of, in un- 
treated patients 
under influence of 
exercise of dis- 
eased parts, 400 INDEX 
659 
Opsonic index, in autoin- 
oculations by mass 
sage, 400 
in sera of normal and 
infected individu- 
al, comparison of, 
395 
in serum therapy, com- 
parison between 
that in exudate of 
infected foci and 
blood serum, 400 
in staphylococcus infec- 
tions, 395 
during vaccine treat- 
ment with dead 
staphylo coccus 
cultures, 396 
in treatment of gonor- 
r h e a 1 arthritis 
with autoinocula- 
tion by massage, 
400 
in vaccine therapy, im- 
provement and, 
401 
of acne, 399 
of staphylococcus fu- 
runculosis, 396 
of tuberculosis, 401, 
403 
value of, in control- 
ling therapeutic 
vaccinations, 403 
relation of, to clinical 
condition, 394 
vaccine therapy and, 389, 
394, et seq. 
value of, in therapeusis, 
397 
Opsonic powers of normal 
serum, 376 
reduction of, by heat, 
377 
Opsonins, 373. See also 
Phagocytosis, 
definition of, 375 
immune bactericidal sen- 
sitizers and, 383, 
et seq. 
increase of, 377 
heated, increase of 
power of, by addi- 
tion of fresh nor- 
mal serum, 380 
reactivation of, by 
addition of alexin, 
380 
normal and, 382, et 
seq. 
resistance of, to heat, 
377 
specificity of, 383 
thermostability of, 382 
normal, 378 
cooperation of heat- 
stable and heat- 
sensitive body in, 
381 
instability of. 377 
nature of, 378 
similarity of, to alexin 
or complement, 
378, 379 
specificity of, 380 
qualitative difference be- 
tween normal and 
immune, 377, 378 
specific thermostable, in 
normal serum, 
379 
Organ specificity, of anti- 
gens, 114 
of precipitins, 286, 287 
“Osmotic theory” of im- 
munity, 94 
Ottenberg, work on inheri- 
tance of iso-anti- 
bodies according 
to Mendelian laws, 
301, 303 
Otto’s work on anaphy- 
laxis, 420 
on anti-anaphylaxis, 439 
on passive anaphylaxis, 
424, 425, 426 
Oxidase, in proteolytic en- 
zymes, 633 
Pancreas cytotoxin, 102 
Panum’s theory of intra- 
cellular destruc- 
tion of bacteria 
in natural im- 
munity, 88 
Parasites, biological transi- 
tion of sapro- 
phytes to, 5 
Bail’s classification of, 
10 
Parasitic bacteria, 3, 4, 5, 
20 
Parasitism, as form of spe- 
cific adaptation, 5 
Paratyphoid fever, agglu- 
tination reaction 
in diagnosis of, 
243 
immunization against, 
580 
vaccination against, 583 
Paroxysmal hemoglobin- 
uria, 298 
hemolysis in, 298 
Pasteur, “exhaustion the- 
ory” of, 93 
experimentation, of, on 
immu nization 
against chicken 
cholera, 73 
experiments of, with 
“rouget” organ- 
ism. 65, 75 
on immunization against 
anthrax, 74 1 
on prophylactic immhni- 
zation in rabies, 
597 
work of, on rabies, 12, 
65 
work of, on bacterial fer- 
ments, 146 
Partial absorption method 
of Ehrlich in 
measurement of 
toxin - antitoxin 
combination, 131 
Pathogenic bacteria, adap- 
tation of, in tis- 
sues, 7, 9, 11, 12, 
19, 20 
antitoxic immunity for, 
97 
entrance of, in body tis- 
sues, 6, 21, 22 
formation of toxic prod- 
ucts by, 38 
ptomains and, 35, 36 
saprophytic nature of 
certain, 3, 4, 5 
specificity of, 25, 26 
Pathogenic microorganisms, 
definition of, 2, 3 
formation of toxic prod- 
ucts by, 18, 38, 45 
occurrence of, 3, 4, 5, 
20 
ptomains and, 35, 36 
Pathogenic microorganisms, 
resistance of liv- 
ing cell to, 6 
specificity of, 25, 26 
uncultivated, 28 
Pearce and Bisenbrey’s 
work on anaphy- 
laxis, 436, 437 
transfusion experiments 
on dogs, strength- 
ening cellular the- 
ory of anaphylax- 
is, 452 
Persensitized cells, 197, 200 
Petersen’s study on non- 
specific resistance 
and protein ther- 
apy, 626, 629 
Peterson’s investigations 
on therapeutic use 
of leukocyte ex- 
tracts, 369 
Pfaundler’s thread reaction 
in agglutination, 
245 
Pfeiffer, endotoxin theory 
of, 40, 45, 506 
immunization with dead 
bacteria practiced 
by, 79 
on causes of bacterial 
anaphylaxis, 506 
“Pfeiffer phenomenon” in 
active immuniza- 
tion, 98, 99 
in bacteriolysis, tech- 
nique of, 156 et seq 
Metchnikoff’s view of 
phagocytosis in 
peritoneal exudate 
and, 363 
work of, on anaphylaxis 
433 
Phagocytes, 337 
as cellular defenses, 15 
destruction of, 343 
fixed, 337 
macrophages, 338 
microphages, 338 
motile, 337 
Phagocytic index, 391 
Phagocytosis, 333 
acquired immunity, and, 
359 
alexin extraction in, from 
leukocytes and 
lymphatic organs, 
365, et seq. 
analogy of, with intra- 
cellular digestion 
among unicellular 
forms, 334 
chemotaxis in, 346 
influence of bacteria 
in, 349, 350 
influence of bacterial 
extracts in, 39, 
349, 350 
malic acid in, 347 
of slime-molds or 
myxomycetes, 347 
of spermatozoa of ferns, 
347 
Pfeflfer’s technique in, 
347 
destruction of bacteria 
in 358, 361 
by alexin (or cytase) 
in leukocytes, 362, 
363 
action of, 363 
Metchnikoflf’s inter- 
pretation of, 363 
by exudates, 361 660 
INDEX 
Phagocytosis, destruction 
of bacteria in, by 
phagocytes, 361, 
633 
destruction of red blood 
cells by, 337 
difference in degree of, 
due to bacteria, 
387 
differences in phagocytic 
energy in, due to 
leukocytes in, 386 
differences in virulence 
of bacteria, de- 
pendent on their 
resistance to leu- 
kocytes in, 18, 387 
digestion among proto- 
zoa and, 335 
“dust cells” in, 340 
early investigations in, 
333 
enzymes in, endocellular 
and extracellular, 
362, 366 
eosinophile cells in, 339 
“epithelioid cells” in, 
345 
factors determining, 373 
fixateur or sensitizer in, 
action of, in im- 
munized animals, 
362, 363 
giant cells In, 118, 341 
foreign body, 118, 341 
tuberculosis, 341 
in daphnia, 89, 334, 357 
in higher animals, 357 
in immune serum, 377 
bacteriolysins in, bac- 
tericidal sensi- 
tizers and, 383, et 
seq. 
heated, opsonic action 
in. increase of, by 
addition of fresh 
normal serum, 380 
increase of, 373 
attributed to “stimu- 
lins,” 373 
with addition of leu- 
kocytes, 374 
opsonin contents a fac- 
tor in, 375 
opsonins in, increase 
of, 377 
normal opsonins 
and, 382, et seq. 
specificity of, 383 
thermostability of, 
382 
in immunity, 101 
in normal serum, op- 
sonins in, cooper- 
ation of heat-sen- 
sitive body in, 
381 
nature of, 378 
similarity of, to 
alexin, 378, 379 
specific thermos- 
table. 379 
specificity of, 380 
in process of inflamma- 
tion, 841, ct seq. 
with pyogenic staphy- 
lococci, 342 
with tubercle bacilli, 
344, 346 
increase of, by injection 
of leukocyte ex- 
tracts, 369 
jn increased resistance, 
.390 
Phagocytosis, increase ot, 
with acquisition 
of immunity, 360 
intracellular digestion 
and, 335 
in vertebrates, 336 
Kupfer cells in, 338, 341 
leukocytes in, 386 
action of, 336, 337 
polynuclear, 339 
leukocytosis in, 351 
bacteria decreasing, 18, 
351 
bacteria increasing, 18, 
351 
lymphocytes in, large, 
339 
measure of degree of, in 
active immuniza- 
tion, 389 
Leishmann’s technique 
for, 390 
Simon, Lamar and Bis- 
pham’s technique 
for, 393, 394 
value of, in thera- 
peusis, 397 
Wright’s technique for, 
391 
difficulties in, 393 
value of, 394 
mechanism of process 
of, 341, ct seq. 
Metchnikoff’s early in- 
vestigations on, 
89, 91, 333, et seq. 
normal and immune 
opsonic action in, 
quantitative dif- 
ferences between, 
385 
normal degenerative and 
retrogressive proc- 
esses and, 337 
observation of, in vitro, 
375 
of microorganisms, with 
or without culture 
media, 358 
of undissolved foreign 
particles in circu- 
lation, 111 
opsonins in, 373. See 
also Opsonins. 
phagocytes engaged in, 
varieties of, 337 
fixed, 337 
macrophages, 338 
microphages, 338 
motile, 337 
process of inflammation 
in. 341. et 8Cq. 
proteolytic enzymes from 
leukocytes in, 367, 
368 
qualitative difference be- 
tween normal and 
immune opsonic 
substances in, 377, 
378 
quantitative measure- 
ment of, 375 
reduction of phagocytic 
activity in, 359 
by capsule formation, 
15 
by growth of bacteria 
with leukocytes, 
359 
by protection of bac- 
teria from phago- 
cytes, 18, 360 
relation of, to active im- 
munity, 389, 390 
Phagocytosis, relation of 
virulence to, 374 
removal ot extravasa- 
tions of blood and, 
336 
resistance of bacteria to, 
due to non-absorp- 
tion otf opsonin, 
388 
resistance of encapsu- 
lated bacteria to, 
388 
resistance of infected 
subject and, 357, 
358 
resistance of virulent 
bacteria to, in 
normal serum, 387 
spontaneous, 375, 376 
tissue cells in, 92, 118, 
339 
varieties of body cells 
engaged in, 339 
dependent on nature 
of invading sub- 
stance, 340 
Phagolysis, 186 
Photoinactivation of alexin, 
203 
Pick and Yamanouchi’s ex- 
periments on two 
separate sub- 
stances in ana- 
phylactic antigen, 
448 
work on anaphylaxis, 414 
von Pirquet and Schick’s 
studies of serum 
sickness, 470, et 
8eq. 
von Pirquet’s tuberculin 
skin reaction, 499 
I’lacentar eytotoxin, 102 
Plague, active prophylactic 
immunizat ion 
against, 587 
Besredka’s vaccines in, 
588 
Ilaffkine’s early work 
on, 587 
Kolle and Otto’s inves- 
tigations in, 587 
Rowland’s vaccine in, 
588 
Strong’s investigations 
in, 587, 588 
relative susceptibility of 
man and animals 
to, 62 
Plague, serum therapy of, 
Dean’s work on, 566 
Dustig’s work on, 566 
Markl’s serum in, 565, 
566 
Rowland’s serum in, 
566 
value of. 567 
Yersin, Calmette nnd 
Borrel’s investiga- 
tions in, 564 
swine, virulence of, 12 
Yersin’s serum in, 564 
value of, 564, 565 
“Plasmines,” 82 
Pneumococci, agglutinating 
antibodies with, 
140 
Pueumococcus antibody ex- 
tracts of Iluntoon, 
in treatment of 
pneumonia, 560 
severity of immediate 
reaction in, 561 
therapeutic value of, 561 INDEX 
661 
Pneumococcus infection, 
relative suscepti- 
bility of man and 
animals to, 63 
Pneumonia, agglutination 
reaction in type 
diagnosis of, 243 
antibody extracts of Hun- 
toon, in treatment 
of, 560, 561 
prophylactic vaccination 
in, 589 
Cecil’s work on, 589, 
590 
Lister’s work on, 582 
serum therapy in, 556 
Cole’s work on, 556 
Dochez and Gillespie’s 
work on, 557 
nature of action in, 
556, 560 
Neufeld and Haendel’s 
work on, 556 
results of, 559 
serum sickness in, 559, 
560 
standardization of 
serum used in, 558 
typing of pneumococci 
essential in, 557 
Poison Ivy, specific anti- 
toxin, 112 
Poliomyelitis, Amoss’s work 
on, 568, 569 
cause of, 49, 568 
Mexner and Lewis’s work 
on, 567, 568 
infection and immunity 
in, 567 
Landsteiner and Popper’s 
work on, 567 
relative susceptibility of 
man and animals 
to, 64 
Polyceptors (Ehrlich), 171, 
172, 203 
Precipitation, 272 
inhibition zones in, 177 
syphilitic sera and anti- 
gens giving, 222, 
229, 231 
Precipitinogen, 273 
chemical nature of, 273, 
275 
effect of heating on, 273 
non-protein, 273, 274 
nature of, 274 
Precipitinoids, 289 
Precipitin reaction, 272 
against heated proteins, 
282 
coctoprecip itin in, 
283, 284 
experiments on, 285 
heat - alkali - precipitin 
in, 284 
agglutination reaction 
analogous to, 288 
analogy of, to colloidal 
flocculation, 289, 
et seq. 
autocytotoxins in, 288 
bacterial precipitins in, 
partial or minor, 
276. 278 
specificity of, 275 
von Dungern’s views of, 
292, 293 
effect of mixed sera in, 
293 
Ehrlich’s conception of, 
288 
electrolytes in, effect of, 
289 
Precipitin reaction, foren- 
sic blood test in, 
281 
ring test of Pornet 
and Muller in, 281 
group reactions of bac- 
terial precipitins 
in, 275 
diagnostic value of, 
276, 277 
heated precipitating 
serum in, 290, 291 
protective action of, 
290 
inhibition zones in, 290 
isoprecipitins in, 279 
medico-legal value of, 
278 
non-specific partial reac- 
tions in, elimina- 
tion of, 278 
precipitins in, delicacy 
of, 277, 278 
determination of po- 
tency of, 277, 278 
inactivation of, by 
heat, 289 
effect of in bacterial 
filtrates, 289 
production of, against 
unformed pro- 
teins, 277, 278 
methods of, 275 
of specific, 275 
by peptone, 275 
effect of heating on, 
273 
in animal sera by 
foreign protein, 
272 
structure of (Ehrlich), 
288 
zymophore group in, 
289 
effect of heat on, 
289 
production of antigen 
for, 80 
quantitative proportions 
in, effect of, 289, 
290 ' 
relative concentration of 
reacting bodies a 
factor in, 289, 290 
residue antigen and anti- 
body in, 291, et 
seq. 
explanations of, 292, 
et seq. 
experiment on, 294 
salts in, effect of, 289 
species determination by 
means of, 277, 278 
species specificity in, 
285, 286 
specificity of, 272, 275, 
278 
vegetable proteins de- 
termined by, 280 
zoological classifications 
by means of, 279 
Precipitin tests, methods 
of performing, 280 
forensic blood test in, 
281 
in diagnosis of glanders, 
277 
ring test of Fornet and 
Miiller in, 281 
Precipitins, 272 
against heated protein, 
282, et seq. 
bacterial, group reactions 
in, 275 
Precipitins, Partial or 
minor, 276, 278 
specificity of, 100, 275 
delicacy of, 277, 278 
determination of potency 
of, 277, 278 
formation of, in serum 
sickness, 473 
heat, 283 
in hemolytic serum, 104 
inactivation of, by heat, 
289 
effect of, in bacterial 
filtrates, 289 
isoprecipitins, 279 
organ specificity of, 286, 
288 
production of, 149, 273 
against unformed pro- 
teins, 277, 278 
methods of, 275 
specific, by peptone, 
275 
effect of heating on, 
273 
in animal sera by 
foreign protein, 
272 
“species,” specificity 
of, 285, 286 
specific, 272 
discovery of, by Ru- 
dolf Kraus, 80, 
272 
structure of (Ehrlich), 
288 
“Precipitoids,” 291 
Pregnancy, diagnostic value 
of Abderhalden’s 
protective fer- 
ments in, 640 
Pro-agglutinoid phenome- 
non in agglutina- 
tion explained as 
protective colloid 
reaction, 259 
Pro-agglutinoids, 258 
Pro-agglutinoid zone in 
agglutination, 177 
complement deviation re- 
action analogous 
to, 177 
Profeta’s law, 603 
Prophylactic immunization, 
active, in man, 
against anthrax, 
593 
against cholera, 585. 
See also under 
Cholera. 
against influenza, 590. 
See also under 
Influenza. 
against paratyphoid 
fever, 580 
against plague, 587. See 
also under Plague, 
against pneumonia, 589. 
See also under 
Pneumonia. 
against rabies, 597. See 
also under Rabies, 
against small-pox, 594. 
See also under 
Small-pox. 
against syphilis, 602, 
613, 614. See also 
under Syphilis, 
against typhoid fever, 
580. See also un- 
der Typhoid fever. 
Prophylactic immunization 
of dogs against 
rabies, 601, 602 662 
INDEX 
“Protection,” of Noguchi, 
213 
Protein anaphylaxis, anal- 
ogy of food idio- 
syncrasy with, 
481, 485 
“Protein fever,” 434 
Protein hypersensitiveness, 
<or true anaphy- 
laxis, 411 
Proteins, altered, in ana- 
phylaxis, 413 
cleavage products, 412 
Proteins, alexin fixation 
test in determina- 
tion of nature of, 
231 
delicacy of, 232 
technique of, 232, 233 
“altered,” 413, 414 
antigenic properties and, 
112 
cleavage products of, 
107 
in anaphylaxis, 412, 
et seq. 
non-specific effects of, 
625 
species specificity of, 
changed to speci- 
ficity of chemical 
structure, 108, 
109 
true, without antigenic 
properties, 106, 
412 
Protein therapy, 626 
choice of substances for 
use in, 628 
contra-indications in, 628 
favorable effects of, 627 
proper dosage in, 628 
protein-split products in, 
preferable to vac- 
cines, 628 
symptoms following in- 
jections in, 627 
Proteolysis, 630 
Proteotoxin, proteolysis in 
formation of, 630, 
635 
Protoxoids, 131 
Protozoa, digestion in, and 
its relation to 
phagocytosis, 335 
Ptomains, as indirect cause 
of infection, 34 
chemistry of, 34 
classification of, 34 
definition of, 36 
non-specificity of, 36, 37 
relation of, to infection, 
34 
Pus, formation, 343 
digestive action of, 630, 
631 
Putrefaction, bacterial 
metabolism and, 
93 
cnemistry of, 34 
microorganism causing, 
1 
Pyemia, 28 
Rabies, active prophylactic 
iminunizat ion 
against, 597 
Harris and Shackell’s 
method of treat- 
ment in, 600, 601 
Hogyes method of treat- 
ment in, 600 
in dogs, 601 
Rabies, Pasteur’s work on, 
12, 597 
preparation and attenua- 
tion of virus for, 
12, 598 
Stimson’s scheme of 
treatment in, 600 
treatment of patients in, 
599 
Ranzi’s work on anaphy- 
laxis, 434 
Rattlesnake poison, anti- 
toxin for, 542 
Reaction, colloidal gold, 
329, 332 
epiphanin, 327, 328 
immunity, 145 
meiostagmin, 328, 329 
toxin-antitoxin, 120 
Receptors, cell, overproduc- 
tion of, 148, 168 
complementophile, 164, 
167 
cytophile, 164, 167 
definition of, 146 
“double,” 165, 167 
of third order, 165 
sessile, in anaphylaxis, 
450 
in toxin hypersuscepti- 
bility, 511 
structure of, 147 
Resistance. See also under 
Immunity, 
acquired, 70 
bactericidal properties in 
serum and, 16, 59, 
89, 358 
body temperature and, 
20, 59, 60 
cellular theory of, 90 
cultural conditions for 
bacteria in body 
and, 11, 65 
degree of phagocytosis 
and, 29, 357, 358 
humoral theory of, 90 
increased invasive powers 
of bacteria and, 
13, 14, 29, 65 
individual differences 
and, 68 
inheritance a factor in, 
65, 66 
local, in skin infections, 
118 
natural, against infec- 
tion, 59 
non-specific, 626 
racial differences in, 64 
species, to infection, 59 
Resistance, vital, 629, 
630 
“Retention theory of im- 
munity,” 93 
Rhus toxicodendron, specif- 
ic antotoxin from, 
112 
Richet and Hericourt’s 
work, on anaphy- 
laxis, 405, 406 
on passive immunization, 
84 
Richet and Portier’s work 
on anpliylaxis, 
405, 406 
Richet’s work on passive 
anaphylaxis, 424 
Ricin, “protein-free,” 112 
“Root-tubercule” bacilli, 8 
Rosenau and Anderson, re- 
searches of, in 
bacterial anaphy- 
laxis, 493 
Rosenau and Anderson, 
work on anaphy- 
laxis, 407, 413, 
415, 418, 420 
Ilosenow, standardization 
of antitoxin in 
United States by, 
124 
Rowland’s antiplague 
serum, 566 
vaccine in prophylactic 
iminunizat ion 
against plague, 
588 
Russell’s vaccine for pro- 
phylactic immuni- 
zation against 
typhoid fever, 581 
Sachs-Georgi reaction for 
syphilis, 230, 231 
Salt-inactivation of alexin, 
193, 194, 201 
Salts, effect of, in ag- 
glutination, 264 
Saprophytes, biological 
transition of, to 
parasites, 5 
classification, 10 
formation of ptomains 
by, 36, 37 
occurrence of, 4, 5 
pathogenic powers of cer- 
tain, 3, 4 
pure, 10 
Saprophytic microorgan- 
isms, definition of, 
2, 10 
Scarlet fever, relative sus- 
ceptibility of man 
and animals to, 63 
Schick reaction, 539 
Schloss’s work on food idio- 
syncrasy, 479, et 
seq. 
Secondary invasion, 24, 25, 
31 
Sensibilisin, 447 
Sensibilisinogen, 447 
Sensitization, 153, 405 
alexin splitting and, 197, 
et seq. 
Bordet’s views on, 174, 
175 
complement deviation in. 
175 
Ehrlich and Sachs’s 
phenomenon in, 
180, 181 
Bordet and Gay’s in- 
terpretation of, 
181, 182 
Ehrlich and Sachs’s 
views on, 180-81 
Ferrata, work of, on, 
195, 196 
local, 479 
Neisser-Wechsberg phe- 
nomenon in, 175 
varying degrees of in- 
tensity of, 140 
Sensitization in anaphy- 
laxis, active, 517 
difference in methods of, 
420 
duration of hypersensi- 
tive state, 423 
in dogs, 421, 422 
in guinea pigs, 421 
in rabbits, 421 
quantity of antigen for, 
423 
passive, 423 INDEX 
663 
Sensitization in anaphy- 
laxis, passive, defi- 
nition of, 424 
direct relationship be- 
tween concentra- 
tion of antibodies 
in a serum and its 
ability for, 429 
Sensitized bacteria, im- 
munization with, 
78 
Sensitized vaccines, 576, 
580 
Sensitizer. See Amboceptor 
Bordet’s definition of, 
174, 175 
quantitative determina- 
tion of, in immune 
serum, 175, 176 
specific, 205 
“Sepsins,” 87 
Septicemia, chronic, adap- 
tation of bacteria 
in, 8 
clinical uniformity of, 26 
definition of. 27 
result of, 27 
secondary foci in, 8 
therapeutic immunization 
in cases of, 547, 
548 
vaccine therapy in, 572, 
573 
Serotoxin, proteolysis in 
formation of, 630, 
635 
Serum. See also Blood 
serum. 
activities of, 629 
antitoxic, direct effect of, 
on toxin, 120 
indirect protective ac- 
tion of, against 
toxin, 120 
bactericidal powers of, 
59, 89 
bactericidal properties of, 
in immunity, 91, 
153 
resistance and, 358 
cell-free, bacterial growth 
in, 91 
enzymes in, 534 
immune, bacteriotropins 
in, without lysins, 
384 
heated, reactivation of, 
by addition of 
alexin, 380 
opsonins in, bacteri- 
cidal sensitizers 
and, 383, et seq. 
heated, increase of 
power of, by addi- 
tion of fresh nor- 
mal serum, 380 
increase of, 377 
normal opsonins 
and, 382, et seq. 
specificity of, 383 
phagocytosis in, 101, 
377 
increase of, 373. See 
also under Phago- 
cytosis. 
normal, agglutinins in, 
100 
a n t i c o m plementary 
properties of, 214 
hemolytic properties 
of, 101 
opsonic powers of, 376 
reduction of, by 
heat, 377 
Serum, normal, opsonins 
in, 101 
cooperation of heat- 
stable and heat- 
sensitive body in, 
380, 381 
nature of, 378 
resistance of, to 
heat, 377 
similarity of, to 
alexin, 378, 379 
specificity of, 380 
thermostability of, 
382 
Serum, normal, specific 
thermostable op- 
sonins in, 379 
normal antitoxic, diph- 
theria, 123 
opsonins in normal and 
immune, qualita- 
tive differences in, 
377, 378 
reactions, physical fac- 
tors in, 325 
Serum, sickness, 469 
accidental sensitization 
and, 475 
analogy of anaphylaxis 
with, 472 
antibody formation in, 
473 
desensitization in, 476 
apparent difficulty of, 
in humans, 476 
intradermal skin test 
showing necessity 
of, in individuals, 
477 
factors determining sev- 
erity of, 471, 472 
incubation time in, 473 
methods of administra- 
tion of antitoxin 
to avoid, 441 
von Pirquet and Schick’s 
studies of, 470, et 
seq. 
symptoms of, 469, 470 
accelerated reactioh of 
von Pirquet and 
Schick in, 470 
after first injection, 
469 
after second injection, 
470, 471 
immediate reaction in, 
471 
analogy of, to ana- 
phylaxis, 471 
Serum therapy, anaphy- 
laxis in. See 
Serum sickness, 
in anthrax, 567 
in diphtheria, 519. See 
also under Diph- 
theria antitoxin, 
in diseases caused by 
bacteria which do 
not form soluble 
toxins, 545 
action of serum upon 
extensive infection 
in, 546 
antibacterial action in, 
546 
in epidmie cerebrospinal 
meningitis, 549. 
See also under 
Cerebrospinal men 
ingitis epidemic, 
in plague, 564. See also 
Plague, serum 
therapy of. 
Serum therapy, in pneu- 
monia, 556. See 
also Pneumonia, 
serum therapy in. 
in streptococcus infec- 
tions, 553. See 
also under Strep- 
tococcus infections 
in typhoid fever, 561. 
See also Typhoid 
fever, serum 
therapy of. 
Side chain theory, 143 
antibody production in 
body cells in, 149 
body cell in, 144 
chemical nature of, 
145 
“Leistungskern” In, 
145 
side chains or recep- 
tors in, 145, 146 
chemical action of anti- 
gens in, 148 
concerning tissue cell in 
formation of anti- 
bodies, 309 
definition of side chains 
in, 145 
diagram showing cell- re- 
ceptors and im- 
mune bodies (Ehr- 
lich) in, 147 
in toxin-antitoxin reac- 
tion, 143, 144 
overproduction of recep- 
tors in, 148 
flow of, into blood a 
cause of immu- 
nity, 148 
physical mechanism of, 
144 
recapitulation of, 149 
relationship between sus- 
ceptibility of tis- 
sue and toxin- 
binding properties 
in, 150 
Skin, defense of, in in- 
fection, 21, 22 
Skin infections, local im- 
munity in, 118 
Smallpox, active prophy- 
lactic immuniza- 
tion against, 594 
efficiency of, 597 
“green” virus in, 596 
Jenner’s discovery of, 
594 
method of vaccination 
important in, 596, 
597 
production and prepa- 
ration of vaccine 
for, 595, 596 
relative susceptibility of 
man and animals 
to, 63 
vaccination for, history 
of, 72 
principles of, 72 
Smith, Theobald, phenome- 
non of, in anaphy- 
laxis, 407 
Snake venolns, 41 
action of, 541, 542 
activation of, by endo- 
eomplement in 
blood cells, 190 
by sera, 190 
antitoxins for, 41, 96, 
541 
effect of heat on, 42, 
121 664 
INDEX 
Snake venoms, immuniza- 
tion against, 85, 
541, 542 
Kyes’ experiments in, 
190, 191 
neutrotoxins in, 542 
toxin-antitoxin combina- 
tion in, stability 
of, 121 
toxin-antitoxin reaction 
with, 121, 541 
filtration experiments 
in, 121 
neutralization theory 
of, 121 
time element in, 121 
toxin-HCl modification 
of, effect of heat 
on, 122 
Sol state, 262 
Species resistance, 59, 60 
Specific hyperleukocytosis, 
370 
Specificity, definition of, 86 
immunization and, 165 
of species, 108, 109 
Spermatotoxins, 102 
Spleen, removal of, and 
antibody forma- 
tion in active im- 
munity, 116, 117 
and susceptibility to in- 
fection, 116, 117 
Standardization of anti- 
toxin, guinea pigs 
used in, 123 
minimum lethal.dose in, 
123 
Standardization of diph- 
theria antitoxin, 
by means of toxin, 
123, 124 
early attempts at, 123 
unit in, 123, 126 
Standardization of teta- 
nus antitoxin by 
means of toxin, 
123 
Standardization of vac- 
cines, 577 
Hopkins’ method of, 579 
Wright’s method of, 577, 
578 
Staphylococcus furuncu- 
losis, opsonic in- 
dex in vaccine 
treatment of, 396 
Staphylococcus (infections, 
opsonic index in, 
395, et seq. 
relative susceptibility of 
man and animals 
to, 63 
use of vaccines in, 395, 
591, 593 
Stimulins, 373 
Strauss Test, 28 
Streptococci, in latent con- 
dition, 32 
symbiosis and, 30, 31 
variations in, 554, 556 
variations in viru- 
lence of, 12 
“X” substances formed 
by, 46 
Streptococcus infections, 
agglutination re- 
action in, 244 
injured tissue and, 24 
latent state of, 32 
relative, of man and ani- 
mals, 63 
secondary, 25 
Streptococcus infections, 
serum therapy in, 
553, et seq. 
nature of action in, 
555 
standardization of serum 
in, 556 
symbiosis and, 30, 31 
toxin formation' in, 46 
use of vaccines in 591 
value of, 556 
Sub-infection, 27, 256 
Surface tension in chemo- 
taxis, 353, 354 
Susceptibility, body tem- 
perature and, 20, 
59, 60 
cultural conditions for 
bacteria in body 
and, 20 
increased invasive powers 
of bacteria and, 
65 
individual differences 
and, 68 
inheritance a factor in, 
65, ct seq. 
racial differences in, 64 
relative, of animals to 
infection, 59, 60 
species resistance to in- 
fection and, 59, 60 
Syntoxoids, 131 
Syphilis, antibodies in, 619 
autoinoculation in, 604, 
605 
Colles’ law, 602 
colloidal gold reaction in, 
329 
diagnosis of by alexin 
fixation, 216, 217, 
616 
by direct precipitation 
of syphilitic serum 
by emulsions of 
lecithin and of 
sodium glycoeho- 
late, 222 
direct precipitation meth- 
ods in diagnosis 
of, 229 
immunity in, 602, 606, 
612, 615 
infection in, 22 
Meinecke reaction for, 
231 
phagocytosis in, 618 
Profeta’s law, 603 
reinfection in, 606 
reinoculation in, 603 
relative susceptibility of 
man and animals 
to, 61, 609, 611 
Sachs-Georgi reaction 
for, 230, 231 
superinfection in, 603 
transmission of, to ani- 
mals, 608, 609 
ultramicroscopic method 
of finding precipi- 
tates in sera of, 
222 
vaccination against, 615 
Wassermann reaction in, 
diagnostic value 
of, 228, 616, 617 
technique of, 225 
Bauer’s modification 
of, 227 
Noguchi’s modifica- 
tion of, 226 
schematic presen- 
tation of, 227 
Syphilis, Wassermann re- 
action in, technque 
of, Stern’s modifi- 
cation of, 227 
Temperature, body, and re- 
sistance to infec- 
tion, 20 
Tetanolysin, experiments 
with, 136, 138 
Tetanus antitoxin, produc- 
tion of, 538 
standardization of, 538 
by means of toxin, 123 
therapeutic use of, 539 
Tetanus bacillus, action of, 
4, 5 
injured tissue and, 24 
symbiosis and, 31 
Tetanus, “cryptogenic,” 5 
relative susceptibility of 
man and animals 
to, 62 
toxin, fixation in, by 
brain tissues, 50, 
54, 150, 151 
lipoidal substances a 
factor in, 152 
proteolytic enzymes 
a factor in, 152 
temperature a factor 
in, 151 
Tetanus toxin, action of, 
43, 50 
immunization with, 85 
Thread reaction in agglu- 
tination, 245 
Thymotoxin, 102 
Thyroid gland, production 
of alexin in, 188, 
189 
production of opsonins 
in, 189 
Toxalbumins, 44 
Toxemia, 9, 98 
bacterial, 4, 9, 47, 545 
sensitized animal and, 47 
Toxicity, chemical struc- 
ture and, 52 
definition of, 11 
increase of, in fresh mix- 
tures of toxin and 
antitoxin, 140 
Toxicity of normal sera, 
513 
anaphylatoxin formation 
a possible explana- 
tion of, 514, 515 
Toxic unit, definition of, 
124, 125 
Toxin-antitoxin combina- 
tion, chemical re- 
lations of, 130 
effect of heat on, 121, 
122 
in snake venom, stability 
of, 121, 122 
toxin-HCl modification 
of, 122 
effect of heat on, 
121, 122 
measurement of, by par- 
til absorption 
method of Ehrlich, 
131 
stability of, 121 
valency of component 
parts of, 130, 131 
Toxin-antitoxin, diphtheria. 
See Diphtheria 
toxin-antitoxin. 
Toxin-antitoxin reaction, 
120, et seq. INDEX 
665 
Toxin-antitoxin reaction, 
analogy between 
chemical reactions 
and, 122, 130, 
134, ct seq. 
antibody production in 
body cells in, 149 
concentration of reagents 
in, 122 
direct neutralization, the 
protective power 
of, 143, 144 
effect of heat on, 121, 122 
neutralization in, adsorp- 
tion theory of, 139 
Arrhenius and Madsen 
on, 136, 140 
Bordet on, 138 
Bordet-Danysz phe- 
nomenon in, 141 
von Dungern’s views 
on, 141 
Danysz effect in, 141, 
143 
Northrop’s views on, 
142, 143 
phenomena of, 135, et 
seq. 
overproduction of recep- 
tors in, 148 
physical mechanism of, 
144 
quantitative relations in, 
122, 139, 140, 142 
relationship between sus- 
ceptibility of tis- 
sue and toxin- 
binding properties 
in, 150 
side chain theory in, 144 
specificity of, 145 
speed of action of, 122 
time element in, 121 
with snake venom, 121, 
et seq. 
Toxin, bacterial. See Bac- 
terial toxins, 
classes of, 39 
definition of, 38 
detoxification of, by ad- 
dition of anti- 
toxin, 532 
differences in combining 
avidity of, 127 
diphtheria, construction 
of, 313 
normal, 123 
direct effect of antitoxic 
serum on, 120 
epitoxoid form of, 128 
immunization with, 82 
indirect effect of anti- 
toxic serum on, 
120 
non-specific effects of, 37 
“seasoning,” 528 
toxoid, 126 
protoxoids in, 131 
syntoxoids in, 131 
toxon, 129 
true, 39 
action of, 43, 96 
analogy of, with en- 
zymes, 17, 40, 105, 
141, et seq. 
bacteria producing, 40, 
44 
chemical nature of, 42, 
51 
heat sensitiveness of, 
42, 43 
incubation time of, 30, 
43 
Toxin, true, production of 
antitoxin by, 41, 
96 
specificity of formation 
of, 38, 87 
Toxin hypersusceptibility, 
493, 510, ct seq. 
anaphylaxis, and, 511, 
512 
“histogenic hypersuscep- 
tibility,” 511 
“paradox phenomenon” 
in, 511 
“sessile receptors” in, 
511 
Toxin “spectra,” construc- 
tion of, 132 
Toxin unit, definition of, 
125 
diphtheria, 123 
Toxoids, 126, 130 
Toxons, 129 
Toxophore group of toxins, 
126 
Trauma, cryptogenic infec- 
tions and, 32 
infection in relation to, 
4, 5, 24 
localization of infection 
dependent upon,28 
Tropins, 373 
Tubercle bacilli, effect of 
body temperature 
on virulence of, 
20 
extraction of, 16 
phagocytosis of, 344, 346 
Tuberculin ophthalmo-reac- 
tion, 497, 499 
Tuberculin reaction, anal- 
ogy of, to ana- 
phylaxis, 500 
Bail’s experiments with 
' passive sensitiza- 
tion in, 501 
diagnostic value of, 501 
Koch’s experiments in, 
497, 498 , 
nature of, 498 1 
Babes’ interpretation 
of, 498 
Koch’s Interpretation 
of, 498 
Wassermann and 
Bruck’s interpre- 
tation of, 498, 499 
“Residue” antigen in, 504 
Tuberculins, 497 
New Tuberculin (TR and 
TO), 82 
Old Tuberculin (Koch), 
79. 499, 501 
“Residue” antigen as, 
504 
Tuberculin skin reaction, 
497, 499 
Tuberculosis, avian type of, 
relative suscepti- 
bility of animals 
to, 61 
bovine type, relative sus- 
ceptibility of man 
and animals to, 61 
complement fixation test 
in diagnosis of, 
236 
human type, relative sus- 
ceptibility of man 
and animals to, 61 
of cold-blooded animals, 
immunity of 
warm-blooded ani- 
mals to, 61 
Tuberculosis, opsonic index 
in, 401, 403 
Tumors, malignant, alexin 
fixation in diag- 
nosis of, 233 
Typhoid bacilli, anaphy- 
latoxins produced 
from, 18 
extraction of, 81 
gain in virulence and loss 
of agglutinability, 
13 
Typhoid carriers, 3 
rabbits experimentally 
changed to, 16, 
249 
Typhoid fever, adaptation • 
of bacteria in, 9 
agglutination reaction 
for diagnosis of, 
241, 242 
effect of path of intro- 
duction of bac- 
teria of, on infec- 
tion, 22 
prophylactic immuniza- 
tion against, ac- 
tive, 580 
early experimentation 
in, 580 
living sensitized vac- 
cines used in, 582 
results of, in United 
States Army, 582 
Russell’s vaccines in, 
581 
sensitized killed vac- 
cines in, 583 
relative susceptibility of 
man and animals 
to, 63 
serum therapy in, 561, et 
seq. 
Vaccine treatment in, 
584 
Typhoidin reaction, 497, 
502, 506 
analogous to tuberculin 
reaction, 502 
Typhus fever, relative sus- 
ceptibility of man 
and animals to, 63 
Ultramicroscope, in Was- 
sermann reaction, 
326 
Ultrainlcroscopic method of 
finding precipi- 
tates in Wasser- 
mann reactions, 
222, 229 
Unitarian view of anti- 
bodies, 318, 320, 
321 
Vaccination, method of 
Kolle, 21 
Vaccination, prophylactic, 
in man, 570 
in anthrax, 74 
in cholera, 585. See also 
under Cholera, 
in influenza, 590. See 
also under Influ- 
enza. 
in paratyphoid fever, 580 
in plague, 587. See also 
under Plague, 
in pneumonia, 589. See 
also under Pneu- 
monia. 666 
INDEX 
Vaccination, prophylactic, 
in rabies, 597. See 
also under Rabies, 
in small-pox, 594. See 
also under Small- 
pox. 
history and general 
principles of, 72 
in syphilis, 615. See also 
under Syphilis, 
in typhoid fever, 580. 
See also under 
Typhoid fever. 
Vaccines, autogenous, 395, 
576 
production of, 576 
with dead bacteria, 
576 
with living bacteria, 
576 
sensitized, 576, 580 
standardization of, 577 
Hopkins’ method of, 
579 
Wright’s method of, 
577, 578 
use of, in streptococcus 
infections, 591 
in staphylococcus pyo- 
genes aureus, 395, 
591 
in gonorrheal infections, 
593 
Vaccine therapy. See also 
Immunization, ac- 
tive. 
as a therapeutic measure, 
action of, in local 
infections, 571 
in generalized systemic 
infections, 572 
in successive local in- 
fections, 572 
value of, in acute dis- 
eases, 574 
in subacute or chronic 
cases, 574, o75 
“negative!” phase in, 
396, 397 
second injection in, 
396, 397 
successive, inoculations 
in, 398 
summation of, 398 
opsonic index in, 389, 
394, et seq. 
improvement and, 401 
relation of phagocytosis 
to, 389, 390 
value of, as prophylactic 
measure, 570, 575 
as a therapeutic meas- 
ure, 571 
Vaughan and Wheeler, 
theory of, on 
mechanism of ana- 
phylaxis, 46i, 462 
work of, on toxic frac- 
tion of protein 
molecule in ana- 
phylaxis, 461 
Vaughan’s, work on ana- 
phylaxis, 434 
on bacterial anaphylaxis, 
506 
split products, 630, 635 
Virulence, accidental fac- 
tors favoring, 24 
Virulence, “aggressin” se- 
cretion of bacteria 
in body and, 18 
capsule formation of 
bacteria and, 13, 
14, 18 
decrease of, by attenua- 
tion of bacteria, 
12, 24, 75 
differentiated from toxic- 
ity, 11 
ectoplasmic hypertrophy 
of bacteria in re- 
lation to, 15, 18 
effect of body tempera- 
ture on, 20 
effect of cultural adapta- 
tion of bacteria 
on, 11, 17, 19, 24 
effect of path of intro- 
duction of bac- 
teria on, 3, 21, 
22 
effect of quantity of bac- 
teria introduced 
on, 3, 23 
enhancement of, 12, 13, 
16, 17, 30, 31 
incubation time of bac- 
teria and, 29, 30 
measurement of relative 
degrees of, 24 
of capsulated bacteria, 
13, 14, 18 
relation of. to phagocy- 
tosis, 374 
relative to number of 
bacteria intro- 
duced, 23 
specificity of bacteria 
and, 25, 26 
symbiosis in, of bacteria 
successively passed 
through animals, 
12, 75, 249 
of different strains of 
some bacteria, 12, 
24 
“Virus fixe,” in treatment 
of rabies, 598 
virulence of, 12 
Virus, “green,” in prophy- 
lactic immuniza- 
tion against small- 
pox, 596 
Vital resistance, 629 
Wassermann reaction, 216, 
et seq. 
alexin fixation principle 
in, 216, 223 
alexin titration in, 224 
antigen preparation for, 
218 
by addition of choles- 
terin, 219 
by method of Browning 
and Cruikshank, 
219 
by method of Noguchi, 
218, 219 
by methods of Porges 
and Meier, 218, 
219 
by methods of Weil 
and Braun, 218 
titration of, 220 
Was sermann reaction, as 
a colloidal reac- 
tion, 222 
diagnostic value of, 228 
in diagnosis of syphilis, 
216 
in diseases other than 
syphilis, 228 
precipitation in, by addi- 
tion of syphilitic 
serum to lecithin 
emulsions, 222 
produced with syphilitic 
serum in antigens 
from normal or- 
gans, 218 
specific antigen from 
spirochaeta pal- 
lida cultures un- 
suitable in, 221-2 
spinal fluid used in per- 
formance of, 217, 
228 
technique of performance 
of, 225, et seq. 
modification of, 226, 
227 
refrigerator method in, 
226 
schematic presentation 
of, 225 
theories of, 221 
titration of hemolytic 
amboceptor or 
sensitizer in, 223 
Weichardt’s epiphanin re- 
action, 327, 328 
Weigert’s law of overcom- 
pensation, 148 
Wells, review of true ana- 
phylaxis, 410 
Widal’s reaction, 242 
work on agglutination 
with dead bac- 
teria, 100 
Wilde, experiments on 
singleness of 
alexin, 171 
Wright’s, method of stand- 
ardization of vac- 
cines, 577, 578 
studies of bactericidal 
and agglutinating 
powers of blood 
serum, 385) 
Wright’s technique for de- 
termination of 
opsonic index, 391 
et seq. 
difficulties in, 393, 394 
value of, 394 
“X” substances, 46 
Yellow fever, susceptibility 
to, 64 
Yersin anti-plague serum, 
564, 565 
Zone phenomena, 175 
Zsigmondi’s colloidal gold 
method, 329 
Zymophore group, of com- 
plement, 173 
in agglutination, 256, 257