THE CLINICAL EXAMINATION OF 
BREAST MILK. 
L. EMMETT HOLT, M.D. 
Reprinted from 
The Archives of Pediatrics, 
March, 1893. 
New York : 
M. J. Rooney, Printer and Publisher, 
114-120 W. 30th St. THE CLINICAL EXAMINATION OE BREAST- 
MILK* 
L. EMMETT HOLT, M.D., 
New York. 
The importance of obtaining more exact knowledge of 
the variations in breast-milk has been for some years 
keenly felt by all those interested in the nutrition of 
young infants. The decision of the question regarding 
the ability or inability of the mother to nurse her child 
has so frequently to be decided by the general practi- 
tioner without special advice, that it is extremely import- 
ant that there should come into general use a simple 
method for the examination of breast-milk, in order that 
infantile life and health may not be jeopardized by con- 
tinuing lactation under circumstances when the milk is 
unfitted for the child’s nutrition, and on the other hand 
in order that a recognition of what is wrong in the milk 
may enable the physician in some cases at least, so to 
modify the secretion by diet, habits of life, etc., as to 
make it meet the requirements of the child. 
Pioneer work in the study of breast-milk has been done 
in this country by the former president of this society, Dr. 
T. M. Rotch, of Boston, and the importance of his work 
can hardly be over estimated. But not every physician is 
* Read before the American Pediatric Society, Boston, May 4,1892. 2 
HOLT : The Clinical Examination of Breast-Milk. 
within reach of a competent analytical chemist, and not 
every family can afford the expense of such analyses. 
It is to make the results of such work available in 
every-day practice that I have been experimenting for 
the past two and a half years upon the subject of clinical 
milk examination, the results of which work are here sub- 
mitted. It is therefore the purpose of the present paper 
to set forth a simple means of breast-milk examination, 
such as can be readily applied by any intelligent practi- 
tioner; and while not pretending to the accuracy of the 
results afforded by complete quantitative analysis, still 
for practical purposes they will be found sufficient in 
ninety per cent, of the cases,' at the same time being 
as trustworthy as methods commonly in use by the gen- 
eral practitioner in the examination of urine. 
During the last twenty or thirty years considerable liter- 
ature has accumulated in France and Germany upon the 
subject of breast-milk examination, but up to the date of 
Dr. Rotch’s work nothing had been done in this country, 
and practically nothing in England. 
In the examination of breast-milk, information is de- 
sirable upon the following points : 
Reaction ; specific gravity ; microscopical appearances ; 
fat; proteids ; sugar ; inorganic salts. 
Reaction.—The reaction of breast-milk if freshly drawn, 
is alkaline in the great proportion of cases, seldom neu- 
tral, almost never acid. It can be easily tested by ordi- 
nary litmus paper. 
Specific Gravity.—The specific gravity varies between 
1029 and 1032, the average being 1031 at yo°F. in a good 
specimen ; abnormal variations are seen between the 
limits of 1017 and 1036. This may be taken with any 
small hydrometer. A urinometer will answer the pur- 
pose, although the instrument described below will 
answer better, since it can be employed with a smaller 
amount of milk. An increase in the fat lowers the speci- 
fic gravity ; an increase in the other solids raises it. 
Microscopical Examination.—Although much less impor- 
tant than other methods of examination, this is still useful Holt : The Clinical Examination of Breast-Milk. 
3 
chiefly in the recognition of colostrum corpuscles and 
foreign matters occurring in milk, such as pus, blood, 
epithelial cells. Observations k> 
made upon the size and num- 
ber of the fat globules have 
little or no practical value. 
Pus, blood and epithelial cells 
appear in various patholog- 
ical conditions of the mam- 
mary gland, and their recog- 
nition is a matter of impor- 
tance, since if present in any 
considerable quantity they 
make nursing, for the time at 
least, improper. Other signs 
of disease in the gland or the 
surrounding tissues are usual- 
ly present, so that one hardly 
needs to resort to the micro- 
scopical examination of milk 
in cases of mastitis to show 
that it is abnormal. 
Colostrum corpuscles are 
normally present in milk dur- 
ing the first week of lactation 
in considerable numbers. 
From this time they diminish 
rapidly, and their appearance 
after the second week may 
be looked upon as patholog- 
ical. These corpuscles do 
not seem to be in themselves 
important, but their pres- 
ence is always an indication 
of other changes in the milk, 
not yet well understood, 
Fig. i. 
which makes it disagree with the infant. The causes of 
this colostrum-milk are many. It is seen occasionally 
during menstruation, also in cases of marked anaemia, and 4 
HOLT : The Clinical Examination of Breast-Milk. 
frequently as a result of strong nervous impressions of 
any kind, such as fatigue, grief, fright and sexual ex- 
citement. 
The presence of colostrum corpuscles in numbers in 
milk after the second week calls for a suspension of lac- 
tation, at least temporarily. 
In determining the composition of milk what we most 
need is a knowledge of the proportions in which the two ele- 
ments which vary most widely, viz.; the proteids and the 
fat, are present. 
Salts.—The inorganic salts do not vary much in their 
amount in ordinary milk. As yet but little is positively 
known of the physiological effects of these variations. 
The salts are in too small amount to affect the specific 
gravity, and in the clinical examination of milk they may 
be dropped from consideration altogether. 
Constancy of Sugar.—The proportion of the sugar is 
very constant in breast-milk under all circumstances. 
This point has been emphasized by all the chemists who 
have made many milk analyses. 
In fifty-seven specimens analysed by Pfeiffer the varia- 
tion was less than one-half per cent, from the average in 
forty-four specimens. 
In forty-three examinations made by Leeds, he states 
that the variations in the proportions of sugar were very 
slight. 
In twenty-five analyses made for me of samples, most 
of which were selected because they were abnormal the 
variation from the average was one-half per cent., or less 
in twenty-one of the cases, 
From these data it may be considered established that 
the percentage of sugar in milk is remarkably constant 
under all circumstances. 
Determination of Fat.—This is the most important 
feature in the analysis. Three methods for estimating 
fat may be employed. The first depends upon the fact 
that the opacity of milk is due to the floating fat globules, 
and the estimates of fat is made by a special instrument 
designed to measure the degree of translucency. Holt : The Clinical Examination of Breast-Milk. 
5 
Several forms of apparatus have been devised and are 
used in Germany for this purpose. One of the best known 
is Vogel’s instrument, but in my opinion it is greatly inferior 
to Feser’s lactoscope. (See Fig. I.) This instrument 
consists of a glass cylinder holding about three ounces, 
from the bottom of which rises a white porcelain stem 
A marked with strong black lines. Four cc. of milk is 
put into the cylinder with a graduated pipette, and water 
is added gradually, the instrument being shaken mean- 
while until the black lines on the porcelain stem are 
faintly visible through the milky water. When the proper 
degree of translucency is reached the percentage of fat 
is read off at the level of the water upon the graduated 
scale marked on the cylinder, e.g., water up to a certain 
part indicates four per cent, fat ; up to another three per 
cent., etc. 
The objections to this apparatus are that it is some- 
what expensive, costing about $3.50, but most of all that 
it is not easy to know just at what point after the addi- 
tion of water the reading is to be made. Without instruc- 
tion in regard to this the observer is quite at sea, and may 
easily make a mistake of one or two per cent. These 
objections make an instrument of this kind rather too 
complicated for general use, although in the hands of an 
expert, the instrument if properly made is an excellent 
one for cow’s milk but not quite so reliable for breast-milk. 
The second method of estimating fat is that first pub- 
lished by Marchand, more than thirty years ago. In 
this the fat is separated from the milk by means of ether, 
and then precipitated from the ethereal solution by strong 
alcohol. Marchand’s apparatus, modified by Conrad to 
enable the test to be made with a smaller quantity of milk 
is shown in the accompanying cut. (Fig 2.) 
The test is applied as follows: The tube is filled to the 
line M with milk, then four or five drops of liquor sodae 
are added, and then ether up to the line E. The tube is 
now corked and shaken strongly fifteen or twenty times; 
alcohol of ninety-five per cent, strength is now added 
up to line A. A slight excess is unimportant; the tube 6 
HOLT : The Clinical Examination of Breast-Milk. 
is now tightly corked and shaken fifteen or 
twenty times, then placed upright in a vessel ol 
water at a temperature of 130° to I4O°F. The 
fat globules in a few moments begin to rise rap- 
idly and form a distinct layer at the top. At 
the end of half an hour the amount of fat may 
be read off on the graduations on the tube, and 
by reference to the accompanying table, taken 
from Conrad’s article, the exact percentage of 
fat is calculated : 
Degrees Marchand. Fat. 
g 1.37 per cent. 
1 i-49 
3 1-95 
5 2.42 
7 2.89 
9 3-35 
11 3.82 “ 
13 4.28 
15 4-75 
17 5.22 
19 5-56 
21 6.15 “ 
23 6.61 “ 
25 7.07 
27 7.54 
29 8.00 “ 
31 8-47 
This method of testing the fat is a fairly good 
one, but it is somewhat complicated and seems 
liable to quite wide variations, due to conditions 
in the milk not readily understood ; certainly 
it is not so accurate as stated by Conrad in his 
monograph, and after employing it somewhat 
over two hundred times, I have not been 
wholly satisfied with it. 
Neither of the above methods have seemed 
to fulfill the requisite simplicity necessary for 
general adoption, I have therefore made a num- 
ber of experiments to determine whether the 
volume of cream rising in a given time upon 
Fig. 2 Holt : The Clinical Examination of Breast-Milk, 
7 
a specimen of milk bore sufficiently constant relation to 
the percentage of fat to make this a reliable guide for 
the calculation of the proportion of the fat. 
The cream test has been made as follows : A glass- 
stoppered cylinder, graduated in one hundred parts and 
holding about ten cc. is filled with milk exactly to the 
upper line marked zero. A pipette should be used for 
the last few drops, care being taken not to allow the milk 
to run down the inner side of the tube since this obscures 
somewhat an exact reading. The cylinder is then corked 
and allowed to stand for twenty-four hours, as nearly at 
70° F. as practicable. A variation of a few degrees on 
either side of this point is unimportant. If, however, the 
variations are wide the rapidity with which the cream 
rises is somewhat modified. 
It is important that the milk should 
be freshly taken and handled as little as 
possible ; also that the graduated cylinder 
should be scrupulously clean, otherwise 
the milk will often sour before the cream 
has had time to rise. This is particularly 
true in summer. After twenty-four hours 
the percentage is read off upon the gradu- 
ated cylinder. 
Fig. 3 shows the shape of the cream 
at the top of the tube. The percentage 
is to be read off between the points A and 
B, and not between C and D, since the lat- 
ter will make an error of from one-half to 
one per cent. 
In the great majority of the cases the 
lower line of the cream BE is sharply de- 
fined at the end of twenty-four hours. If 
this is not the case, the milk should stand 
for six hours longer before the reading 
is made. In rare cases at the end of 
Fig. 3. 
twenty-four hours only a small amount of cream has 
risen, and the whole column of milk is nearly of a 
uniform white. This is due to some condition which 8 
HOLT : The Clinical Examination of Breast-Milk. 
has prevented the cream in rising with the usual rap- 
idity. Under these circumstances also, from six to 
twelve hours additional are required. These are, how- 
ever, very rare exceptions. In nearly all the cases at the 
end of the twenty-four hours the lower line of the cream 
is sharply defined and the milk at the bottom of the tube 
has become translucent for an inch or so. It is these con- 
ditions which are to be secured before the reading is 
made, and it is only very exceptionally that more than 
twenty-four hours are required. 
Relation of the Cream to the Fat.—By comparing the 
percentage of cream with that of the fat as determined 
by chemical analysis of the same specimen, it has been 
discovered that the ratio of the cream to the fat is very 
nearly five to three. 
The following table shows the percentage of cream, 
the amount of fat as calculated in this manner, the 
amount as determined by chemical analysis and the size of 
the error in 
twenty-four cases: 
Ckeam. 
Fat (Calculated) 
Fat by Analysis. 
Error. 
I 
3% 
i-8V„ 
i-35°/0 
•457 
>+ 
2 
6 
3-6 
3-3i 
.29 
+ 
3 
7 
4.1 
3.86 
.24 
+ 
4 
4 
2.26 
2.68 
.42 
— 
5 
2-5 
1.50 
1.23 
• 27 
+ 
6 
6.5 
3-9° 
4.04 
.86 
+ 
7 
5 
3.°° 
2.78 
.22 
+ 
8 
3 
1.80 
1.52 
..28 
+ 
9 
3 
1.80 
1.74 
.06 
+ 
IO 
6 
3.60 
* 3.62 
.02 
— 
n 
4 
2.26 
2 95 
.69 
— 
12 
7 
4.20 
423 
•03 
— 
13 
6 
3.60 
3-48 
. 12 
+ 
14 
3-5 
2.10 
2.12 
.02 
— 
15 
5 
3.00 
3.78 
■78 
— 
lb* 
4 
2.40 
2.79 
•39 
— 
17 
5 
3.0° 
3-3i 
•39 
— 
18 
2-5 
1.50 
'•54 
.04 
— 
19 
4 
2.26 
2.26 
'.oo 
20 
3-5 
2.10 
2.17 
.07 
— 
21 
6 
3.60 
2-97 
•63 
+ 
22 
6-5 
3-9° 
3-53 
•37 
4- 
23 
5 
,3-oo 
3-i3 
•r3 
— 
24 
5 
3.°° 
4-5i 
I-5I 
— 
*For cases 16 to 24 I am indebted to Dr. Charles Harrington of 
Boston, who made the analyses. Holt : The Clinical Examination of Breast-Milk. 
9 
Thus in fourteen of these twenty-four cases the error 
was less than three-tenths per cent. In only five cases 
was it greater than one-half per cent. In only one did 
it reach one per cent. In all but two cases, Nos. 12 and 
13, the cream was read off at the end of twenty-four 
hours. In these two tests which were made on the same 
day for some unexplained reason, the cream was very 
slow in rising and thirty-six hours were required before 
the conditions mentioned above as necessary were obtained. 
It follows from the above, that when carefully made 
the volumetric cream test gives a very approximate idea 
of the percentage of fat in breast-milk, and that this is 
accurate enough for all practical purposes. 
Variations in Fat and Proteids.—In striking contrast to 
this uniformity in the sugar are the wide variations met 
with in the fat and proteids, as is shown by the following 
table : 
VARIATIONS IN FAT. 
From 43 analyses by Leedsl 2.11 to 6.89 per cent. 
“ “ Koenig 1.71 to 7.60 “ “ 
“ 29 “ “ Personal2 1.12 to 5.02 “ “ 
VARIATIONS IN PROTEIDS. 
From 43 analysis by Leeds1 85 to 4.86 per cent. 
“ “ Koenig. 57 to 4.25 “ “ 
“29 “ “ Personal2 1.10 to 3.62 “ “ 
It is just these wide variations in the composition of 
milk that are important for the practitioner to recognize. 
Estimation of Proteids.—There is no known method 
of determining directly the percentage of the proteids in 
milk by a clinical examination. Nothing short of a full 
chemical analysis, and that by an expert, can be looked 
upon as absolutely accurate. It is possible, however, 
from a knowledge of the specific gravity and the percent- 
age of fat to make an approximate calculation in regard 
to the percentage of proteids ; enough to determine 
whether in a given case they are near the normal, or in 
very large, or in very small proportions. 
IMedical News, July 21, 1883. 
2The greater part of the analyses were made for me at the Chemical 
Laboratory of the College of Physicians and Surgeons, New York. lo HOLT : The Clinical Examination of Breast-Milk. 
From the fact that the proportion of sugar is so nearly 
constant, and that the inorganic salts are in such small 
amount we may for clinical purposes 
consider the specific gravity as mod- 
ified solely by the fat and the proteids. 
Supposing now the proteids to re- 
main the same, if the fat is high, the 
specific gravity will be high ; but if 
the fat is low the spe- 
cific gravity will be 
low. 
Supposing now the 
fat remains constant, 
if the pjoteids are 
high the specific 
gravity will be high; 
if the proteids are 
'low the specific 
gravity will be low. 
If therefore the fat 
and the specific grav- 
ity are known the 
proteids may be es- 
timated as follows: 
If the fat is found 
to be high, e.g., eight 
to ten per cent, and 
the specific gravity 
is high, e.g., 1033 
to 1034 we may as- 
sume that the pro- 
teids are also high, 
otherwise the exces- 
sive fat would bring 
Fig. 4a. 
Fig. 4b. 
the specific gravity 
below the normal 
average. 
If the fat is found to be low, three to four per cent., 
and the specific gravity high we may assume the proteids Holt: The Clinical Examination of Breast-Milk, n 
to be nearly normal, since the high specific gravity is ex- 
plained by the low fat. 
If the fat is high, and the specific gravity low, the pro- 
teids may be assumed to be normal, since the variation 
in the specific gravity is explained by the low fat. 
If the fat is low and the specific gravity is low, the pro- 
teids are also low, since otherwise the low fat would make 
the specific gravity above the average. 
Of course it is only the wide variations of the proteids 
which can be recognized, but these are the most import- 
ant in practice. 
The application of the foregoing principles will readily 
be seen by reference to the accompanying table. It has 
been constructed from an experience with these methods 
in about four hundred 
samples 
of milk. 
SPECIFIC GRAVITY 
7O0 P. 
Cream—24 Hours. 
Proteids—(Calculated.) 
Normal average. 
I.O31 
7% 
'.5% 
Healthy variations. 
1.028—1.029 
8%-, 2% 
Normal (rich milk) 
I.032—I.O33 
S%- 6% 
“ (fair milk) 
Unhealthy “ 
Below 1.028 
High (above 10%) 
“ or slightly below 
“ 
“ “ 
Normal {5%—10%) 
Low 
“ 
“ “ 
Low (below 5%) 
Very low (very poor milk) 
Above 1.033 
High 
Very high (very rich milk) 
“ * 
Normal 
High 
“ “ 
Low 
Normal (or nearly so) 
Method of Determming the Composition of Milk.—This 
may briefly be summarized as follows: The specimen 
taken should be after the breast has been about half 
emptied, or else the entire amount yielded by the breast, 
since the first milk is more watery and the last is richer 
than the average. The specific gravity is taken with a 
small hydrometer* shown in Fig. 4 (b). With this in- 
strument only one half ounce is required for the test. The 
cylindrical tube or creamometer* shown in Fig. 4 (a), is 
filled to the zero line and the cream read off at the end 
of twenty-four hours with the precautions previously 
mentioned. The percentage of fat may be calculated to 
be three-fifths of the cream. The knowledge of these 
*This apparatus, consisting of two graduated cylinders and hydrom- 
eter, and printed directions, can be obtained from Eimer and Amend, 
18th street and Third Avenue, New York City. Price $1.60. 12 
Holt: The Clinical Exa7nination of Breast-Milk. 
two factors enables us to judge whether the proteids are 
nearly normal, very high or in very small amount. 
Special attention is called to the fact that neither this 
nor even a chemical analysis will give any idea regarding 
the quantity of milk. It not unfrequently happens that 
the quality may be excellent and yet the nursing child 
doing wretchedly, because the quantity is very small. 
This is to be determined by other means. The most re- 
liable symptoms of this condition are breasts which are 
soft and flabby at nursing time, and from which only a 
small quantity can be extracted by pressure or the pump. 
Long nursing i.e., thirty or forty minutes, before the child 
is satisfied should also make one very suspicious that the 
milk is scanty. Weighing before and after nursing is con- 
clusive. This requires scales which are accurate and 
sufficiently sensitive to indicate half ounces. Three or four 
observations should be made, at different hours in the 
day before a conclusion is reached.