Bryn Mawr College 
Monographs 
REPRINT SERIES, Vol. XI 
Contributions from the Psychological Laboratory 
BRYN MAWR, PENNA. 
June, 1916 Table of Contents 
I. A Note on the Determination of the Retina's Sensitivity to Colored 
Light in Terms of Radiometric Units. By C. E. Ferree and 
Gertrude Rand. 
(Reprinted from the American Journal of Psychology, Vol. XXIII, 
April, 1912.) 
II. Colored After-Image and Contrast Sensations from Stimuli in which 
no Color is Sensed. By C. E. Ferree and Gertrude Rand. 
(Reprinted from, the Psychological Review, Vol. XIX, May, 1912.) 
III. An Optics-Room and a Method of Standardizing Its Illumination. 
By C. E. Ferree and Gertrude Rand. 
(Reprinted from the Psychological Review, Vol. XIX, September, 1912.) 
IV. The Effect of Changes in the General Illumination of the Retina 
upon Its Sensitivity to Color. By Gertrude Rand. 
(Reprinted from the Psychological Review, Vol. XIX, November, 1912.) 
V. Tests for the Efficiency of the Eye under Different Systems of Illumi- 
nation and a Preliminary Study of the Causes of Discomfort. 
By C. E. Ferree. 
(Reprinted from the Transactions of the Illuminating Engineering 
Society, Vol. VIII, January, 1913.) 
VI. The Factors that Influence the Sensitivity of the Retina to Color: 
A Quantitative Study and Methods of Standardizing. By 
Gertrude Rand. 
(Reprinted from the Psychological Monographs, Vol. XV, No. 1, 
March, 1913.) 
VII. The Fluctuation of Liminal Visual Stimuli of Point Area. By C. E. 
Ferree. 
(Reprinted from the American Journal of Psychology, Vol. XXIV, 
July, 1913.) 
VIII. The Problem of Lighting in Its Relation to the Eye. By C. E. Ferree. 
(Reprinted from Science, N. S., Vol. XL, July 17, 1914.) ■ IX. A Preliminary Study of the Deficiencies of the Method of Flicker 
for the Photometry of Lights of Different Color. By C. E. 
Ferree and Gertrude Rand. 
(Reprinted from the Psychological Review, Vol. XXII, March, 1915.) 
X. The Efficiency of the Eye under Different Conditions of Lighting: 
The Effect of Varying the Distribution Factors and Intensity. 
By C. E. Ferree and Gertrude Rand. 
(Reprinted from the Transactions oj the Illuminating Engineering 
Society, Vol. X, No. 6, 1915.) 
XI. Further Experiments on the Efficiency of the Eye under Different 
Conditions of Lighting. By C. E. Ferree and Gertrude Rand. 
(Reprinted from the Transactions of the Illuminating Engineering 
Society, Vol. X, No. 6, 1915.) 
XII. A Resume of Experiments on the Problem of Lighting in Its Relation 
to the Eye. By C. E. Ferree and Gertrude Rand. 
(Reprinted from the Journal of Philosophy, Psychology and Scientific 
Methods, Vol. XII, November 25, 1915.) 
XIII. Some Experiments on the Eye with Inverted Reflectors of Different 
Densities. By C. E. Ferree and Gertrude Rand. 
(Reprinted from the Transactions of the Illuminating Engineering 
Society, Vol. X, No. 9, 1915.) 
XIV. A New Method of Heterochromatic Photometry. By C. E. Ferree 
and Gertrude Rand. 
(Reprinted from the Journal of Experimental Psychology. Vol. I, No. 1, 
February, 1916.) 
XV. A Spectroscopic Apparatus for the Investigation of the Color Sensi- 
tivity of the Retina, Central and Peripheral. By C. E. Ferree 
and Gertrude Rand. 
(Reprinted from the Journal of Experimental Psychology, Vol. I, 
No. 3, June, 1916.) PREFACE 
From the standpoint of the worker in the subject the object 
of a volume of this kind is to collect under one cover from 
sources more or less scattered the studies made by a group of 
investigators working in general from the same point of view 
and with similar technique and laboratory equipment. Such 
a compilation in part defeats its own purpose unless complete. 
We regret, therefore, to be compelled to omit from the initial 
volume so many of the studies issued from the laboratory 
during the period covered. 
C. E. Ferree, 
Director of Laboratory. A NOTE ON THE DETERMINATION OF THE 
RETINA'S SENSITIVITY TO COLORED LIGHT 
IN TERMS OF RADIOMETRIC UNITS 
By C. E. Ferree and Gertrude Rand 
Bryn Mawr College 
About a year ago1 the writers undertook to determine the 
retina's sensitivity, relative and absolute, to colored light 
in terms of units that can be compared. Since several years 
will be required to complete this work, they have thought 
it best to publish a preliminary note showing briefly the pur- 
pose and scope of the investigation. The following points 
will serve to indicate what is being attempted in this study. 
(1) All measurements of sensitivity will be made in radio- 
metric terms. This will give an expression of the sensitivity 
of the retina in units which are directly comparable with one 
another. At present we have no direct estimate of the com- 
parative sensitivity of the retina to the different colors further 
than is expressed, for example, by the relative width of the 
collimator-slit that has to be used to arouse color sensation 
when a light-source of a given candle-power is used. This 
kind of comparison is obviously unfair because such different 
amounts of energy are represented from point to point in the 
spectrum that a given width of slit would admit many times 
the amount of energy at one part of the spectrum that it 
would at another. In short, no adequate estimation and ex- 
pression of the retina's sensitivity to color, comparative or 
absolute, can be made by means of the methods now in 
common use.2 
1 The first public statement of our intention to use radiometric units 
in the investigation of the retina's sensitivity to color was made to 
the committee in charge of the Sarah Berliner Research Fellowship, 
February 1, 1911. 
2 Two criticisms have been received from private sources which it 
may be well to take account of here. In one the possibility of a point 
of view is implied, in the other a point of view is stated. The point of 
view, the possibility of which is implied in the first criticism, is that it 
is not proper to estimate the sensitivity of the retina in terms of physi- 
cal units, because it is generally conceded by modern investigators of 
color vision that the retinal processes which transform the physical 
energy of the color stimulus into nervous energy is essentially chemical 
in its nature; and one can not assume that a certain amount of physical 
energy arouses an equal amount of chemical energy in the retina, or 
that equal amounts of physical energy arouse equal amounts of chem- 
ical energy. In answer to this, the writers would point out that these 
chemical substances are a part of the retina and their respective iner- 329 
RETINAL SENSITIVITY 
(2) Comparisons of results on many other points with 
such disparate stimuli seem equally inadequate: the relative 
time required for the different color sensations to attain their 
maximum of intensity, or retinal inertia; the relative rate of 
fatigue to the different colors; after-image and contrast sen- 
sitivity, etc.3 In fact there is not a quantitative problem 
tias constitute one set of factors that determines the sensitivity of the 
retina to the different colored lights. It is not necessary to assume, 
therefore, that a given amount of physical energy arouses an equal 
amount of chemical energy, etc., in order to make our determinations 
of the comparative sensitivity of the retina to the different colors in 
terms of physical units. That would be necessary only if we were try- 
ing to separate out the nerve filaments, and to measure or compare 
their sensitivity to the different colors in terms of physical units. But 
even in chemical theories when speaking of the comparative sensitivity 
of the retina to the different colors, we do not mean the comparative 
sensitivity of the nerve filaments alone. We include the reaction of 
the chemical substances as well. Our contention, then, is that if the 
determination of the comparative sensitivity of the retina to the differ- 
ent colors is a proper problem, the determination should be made in 
terms of quantities that can be compared. This can be done either 
a, by using lights equalized in energy and determining by means of a 
sectored disc the relative amounts of these lights that are required to 
arouse sensation; or b, by using lights representing different amounts 
of energy and measuring directly in terms of radiometric units the 
amounts required to arouse sensation. We scarcely need point out that 
in speaking of the comparative sensitivity of the retina to the different 
colors we are not raising a new problem, but are merely recognizing a 
very old one. 
The second criticism is in substance that a quantitative comparison 
of the effect of the different wave-lengths on the retina is improper 
because the different wave-lengths constitute stimuli too different in 
kind to permit of such comparison. This criticism we leave open, be- 
cause we do not wish to discuss in this paper the propriety of the 
problem of comparing sensitivities. 
3 It is conceivable that two points of view may be held with regard 
to what is meant by after-image and contrast sensitivity. (1) After- 
image and contrast sensitivity may express a relation between the 
amount of light required to arouse after-image and contrast sensations 
and the unit of light used. (2) It may express a relation between the 
amount of light required to arouse the after-image and contrast sensa- 
tions and the amount required to arouse positive sensation. If the 
former view should be held it will be convenient to start with stimuli 
equalized in energy, and to determine the relative amounts of light 
required to arouse the after-image or contrast sensation by means of 
a sectored disc. If the second view should be held, the energy of the 
lights used may first be rendered proportional to the sensitivity of the 
eye to the colors in question; and the liminal values may then be deter- 
mined by means of the sectored disc. In each case the relative sensi- 
tivity may be expressed by the inverse ratio of the open to the closed 
sectors. 
Similarly two views may be held with regard to the determination of 
the comparative rates of fatigue, and of the development-time of sen- 
sation. (1) Lights equalized in energy may be used. (2) The energy 330 
dealing with the comparative functioning of the retina to the 
different colors in which there does not seem to be a need for 
the regulation and estimation of the stimulus in terms of a 
common unit of measurement. It is the purpose of the writers 
to extend the work as fast as possible into these related fields. 
(3) We wish to make a careful study of the sensitivity of 
the peripheral retina, quantitative4 and qualitative, in a large 
FERREE AND RAND 
of the lights may be made proportional to the sensitivity of the eye to 
the different colors. 
The need in both the above cases is equally great for a method of 
regulating and determining the amounts of light to be used in terms 
of a common unit of measurement. 
4 The following are two of the points we wish to take up: (1) A de- 
termination will be made of the ratio of sensitivity of peripheral to 
central retina from point to point for a single color in several meridi- 
ans. This will show at what rate the retina falls off in sensitivity in a 
single meridian, and how uniform this decrease is in the different 
meridians. We have found in a preliminary study that this knowledge 
is greatly needed in explaining certain phenomena of the peripheral 
retina. Furthermore, when this determination is made for each of the 
colors with which we wish to work, the ratios of sensitivity for these 
colors at all the points can be calculated and a definite answer can be 
given to the question whether or not uniformity of ratio obtains 
throughout the retina. This question has been given considerable im- 
portance in the discussion of color theories. (2) The limits of sensi- 
tivity will be investigated. In general two problems are involved here, 
(a) The limits may be considered in relation to the comparative sensi- 
tivity of the retina to the different colors. (5) They may be consid- 
ered in relation to existing color theories. In the first of these prob- 
lems the limits should be obtained with stimuli equalized in energy. 
So obtained the results will constitute merely another expression of 
the comparative sensitivity of the retina to the different colors. The 
second problem is more complicated and will later be made the subject 
of a separate paper. A word indicating its relation to our present plan 
of work may, however, not be out of place here. It may be logically 
assumed, for example, that the Hering theory demands that wherever 
the blue-sensing substance is found, the yellow-sensing substance must 
also be found. We have no means of knowing where these substances 
are except by the sensations aroused. Speaking in terms of the theory, 
then, we have a right to assume that wherever the blue sensation can 
be aroused the yellow sensation should be able to be aroused also, pro- 
vided a sufficiently intensive stimulus be used. If, therefore, in pass- 
ing towards the periphery of the retina, a point be found where blue 
can be aroused and yellow can not, the evidence will be strongly in 
favor of the conclusion that no yellow substance is present, unless it 
can be shown that elsewhere in the retina so much greater energy of 
yellow light than of blue is required to arouse sensation that the amount 
needed for this far peripheral point is greater than can be obtained. To 
establish this point the comparative sensitivity to these colors would 
have to be obtained at various points in the retina. This would involve 
the determination of a ratio based upon the amounts of blue and yellow 
light required to arouse sensation. Two methods of measurement may 
be used, (a) The amounts needed may be measured directly by means of 
a thermopile of the type we use, or other sensitive radiometer. In a deter- 331 
RETINAL SENSITIVITY 
number of meridians. In general too much uniformity has 
been assumed with regard to the sensitivity of the peripheral 
retina. Generalizations of great importance to color theory 
have frequently been based upon the results of work in which 
careful investigation was made in only one or two meridians. 
The conception of stable colors, and its application in support 
of the Hering Urfarben may be taken as a fair example of a 
sweeping conclusion which is based upon work too limited 
in its range. With a careful standardization of factors, an 
investigation in any considerable number of meridians shows 
that stable colors do not exist.5 Many other points of interest 
have come out in our more detailed study of the peripheral 
retina. For example, we find in the periphery of the normal 
retina small areas which are exact replicas of the Schumann 
case of color-blindness. 
(4) We wish to conduct our investigation in full daylight 
instead of in the dark room. This is to eliminate the influence 
of the field surrounding the colored stimulus and of the pre- 
exposure. When the surrounding field is black, white is 
induced by contrast across the stimulus color. Since the 
colors all differ in brightness,6 the induction takes place in 
different amounts for the different colors. This white, in 
proportion to its amount, reduces the action of the colors 
on the retina. Further, a given amount of white affects to 
different degrees the action of the different colors on the retina. 
To eliminate this twofold unequal action, the surrounding 
mination of limens the number of readings required would render this 
method tedious. (&) The energy of the two lights may be made equal by 
means of a thermopile and the final amounts required to arouse sensation 
may be secured by means of a sectored disc. From the ratio of open to 
closed sectors the amount the light is cut down in each case may be calcu- 
lated and the ratios of energy may be determined from these amounts. 
5 The following points are offered in support of the above statement. 
(1) A red and green cannot be obtained which in every meridian of the 
peripheral retina will pass into gray without an intermediate change into 
yellow or blue. (2) The amount of blue that has to be added to a mixture 
of red and green to produce gray varies from point to point in a given 
meridian even where the extramacular region alone is considered. Further, 
a series of determinations made for a given meridian will not hold for the 
remaining meridians. (3) A red, green, and yellow can not be obtained 
which will not change in color-tone in passing from the center to the 
periphery of the retina in a single meridian. 
Blue alone of the four principal colors is stable in tone for all parts 
of the retina. 
6 In a later paper one of the writers (Rand) will show that it is of no 
advantage to equate in brightness in determining the limits of color sensi- 
tivity, and that harm results in so many ways from the attempt to equate 
that it is doubtful whether it should be done even in determining the 
limens of color in the more sensitive parts of the retina. FERREE AND RAND 
332 
field should be made in each case of the brightness of the color 
to be used. This can be done by working in a light room of 
constant intensity of illumination and making the surrounding 
field of a gray paper of the brightness of the stimulus color. 
In order to accomplish this, and at the same time be able to 
work upon any meridian of the retina we choose, we have 
constructed a special piece of apparatus which we call a rotary 
campimeter. The influence of pre-exposure is even more 
important than of surrounding field. If the pre-exposure is 
to black, white is added as after-image to the stimulus color. 
The effect of a black pre-exposure upon the stimulus color 
is greater than the effect of a surrounding field of black, because 
more white is added as after-image of pre-exposure than is 
induced by contrast from the surrounding field. This effect 
also can be eliminated only by working in a light room of 
constant intensity of illumination and by choosing as pre- 
exposure a gray of the brightness of the color to be used. 
We began a quantitative study of the factors that influence 
the sensitivity of the retina to color three years ago. With 
the control of factors we had at that time, we could not, for 
example, duplicate by several degrees at any two consecutive 
determinations the limits of the zone of sensitivity to any color. 
The result of our study has been that we are now able with a 
given light-source to duplicate, within a degree, the results 
obtained at a previous sitting. We can also duplicate, almost 
as closely, the threshold values or the amounts of light required 
to arouse color sensation in the more sensitive parts of the 
retina. Details of this work will not be given here. They 
will appear in a series of papers to be published in the course 
of the present year. 
Having completed our work of standardizing the factors 
extraneous to the source of light, we are trying now to secure 
a better control of the source. Standardization, so far, can 
be considered successful only with regard to the quality of 
the light. No adequate work has been done upon the stand- 
ardization of the quantity of light. We believe this can be 
accomplished Qnly by means of energy determinations. We 
expect to do our radiometric work by means of a surface 
thermopile (Coblentz model), and a DuBois-Rubens Panzer- 
galvanometer, unless future results show that some other com- 
bination of radiometer and galvanometer is more satisfactory. [Reprinted from The Psychological Review, Vol. XIX., No. 3, May, 1912.] 
COLORED AFTER-IMAGE AND CONTRAST 
SENSATIONS FROM STIMULI IN WHICH 
NO COLOR IS SENSED 
C. E. FERREE AND GERTRUDE RAND 
Bryn Mawr College 
I. Introduction 195 
II. Cases in which colored after-images and contrast sensations are aroused by 
retinal excitations which do not directly condition color sensation. .. . 204 
1. After-images 204 
i. In central vision 205 
ii. In peripheral vision 208 
2. Contrast 221 
3. The Purkinje-Briicke phenomenon 224 
i. Evidence that it is an after-image of a previous contrast sensation, 
rather than contrast in the after-image 228 
ii. But if a contrast effect, evidence that it may take place when the 
inducing color is unsensed 234 
III. Explanation 236 
* 
I. Introduction 
In the March number of the Psychological Review, 1907, 
Thompson and Gordon1 describe a series of experiments in 
which colored after-images are obtained in the peripheral 
retina from stimuli in which no color was sensed. In the 
November number, 1905, and in the January number, 1908, 
of the same journal, Fernald2 working under approximately 
the same condition reports similar results. In the Proceedings 
of the American Philosophical Society, 1908, however, Titchener 
and Pyle3 deny the phenomenon, affirming their complete in- 
ability to get colored after-images when no color is sensed in the 
stimulus. 
1 Thompson, H. B., and Gordon, K., 'A Study of After-images on the Peripheral 
Retina,' Psychol. Rev., 1907, XIV., pp. 122-167. 
2 Fernald, G. M., 'The Effect of the Brightness of Background on the Extent of 
the Color Fields and on the Color Tone in Peripheral Vision,' Psychol. Rev., 1905, 
XII., p. 405; 'The Effect of the Brightness of Background on the Appearance of Color 
Stimuli in Peripheral Vision,' Psychol. Rev., 1908, XV., pp. 33-35. 
3 Titchener, E. B., and Pyle, W. H., 'On the After-images of Subliminally Colored 
Stimuli,' Proc, of the Amer. Philos. Soc., 1908, XLVII., No. 189, pp. 366-384. 
195 196 
C. E. FERREE AND GERTRUDE RAND 
With regard to these discrepant reports, nothing more will 
be attempted at this place than to point out clearly the position 
held by each investigator. Titchener and Pyle contend that 
under no condition known to them can colored after-images 
be obtained from stimuli in which no color is sensed. Thomp- 
son and Gordon, and Fernald claim that under the conditions 
described by them, after-images may be obtained in which the 
color is clearly distinguished. The point at issue between 
them, then, is not whether the phenomenon can be gotten 
under this or that condition usually obtaining in the work on 
after-images, or under any given percentage of possible condi- 
tions, but whether it may be gotten by any experimental device 
whatsoever. Nor is it meant to extend the phenomenon to 
brightness sensation.1 So far as the writers know, no 
one has ever claimed to be able to get an after-image 
from a brightness stimulus too weak to arouse sensation. 
The question, therefore, whether a subliminal sensory exci- 
tation can produce a supraliminal after-effect, considered 
in a general sense, is in no fashion under discussion. With 
the issue thus stated, the writers are impelled to take the 
affirmative side of the question by the experimental results 
they have obtained, and scarcely less strongly by theo- 
retical considerations of the difference in the effect of different 
members of the brightness series upon all the colors when 
fused with them, and in the effect of a given brightness upon 
different colors. Working under the right conditions, the 
phenomenon is easily obtained. That its explanation is not 
essentially difficult will be shown in a later section of this paper. 
The work on the peripheral retina has been repeated by the 
writers and the investigation extended to include other cases 
in which colored after-image and contrast sensations may be 
aroused by stimuli in which no color is sensed. It is the purpose 
of this report to describe these experiments and to explain the 
results obtained in terms of visual phenomena already known. 
Before passing to a description of our own experiments it 
may be helpful to examine the work of previous investigators in 
1 For the sake of brevity, brightness is used here as a general term for the colorless 
sensation series, white, black and the grays. AFTER-IMAGE AND CONTRAST SENSATIONS 
197 
order to determine from neutral ground if possible the cause of 
the disagreement in the results they have obtained.1 In general 
two types of method have been used in arousing the color 
excitation. In one the color was kept below the limen of 
sensation by adaptation; in the other it was obscured by the 
action of an unfavorable brightness excitation. The former 
method was first used by Tschermak.2 Tschermak's article is 
not a description of experimental work. It is an essay in 
which the author in part seeks to trace the analogies to visual 
adaptation found in the reactions of other tissue: muscle, nerve, 
secretory, etc.,' to external stimuli. The reference to the phe- 
nomenon we are discussing is very brief in this article and the 
description of the conditions under which it was obtained is 
quite inadequate to serve as a guide for future work. He says: 
"Haben wir doch gerade in der Anwendung des constanten 
Stromes auf Nerv und Muskel ein vorzugliches didaktisches 
Mittel, um die Grundbegriffe der allgemeinen Reiz- und 
Adaptationslehre zu veranschaulichen und einzupragen. Am 
besten demonstrieren wir als Gegenstiick zugleich die Wirkung 
eines massig satten Farbglases auf das Auge: die Phase der 
Reizwirkung, individuell verschieden lang, und dadurch erin- 
nernd an die verschiedenrasche Adaptation des Praparates vom 
Warmfrosch und Kaltfrosch an den constanten Strom-weiter- 
hin das Stadium der vollendeten Adaptation, endlich den 
gegensinnigen Oeffnungseffect. Nicht minder lehrreich ist die 
Parallele des subjectiven und des objectiven Erscheinungsge- 
bietes fur das Phanomen des Einschleichens d. h. des Ausblei- 
bens einer sinnfalligen Reizwirkung, wenn der Reiz so lang- 
sam anwachst, dass das Adaptationsvermogen folgen kann- 
gleichwohl hat auch nunmehr Wegfall des 'Reizes' eine gegen- 
sinnige Oeffnungswirkung. Analoges gilt vom Ausschleichen, 
also vom Ausbleiben eines sinnfalligen Oeffnungseffectes. Zum 
optischen Versuche schiebt man zweckmassig successive eine 
schwach tingierte Glasplatte vor die andere oder beniitzt einen 
Keil farbigen Glases." 
1 Because of the inadequacy of Tschermak's description of his method of working 
this can not be attempted for the work in the central retina. 
2Tschermak, A., 'Das Anspannungsproblem in der Physiologic der Gegenwart,' 
Archives des Sciences biologiques, Sup. Band, 1904, XL, pp. 79-97. 198 
C. E. FERREE AND GERTRUDE RAND 
In 1901 Titchener and Pyle carefully repeated and elabo- 
rated upon Tschermak's experiment on the central retina with 
entirely negative results. The writers can only commend the 
thoroughness with which they seem to have done this work. 
Their work on the peripheral retina is introduced with the 
following words. "We have already mentioned the experi- 
ments made by Titchener in 1906 with the view of testing the 
conclusions of Miss Fernald's first paper. The observations 
were rigorously confined to the black-white zone, and the out- 
come was definitely negative. In the meantime, however, the 
arousal of a colored after-image by a subliminally colored 
stimulus had been maintained for both the blue-yellow and the 
red-green zones. Unsystematic observations made in the 
Cornell Laboratory failed to confirm this result. It seemed 
worth while, however, to obtain further testimony; and Pro- 
fessor Baird, of the University of Illinois, very kindly consented 
to investigate the subject. The experiments were carried out 
by means of a simplified form of the Zimmermann perimeter, 
which permitted an accurate record of the degree of eccentricity 
at which the stimulus was exposed. Exploration was confined 
to the horizontal nasal meridian of each eye. The stimulus 
was a beam of light from an electric (16 c.p.) lamp, transmitted 
through appropriate combinations of gelatines and colored 
glasses; the colors employed were (non-equated) blue and 
yellow, red and green. Six of the most reliable laboratory 
students acted as observers, and Professor Baird had personal 
charge of the entire work. The after-images were projected 
upon white, gray and black grounds. The experiments proper 
were preceded by a careful determination of the outermost 
limits of color vision for the stimuli used, and all pains were 
taken to avoid chromatic adaptation" (pp. 376-377). 
Professor Baird reports negative results in every instance. 
With regard to this work the writers can not help but observe 
that Baird has failed to conform to the conditions which 
Fernald had said are essential for getting the phenomenon. 
Without drawing upon their own experiments for a knowledge 
of essential conditions, they will point out three conditions 
which Baird has apparently failed to fulfill, the neglect of any AFTER-IMAGE AND CONTRAST SENSATIONS 
199 
one of which is amply sufficient to account for his results being 
negative, (i) Fernald lays great stress upon the use of a cam- 
pimeter screen by means of the induction from which the 
brightness conditions were obtained which obscured the color 
in her stimulus.1 Baird used a simplified form of perimeter, 
how simplified Titchener and Pyle do not state. (2) Baird 
uses for the duration of the stimulation intervals of 30 to 40 
seconds. Fernald is careful to state that the interval of stimu- 
lation should not exceed three seconds.2 (3) In her descrip- 
tion of conditions Fernald states that the color should be ex- 
posed behind the opening in a campimeter screen, and the 
card upon which the after-image is projected should be slipped 
in between the colored surface and the stimulus opening. 
Thereby the campimeter screen and thus the larger part of 
the field of vision remains unmoved and the least possible 
incentive is given for involuntary eye-movement. With 
Baird's apparatus, however, we would judge that the ground 
upon which the after-image was projected must have been 
moved in between the stimulus and the observer's eye, thus 
exerting a strong incentive to drag the fixation with it. A 
very slight eye-movement indeed is amply sufficient to blot 
out or to prevent from • developing the instable peripheral 
after-image. 
1 While, as will be shown later in the article, the writers do not hold as Miss Fernald 
does that the inductive action of the campimeter screen is an essential or even a favor- 
able condition, still they do insist that a campimeter screen or its equivalent is neces- 
sary in order to be able to give a projection field for the after-image without causing 
an amount of involuntary eye-movement that would prevent the momentary and 
instable peripheral after-image from developing. 
2 With regard to the importance of this point the writers are in entire agreement 
with Fernald. In fact one does not need to work long with peripheral after-images 
to be convinced that so long an interval as Baird used is absolutely prohibitive of 
colored after-images even when a less excentric portion of the retina is investigated than 
was the case here. A short interval of stimulation is necessary because of the rapid 
adaptation of the peripheral retina to color. It is well known that adaptation to 
color in any part of the retina takes place rapidly at first and then progressively more 
slowly until a stationary point is reached. Working in the central retina the writers 
have found a similar curve of effectiveness for after-images. After-images seem to 
occur most intensively when the stimulus is removed while adaptation is still going on. 
If one carries the stimulus beyond a stationary point in adaptation, the after-image 
will weaken roughly in proportion to the length of time during which the stimulus is 
regarded after the stationary point has been reached. This is true with both intensive 
and slightly supraliminal stimuli. 200 
C. E. FERREE AND GERTRUDE RAND 
Failing to obtain the after-image in the central retina and 
to get confirmation by Baird that it may be obtained in the 
peripheral retina, Titchener and Pyle suggest explanations for 
the positive results gotten by Tschermak and Fernald. "The 
outcome of Tschermak's observations with the glass wedge 
must then in our opinion be explained by the prepossession of 
the observer and the roughness of the method employed. . . . 
It is less easy to account for the peripheral results." For a 
full statement of their explanation of the peripheral results 
the reader is referred to their article, pp. 378 ff. Brief mention 
only can be made of it here. We wish to comment on three 
points. (1) Their conception of the problem of getting color 
in the after-image when none is sensed in the stimulus is, we 
believe, different from that held and stated by Fernald. "The 
experimentum crucis" they say, "would be the production of 
a colored after-image in the achromatically adapted eye at a 
point lying well beyond the limits of blue-yellow vision." It 
is conceivable that two interpretations might be given to "well 
beyond the limits of blue-yellow vision." (a) The meaning 
might be well beyond the limits of blue-yellow vision whatever 
area and intensity of stimulus be used. And (b) it might be 
well beyond the limits at which these colors are seen when the 
stimuli used in the after-image experiments are employed. 
Wishing in every case to favor the point criticised, we choose 
the second interpretation. Thus to obtain a colored after- 
image well beyond the limits of blue-yellow vision would imply 
that a negative color excitation sufficiently strong to arouse 
sensation could occur in response to a stimulus which can under 
no condition arouse a positive color sensation. This is not at 
all the claim of Fernald nor of Thompson and Gordon as we 
understand their claim. They believe, in the cases cited by 
them, that the stimulus was prevented from arousing positive 
sensation by unfavorable brightness conditions. While it may 
not be clear from their work how these unfavorable brightness 
conditions prevent the positive sensation and permit the nega- 
tive, there can hardly be any doubt that they would not claim 
that colored after-images can be aroused in the retina at a 
point "well beyond the limits of color vision." In fact in her AFTER-IMAGE AND CONTRAST SENSATIONS 
201 
paper of 1905 Fernald specifically states that one should work 
just within the limits of color vision.1 
(2) It is pointed out by Titchener and Pyle that the limits 
obtained by Miss Fernald show a considerable range of varia- 
bility. The "experimentum crucis" thus has been inade- 
quately performed. She has in reality obtained the after- 
image within the limits of blue-yellow vision. As pointed out 
in (1), this, we believe, is not the point at issue. The question 
is not whether the after-image can be obtained beyond the 
limits of color vision, but whether it can be obtained anywhere 
in the retina when color is not sensed in the stimulus in a per- 
centage of cases large enough to preclude the possibility of its 
being due to chance happening, to error of observation, or what 
not. Apparently neither Fernald nor Thompson and Gordon, 
and certainly not Tschermak, have entertained the idea that 
the colored after-image can be obtained at any point on the 
retina where experiment shows that color can not be obtained 
in the positive sensation for the given stimulus under any con- 
ditions whatever. 
(3) Titchener and Pyle quote the following observations 
furnished by Miss Fernald by private correspondence.2 
1 In her paper of 1908 Fernald states that the after-images are perceived most 
frequently just inside or just beyond the regular limits for the color. As compared 
with her first statement, this may seen somewhat loose and might lead to misunder- 
standing. There can be little doubt, however, that beyond the regular limits means 
for her beyond the limits obtaining for some given set of brightness conditions 
taken as standard-not beyond the limits for all brightness conditions. Fernald, 
it will be remembered, worked with very little control of the factors extraneous to the 
source of light that influence the sensitivity of the retina to color. The boundaries 
of her color zones thus varied within quite wide limits. In this region through which 
the boundaries of the zones varied, she was able to obtain color in the after-image 
when none was sensed in the stimulus. Her ' regular ' limits fell somewhere within 
this region. She was thus able to obtain the after-image sometimes just within, some- 
times just beyond the regular limits. She apparently never obtained the after-image 
at a point beyond her widest limits of sensitivity to a given color. The present writers, 
working with a control of brightness conditions that enabled them to duplicate their 
limits from sitting to sitting within a degree of variation, never obtained the after- 
image beyond the limits determined under their most favorable brightness condi- 
tions. 
2 A request was sent to one of the present writers (Ferree) to determine whether 
these after-images could be gotten, at the same time the similar request was sent to 
Dr. Baird. The results quoted are samples of the observations made at that time. 202 
C. E. FERREE AND GERTRUDE RAND 
Observer: C. E. Ferree. Full illumination on bright day (May 17, 1908). Nasal 
meridian, right. White ground. Projection field white, except in obs. 14-17, when 
it was black. Stimulus, 13 sq. mm. Distance from eye to stimulus, 25 cm. 
Fixation 
Point 
Stimulus 
Color Seen 
After-image 
8o° 
0 
Dark gray 
Unsaturated light blue 
85° 
B 
Just dark 
Wash of unsaturated yellow 
85° 
Y 
Nothing 
Nothing 
8o° 
Y 
Tinge of dirty yellow 
Very pale blue 
8o° 
Medium gray 
Dark 
White 
8o° 
0 
Indefinite gray 
Nothing 
8o° 
Light gray 
Dark 
White 
75° 
Y 
Reddish yellow 
Good blue 
75° 
B 
Good blue 
Good yellow 
75° 
B 
Good blue 
Good yellow 
65° 
0 
Yellowish red 
Unsaturated blue 
65° 
Y 
Reddish yellow 
Blue 
6o° 
G 
Indefinite greenish gray 
Uncertain 
65° 
G 
Greenish yellow 
Dark red, more saturated 
than stimulus 
8o° 
Medium gray 
Dark 
Nothing 
8o° 
Medium gray 
Dark 
Nothing 
65° 
G 
No color 
Flash of red 
65° 
R 
No color 
Blue 
Commenting on these observations Titchener and Pyle say: 
"Positive results occur in the first two and last two observations 
of the series. The former may be explained in terms of chromatic 
adaptation. If as the illumination suggests the observer began 
the work in yellow adaptation, the first blue after-image would 
naturally follow." If the above explanation were adequate, 
the blue after-image should have been obtained also when gray 
was used as stimulus. As a check experiment gray was re- 
peatedly used as stimulus by Fernald, and in no case was the 
characteristic blue after-image obtained.1 Furthermore the 
same yellow light which in terms of the explanation gave the 
stimulation for the blue after-image fell with undiminished 
intensity on the projection field, hence there was no shutting 
off and, since the projection field was white, even no diminution 
1 In a footnote p. 382 we find: "These observations were taken after the limits 
had been roughly determined in previous experiments. If the determination of limits 
was made at the same sitting, and if the last test color employed was orange, there 
would be additional reason for an initial yellow-adaptation." In reply to this we would 
say that in case of our own experiments orange and yellow were purposely not the 
last colors used in determining the limits; that had they been, abundance of time was 
given for the complete recovery of the eye before the after-image experiments were 
performed; and that if chromatic adaptation had been present, its effect should have 
shown in the after-image when gray was used as stimulus. AFTER-IMAGE AND CONTRAST SENSATIONS 
203 
in the amount of the yellow daylight reflected to the eye for 
the period during which the after-image was observed. Thus 
no chance was given for a blue after-image to develop as the 
result of stimulation by the yellow in the daylight. Continuing 
Titchener and Pyle say: "If the second observation was taken 
at too short an interval of time the resulting blue-adaptation 
should show itself as a yellow after-image." We interpret this 
statement with some hesitation. It seems to mean that the 
first observation gave a blue after-image, and that if the second 
observation followed too closely upon the first, this after- 
image excitation in turn aroused a negative excitation which 
formed the physiological basis of the yellow after-image ob- 
served. If this interpretation is correct two conclusions would 
logically follow, (i) A negative after-image can itself give a 
negative after-image,-a phenomenon which, so far as the 
writers know, has never yet been observed. And (2) although 
a colored stimulus too weak to arouse positive sensation can 
not arouse a colored after-image, still the excess of yellow in 
clear daylight, which, reflected from the yellow stimulus, is too 
weak to be sensed as color, can arouse an after-image which 
not only can be sensed as color but which in turn can arouse 
a second after-image in which color can be sensed. "The two 
final observations suggest a shift of conditions. Green is seen 
at 65° as greenish yellow and at 6o° as indefinite greenish gray. 
It is possible that in the case in which 'no color' is reported 
the green simply escaped notice; peripheral colors at the limit 
of vision often appear as momentary flashes. Again red is 
reported at 65° as 'no color' although reddish yellow had been 
seen as far out as 750. It is possible that the flash of red 
escaped notice; it is also possible that red-adaptation from the 
preceding after-image brought out the blue." With regard 
to the possibility of the red and green escaping notice, the 
following points may be noted. (1) The stimulus color in 
this region of the retina has not so much a momentary character 
as the after-image color and would, therefore, not be so likely 
to escape notice as the after-image color. (2) The observations 
quoted above were made by one of the writers (Ferree), who 
here positively asserts that to the best of his knowledge this 204 
C. E. FERREE AND GERTRUDE RAND 
was not the case. (3) In our own experiments to be described 
later in this article, conditions were obtained under which the 
after-images of red and green were obtained in practically 100 
per cent, of the observations made. It seems scarcely possible 
that the color should have escaped notice in the stimulus in all 
of these cases and have been observed in the after-image. In 
case it be held that " red-adaptation from the preceding after- 
image brought out the blue," we are again asked to accept the 
thesis that an after-image excitation may in turn give rise 
to an after-image excitation strong enough to be sensed as color. 
II. Cases in Which After-Images and Contrast Sensa- 
tions are Aroused by Retinal Excitations Which 
Do Not Directly Condition Sensation 
1. After-Images 
The problem presented is: Can the color in a stimulus be 
obscured directly for sensation, and still set up an excitation 
upon the retina which will give an after-image? Our answer 
is that it can be done both in the central and the peripheral 
retina, but possibly more readily in the latter than in the 
former. To accomplish it, an experimental condition must be 
devised which will work against the color in the stimulus and 
relatively favor it in the after-image. A working principle is 
found in the difference in effect of brightness changes upon the 
saturation of the colors. 
This effect may be expressed as follows. (1) Brightness 
fused with color inhibits the color sensation. With the excep- 
tion of the region just within the limit of sensitivity for two 
colors,1 the following may be stated roughly as the law of this 
action for all colors for all parts of the retina. White inhibits 
1 Over a region 30 in width just within the limits of sensitivity as determined with 
the Hering pigment papers at full illumination, red and yellow have a higher limen in 
black than in white. In this zone red and yellow darken as compared with their 
brightness values in the central and paracentral retina. Added to the black of the 
disc, then, we have the black due to the darkening of the color. Thus the figures 
which are read from the disc do not express the actual amount of black added to the 
color. Whether or not we have an exception here to the law of the action of bright- 
ness upon color which obtains in the rest of the retina is thus open to question. At 
least, the exception as expressed by the measurements is exaggerated. AFTER-IMAGE AND CONTRAST SENSATIONS 
205 
most, the grays in the order from light to dark next, and black 
the least. (2) A given brightness change does not affect all 
colors to the same extent. If local and towards either black 
or white, blue and green lose their saturation completely before 
red, orange, and yellow; when the change is general and toward 
black, produced by decrease of illumination, yellow, red, and 
dark orange are obscured before green and blue. All of these 
changes were utilized at different times in our experiments both 
on the central and peripheral retina. It is readily seen that 
the technique of getting a colored after-image from a stimulus 
in which no color is sensed becomes merely a matter of fusing 
the least favorable brightness quality with the stimulus color 
and the most favorable with the after-image color. When this 
technique was carried out in its best form, the colored after- 
image was obtained in practically every case. 
i. In Central Vision 
Two methods of working were used in these experiments, 
one in which the brightness control came through changes in 
the general illumination of the field of vision, and the other in 
which the changes were local. The former method will be 
described first. 
After-images were obtained of red, yellow, and orange 
stimuli of the Hering series of colored papers. The work was 
done in a long narrow dark-room with a small window at one 
end near the center, darkened by a solid and carefully padded 
door. By opening and closing this door, the general illumina- 
tion of the room could be varied at will. The observer was 
seated in front of the window and to the left, so that the light 
coming in from above and behind fell free from shadow upon a 
screen placed in front of him. 
The intensity of the stimulus was decreased by decreasing 
the illumination. Two methods were used in doing this, differ- 
ing in the amount of dark-adaptation the eye had experienced 
before the stimulus was given. In the first, a preliminary 
determination was made of the illumination at which the 
stimulus was sensed as colorless. The experiment was then 
conducted at that illumination. The stimulus was exposed 206 
C. E. FERREE AND GERTRUDE RAND 
from 5 to 15 seconds, varying with the subject, and the after- 
image observed. In the second method, the experiment was 
started at full illumination. The door was opened and the 
stimulus covered with a card of the same gray as the back- 
ground. The door was then closed rather quickly until the 
point was reached at which the color is unperceived, this point 
having previously been carefully determined. The stimulus 
was exposed before the eye had become adapted to the de- 
creased illumination, and the after-image was observed. The 
advantage of the second method is that with it the eye loses 
its sensitivity to color at a more intensive illumination than 
when the first method is used. That is, the eye gains in sensi- 
tivity to color at any point in a series of decreasing illuminations 
if it is allowed to adapt to that illumination before the exposure 
is made. Thus, by taking advantage of the lessened sensitivity 
of the eye to color with decreased illumination before adapta- 
tion sets in, we were able to work at an illumination that gave a 
stronger objective stimulus, and was more favorable for the 
development of the after-image. Either method, however, gave 
us unmistakable color in the after-effect. In order that we 
might be sure that the observer worked with stimuli in which 
the color could not be sensed and that color really was dis- 
tinguished in the after-image, he was in the first place kept in 
ignorance of the color that was to be used in a given test, and 
in the second place was given at intervals in the series a gray 
for stimulus which could not be distinguished from the color 
at that illumination. Thus having run blank tests, and having 
used every precaution known to us to eliminate the influence 
of expectation, etc., we feel a reasonable degree of confidence 
in the results obtained. 
The after-images gotten in this way in central vision are not 
so saturated as those obtained in peripheral vision, but in pro- 
portion to their saturation last much longer than peripheral 
after-images. The after-image of red shows greatest saturation, 
dark orange next, and yellow least. As one would expect from 
its quicker rate of adaptation and from the fact that after- 
images decrease in intensity after the stationary-point in adap- 
tation is reached, the time of stimulation most favorable for AFTER-IMAGE AND CONTRAST SENSATIONS 
207 
producing the after-image was shorter for the red stimulus 
than for the other colors. The time for orange was next 
shortest, and for yellow the longest. 
Local changes in the brightness of color may be produced 
by objective mixing, by contrast, and by after-image. Of these 
three methods the addition of brightness as after-image is by far 
the most successful. By objective mixing the amount of colored 
light coming to the eye is reduced in proportion to the amount 
of brightness added. The color thus loses saturation from 
two causes: the reduction of the amount of colored light coming 
to the eye, and the inhibitive action of the brightness excitation 
upon the color excitation. For our purpose, it is of advantage 
to obtain the loss in saturation without any reduction of the 
amount of colored light coming to the eye, i. e., by means of the 
inhibitive action of the brightness excitation alone.1 This can 
be done by the addition of the brightness either as contrast 
or as after-image. Of these two, the after-image method yields 
the better results, as much more brightness can be added as 
after-image than can be induced as contrast. This is especially 
true of the central retina. The amount of inhibitive action 
that can be obtained by this means may be shown as follows. 
The limen of Hering red in a gray of its own brightness is for 
one observer 270. When the after-image obtained by a stimu- 
lation of 10 seconds to black is projected on this color, the 
limen is raised to 6o°. This after-image decreases in intensity 
very little for the first 8 to 10 seconds, and disappears after 
from 12 to 15 seconds. 10 to 12 seconds stimulation to color 
is quite sufficient to arouse an intensive colored after-image. 
Hence by a preexposure of 10 seconds to black and by the 
projection of the after-image upon a colored stimulus, it is 
possible to keep the color in a fairly intensive colored stimulus 
below the limen long enough to arouse the excitation necessary 
to give a colored after-image of considerable intensity. 
The method of working is as follows. A small colored disc 
composed of sectors of the color to be used and a gray of the 
brightness of this color, is set up on a color-mixer. A second 
1 As will be shown in a later paper, this action inhibits the effect of the color 
excitation, although it does not reduce its power to arouse after-image and contrast. 208 
C. E. FERREE AND GERTRUDE RAND 
disc of black of the same dimensions is set up beside it. As 
much color is put in the first disc as will be rendered just sub- 
liminal by the after-image obtained by a io second stimulation 
to the black disc. The colored after-image is then projected 
on a black field. The addition of white to the stimulus works 
against its color, and the black projection field favors the 
color of the after-image. Proceeding by this method, a col- 
ored after-image of good saturation and duration can be 
obtained in every case. The method can be made still more 
effective by adding the brightness to the stimulus color by both 
contrast and after-image. For example, the colored disc may 
be exposed through an opening in a black screen. The white 
induced by this screen would then add to the effect of the 
white after-image and in consequence still more color may be 
used as stimulus without arousing a color sensation. 
ii. In Peripheral Vision 
The color in the stimulus may be obscured in three ways. 
(a) It may be carried to the peripheral retina near the limit 
of sensitivity for the color used and a brightness be induced 
across it that works against its saturation, (b) The unfavor- 
able brightness may be mixed with it as after-image aroused by 
preexposure to a brightness stimulus, (c) The stimulus may 
be carried to some angle of indirect vision not too remote from 
the fovea and the general illumination be decreased until the 
color is lost. In (a) and (b) the brightness is mixed with the 
color by contrast and after-image rather than added to it on 
the color-wheel or by some other means of objective mixing, 
because, as stated before, objective mixing decreases the phys- 
ical intensity of the stimulus. This would decrease the energy 
of the positive color excitation, which would in turn decrease 
the energy of the negative excitation, and thus defeat the 
purpose of the experiment. But the addition of the brightness 
as contrast or after-image does not affect the energy of the 
stimulus; and while it reduces the effect of the positive excita- 
tion upon sensation, it apparently does not decrease its energy 
as retinal excitation, for it does not lessen its power to arouse 
after-images. This strongly indicates that the action of bright- AFTER-IMAGE AND CONTRAST SENSATIONS 
209 
ness upon color takes place at some physiological level posterior 
to the seat of the positive and negative color processes, as 
will be shown in the next paper of this series.1 Consider 
method (b} for example. Here the brightness is added as after- 
image. Its effect is to blot out the weakly supraliminal color 
in the stimulus, but it does not prevent the complementary 
color from appearing in the after-image. This is readily ex- 
plained in terms of the above hypothesis as to the level at which 
this action takes place. If, as we suppose, the inhibition takes 
place posterior to the level of the positive excitation, the nega- 
tive excitation is not weakened thereby. And since the bright- 
ness after-image which was added cannot itself leave behind a 
brightness after-excitation, nothing is carried over to the nega- 
tive color excitation to weaken its effect upon sensation. 
Hence by this method of working we should get as much color 
in the after-image as if no brightness quality had been added 
to the colored stimulus. 
While method (a) can also be used to obscure the stimulus 
color, still it is not so effective as (0 for obtaining the colored 
after-image; because (i) by means of it the zone of sensitivity 
cannot be narrowed nearly so much as by (Z?), and (2) the 
inducing field throws by contrast the same brightness quality 
across the after-image as it does across the stimulus, and so, 
roughly speaking, it inhibits the after-image as much as it 
inhibits the stimulus.2 In fact, in the way it has to be used in 
the peripheral retina method, (a) alone is of little use in getting 
our phenomenon. To make this method work successfully, 
some means would have to be devised for changing the quality 
of the inducing surface in the interval between the end of the 
stimulation and the beginning of the after-image. That is, 
the stimulation would have to be given on a field which induced 
a brightness quality unfavorable to the saturation of its color, 
1 See 'An Experimental Study of the Fusion of Colored and Colorless Light 
Sensations: The Physiological Level at which this Action Takes Place.' (In press.) 
See also abstract in Journ. of Phil., Psych., and Scientific Methods, 1911, VIII., p. 294. 
2 The second objection to this method does not apply to the use of the contrast 
method in the central retina (see p. 207), for in the central retina the colored after- 
image is sufficiently stable and durable so that an entirely new projection can be 
substituted for the screen which induces the contrast. In this case the effect of the 
induction of this screen can be eliminated from the after-image. 210 
C. E. FERREE AND GERTRUDE RAND 
and the after-image observed on a field whose inductive action 
was favorable to the saturation of its color. This change of the 
inducing field, however, can not be made for two reasons. (1) 
The duration of the peripheral after-image is very short. It 
comes as a momentary flash of color immediately after the 
stimulus light is shut off, and disappears before a change in the 
inducing field can be made. And (2) the after-image would be 
completely extinguished by the eye movement set up by shifting 
so large a part of the field of vision.1 
The investigation of favorable conditions was conducted in 
part by means of the vertical campimeter and in part by means 
of the rotary campimeter.2 The campimeter provides three 
possibilities for the variation of brightness conditions which 
was needed in our problem. By means of it, the brightness of 
the local preexposure, the brightness of the inducing field, and 
the brightness of the field on which the after-image is projected 
may be changed at will. The brightness of the local preex- 
posure and of the field on which the after-image is projected 
may be varied by changing the cards held in the frame behind 
the stimulus opening in the campimeter screen. The bright- 
ness of the inducing field may be varied by changing the paper 
covering the campimeter screen. All of these variations were 
used in our investigation. Fernald recognized the influence of 
only two of these variations: the campimeter screen and the 
projection field.3 And of the two the importance of cam- 
pimeter screen was very much overestimated, while the impor- 
1 See Ferree, C. E., 'The Fluctuation and Duration of the Negative After-image,' 
Amer. Jour, of Psychol., 1908, XIX., pp. 68, 87-97. 
2 For a description of the vertical campimeter see Fernald,"G. M., 'The Effect of 
the Brightness of Background on the Appearance of Color Stimuli in Peripheral 
Vision,' Psychol. Rev., 1908, XV., pp. 27-29. For a description of the rotary cam- 
pimeter see Ferree, C. E., 'A Description of a Rotary Campimeter,' Amer. Jour, of 
Psychol., 1912, XXIII. (In press.) 
3 The influence of preexposure under the ordinary conditions of working was 
apparently not at all realized by her. Her preexposure and projection field were in 
most cases made of the same brightness as the campimeter screen. No reason is 
assigned for doing this. Some indication of her reason may perhaps be had from 
Thompson and Gordon who also use a preexposure and projection field of the same 
brightness as the campimeter screen. They say: "When the stimulation had lasted 
the desired time, the screen was again put over the color, thus making with the 
campimeter a uniform gray surface upon which the after-image could be observed" 
(Ps/chol. Rev., 1907, XIV., p. 124). AFTER-IMAGE AND CONTRAST SENSATIONS 
211 
tance of projection field was relatively underestimated. More- 
over she seemed to entertain no clear notion of just how these 
factors produce their results. For example, the brightness 
induced by the campimeter screen was supposed to contribute 
in some fashion to the production of her phenomenon by 
working against the color in the stimulus and favoring it in the 
after-image; but just how it operates to do this was left in 
question. Apparently she had made no quantitative study 
of the fusion of brightness with color and of the relative in- 
fluence of the different brightness qualities upon the saturation 
of color. From the side of explanation, then, she can scarcely 
be regarded as having a definite point from which to start. 
Since the writers report the after-image in so much greater percentage of cases 
than Miss Fernald, and differ from her so much in their explanations of results, it 
may be well to discuss her work at this point in greater detail.1 They have the follow- 
ing comments to make on her general technique and her explanation of how it pro- 
duced the results she obtained. 
I. No systematic attempt was made to determine the factors influencing her 
phenomenon. Of the full list of achromatic factors, for exainple: brightness of the 
surrounding field, brightness of the preexposure,2 brightness of the projection field, 
and the degree of general illumination, she recognizes only brightness of the sur- 
rounding field, brightness of the projection field, and the degree of general illumi- 
nation. Especial stress is laid on the brightness of the surrounding field. As we have 
already shown, the influence of brightness of surrounding field is comparatively 
insignificant because it is exerted both upon the color of the stimulus and the color of 
the after-image. Thus if it is unfavorable to the one it will also be unfavorable to 
the other. It has a margin of influence, however, due to the fact that it does not act 
with equal strength upon the stimulus and the after-image. Its action is unequal 
because stimulus and after-image differ both in color and brightness. There would 
be on this account a small difference in the amount of induction by the screen in the 
two cases, due to the difference between the brightness relation of stimulus to back- 
ground and that of after-image to background; and further a probable difference in the 
1 Fernald's work rather than Thompson and Gordon's is singled out at this point 
because Fernald has made a more extensive investigation of the brightness factors 
and their relation to the phenomenon in question than have Thompson and Gordon. 
2 Miss Fernald always used some kind of colorless preexposure, but apparently 
she did not recognize its importance as a brightness factor. For her the preexposure 
card served in the main merely as a neutral covering for the colored stimulus until a 
steady fixation could be obtained and the observer be otherwise made ready for the 
exposure to color. Towards the end of the work reported in 1909, however (p. 29), 
she apparently began to realize that if the fixation were held for a period varying from 
three to ten seconds, the preexposure card would give a brightness after-image which 
would mix with the stimulus color and lighten or darken it, as the case might be. 
But under the ordinary conditions of working, she took no account of it as a brightness 
factor. 212 
C. E. FERREE AND GERTRUDE RAND 
inhibitive action of this induced brightness upon the color, due to the fact that the 
stimulus and after-image are different colors. The factors of predominant importance 
are the brightness of the preexposure and of the projection field. These are of greater 
importance (a) because more brightness influence can be exerted by means of them; 
and (&) because their action is more differential, i. e., the brightness of the preexposure 
acts upon the stimulus alone, and the brightness of the projection field acts upon 
the after-image alone. The brightness of the preexposure can thus be made to work 
against the color in the stimulus and to have no effect upon the color in the after-image, 
and the brightness of the projection field can be made to favor the color in the after- 
image and to have no effect on the color of the stimulus. 
2. No attempt was made to separate the achromatic factors and to study directly 
the relative importance of their effect upon the frequency with which the after-image 
may be obtained. However, in her study of the effect of achromatic conditions upon 
the limits of color sensitivity and upon color quality in stimulus and after-image, 
some attempt at separation was made, and in these experiments the observer was 
required to report cases in which color was sensed in the after-image when none was 
sensed in the stimulus. In order to study the effect of a given factor, the influence of 
all the factors but the one to be studied should be eliminated from the conditions of the 
experiment. If, for example, it were wanted to study the effect of the projection field, 
a gray of the brightness of the colored stimulus should have been chosen for the cam- 
pimeter screen and preexposure. In no case was this method of working employed. 
In her experiments to determine the effect of projection-field1 the following conditions 
were used: The campimeter screens were platinum white, and black; the projection 
fields for each screen were in turn of black, white, and medium gray; and the pre- 
exposure was in each case of the same brightness as the projection field. Since neither 
the preexposure nor the screen was chosen of the brightness of the colored stimulus, 
all the factors were present in each case, instead of one. And in her experiments to 
determine the effect of surrounding field, screens of white, black, and gray of the 
brightness of the color were used with preexposures and projection fields to match the 
screen in each case. In this case also it will be observed that when the brightness of 
the screen was varied the influence of neither of the other two factors was ruled out. 
3. Her statement of the most favorable brightness conditions for getting the 
after-image is in complete contradiction to all we have been able to find out directly 
concerning the phenomenon, or to infer from our investigations of the action of bright- 
ness upon color either in the positive sensation or in the after-image. For example, 
she states that white campimeter screen and a white projection field give the most 
favorable conditions for obtaining after-images of blue and yellow when no color was 
sensed in the stimulus. Stating her most favorable conditions for the after-images of 
red and green, she says: "In several instances in our later work a red after-image has 
followed an unperceived green when the stimulus was given on a white background 
and the dark screen [projection field] pushed over the color, and a green after-image 
was obtained for red and orange when the projection-screen was middle gray or black."2 
The writers were able to obtain after-images of the above colors under the conditions 
cited by Miss Fernald as most favorable, but instead of finding them to be the most 
favorable they found them to be almost the most unfavorable. As stated above, their 
most favorable conditions were black preexposure, black campimeter screen, black 
projection field; and their most unfavorable conditions were white preexposure, white 
1 Fernald, G. M., Psychol. Rev. Monog., 1909, X., pp. 37-45. 
2 Op. cit., p. 82. AFTER-IMAGE AND CONTRAST SENSATIONS 
213 
campimeter screen, and white projection field. The variation of the factors between 
these extremes gave the phenomenon with a frequency ranging between maximal and 
minimal. Working with the conditions she found to be most favorable, Miss Fernald 
was able to obtain the after-image in only approximately 31 per cent, of the total 
number of cases.1 With the conditions we have found to be most favorable the after- 
image may be obtained in practically every case. 
4. Miss Fernald's explanation of her results is also in contradiction to the facts 
as we find them. (^) She says that the most favorable conditions are gotten when 
the "projection-screens were determined so as most to emphasize the after-image color 
and the background least to favor the stimulus color." 2 The campimeter screen, 
then, that narrows the zone of sensitivity for a given color furnishes the most favorable 
conditions to be obtained for that color. The white screen is most favorable because 
"in so far as our results justify any conclusions concerning the color limit they seem 
to show that all the colors except the reds are perceived at a greater degree of eccentricity 
with the dark than with the light backgrounds. Red is seen as red to about the same 
degree of eccentricity with the dark and light backgrounds, but it is seen as yellow or 
orange with the dark background at the same points at which it is seen as colorless 
with the light background." 8 Her exception in case of red seems due to the fact that 
the limit for red is for her where first a trace of yellow comes in. If the limit for red 
had been taken at the point where the last trace of red is seen, as is usually done, red 
would have proved no exception to the other colors, and we could derive from her 
results the law that all the colors have a narrower limit with the white screen than with 
the dark or black screens. We do not find this to be true. For us blue and green 
have a wider limit with the black screen than with the white, and red and yellow a 
narrower limit. These results were obtained with the effect of preexposure eliminated 
and with the illumination of the optics room carefully standardized by a method to 
be described in a later paper. Neither of these precautions was observed in Fernald's 
work. Their importance may be shown by the following results. As compared with 
the gray of the brightness of the color, a preexposure of three seconds to black was 
found by one of the writers (Rand) to narrow the limits of sensitivity for red, green, 
and yellow 6°, and for blue xx°; a preexposure to white to narrow the limits for red 50, 
for green 2°, for yellow 40, for blue 70. At an eccentricity of 350 on the temporal 
meridian, a preexposure of three seconds to black raised the limen for red 58°, for green 
8o°, for yellow 320, and for blue 130; a preexposure to white raised the limen for red 
450, for green 55°, for yellow 120, and for blue 70. When the white campimeter screen 
was used, the changes in illumination from n A.M. until 4 P.M. on a bright day in 
September were found to vary the limits for the different colors from 40 to 6°. The 
change in the illumination of our optics room, lighted by skylight and provided with 
diffusion sashes to lessen the effect of external changes, from a bright morning to a 
cloudy afternoon was sufficient to vary the limits with a white screen from 6° to 140 
for the different colors; and with a black screen from 20 to 70. The greater variability 
is found for the white screen because the amount of contrast it induces is found to vary 
more with the change of illumination. (i) Miss Fernald seems to believe that the 
brightness process may have either a stimulating or an inhibiting effect on the color 
process. She says: "There seem to be two possible ways of explaining the action of 
brightness: Either the brightness of the stimulus has a direct inhibitory or stimulating 
1 Fernald, G., Psychol. Rev., 1908, XV., p. 33. 
2 Fernald, G., Psychol. Rev. Monog. Sup., 1909, X., p. 82. 
3 Op. cit., p. 23. 214 
C. E. FERREE AND GERTRUDE RAND 
effect on the color processes, or the brightness primarily affects the brightness sub- 
stance, and the activity in brightness substance has some differential effect on the 
color activity. . . . The fact that our most striking effects were obtained when the 
brightness is superimposed on the color, i. e., when the brightness is largely determined 
by contrast with a brightness background, or when the after-image is projected on a 
light or dark ground, seems at least to justify the statement that the superimposed 
brightness acts in such a way as to inhibit, increase, or modify the color activity." 1 
Her claim that all the colors have their widest limits with the black screen and their 
narrowest with the white seems to indicate that she considers that white increases the 
color activity and that black decreases it. She says: "This brightness factor is effective 
to a very limited extent in central vision, to a much greater extent in peripheral vision. 
As we go out into the peripheral retina, the action of the white process is needed more 
and more, i. e., the colors must be made lighter to be most strongly sensed." 2 Our 
results show that the brightness process, white, black, or gray, must be considered as 
always having an inhibitive action on the color processes. When the brightness 
process is added to the color process either as after-image or contrast, the amount of 
colored light coming to the eye remains unchanged and yet the saturation of the color 
is considerably decreased. Instead of being increased by the action of white and 
decreased by the action of black the saturation is decreased by the action of both, 
but much the most by the action of white for all colors for all parts of the retina with 
the exception of for red and yellow within a very narrow zone just within the limits 
of sensitivity. In this zone red and yellow have a higher limen in black than in white. 
It is not certain, however, that even in this zone black exerts the greater inhibitive 
action, for it is difficult to add the same amounts of black and white to red and yellow 
because these colors darken in the peripheral retina and it is hard to get a measure of 
how much black is added by this process alone. The relative amounts of inhibitive 
action by white, black, and gray of the brightness of the color can also be shown by 
the method of objective mixing. When equal amounts of white, black, and the gray 
are added in turn to a disc of color on the color mixer, the colored light coming to the 
eye is reduced an equal amount in each case, yet the apparent saturation of the color is 
very different on the three discs. It is much the greatest on the disc to which black 
has been added, next greatest on the disc to which gray has been added, and the least 
on the disc to which white has been added. Or to get a more exact measurement of 
the actions in each case, the color limen may be determined. An amount of color 
which is just noticeable when added to the black disc will give no color at all when 
added to the white or gray disc. A considerably larger amount must be added to the 
gray disc to give color, and a still larger amount must be added to the white disc than 
was added to the gray disc. 
Such a confusion as Miss Fernald and many others before her have fallen into 
with regard to the relation of brightness to intensity or strength of color is apt to come 
from the failure to separate the action of the brightness or achromatic excitation from 
the action of physical intensity or energy of the light waves coming to the eye. When 
a colored light of low energy strikes the eye it is both unsaturated as to color and for 
the most of the colors of low brightness. As its energy is increased it becomes more 
strongly sensed as color and in most cases lighter. That the two changes go hand in 
hand when the physical energy of the colored light is changed, however, does not jus- 
tify the inference that they will go hand in hand when there is no change in the energy 
1 Op. cit., p. 79. 
2 Op. cit., p. 73. AFTER-IMAGE AND CONTRAST SENSATIONS 
215 
of the colored light-in other words does not justify the inference that an increase in 
the white excitation will heighten the color excitation. The central principle of Miss 
Fernald's method of working is to change the achromatic excitation without affecting 
the amount of colored light coming to the eye. In all such cases, with the possible 
exception of red and yellow for the small zone mentioned above, the saturation of the 
color decreases rapidly as the quality of the achromatic component is made lighter. 
In addition it is scarcely necessary to point out that just because a color reaches its max- 
imal saturation at a given brightness as the energy of the colored light is increased, 
it does not follow that the achromatic excitation corresponding to this brightness is 
the most favorable in its action upon the color process. When the amount of colored 
light is rendered constant and the achromatic factor is varied, it is found that black is 
the most favorable quality of the achromatic series and the specific grays are favorable 
in proportion as they are near to black in quality. The gray of the brightness of 
yellow, for example, kills out the color in yellow when mixed with it almost as rapidly 
as does white and much more rapidly than does black and the darker grays. Thompson 
and Gordon, while in general agreeing with Fernald with regard to the effect of bright- 
ness relations, in one place seem both in addition and in contradiction to have fallen 
into this latter error. They say: "The effect of the background then seems to be this; 
that in a colored after-image, that color element is emphasized which in brightness 
approaches the brightness of the background [by background is meant here the field 
on which the after-image is projected], that is, on the lighter grounds the brighter 
element comes out and on the darker grounds the darker color element."1 
By way of clearing the ground for explanation, the 
writers have made a detailed study of the influence of the 
various brightness qualities upon color in the central retina, 
and for a large number of meridians in the peripheral retina. 
Since a statement of the results obtained will be given in full 
in a later paper, nothing more than a general statement will be 
attempted here. In the central retina, white reduces the 
saturation of colors the most, the grays in the order from light 
to dark next, and black the least. In the peripheral retina, this 
law holds for blue and green out to the limit of sensitivity for 
all of the observers worked with; and also for red and yellow 
for all of the peripheral retina except a very narrow zone just 
within the limit of sensitivity. In this zone, black apparently 
inhibits red and yellow most strongly-at least, red and yellow 
have a higher limen here when mixed with black than when 
mixed with white or the grays. An ultimate statement of 
most favorable conditions for all parts of the retina, then, 
would have to take into consideration all of the facts. It would 
involve a more detailed consideration of the topography of the 
retina than can be gone into in this paper. In formulating our 
1 Thompson and Gordon, op. cit., p. 128. 216 
C. E. FERREE AND GERTRUDE RAND 
experimental technique, however, we have avoided, we believe, 
the difficulty raised by the exception to the law found just 
within the limit of sensitivity by employing inhibitive condi- 
tions sufficiently strong to allow us to work nearer the center of 
the retina, where the exception has never been found. Thus we 
have been able to use as our working principle the law that 
white inhibits the most, grays in the order from light to dark 
next, and black the least. But to determine in accord with 
the law even thus simply stated, the brightness quality of 
preexposure, campimeter screen, and projection field that will 
give the conditions most unfavorable for the stimulus color 
and most favorable for the after-image color, is not easy. For 
example, in order to inhibit the stimulus color most strongly, 
the preexposure must be black, so as to give a white after- 
image to fuse with the color. This effect could be strongly 
intensified by having the black preexposure made through an 
opening in a white campimeter screen, and the color exposure, 
which comes immediately after and simultaneously with the 
after-image of the black preexposure, made through a black 
screen. This would secure the greatest possible intensification 
of the white after-image, and, therefore, the greatest possible 
amount of inhibition of the stimulus color. In order to favor 
maximally the saturation of the after-image color, the after- 
image should be projected on a field of very dark gray or black.1 
As to campimeter screens to be used during the projection of 
the after-image, we have a choice again of brightness qualities 
ranging from white to black. White and the grays in propor- 
tion to their whiteness would intensify by contrast the blackness 
of the projection field, while black would exert little if any 
influence. The black screen, then, is probably the safest to 
use while the after-image is being observed for two reasons, 
(i) If one is working with a general illumination at all inten- 
sive, white and the grays in proportion to their whiteness 
induce enormously in the peripheral retina. This amount of 
induction of black carries the brightness quality beyond that 
1 It may be well to state that in our study of the effect of brightness upon color, 
the black used was the matt black of the Hering papers. When we state that black 
favors the saturation of colors, this black is referred to. AFTER-IMAGE AND CONTRAST SENSATIONS 
217 
specified in our law as most favorable, namely, the blackness of 
the pigment of the Hering paper. (2) There is always danger 
that we may not be working far enough within the limit of 
sensitivity to escape the exception to our law of most favorable 
action. If, then, our formulation of conditions be correct, we 
should have a white campimeter screen during preexposure to 
black, and a black screen during both the exposure of the 
stimulus color and the projection of the after-image. This 
would involve two changes of the card behind the stimulus 
opening in the screen, and one change of the screen. The 
change of cards causes no disturbance in our phenomenon, but 
a change of the campimeter screen in the interval between the 
preexposure and the exposure to color would be fatal to the suc- 
cess of our experiments. This is because the white after- 
image would not last through the change, for reasons that have 
already been discussed; and the effect of the preexposure on the 
stimulus color would therefore be lost. We are thus limited 
to one screen for a single experiment, and our problem becomes 
to determine which of the brightness qualities acting con- 
tinuously through all three stages of the experiment will be 
the most favorable for our phenomenon. After rough pre- 
liminary tests, three screens were selected as representative 
of the action of all, namely, white, black, and a gray of the 
brightness of the color to be used. Of these three, the black 
screen was found to be much the most favorable. A considera- 
tion of the action of the three brightness qualities upon pre- 
exposure, stimulus color, and after-image shows sufficient rea- 
son for this. The effect of each is as follows: (1) The white 
screen intensifies the blackness of the preexposure, darkens 
the stimulus card, and, in the peripheral retina, especially if 
the general illumination is intensive, piles up the blackness on 
the projection field to a degree that is unfavorable to the 
saturation of the after-image color. (2) The gray screen of 
the same brightness as the color adds blackness both to the 
preexposure and to the projection field, but not so much as 
is added by the white screen. It has no effect on the stimulus 
card. (3) The black screen has no effect on the preexposure; 
it induces white on the stimulus card and thus adds to the 218 
C. E. FERREE AND GERTRUDE RAND 
effect of the preexposure on the stimulus color; and it exerts 
little or no effect on the projection field. A comparison of these 
effects shows that the black screen in all probability inhibits 
the stimulus color more than any of the others, and is less 
unfavorable in its action upon the after-image color. For this 
reason, it gives the most favorable brightness conditions for 
obtaining our phenomenon. Considering all the factors, then, 
we find that apparently the most favorable combination that 
can be made for our purpose is black preexposure, black cam- 
pimeter screen, and black projection field. The results of our 
experiments show this to be true. Under these conditions, 
color was obtained in the after-image in practically every case., 
The next most favorable condition was given by the white or 
gray screen with black preexposure and black projection field. 
The poorest results were obtained with a white preexposure and 
white projection field. With this combination, color could not 
be gotten in the after-image with any consistency of result, 
whatever brightness quality was used in the campimeter screen. 
Having thus worked our way through an explanation of 
our phenomenon and a determination of the conditions 
under which it can best be obtained, we will devote the re- 
mainder of the report of the work done by methods (<2) and 
(&) to a brief review of the results obtained in support of the 
various points that have been made. Our general thesis was 
that the phenomenon under consideration is but a special 
case of the difference in the inhibitive action exerted upon 
color by the different brightness qualities. Color may be 
obtained in the after-image when none is sensed in the stimu- 
lus, if an unfavorable brightness quality is fused with the 
stimulus color and a favorable one with the after-image color. 
In detail, our first point was that for all parts of the retina 
and for all colors, with the exception of two over a narrow zone 
just within the limits of sensitivity, white reduces the satura- 
tion of color the most, the grays in the order from light to dark 
next, and black the least. This law was generalized from the 
results of fusion and limen experiments in a large number of 
meridians of the retina. Our second point was that the 
combination of the preexposure and campimeter screen AFTER-IMAGE AND CONTRAST SENSATIONS 
219 
most unfavorable to the saturation of the stimulus color was 
the most favorable condition for obtaining our phenomenon. 
The effect of preexposure and campimeter screen upon the 
saturation of the stimulus color was measured in two ways: 
(i) by the effect on the limen color; and (2) by the effect 
on the limits of sensitivity.1 Data were thus obtained 
which we could directly correlate with the frequency with 
which color was obtained in the after-image, and so deter- 
mine whether our position was correctly taken. The results 
of this correlation show that, estimated in both these ways, 
the combination that proved the least favorable to the satura- 
tion of the stimulus gave color in the after-image in the largest 
percentage of cases; and, conversely, that the combination 
most favorable in its action on the stimulus gave color in the 
after-image in the smallest percentage of cases. A third 
point was that a black preexposure and black screen gave 
the brightness conditions that were most unfavorable to the 
saturation of the stimulus color. An estimation of the effect 
of the different combinations of screen and preexposure by 
either of the methods mentioned above brings out this point 
strongly. The least effect was found when preexposure and 
screen were both of the brightness of the stimulus color. A 
fourth point was that preexposure provides a stronger means of 
reducing the saturation of the stimulus color than does the 
screen. To make the test of this point absolute, the influence 
of one should be completely eliminated while the influence 
of the other is being determined. This can not be done. 
The eye must always have some preexposure, and there will 
1 See Rand, G., 'The Factors Which Influence the Campimetrical Observation: 
A Quantitative Examination and Methods of Standardizing.' (In press.) 
Of these two tests, the limen test has a much broader application, and measures 
much more directly what needs to be measured. It has a broader application because 
it can be made anywhere in the zone of sensitivity. It measures more directly what 
needs to be measured, because the results obtained show just how much color has to be 
present under a given condition to be sensed as color, while the results in the method 
of limits only express in terms of degrees how much the limit of sensitivity has been 
changed. This is a poor measure of how much the color in the stimulus has been in- 
hibited under a given condition: because, in the first place, it is not a direct measure of 
this action; and, in the second place, the results obtained cannot be even roughly ren- 
dered into terms of direct measurement, owing to the fact that the sensitivity of the 
retina near the limits does not fall off either gradually or regularly. 220 
C. E. FERREE AND GERTRUDE RAND 
always be a surrounding field. The effect of preexposure, 
however, can be minimized by choosing it of the gray of the 
brightness of the color to be used as stimulus, and by making 
the stimulation to it extremely short. Working in this fashion, 
we have only a slight local brightness adaptation to modify 
the color excitation immediately following, the effect of which 
can be taken as practically negligible. The influence of the 
screen also can be minimized in a similar way, by having it 
always of the brightness of the color to be used as stimulus. 
Isolating the action of preexposure and screen by this method, 
we estimated the effect of each in turn upon the saturation 
of the stimulus color, both by the effect on the limen of color 
and on the limits of sensitivity. Our results in both cases show 
that preexposure can be made much the stronger factor. For 
example, the limits of sensitivity were never made to vary 
more than 40 by the most extreme changes that could be made 
in screens, while it could be varied as much as 140 by changes 
in the preexposure. The difference stands out still more 
strongly in the effect on the limen, as would naturally be 
expected, since changes in the limen more directly express the 
differences in the inhibitive action than changes in the limits 
of sensitivity, as was shown in the footnote, p. 219. A fifth 
point was that black is the most favorable brightness quality 
for the projection field. This was shown very clearly by using 
projection fields of white, gray of the brightness of the stimulus 
color, and black, with each of the combinations of preexposure 
and screen, and comparing the results obtained. In addition, 
these results, when compared with those obtained by varying 
the preexposure and the campimeter screen, show that the 
brightness of the projection field is a very important factor- 
much more important than the brightness of the campimeter 
screen, and just as important, possibly more so than the bright- 
ness of the preexposure. 
Results in support of these points in general are given in 
Table I. In this table is shown the percentage of cases, based 
on twenty trials, in which color was obtained in the after-image 
when none was sensed in the stimulus. The object of the 
experiments was to determine the relative importance of the AFTER-IMAGE. AND CONTRAST SENSATIONS 
221 
three factors, preexposure, campimeter screen, and projection 
field. The method of experimenting in the following cases 
was to keep two of the factors constant and find the effect of 
varying the third. However, in order to show most effect- 
ively the most favorable conditions in decreasing order, the 
results have been grouped as follows. If the individual, 
horizontal columns are compared, the effect of campimeter 
screen is shown. If groups of three are compared, the 
effect of preexposure is seen. If compared in groups of nine, 
the effect of projection is seen. 
From the table it will be seen that projection field and 
preexposure are the most important factors. Of the pre- 
exposures, black is seen to be the most important. Its effect 
is greater on green and blue than on red and yellow. This 
is because the inhibitive action of the white after-image follow- 
ing the preexposure is greater for blue and green than for red 
and yellow. The full effect of preexposure was not obtained 
in our experiments because with a given black preexposure 
we did not work as near the center of the retina as we might 
have done. 
In method (c) also (see p. 208), the vertical campimeter 
was used to give the stimulus. Black was chosen both for 
the campimeter screen and for the projection screen in each 
case. Otherwise the procedure was the same as for after- 
images in direct vision. The results obtained were also similar, 
with the exception that not so much decrease of illumination was 
needed to obscure the stimulus color, and more saturated 
after-images were obtained. • The stimulus time in indirect 
vision must always be shorter than in direct vision. (From 
2 to 3 seconds was used.) This is due to the rapid exhaustion 
to color in indirect vision. After-images fall off in saturation 
if exhaustion to the stimulus color is carried beyond the station- 
ary-point. 
2. Contrast 
It was found that contrast could be induced for certain 
colors when the general illumination was sufficiently reduced 
to obscure the color in the inducing stimulus. Very strong 222 
C. E. FERREE AND GERTRUDE RAND 
Table I 
Showing the Percentage of Cases in which the Colored After-Image was 
Obtained when No Color was Sensed in the Stimulus, under All 
Variations of Screen, Preexposure, and Projection Field 
Campimeter 
Screen 
Preexposure 
Projection 
Red 
Yellow 
Green 
Blue 
Black 
Black 
Black 
IOO 
IOO 
IOO 
IOO 
Gray 
Black 
Black 
IOO 
IOO 
80 
70 
White 
Black 
Black 
SO 
4° 
So 
40 
Black 
Gray 
Black 
40 
30 
30 
20 
Gray 
Gray 
Black 
5° 
30 
0 
0 
White 
Gray 
Black 
SO 
4° 
0 
0 
Black 
White 
Black 
5° 
30 
40 
0 
Gray 
White 
Black 
4S 
20 
0 
0 
White 
White 
Black 
4S 
30 
0 
0 
Black 
Black 
Gray 
40 
70 
So 
4° 
Gray 
Black 
Gray 
20 
40 
20 
20 
White 
Black 
Gray 
20 
30 
0 
O 
Black 
Gray 
Gray 
0 
0 
0 
0 
Gray 
Gray 
Gray 
0 
0 
0 
O 
White 
Gray 
Gray 
0 
0 
0 
O 
Black 
White 
Gray 
0 
10 
0 
0 
Gray 
White 
Gray 
0 
0 
0 
0 
White 
White 
Gray 
0 
10 
0 
0 
Black 
Black 
White 
0 
0 
10 
0 
Gray 
Black 
White 
0 
0 
0 
0 
White 
Black 
White 
0 
20 
0 
0 
Black 
Gray 
White 
0 
O 
0 
0 
Gray 
Gray 
White 
0 
O 
0 
0 
White 
Gray 
White 
0 
0 
0 
0 
Black 
White 
White 
0 
0 
0 
0 
Gray 
White 
White 
0 
O 
0 
0 
White 
White 
White 
0 
0 
0 
0 
contrast was aroused under these conditions by standard red, 
dark orange, and yellow of the Hering series of colored papers. 
Contrast discs were cut as follows. The inducing surface 
was made of two colored discs; one 25 cm., the other 10.5 cm. 
in diameter. The contrast surface was made by placing be- 
tween these discs on the color-mixer a black and white disc, 
11.5 cm. in diameter. When these were rotated, the black 
and white mixed to give a gray ring 1 cm. in width, separating 
the two colored surfaces. The proportions of black and white 
were taken so that the gray ring matched in brightness the AFTER-IMAGE AND CONTRAST SENSATIONS 
223 
color used, the brightness of the color having been determined 
by means of Schenck's flicker photometer. 
The observations were made in the dark-room described 
above. The color of the inducing surface was obscured by a 
decrease of the illumination. As in the after-image experi- 
ments, this was done by two methods, one in which the observa- 
tion was made after the eye had adapted to the illumination 
chosen; the other before adaptation had set in. The latter 
method gave a much stronger effect, for although decrease of 
illumination within wide limits increases contrast effect in 
general, this increase is very much greater while the illumina- 
tion is decreasing. That is, if two determinations of contrast 
are made for the same inducing and contrast surfaces, one in 
which the judgment is passed while the illumination is decreas- 
ing, the other after the eye has become adapted to the illumina- 
tion at which the former judgment was made, it will be found 
that the former determination greatly exceeds the latter. Thus 
it appears that color induction is greatly enhanced while the 
retinal change corresponding to dark-adaptation is going on. 
This phenomenon will be treated more fully in a later paper 
on contrast. Our present purpose is satisfied with the con- 
sideration of the phenomenon at one point in the series of 
decreasing illumination; namely, the point at which the color 
in the inducing surface is obscured. This point varies greatly 
for the different colors used. It is the highest for red, next 
highest for yellow and dark orange, lowest for green, and next 
lowest for blue. 
The amount of contrast induced in each case was deter- 
mined by two methods. In each method a measuring-disc was 
used, compounded from discs of the proper colors and of black 
and white. In the first method, the comparison-judgment be- 
tween the contrast ring and the measuring-disc was made at the 
illumination at which the inducing color was obscured. This 
method may not be clear to the reader. It may seem, for 
example, that a degree of illumination that wholly obscures 
the inducing color would also wholly obscure the color of the 
measuring-disc. This is not true because of the different effect 
of the decrease of illumination upon the saturation of the dif- 224 
C. E. FERREE AND GERTRUDE RAND 
ferent colors, i. e., blue, green, and blue-green, the colors used 
for the measuring-discs, retain considerable color at the illumi- 
nation at which red, yellow, and dark orange, the inducing 
colors used, lose their saturation. The difficulty with the 
method is, that the measurements on the comparison-disc at a 
low illumination are not of standard value because of decrease 
of saturation, and thus convey little meaning to the mind of 
the reader. In order to get measurements in standard terms, 
a second method was resorted to, the results of which are more 
intelligible although the method of judgment is less accurate. 
In this method the comparison was made in terms of the satura- 
tion of colors on the measuring-disc at full illumination. Since, 
one of the terms of comparison is a memory-image, a time 
error was involved in this method. This was to some extent 
compensated for, however, by working both ways, i. e., a part 
of the judgments were made by first getting a memory-image 
of the contrast sensation at decreased illumination and com- 
paring that with the measuring-disc at full illumination, and 
a part were made by the inverse procedure. 
The tables recorded in the report are compiled from the 
results of Misses Chamberlain (C) and Rand (7?), fellows in 
psychology, of Bryn Mawr College, and Bunker (7?), graduate 
student. 
3. The Purkinje-Briicke Phenomenon 
The Purkinje-Briicke phenomenon was found by us to 
demonstrate in a very striking fashion that it is not necessary 
for the inducing excitation to condition sensation directly in 
order that color induction may take place. 
This phenomenon was first described by Purkinje in 1825.1 
He says: "Man liege ein weisses Quadratchen von der Breite 
zweier Linien auf einen schwarzen Grund, starre es 20-30 
Secunden an, und blicke sodann ins Schwarze hinein, so wird 
man ein noch dunkleres Viereck sehen, dessen Randes mit 
einem graulichen, sich allmahlich verlierenden Scheine umgeben 
sind. Lagt man auf den schwarzen Grund statt des weissen 
Quadratchens ein rothes, so zeigt sein griines Spectrum einen 
1 Purkinje, J., ' Beobachtungen und Versuche zur Physiologic die Sinne,' 1825, 
II., p. 107. AFTER-IMAGE AND CONTRAST SENSATIONS 
225 
Table II 
C. Showing the Amount of Contrast Induced when the Color in the 
Inducing Surface is Unsensed 
Stimulus 
Contrast Ring 
Direct Judgment 
Memory-image 
Judgment 
Red 
White 410 
Black 3190 
Green 1300 
Black 2300 
Green 1710 
Black 119° 
White 700 
Yellow 
White 236° 
Black 1240 
Blue 1510 
Black 1300 
White 790 
Blue 76° 
Black 2230 
White 6i° 
Dark orange 
White 82° 
Green 2100 
Green 1680 
Black 278° 
Blue 56° 
Black 940 
Blue 140 
Black 1720 
White 6° 
Table III 
Observer 5 
Stimulus 
Contrast Ring 
Direct Judgment 
Memory-image 
Judgment 
Red 
White 410 
Black 3190 
White 236° 
Black 1240 
White 82° 
Green 228° 
Green 2140 
Black 940 
White 5 2° 
Blue 163° 
Yellow 
Black 132° 
Blue 1210 
Dark orange 
Black 1420 
White 970 
Green 136° 
Blue 1180 
Black 1060 
Black 590 
White 138° 
Green 1520 
Blue 1060 
Black 1020 
Black 278° 
Table IV 
Observer R 
Stimulus 
Contrast Ring 
Direct Judgment 
Memory-image 
Judgment 
Red 
White 410 
Green 2250 
Green 1660 
Black 3190 
Black 1350 
Black 126° 
White 68° 
Yellow 
White 236° 
Blue 1370 
Blue 82° 
Black 1240 
Black 189° 
White 340 
Black 1590 
White 1190 
Dark orange 
White 82° 
Green 1950 
Green 1270 
Black 278° 
Blue 64° 
Black 520 
White 490 
Blue 440 
Black 1590 
White 300 226 
C. E. FERREE AND GERTRUDE RAND 
rothlichen Schein; auf gleiche Weise das blaue Spectrum einen 
orangen Schein, u.s.w. Man sieht heraus dass die objective 
Farbe nicht bloss in die Tiefe der Retina, sondern auch in die 
Breite einwirkt jedoch nicht gleich massig nach ihrer ganzen 
Ausbreitung, sondern zunachst an der Granze der heterogenen 
Beleuchtungen am intensivesten." 
The phenomenon with some modifications was next de- 
scribed by Brucke in 1851 / and by Aubert in 1862.2 The color 
or added brightness in the after-image was called by Brucke an 
after-effect of induction;3 by Aubert an after-image of contrast. 
Attention was again called to the phenomenon by Hering in 
1 Brucke (Pogg. Ann., 1851, LXXXIV., pp. 418-448) says (p. 47): "Einen dritten 
Beweis endlich kaum man aus der Beobachtung der negativen Nachbilder entnehmem 
welche nach diesem Versuchen zur Erscheinung kommen welche zeigen, dass die indu- 
cirten Farben als solche im Stande sind, complementaren gefarbte Nachbilder hervor- 
zurufen." His demonstration consists in getting the after-effect of looking towards 
the light through squares of red, green and violet glass with small discs of black paper 
placed at their centers. He describes the after-effect as follows: "Bei Anwendung 
des rothen Glases erscheint als negatives Nachbild eine helle rothe Scheibe auf dpnkel 
griinem Grunde. Hierein liegt nichts Auffallendes und dieser Erfolg wfirde sich nach 
Analogic der Versuche von Fechner erklaren lassen, auch ohne dass man eine Nach- 
wirkung der inducirten Farbe voraussetze. Wende ich aber das griine Glas an, so 
habe ich von der dunklen Scheibe ebenfalls ein helles rothes Nachbild und der Grund 
ist Schwarz, oder wenigstens so dunkel, dass ich seine Farbe nicht mit Sicherkeit 
habe unterscheiden konnen. Hier hat also das inducirte Grun Roth hervorgebracht. 
wahrend das inducirende gleichzeitig kein deutlich gefarbtes Nachbild erzeugte, 
In derselben Weise zeigte sich mir bei Anwendung des violetten Glases das negative 
Nachbild als eine gelbgrune Scheibe auf schwarzem Grunde." In two cases here 
Brucke apparently has the positive excitation induced across the black disc, and 
in two cases the negative excitation. But since he does not make induction apply 
to the negative excitation, as has been done later, he does not explain this case as 
after-image of contrast, but makes it instead analogous to the phenomenon described 
by Fechner. 
2 Aubert (Pogg. Ann., 1862, CXVI., pp. 249-279) says on p. 259: "Ausserdem ist 
es aber auffallend dass der simultane Contrast selbst noch einen successiven Contrast 
hervorruft, indem die durch den simultanen Contrast complementar gefarbten weissen 
Quadrate noch einmal complementare Nachbilder hervorrufen, so dass jene Quadrate 
im Nachbilde dieselbe Farbe, nur sehr bedeutend abgeschwacht haben wie der Grund." 
Aubert observed white squares on a colored ground. In the positive sensation the 
squares took on a tinge of color complementary to the background, and in the after- 
image of the same color as the background. That is, a white square on a red back- 
ground appeared greenish in the stimulus and red in the after-image. He says: 
"Besonders schon und mit einem eigenthumlichen Glanze erschienen die Nachbilder 
der weissen Quadrate auf dem blauen Streifen." 
3 Brucke applies the term induction to the positive excitation only. AFTER-IMAGE AND CONTRAST SENSATIONS 
227 
1878.1 Hering worked with achromatic sensation alone. He 
also called the phenomenon an after-image of contrast sensa- 
tion,2 and upon his observations based his arguments against 
Helmholtz's theory of contrast. Continuing the discussion, 
the experiments were extended to color sensations by Ebbing- 
haus3 who considered the phenomenon a combination both of 
after-image of contrast and of contrast induced by an after- 
image. 
The following form of experiment was adapted by us from 
Ebbinghaus. Squares 20X20 cm. of red, green, blue, and 
yellow Hering papers were fastened upon neutral gray back- 
grounds. Passing vertically through the center of these 
squares, gray strips 2X20 cm. of Hering papers numbers 8, 
24, 2, 41, respectively, were pasted. The after-effect of stimu- 
lation by this combination is a square of a color complementary 
to the color of the stimulus square, traversed by a strip of the 
same color as the stimulus square. The red square, for ex- 
ample, gives a green square traversed by a strongly saturated 
1 Hering, E., 'Zur Lehre vom Lichtsinn,' Wien, 1878, pp. 5-18. 
2 On pp. 5-18 Hering describes experimental conditions similar to those described 
by Purkinje. Here he calls the phenomenon successive light induction. On pp. 24-29 
he describes slightly different experimental conditions. Two dark gray strips of the 
same brightness, 3 to 4 cm. long and .5 cm. broad, are fastened parallel to each other 
2 cm. apart, one upon a white and the other upon a black ground. A fixation-point 
is taken on the boundary between the white and black backgrounds midway between 
the strips. By contrast the strip on the white ground looks darker than the strip on 
the black ground. When an after-image of the strips, the one lightened, the other 
darkened by contrast, is obtained, their brightness values are reversed, and a still 
greater brightness difference between them is found. This difference toward black 
on one hand, and toward white on the other, Hering calls an after-image effect of the 
brightness contrast induced in the stimulus. The phenomena obtained by these 
two sets of conditions are in every sense identical. One can scarcely see a reason for 
the separate treatment they have received by Hering. 
3 Ebbinghaus (Grundzuge der Psychologic, Erster Band, p. 239) says: "Man lege 
zwei massig grosse Blatter z. B. von sattgriiner Farbe so auf einen grauen Grund, dass 
nur ein schmaler, etwa 5 mm. breiter horizontaler Streifen zwischen ihnen freibleibt, 
und lasse diesen von einer unbefangenen Person eine Weile fixieren. Dann lasse 
man sie das Nachbild auf einem etwas unregelmassig geformten Grunde entwerfen, 
z. B. auf dem Fensterkreus, und frage, was sie sehe. Man wird so gut wie ausnahmlos 
die Antwort erhalten: 'einem griinen Streifen.' Der objektiv vollig neutrale Streifen 
hat durch die zweimalige Kontrastwirkung (im Vor- und Nachbilde), die sich von einer 
ausgedehnten Umgebung auf seine schmale Flache konzentriert, eine so intensive 
Farbung bekommen, dass er sofort die Aufmerksamkeit auf sich zieht, wahrend die 
rbthliche Nachbildfarbung seiner Nachbarschaft bei den Unregelmassigkeiten der 
reagierenden Flache in der Regel gar nicht beachtet wird." 228 
C. E. FERREE AND GERTRUDE RAND 
red strip; the green square, a red square traversed by a green 
strip. Now if the color of the strip is an after-effect of a pre- 
viously induced contrast, we have a strongly saturated, long- 
enduring after-image of an unsensed stimulus, for the brightness 
opposition of the gray strip to the inducing color, and the rather 
intensive illumination under which we worked, both combined to 
inhibit all contrast color in the stimulus. At this stage of the 
work we do not feel prepared to take positive ground on the ques- 
tion of explanation, but have the following evidence to offer 
that the Briicke interpretation is correct. 
i. Evidence that the Color in the Strip is an After-Image of a 
Previous Contrast Sensation, rather than 
Contrast in the After-Image 
(a) In the after-effect, the strip and square apparently 
develop, fluctuate, and die away independently of each other; 
the strip frequently develops before the square, especially if 
the stimulation has been very short; it invariably lasts longer 
than the square, returning several times after the square has 
finally disappeared; and in fluctuating, the two figures behave 
much as two after-images are observed to do, so far removed 
from each other as to be wholly without the sphere of reciprocal 
influence. The strip is frequently present when the square has 
disappeared, and vice versa. It rarely happens that their 
phases coincide, and when they do, the connection is obviously 
a chance one. 
Records on this and the following points were taken from a 
number of observers, both experienced and inexperienced. The 
results given in the following tables are typical. The work was 
done in a large optics room, lighted on one side by a bank of 
windows extending nearly to the ceiling. The observer, head 
in rest, was seated in front of these windows so that the light 
coming from above and either side fell uniformly upon the 
projection-field of engine-gray cardboard, I meter distant. 
The time of stimulation, unless otherwise stated, was 30 
seconds, and the unit of record was one second. The recording 
apparatus used throughout consisted of a Ludwig-Baltzar 
kymograph; a double electro-magnetic recorder, and two con- AFTER-IMAGE AND CONTRAST SENSATIONS 
229 
tact keys, one for strip and one for square; a Jacquet chrono- 
graph set to seconds; and a lamp rheostat to reduce the current 
from the lighting circuit. 
Table V 
C. Showing the Independent Phases of Visibility and Invisibility of 
Square and Strip 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis. + 
Invis. 
Red, 20X20 cm. .. . 
8 
5 
6.2 
56 
7-1 
57 
113 
Gray No. 8, 2X20.. 
15 
6 
5 
89-5 
4 
60 
149-5 
Green, 20X20 
2 
6 
5 
IS 
3-8 
7-5 
22.5 
Gray No. 24, 2X20. 
6 
7 
4.6 
32 
3-7 
22 
54 
Yellow, 20X20 
3 
17 
8 
32 
11 
33 
65 
Gray No. 41, 2X20. 
8 
18 
6.4 
58 
2.4 
19 
77 
Blue, 20X20 
4 
12 
6-5 
32.5 
5 
20 
52.5 
Gray No. 2, 2X20.. 
9 
11 
6.1 
6l 
3-7 
33 
94 
Table VI 
Observer B 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis.+ 
Invis. 
Red, 20X20 cm... . 
II 
12 
8-5 
102 
4 
44-5 
146.5 
Gray No. 8, 2X20.. 
18 
9-5 
4.2 ' 
79 
4 
72 
151 
Green, 20X20 
4 
5 
3-i 
15-5 
4-i 
16.5 
32 
Gray No. 24, 2X20. 
6 
7 
4-4 
3°-5 
3-8 
22.5 
53 
Yellow, 20X20 
3 
3-5 
3-i 
12.5 
7-5 
22-S 
35 
Gray No. 41, 2X20. 
7 
3 
3-i 
25 
3-3 
23 
48 
Blue, 20X20 
6 
13 
4.1 
28.5 
7-3 
44 
72-5 
Gray No. 2, 2X20.. 
11 
3 
3-6 
43 
3-3 
36 
79 " 
Table VII 
Observer R 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis. 4- 
Invis. 
Red, 20X20 cm. ,. . 
3 
5 
4.8 
19 
IO.6 
32 
51 
Gray No. 8, 2X20.. 
7 
5 
3-5 
28 
3-3 
23 
51 
Green, 20X20 
3 
2 
6.2 
25 
4 
12 
37 
Gray No. 14, 2X20. 
5 
7 
5-3 
32 
3-6 
18 
SO 
Yellow, 20X20 
2 
12 
7-7 
23 
9-5 
19 
42 
Gray No. 41, 2X20. 
7 
13 
5-5 
44 
2-3 
l6 
60 
Blue, 20X20 
3 
22 
9.8 
39 
4-3 
13 
52 
Gray No. 2, 2X20.. 
IO 
7 
4.2 
46 
3-i 
31 
77 230 
C. E. FERREE AND GERTRUDE RAND 
(&) The mutual independence of strip and square can be 
further indicated by a method of concomitant variations. 
That is, the strip can be made to fluctuate less and last longer 
without any corresponding change in the fluctuation and dura- 
tion of the square. And conversely, the square, or rather 
what in this variation corresponds to it, can be made to change 
its duration and rate of fluctuation, without any corresponding 
change in the fluctuation and duration of the strip. Both of these 
variations are based upon the effect of the arrangement relative 
to the direction of greatest involuntary eye-movement, upon 
the fluctuation and duration of a strip after-image. If the 
observer has more eye-movement in the horizontal than in the 
vertical, a strip after-image with its greater dimension in the 
vertical will fluctuate more and last a shorter time than one of 
the inverse arrangement. It may be stated as a law1 that 
whenever the direction of greatest eye-movement is along the 
shorter dimension of the after-image, the maximal fluctuation 
and minimal duration is attained for that form of after-image. 
Thus, to vary the duration and rate of fluctuation of the 
strip without changing them in the square, we need only to 
arrange the stimulus so that the longer dimension of the strip 
is first in the vertical and then in the horizontal. This rotation 
of the stimulus 900 will obviously have no effect upon the 
duration and rate of fluctuation of the after-image of the square, 
since both of its dimensions are equal. If we wish to make the 
converse variation, i. e., change the duration and rate of fluctua- 
tion of the outer figure without changing them for the inner, 
we shall obviously have to make the outer figure a strip and 
the inner a small square. Then by rotating the stimulus 900 
we shall increase or decrease the duration and rate of fluctua- 
tion of the outer figure, depending upon whether its shorter 
dimension is in the vertical or horizontal, while the duration 
and rate of fluctuation for the inner figure will not be affected. 
(c) The Briicke interpretation seems also to receive nega- 
tive support from the following fact. When one observes the 
1 Ferree, C. E., 'Intermittence of Minimal Visual Sensations,' Amer. Journ. of 
Psychol., 1908, XIX., pp. 101-103. For explanation of this phenomenon see same 
reference, pp. 126-127. AFTER-IMAGE AND CONTRAST SENSATIONS 
231 
Table VIII 
C. Showing by Method of Concomitant Variations a Decrease in Fluctuation 
and an Increase in Duration of the Strip with no Significant 
Change in the Phases of the Square 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis. + 
Invis. 
Blue, 20X20 cm. 
Gray No. 2, 2X20 
(vertical) 
3 
25 
iS-3 
6l 
5-7 
17 
78 
11 
12 
6.8 
8l 
3-3 
36 
117 
Blue, 20X20 
Gray No. 2, 2X20 
(horizontal) 
2 
48 
20.6 
62 
5-8 
n-5 
73-5 
9 
II 
9-4 
94 
4-3 
39 
133 
Table IX 
Observer B 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis.+ 
Invis. 
Blue, 20X20 
7 
6 
6.2 
5° 
3-9 
27 
77 
Gray No. 2, 2X20 
(vertical) 
13 
4 
4-1 
57 
2.2\ 
29 
86 
Blue, 20X20 
IO 
13 
4-4 
48 
2.4 
24 
72 
Gray No. 2, 2X20 
(horizontal) 
9 
5 
7 
70 
2.9 
26 
96 
Table X 
Observer R 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis.+ 
Invis. 
Blue, 20X20 
3 
17 
7-5 
30 
5 
15 
45 
Gray No. 2, 2X20 
(vertical) 
7 
7 
4-4 
35 
2.1 
15 
50 
Blue, 20X20 
3 
14 
7-8 
3i 
3-5 
14 
45 
Gray No. 2, 2X20 
(horizontal) 
3 
14 
10.5 
42 
3-3 
IO 
52 
Table XI 
C. Showing by Method of Concomitant Variations a Decrease in Fluctuation 
and an Increase in the Duration of the Square with no Significant 
Change in the Phases of Strip 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis. + 
Invis. 
Blue, 20X.5 
(vertical) 
Gray No. 2, .5X.5 . 
9 
0 
9 
9 
4-3 
9 
43 
9 
i-5 
14 
57 
9 
Blue, 20X.5 
(horizontal) 
Gray No. 2, .5X.5 . 
5 
0 
15 
7 
9-7 
7 
58 
7 
2.4 
12 
70 
7 232 
C. E. FERREE AND GERTRUDE RAND 
Table XII 
Observer B 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis.+ 
Invis. 
Blue, 2OX-5 
(vertical) 
Gray No. 2, .5X.5 . 
8 
0 
6 
7 
3-3 
7 
30 
7 
2-3 
18 
48 
7 
Blue, 2oX-5 
Gray No. 2, .5X.5 . 
4 
0 
IO 
8 
9 
8 
45 
8 
2-5 
10 
55 
8 
Table XIII 
Observer R 
Stimulus 
No. of Fluc- 
tuations 
1st Vis.|Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis. 4- 
Invis. 
Blue, 20X.5 
7 
5-5 4 
32 
2 
14 
46 
Gray No. 2, .5X.5 . 
0 
5 5 
5 
5 
Blue, 20X.5 
(horizontal) 
3 
43 17 
68 
I 
3 
7i 
Gray No. 2, .5X.5 . 
0 
IO IO 
IO 
IO 
stimulus through tissue-paper or under decreased illumination, 
the relative duration of strip to square in the after-effect is very 
greatly increased. This should not take place if a color in the 
strip after-image is induced by the color in the after-image 
square. The ratio should remain constant or approximately 
so, i. e., since the intensity of the after-image square has been 
reduced by weakening the color of the stimulus square by 
tissue paper, a corresponding weakening would be expected 
in the color induced in the strip. The observation thus seems 
to furnish a negative indication that the after-image interpreta- 
tion is correct. Relative to this observation, however, two 
points need to be taken into account, (i) An inspection of the 
tables shows that although there is an increase in the ratio of 
duration of strip to square, there is an actual decrease in the 
absolute duration of the strip. This might be supposed to 
indicate a tendency for the strip to vary with the square and 
thus to favor the contrast interpretation; but the increase of 
relative duration is too great for this to be probable. Besides, 
the decrease in absolute duration can be readily accounted for 
from the other side by a decreased retinal induction in the AFTER-IMAGE AND CONTRAST SENSATIONS 
233 
stimulus due to the decreased saturation of the square. (2) 
The second point is quite aside from differential evidence as 
between the Briicke and Ebbinghaus interpretations. When 
the stimulus was observed through tissue-paper or under de- 
creased illumination, considerable contrast color developed in 
the stimulus, where before there had been none, yet the after- 
image was less than in the former case. This seems to indi- 
cate the following. (#) The contrast sensation is only an 
equivocal index of the amount of excitation set up on the 
retina by a neighboring surface. This excitation may under 
one set of conditions arouse an intensive sensation, and under 
other conditions be equally strong, at least as far as after- 
effect goes, and excite no sensation. (&) Brightness opposi- 
tion inhibits only the contrast sensation. It apparently does 
not inhibit the corresponding retinal induction due to the 
neighboring surface, at least not its power to give after-effects. 
That is, the brightness opposition between the square and 
strip in the stimulus was greater when the tissue-paper was not 
used and yer the strongest after-image of the contrast excitation 
was obtained in this case. These suggestions are thrown out 
merely tentatively and are meant to apply only within the 
bounds of the evidence offered. 
Since the results are similar for both the tissue-paper device 
and the decrease of illumination, tables will be given only for 
the former. 
Table XIV 
C. Showing the Phases of Invisibility and Visibility of Strip and Square 
AND THEIR RELATIVE DURATIONS, OBSERVED UNDER TlSSUE PAPER 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av.Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis.+ 
Invis. 
Red, 20X20 
0 
3 
4 
3 
2-5 
3 
IO 
3 
19 
Gray No. 8, 2X20. . 
3 
3 
9 
Green, 20X20 
0 
7 
7 
7 
7 
Gray No. 24, 2X20. 
2 
5 
3-3 
IO 
6-5 
13 
23 
Yellow, 20X20 .... 
1 
8 
4.8 
9-5 
3 
3 
12.5 
Gray No. 41, 2X20. 
4 
8 
3-8 
19 
3-6 
14-5 
33-5 
Blue, 20X20 
I 
5 
3 
6 
12 
6 
6 
18 
Gray No. 2, 2X20 . 
4 
2-3 
14 
2-5 
IO 
24 234 
C. E. FERREE AND GERTRUDE RAND 
Table XV 
Observer B 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. In vis. 
Total 
Invis. 
Vis.+ 
Invis. 
Red, 20X20 . 
0 
2 
2 
2 
2 
Gray No. 8, 2X20. . 
3 
I 
i-9 
7-5 
4.8 
14-5 
22 
Green, 20X20 
0 
7 
7 
7 
7 
Gray No. 14, 2X20. 
2 
4 
4 
12 
5-8 
n-5 
23-5 
Yellow, 20X20 .... 
0 
4-5 
4-5 
4-5 
4-5 
Gray No. 41, 2X20. 
3 
4 
3-i 
12.5 
4 
12 
24-5 
Blue, 20X20 
I 
3 
2.5 
5 
7 
7 
12 
Gray No. 2, 2X20 . 
7 
4 
3-5 
28 
3-1 
22 
50 
Table XVI 
Observer R 
Stimulus 
No. of Fluc- 
tuations 
1st Vis. 
Av. Vis. 
Total Vis. 
Av. Invis. 
Total 
Invis. 
Vis. + 
Invis. 
Red, 20X20 
Gray No. 8, 2X20. . 
0 
3 
3 
3 
3 
2 
3 
8 
3-3 
IO 
3 
18 
Green, 20X20 
Gray No. 14, 2X20. 
0 
3 
8 
4 
8 
3-5 
8 
14 
3-3 
IO 
8 
24 
Yellow, 20X20 .... 
1 
6 
4-5 
9 
2 
2 
11 
Gray No. 41, 2X20. 
3 
5 
4-5 
18 
4 
12 
30 
Blue, 20X20 
1 
5 
4 
8 
5 
5 
13 
Gray No. 3 2, 2 X 20. 
4 
5 
4.6 
23 
3 
12 
35 
ii. But if a Contrast Effect, Evidence that it May Take Place 
when the Inducing Color is Unsensed 
As already stated, the strip frequently develops before the 
square; it invariably lasts longer than the square; returning 
several times after the square has finally disappeared; and in 
fluctuating its phases rarely coincide with those of the square, 
i. e., it is frequently visible when the square is invisible, and 
conversely. 
Thus it is immaterial for our thesis which interpretation be 
given to the phenomenon. For (a) if the after-color in the 
strip be a contrast sensation, our results show that it may be 
set up when the inducing excitation is not directly conditioning 
sensation; and (&) if it be an after-image sensation, they show 
that it may be aroused by a previous excitation which did not 
itself directly give rise to sensation. AFTER-IMAGE AND CONTRAST SENSATIONS 
235 
Fig. i. (Observer Ji.) Records showing the fluctuations and duration of strip and square. 236 
C. E. FERREE AND GERTRUDE RAND 
III. Explanation 
A later paper will contain a history of the observations on 
the effect of brightness changes, general and local, upon color 
phenomena, beginning with Purkinje.1 Purkinje's observa- 
tions cover the following points, (i) The relative difference 
in the brightness values of the spectral colors at full and de- 
creased illumination. (2) The difference in the effect of change 
in brightness upon the saturation of colors. (3) The changes 
in color-tone produced by changes in brightness. 
After-Image 
The explanation of the possibility of getting color as an 
after-image from a stimulus in which no color can be sensed 
rests in general with the second point of Purkinje's observa- 
tions; namely, the difference in the effect of change in the 
brightness of different colors upon their saturation. In every 
case in which a colored after-image was obtained from a color- 
less stimulus, it was gotten at a degree of brightness which 
worked against color saturation in the stimulus and relatively 
favored it in the after-image. Take, for example, the case of 
central vision. By the first method the influence of the bright- 
ness factor was introduced by means of a decrease in the general 
illumination. Yellow and red and orange lose their satura- 
tion at an illumination that permits of a supraliminal satura- 
tion of blue and green. The case, then, is simple. There was 
retinal excitation in the cases of red, yellow, and orange, but 
it was obscured for sensation by the brightness factor. The 
after-effect of this excitation, however, was not obscured for 
sensation. It was relatively favored, and, therefore, gave the 
after-images which were observed. There is nothing new or 
strange in principle about this phenomenon. The foregoing 
results show that it depends entirely upon the effect of the bright- 
ness changes in a color upon its saturation. These effects were 
observed and reported as far back as the time of Purkinje, 
and have been discussed sporadically in the literature from 
that time to this. By the second method the unfavorable 
'■Purkinje, 'Beobachtungen und Versuche zur Physiologic der Sinne,' 1823, I., 
p. 109. AFTER-IMAGE AND CONTRAST SENSATIONS 
237 
brightness quality was added to the stimulus color by objective 
mixing, contrast, or after-image; and the favorable brightness 
quality was added to the after-image by means of the projection 
field. The after-image method was found to be the most effect- 
ive for adding the unfavorable brightness quality to the colored 
stimulus because by means of it (a) the amount of colored light 
coming to the eye is not reduced as it is by the method of 
objective mixing; and (b) the unfavorable brightness quality 
added to the stimulus color is not added to the after-image 
color also, as is done under the conditions of the experiment by 
the method of contrast. 
For the after-image in peripheral vision, we have a slightly 
different case. The peripheral retina differs from the central 
with regard to the effect of change in brightness upon both 
the saturation and the quality of colors. With regard to 
saturation, we have in general merely an exaggeration of the 
condition found in the central retina, i. e., brightness changes 
produce greater difference in effect in the case of the different 
colors. With regard to color tone, the change is not in the 
same direction in every case as it is in central vision, i. e., in 
central vision at full illumination. In this paper, however, 
we are concerned with the effects upon saturation alone. More- 
over, under a given set of conditions more of the brightness 
quality can be added as after-image and contrast in this region 
of the peripheral retina than in the central retina because of the 
increased sensitivity of the former to achromatic after-image 
and contrast. We have then in these regards especially favor- 
able conditions in the peripheral retina for obscuring stimulus 
color by brightness changes, and for relatively favoring the de- 
velopment of the after-image. We have as an additional factor 
the enhanced sensitivity of the peripheral retina to adaptation 
and after-image effects. It adapts very rapidly to color stimuli 
and responds quickly with a vivid after-image of short almost 
momentary duration, often described as a vivid flash of color. 
All of these factors make it comparatively easy to get after- 
images of unsensed stimuli in peripheral vision, the only fact 
relative to our problem that needs to be explained. 238 
C. E. FERREE AND GERTRUDE RAND 
Contrast 
It was found to be especially easy to arouse green, blue- 
green, and blue as contrast sensations when their inducing 
stimuli do not directly excite a sensation of color. Two factors 
are involved in this result, (i) A decrease of the illumination 
obscures red, orange, and yellow before it obscures their con- 
trast colors, green, blue-green, and blue. Hence, induction 
and the effect of decrease of illumination upon it aside, there is 
reason in the nature of the color processes themselves why red, 
for example, should not be sensed when its contrast color, green, 
is sensed. But (2), in addition to this, decrease of illumination 
enormously enhances the induction of the contrast color, this 
effect being greater for green, blue, and blue-green than for their 
complementary colors red, yellow, and orange. 
The Purkinje-Briicke Phenomenon 
Until a final decision has been made between the after- 
image and the contrast interpretations, the Purkinje-Briicke 
phenomenon presents a two-fold problem. From the side of 
the after-image interpretation it must be explained how an 
after-effect so intensive and of such long duration can be 
obtained from a stimulus of so little apparent intensity; e. g., 
in our form of the experiment, no color at all could be sensed 
in the strip. From the side of the contrast interpretation, it 
must be shown how contrast color can be gotten in the after- 
effect of the strip when the square is not in sensation; i. e., 
before the square appears, in the invisibility phase of its inter- 
mittence, and after its final disappearance. 
At this stage of the work, an explanation will not be at- 
tempted. From the Briicke side, however, it may be pointed 
out that evidence has already been given that color in the con- 
trast sensation is only an equivocal index of the actual amount 
of the corresponding retinal color excitation. Also, that, while 
this induced excitation may arouse a strong sensation, a weak 
sensation, or even none at all, depending upon concomitant 
brightness conditions, contour, etc., nevertheless, in all of these 
cases, it gives rise to after-sensations of color, and to the most 
intensive in the cases we used, when no color was sensed in the AFTER-IMAGE AND CONTRAST SENSATIONS 
239 
stimulus. In fact, it was just to take advantage of this point 
that our method was devised. We introduced brightness 
opposition between strip and inducing color to prevent the 
induced excitation from arousing a sensation of color, knowing 
that its power to condition color in the after-image was not 
diminished thereby, providing the brightness conditions were 
favorable for the development of the color. The brightness 
conditions were made tolerably favorable by choosing a shade 
of gray for the stimulus strip whose after-image was approxi- 
mately the brightness of the after-image color. From the side 
of the Ebbinghaus interpretation, the writers have not at present 
even a plausible suggestion to offer in explanation. However, 
lack of explanation of this phenomenon does not limit its con- 
firmation of our thesis that stimuli in which no color is sensed 
may arouse after-image and contrast sensations in which color 
is sensed. 
We wish to state in conclusion that our chief interest in the 
problem has been its share in the broader problem presented by 
the Purkinje observations. We do not think that sufficient 
attention has been given to the second and third of these ob- 
servations and the light they may throw on color theory. Up to 
this time, theory has practically ignored the effect of brightness 
changes upon the saturation and quality of color sensation. 
These effects seem to argue a functional connection between 
chromatic and achromatic processes which should not be dis- 
regarded. [Reprinted from The Psychological Review, Vol. XIX., No. 5, Sept., 1912.] 
AN OPTICS-ROOM AND A METHOD OF STANDARD- 
IZING ITS ILLUMINATION 
BY C. E. FERREE AND GERTRUDE RAND 
Bryn Mawr College 
I. Introduction 
In a previous article1 the statement was made by the writers 
that all comparative estimates of the sensitivity of the retina 
to color (limens or limits) should be made in daylight instead 
of in the dark-room. This is to eliminate the influence of the 
field surrounding the colored stimulus, and of the preexposure. 
When the surrounding field is black, white is induced by con- 
trast across the stimulus color. Since the colors all differ in 
brightness,2 the induction takes place in different amounts for 
the different colors. This white, in proportion to its amount, 
reduces the action of the colors on the retina. Further, a 
given amount of white affects to different degrees the action 
of the different colors on the retina. To eliminate this two- 
fold unequal action, the surrounding field should be made in 
each case of the brightness of the color to be used. This can 
be done by working in a light-room of constant intensity of 
illumination and by making the surrounding field of a gray 
paper of the brightness of the stimulus color. In order to ac- 
complish this and at the same time be able to work in any 
meridian of the retina we choose, we have constructed a special 
piece of apparatus which we call a rotary campimeter.3 The 
influence of preexposure is even more important than of sur- 
1 Ferree and Rand, 'A Note on the Determination of the Retina's Sensitivity to 
Colored Light in Terms of Radiometric Units,' Amer. Journ. of Psychol., 1912, XXIII., 
p.33*- 
2 In a later paper, one of the writers (Rand) will show that it is of no advantage 
to equate in brightness in determining the limits of color sensitivity, and that harm 
results in so many ways from the attempt to equate, that it is doubtful whether it 
should be done even in determining the limens of color in the more sensitive parts of 
the retina. 
3 See C. E. Ferree, 'Description of a Rotary Campimeter,' Amer. Journ. of Psychol., 
1912, XXIII., pp. 449-453. 
364 365 
C. E. FERREE AND GERTRUDE RAND 
rounding field. If the preexposure is to black, white is added 
as after-image to the stimulus color. The effect of a black 
preexposure upon the stimulus color is greater than the effect 
of a black surrounding field because more white is added as 
after-image of preexposure than is induced by contrast from 
the surrounding field. This effect also can be eliminated only 
by working in a light-room of constant intensity of illumination 
and choosing as preexposure a gray of the brightness of the 
color to be used. 
Standardization for either one of these factors, however, 
can be accomplished for one degree of illumination only.1 As 
the general illumination changes, the relation of the brightness 
of the preexposure and of the surrounding field to the brightness 
of the colored stimulus changes. It is obvious, then, that 
if standardization is to be accomplished with regard to the 
influence of either of these factors, some means must be devised 
of maintaining the general illumination of the retina constant. 
No satisfactory method has as yet been obtained for keeping 
the illumination of a room by daylight constant. To keep it 
constant presupposes what has not as yet been provided, 
namely, a sensitive means of measurement. Constancy may be 
approximated by artificial illumination, but no artificial source 
of light has yet been devised which gives a light that approaches 
average daylight2 sufficiently closely in composition to warrant 
its use in color work. Of the various sources of light the 
Moore Tube comes nearest to doing this, but spectrophoto- 
metric and colorimetric determinations show that the light 
from it contains an excess of blue,3 and, therefore, although it 
1 When the colored light used to stimulate the retina is independent of the general 
illumination, e. g., when it is obtained from the spectrum, from monochromatic sources, 
or from standard filters, these two factors alone will modify the result of the color 
observation. If, however, light reflected from a pigment surface be used as stimulus, 
a change in the illumination will in addition change the amount of colored light coming 
to the eye. 
2 For results of measurements of the color values of average daylight, see Nichols, 
E. L., Transactions of the Illuminating Engineering Society, 1908, III., p. 301. Ives, 
H. E., 'The Daylight Efficiency of Artificial Illuminants,' The Illuminating Engineer, 
1909, IV., pp. 434-442; and 'Color Measurements of Illuminants,' Transactions of the 
Illuminating Engineering Society, 1910, V., pp. 189-207. 
3 See Ives, H. E., 'Color Measurements of Illuminants,' Transactions of the Il- 
luminating Engineering Society, 1910, V., p. 206; and Rosa, E. B. quoted by Moore, AN OPTICS-ROOM 
366 
has been adopted by various textile concerns for use in color 
matching, its substitution for daylight can scarcely be recom- 
mended for the more exact requirements of color optics. Ives 
and Luckiesh1 attack the problem of producing artificial day- 
light from another side. By their subtraction method they 
claim to have gotten the closest approximation to average 
daylight yet attained. They aim to cut out by absorbing 
screens the excess of red and yellow in artificial light due to the 
comparatively low temperature of artificial illuminants. Tung- 
sten lamps are used by them as the source of light, and two 
kinds of commercial glass approximating in their absorptive 
action cobalt blue and signal green are used as screens. In 
order to correct for the pronounced band of yellow-green trans- 
mitted by the cobalt blue, a film of gelatine dyed with rozazeine 
is also used. Although according to comparative measure- 
ments made by Ives and Luckiesh the light thus gotten is the 
closest approximation to average daylight yet obtained, still it 
shows a deficiency of 15 per cent, in the green and about 25 
per cent, in the blue. Moreover the spectrum of this' light 
does not show the brightness distribution of the spectrum by 
daylight. Since the absorbing screens cut down the light 
emitted by the tungsten lamp to 15 per cent, of its original 
intensity, the spectrum of the light finally given out shows 
the brightness distribution characteristic of lights of low in- 
tensity. We seem thus compelled either to give up the investi- 
gation of the color sensitivity of the retina for daylight illumina- 
tion, or to devise some means of keeping this illumination con- 
stant. At an early stage of our work of standardizing the 
factors extraneous to the source of light, we were compelled 
to take into account the influence of the changes in the illumina- 
tion of the visual field upon the color observation. The 
changes of illumination that took place from day to day, the 
progressive changes during the day, and the many sudden 
changes even in the course of an hour, rendered any constancy, 
or close reproduction of results entirely out of the question. 
D. McF., 'A Standard for Color Values,' Transactions of the Illuminating Engineering 
Society, 1910, IV., p. 224. 
1 Ives, H. E., and Luckiesh, M.,'Subtractive Production of Artificial Daylight,' 
Electrical World, 1911, LVIL, pp. 1092-1094. 367 
C. E. FERREE AND GERTRUDE RAND 
In order to obtain a standard illumination, two things are 
necessary: (a) A means of controlling the illumination must 
be provided, which is sufficiently sensitive to cause small 
changes, (F) A method of measuring the illumination pro- 
duced has to be devised; at least, a means must be secured 
for determining when an illumination has been obtained that 
is equal to a given preceding illumination. It is the purpose of 
this paper to describe an optics-room provided with means of 
control which we have found adequate to meet the above 
requirements; and to state a method of identifying and repro- 
ducing any given illumination of this room. 
II. Description of Optics-Room 
The dimensions of the room are I2j4 X io ft. It is situated 
on the upper floor of an isolated building and is lighted by a 
skylight8 X 7/4 ft- Beneath the skylight two diffusion sashes, 
4 X 7/4 ft., are swung on hinges so that they can be raised or 
lowered as desired. The framework of these sashes is made of 
light-weight iron. For convenience of local control of illu- 
mination, if needed, each sash is divided into four units by 
means of cross-pieces. The sashes are filled with double- 
strength glass ground on one side, so adjusted to the frame 
that they can be removed easily for cleaning or for the substitu- 
tion of some other kind of glass in case that is desired. This 
glass diffuses the light so effectively that local shadows cast 
by the cross-pieces in the framework of the skylight are com- 
pletely eliminated, while the sudden changes of illumination 
produced by the passage of the sun behind a cloud are reduced 
to a minimum. This diffusion seems to have the further 
advantage of reducing the yellowness of direct sunlight below 
the limen of sensation. At least, when working under the 
sash, the observer never judged a gray exposed through the 
campimeter opening as yellow under any local conditions, as 
frequently happened when working under direct sunlight. 
The room is planned also so that small changes of illumina- 
tion can be produced, ranging from the intensive illumination 
of a south-exposure skylight to the blackness of a moderately 
good dark-room. Two provisions are made for this, (i) The AN OPTICS-ROOM 
368 
diffusion-sashes are made so that any or all of the panes of 
ground glass can be quickly and easily taken from the sash, 
and anything can be substituted that is desired; or the illu- 
mination can be varied by placing layers of tissue paper above 
the glass. (2) The room is provided with two curtains 
mounted on heavy spring rollers. One is a white curtain made 
of thin muslin; the other is a black light-proof curtain so 
mounted that, when drawn, its edges are deeply enclosed in 
light-proof boxing extending along the four walls of the room. 
One or both of these curtains can be drawn any distance that 
is desired, and the illumination can thus be changed gradually 
from a very intensive brightness to a fairly good blackness. To 
aid in getting dark-room effects, the doors of the room are care- 
fully boxed and curtained. One requirement of a perfect dark- 
room, however, is lacking, namely, the walls and floor of the 
room are painted white. This is because it is of advantage in 
the light-room work, and bpcause complete blackness is not 
needed in the type of work for which the room is devised. 
III. Method of Standardizing 
As stated earlier in our paper, no satisfactory means of 
determining the amount of daylight illumination in a room has 
been provided by the physicist, so there is little hope at this 
time of solving the problem from that side. The brightness 
induction of the peripheral retina, however, has been found 
by us to be extremely sensitive to changes in the general 
illumination. This phenomenon seems to provide us with a 
sensitive measure of these changes while, at the same time, it 
represents the combined effects for sensation of the principal 
subjective factors that might vary from day to day.1 To apply 
the method in its most sensitive form, the inductive power of 
white was chosen because it is the most strongly affected by 
illumination changes. For example, when No. 14 Hering gray 
was used as stimulus and white as campimeter screen, a 
noticeable change was produced in the induction when the 
white curtain of the optics-room was pulled forward 1 cm.2 
1 This means of identifying the illumination of a room was devised by Rand. 
2 The sensitivity of this method of detecting changes in the general illumination 
was compared with the sensitivity of the Sharpe-Millar portable photometer. In 369 
C. E. FERREE AND GERTRUDE RAND 
from a position in which its edge was directly above the long 
axis of the campimeter. This caused a change in the illumina- 
tion of the room so small that it could not be directly sensed. 
Further, at 11 o'clock in the morning of a bright day in Sep- 
tember, when a point at 250 on the nasal meridian was stimu- 
lated, one of the writers (Rand) reported that the white screen 
induced black across the stimulus No. 14 gray to an amount 
that caused it to equal in brightness 107° of black and 2530 of 
No. 14 gray; at 2 o'clock of the same day the induction was 
increased until the No. 14 gray matched 1500 of black and 2100 
of the gray; at 4 o'clock of the same day the No. 14 gray 
matched 1800 of black and 1800 of the gray.1 Working at 250 
in the temporal meridian, this observer reported at different 
times during one day and on different days, the wide variations 
shown by the following figures: 283° of black, 2250, 1450, 1900, 
238°, etc. Another observer (Miss Campbell) reported less 
induction, but her variations from time to time were equally 
great. At 250 in the temporal meridian, she found at different 
this photometer one of the comparison fields is illuminated by the light of the room and 
the other by a standard tungsten lamp enclosed in the photometer box. When the 
room is illuminated by daylight, the field receiving the light of the room is seen as white 
while the field lighted by the tungsten lamp appears as a saturated orange. The dif- 
ference in color between the two fields renders the photometric judgment difficult and 
makes the instrument very insensitive for daylight tests. For example, our tests 
showed that by the method for identifying an illumination described in the text, a 
change in illumination could be detected which was produced by drawing the white 
curtain I cm. from a position in which its edge was directly above the long axis of the 
campimeter. But with the receiving surface of the portable photometer in precisely 
the same position as the stimulus screen of the campimeter, the edge of the curtain had 
to be moved 11.3 cm. in order that the change of illumination might be detected. 
Moreover, this amount of change could be detected only in case the photometric 
field was continuously observed while the curtain was being drawn, in which case 
the comparison field was observed to be slightly darkened. The judgment was made, 
then, in terms of a just noticeably different brightness of the field which was illuminated 
by the daylight, rather than in terms of a disturbance in the brightness-equality of the 
two fields. When, on the other hand, the judgment was made in terms of a just 
noticeable disturbance in the equality of the two fields, as the judgment would have 
to be made if the photometer were to be employed for the reproduction of any former 
illumination taken as standard, the curtain had to be drawn 44.2 cm. before the change 
could be detected. This j. n. d. represents an amount of illumination equal to 2.5 
foot-candles. 
1 This increase in the inductive action of the screen caused by the decrease in 
illumination, was accompanied by a shrinkage of the zones sensitive to color covering 
an area of 40 to 6°. AN OPTICS-ROOM 
370 
times 8o° of black, 1030, 1600, 1750, etc. After a careful study 
of the phenomenon with different screens and with different 
backgrounds, the inductive action of the white screen upon 
a stimulus of No. 14 Hering gray, at 250 in the temporal 
meridian, was found to provide the best means of detecting 
changes in the illumination of the optics-room. At this point 
on the retina, the induction was by no means minimal, nor 
was it sufficiently great to cause the medium gray chosen for 
our stimulus to appear too dark to give a small j. n. d. of 
sensation. Having thus provided ourselves with a means of 
producing small changes of illumination and a method of 
detecting them, we had in order to complete our work but to 
choose an illumination for each observer, which could be used 
as standard. Since we wished to work on both light days 
and days of medium darkness, an average had to be chosen as 
our standard from the measurements obtained on a number 
of days ranging from light to dark, so that on bright days, the 
i tom could be darkened, and on dark days it could be lightened 
until this value was obtained. For observer A (Rand) an 
illumination was selected which caused an induction of black 
across no. 14 gray stimulus viewed at 250 in the temporal 
meridian to an amount which caused the gray stimulus to 
equal in brightness 2100 of black and 1500 of no. 14 gray; for 
observer B (Ferree) 1800 of black and 1800 of no. 14 gray; 
and for observer C (Campbell) 1450 of black and 2150 of no. 
14 gray. The amount of black induction was identified in 
each case by means of a measuring-disc made up of sectors 
of black paper and no. 14 gray of the Hering series. This 
measuring-disc was carried by a motor and placed just behind 
the 250 point. The observer fixated the 250 point and com- 
pared the gray of the measuring-disc as seen in central vision 
with the gray of the stimulus seen 250 from the fovea. The 
sectors of the measuring-disc were then changed until the two 
grays were of equal brightness. 
Previous to each series of observations the illumination of 
the room was changed until the amount of brightness induc- 
tion was brought to the value chosen as standard. It was 
tested at intervals during the sitting and was readjusted when 371 
C. E. FERREE AND GERTRUDE RAND 
necessary. Details of the method of doing this are as follows. 
When the white screen and the no. 14 gray stimulus had been 
put in place, the observer took his position and adjusted the 
fixation-knot in front of the motor for the 250 point on the 
temporal meridian. The measuring-disc set at the standard 
value was mounted on the motor. The observer reported 
whether the stimulus appeared lighter or darker than the 
measuring-disc, or of a brightness equal to it. If the judgment 
lighter or darker was given, the curtain was drawn one way, 
or the other until the stimulus accurately matched the measur- 
ing-disc in brightness. 
This method not only gives a sensitive measure of the 
changes of illumination of the visual field and a successful 
means of standardizing the illumination of a room by daylight, 
but it has in addition advantages for work in psychological 
optics not possessed by an objective standardization, could 
that be successfully obtained. The problem of standardiza- 
tion includes more for the psychologist than it does for the 
physicist, for the former has variables to take into account 
in addition to the changes that may take place in the energy of 
the stimulus. Even though the illumination of the room be 
made objectively constant, we should expect variations in 
the response of the retina to this illumination because of its 
own changes from time to time. Brightness contrast, for 
example, might be expected to vary from sitting to sitting 
even when the stimulus conditions are kept absolutely con- 
stant. Two factors would be concerned in these variations: 
changes in the inducing power of the surrounding parts of the 
retina, and changes in the sensitivity of the local area. These 
changes would take place even when the usual precautions 
known to the experimenter in this field have been observed. 
Such precautions are commonly limited to fatigue, adaptation, 
etc. These precautions do not provide for the changes that 
occur in the retina from day to day. Moreover, they do not 
adequately guard against the variations to which they are 
intended to apply, for no precaution can adequately guard 
against a change in a factor, unless some measure of that 
factor be had. So far as the writers know, in these general AN OP TICS-ROOM 
372 
precautions intended to keep the state of the retina constant, 
no measure of the variable factor has been provided to test 
the adequacy of the method. The method proposed by us, 
however, is planned with this in view. It takes into account 
not only the objective, but the subjective variables, and 
reduces both to a constant. For example, when no. 14 gray 
surrounded by the white field is made equal to the measuring- 
disc composed of 2100 of black and 1500 of the no. 14 gray 
for observer A, it means that the observation may be begun 
with the assurance that the total result of all the factors-the 
illumination of the room, the local sensitivity of the retina, 
and the inductive action of the surrounding parts of the retina 
-is the same as in the preceding observation. 
What has just been said should not be considered as more 
than a general statement of the application of the principles 
of the method. In actual practice a greater refinement of 
working may be attained. If, for example, one wishes to 
use a preexposure differing in brightness from that of the 
colored stimulus, and doubts whether a test which covers only 
the local sensitivity of the retina and the inductive action of 
the surrounding parts is a sufficient check upon the after- 
image sensitivity, he may make his standard include the effect 
of the pdeexposure he wishes to use. In short, if he does not 
consider adequate the more general test we have described, 
he may duplicate, in establishing his standard, any com- 
bination of brightness factors, due to preexposure, brightness 
of screen, or what not, that he may wish to use in his experi- 
ment proper. 
The test of a method is how well it works. The test of 
this method is that we shall be able closely to duplicate our 
results from sitting to sitting regardless of the changes in the 
outside illumination from day to day or from morning until 
afternoon. The method stands the test. Long series of 
observations in the peripheral retina show a very small M.V. 
■-much less even than is shown in the ordinary color obser- 
vations in the central retina, where, as compared with the 
peripheral retina, the factors extraneous to the stimulus exert 
little influence. 373 
C. E. FERREE AND GERTRUDE RAND 
The following table has been compiled from a number of 
observations to show the variations in the results of color 
limens and color limits (a) when the general illumination was 
controlled according to the method described above, and 
(&) when no more precautions were observed than were used 
by previous investigators. In previous investigations of the 
color sensitivity of the peripheral retina, care has been taken 
to work only at the same hours of days that appeared equally 
bright, or, if on days of different brightness, to make a rough 
approximation of preceding illumination by means of curtains 
without using either a definite standard or means of measuring. 
For our work with the illumination controlled, the gray of the 
brightness of the color at the illumination selected as standard 
was used for the preexposure and the campimeter screen. For 
the work without especial control of the illumination, the 
gray of the brightness of the color for one of the days selected 
as typical was used throughout for preexposure and screen. 
This gave in the first case complete elimination of the effect 
of preexposure and surrounding field, and in the second case 
elimination as complete as could be gotten without accurate 
control of the general illumination. Results are given in the 
table for blue and green only because the sensitivity to these 
colors is affected most by changes of illumination. 
Stimulus 
Illumination 
Screen and Preex- 
posure 
Variations of 
Limits on Differ- 
ent Days 
Variation of 
Limens on Differ- 
ent Days 
Green 
Controlled 
Gray No. 9 
Gray No. 8 
O° 
O01 
Uncontrolled 
4^6° 
6o°-82° 
Blue 
Controlled 
Gray No. 33 
Gray No. 32 
0° 
2°-3°2 
i8°-3o° 
Uncontrolled 
4°-5° 
1 The limen for green was taken in both cases at 250 on the temporal retina. 
2 The limen for blue was taken in both cases at 400 on the temporal retina. [Reprinted from The Psychological Review, Vol. XIX., No. 6, Nov., 1912.] 
THE EFFECT OF CHANGES IN THE GENERAL 
ILLUMINATION OF THE RETINA UPON 
ITS SENSITIVITY TO COLOR1 
BY GERTRUDE RAND 
Bryn Mawr College 
I. Introduction. 
II. Historical. 
III. Experimental. 
(1) Quantitative estimate of the influence of change of illumination upon the 
induction of brightness by the surrounding field. 
(2) The effects of these amounts of induction on the limits of color sensitivity. 
(3) The effect of these amounts of induction on the limens of color at different 
degrees of excentricity. 
(4) The influence of change of illumination upon the effect of the preexposure 
on the limens and limits of color. 
IV. Conclusion. 
I. Introduction 
It is the purpose of this paper to show the effect of changes 
in the intensity of the illumination of the field of vision upon 
the results of investigations which deal with the sensitivity 
of the retina to color.2 
A method of standardizing the illumination of the field of 
vision was described in an earlier paper.3 
II. Historical 
The effect of the general illumination of the retina on color 
'From the Bryn Mawr Psychological Laboratory. 
2 With regard to this effect two cases may be recognized: (i) When the colored 
light used to stimulate the retina is independent of the general illumination, e. g., 
when it is obtained from the spectrum, from monochromatic sources, or from standard 
filters; and (2) when it is obtained by reflection from pigment surfaces. In the first 
case the effect is exerted in the following ways: (a) by changing the brightness relation 
of the preexposure to the colored surface, (b) by changing the brightness relation of 
the surrounding field to the colored stimulus, (c) by altering the sensitivity of the 
retina to brightness after-image and contrast and thus changing the effect of the bright- 
ness of the preexposure and of the surrounding field upon the sensitivity of the retina 
to the colored stimulus. To these effects is added in the second case a change in the 
amount of colored light coming to the eye. 
8 Ferree and Rand, 'An Optics-room and a Method of Standardizing its Illumi- 
nation,' Psychol. Rev., 1912, XIX., pp. 364-373. 
463 464 
GERTRUDE RAND 
sensitivity has been recognized since the time of Purkinje and 
Aubert. It has been studied in some detail by a number of 
experimenters, among whom may be mentioned Kramer and 
Wolffberg. Both have shown that the sensation aroused by 
the colored stimulus is weakened by a reduction of the general 
illumination but neither has given a method of keeping the 
general illumination constant. Kramer's1 purpose was to 
determine the sensitivity of the eye under different intensities 
of daylight and artificial illumination. His method was as 
follows. Stimuli, 4 mm. square, of blue, yellow, red, and 
green paper on a black background were used. The distance 
at which the stimulus had to be placed from the observer to 
be just recognized as colored was tested by sunlight and when 
the sky was obscured by clouds and for three intensities of 
each of the following sources of artificial illumination: candle- 
light, gas, petroleum, sodium, potassium, strontium, and 
calcium lights. His results show the following facts: (1) Red 
is seen at the greatest distance in all lights except calcium, 
in which case green is seen when placed farther away than red. 
The other colors are recognized in the order green, yellow, 
blue. (2) All the colors are recognized at a greater distance 
when seen by sunlight than when illumined by artificial light 
or the dull light from a clouded sky. (3) As the intensity of 
the artificial illumination is decreased, the colors must be 
placed nearer the eye to be recognized. In drawing his con- 
clusions with regard to comparative sensitivity, Kramer ignored 
the white contrast which the black background induced across 
the stimuli. The induction across stimuli whose sizes were 
only 4 mm. square must have been considerable. It was, 
moreover, of different amounts in each case; because brightness 
contrast is greatest when there is maximal brightness opposi- 
tion. The modification of the light colors, as a result of 
contrast induction, must, therefore, have been greater than 
that of the dark colors. Wolffberg's2 interest was in the 
influence of gradual alterations of the general illumination on 
1 Kramer, J., 'Untersuchungen uber die Abhangigkeit der Farbenempfindung 
von der Art und dem Grade der Beleuchtung,' Inaug.-Diss., Marburg, 1882. 
2 Wolffberg, 'Ueber die Priifung des Lichtsinnes,' A A. 0., 1887, XXXI., pp. 1-78. RETINAL SENSITIVITY 
465 
the light and the color sensitivity of the central and of the 
peripheral retina. His room was illuminated by daylight 
entering through a window. Fifteen different degrees of 
illumination were produced by fastening from one to fifteen 
thicknesses of tissue-paper over the window. The illumination 
obtained when the window was uncovered was called 15/15; 
when covered with one thickness of tissue-paper, 14/15, etc. 
His method of determining the effect of variations of illumina- 
tion upon the central retina was as follows: Pigment stimuli 
were placed at a standard distance of 5 meters from the 
observer, and the size of stimulus necessary to render it just 
visible in its true color was determined. In the peripheral 
retina, he investigated to what extent the limits of white 
and of colored stimuli were altered by reducing the illumina- 
tion. In all his experiments, the stimuli were fastened on a 
black background. Wolffberg's results for the central retina 
are shown in the following table. The stimuli were circular 
in shape and of diameters given in columns 2, 3,4, 5, and 6. 
Illumination 
Size of Red 
Stimulus 
Size of Blue 
Size of Green 
Size of Yellow 
Size of White 
15/lS 
.5 mm. 
3 mm. 
3 mm. 
1.5 mm. 
.2 mm. 
I4/I5 
i-5 
5 
4 
2 
•5 
I3/I5 
2 
6 
6 
4 
I 
12/15 
2.5 
12 
12 
4-5 
2 
II/I5 
3 
20 
20 
5 
2-5 
5/15 
10 
5° 
50 
10 
6 
These results show that in the central retina a decrease of 
illumination has greater effect upon the sensation of color 
than upon the sensation of white. Wolffberg next tested the 
effect of a gradual decrease of illumination upon the limits of 
sensitivity to white and to the colors. He found that the 
extent of the visual field was not narrowed for white when 
the illumination was decreased to 1/15. The color limits, 
however, narrowed gradually when the illumination was de- 
creased from 15/15 to 3/15. The narrowing was in no case 
more than 150. The relative extents of the fields remained 
unaltered, i. e., the order of size was in every case blue, red, 
and green. 466 
GERTRUDE RAND 
Although special investigations have been conducted by 
Kramer, Wolffberg, and others to show the effect of changes in 
the general illumination upon color sensitivity, in general little 
if any precautions have been taken by earlier experimenters to 
prevent' such changes when investigating color sensitivity. 
Either the experimenter has not considered the influence of 
the general illumination, or he has been satisfied to take 
the rough precaution to work only on bright days at stated 
hour's. Ole Bull,1 for example, commented at length on the 
factor of general illumination, but suggested no method for its 
standardization. He writes: "The amount and nature of the 
general illumination are of more significance in perimetrical 
observations than one is accustomed to consider. It must 
always be noted whether the sky is clear or cloudy, whether 
it rains or snows. The extreme limits of the visual field for 
mixed light undergo such wide fluctuations that it is of little 
value to establish an average limit on the basis of a number of 
measurements. Changing illumination, conditioned by the 
time of day and of year during which the work is carried on, 
as well as the locality in which it is undertaken, produce varia- 
tions in the same stimulus large enough to cause differences of 
from io° to 20° [in the limit of sensitivity]. Especially in the 
nasal parts of the retina does the illumination influence the 
color limits, while their position remains more constant in the 
temporal retina." Fernald,2 however, did make some attempt 
to obtain a standard illumination. She arranged white cur- 
tains at the windows of her optics-room, which could be lowered 
on bright days and drawn on dark days. This rather crude 
method was used also by Thompson and Gordon.3 It is 
scarcely necessary to point out that the method lacks the first 
essential of standardization, namely, a means of measuring. 
It is surprising that Wolffberg, as the logical corollary of 
his work, did not draw attention to the importance of standard- 
1 Ole Bull, 'Perimetrie,' Bonn, 1895, p. 8. 
2 Fernald, G. M., ' The Effect of the Brightness of Background on the Extent of 
the Color Fields and on the Color Tone in Peripheral Retina,' Psychol. Rev., 1905, 
XII., p. 392. 
3 Thompson and Gordon, 'A Study of After-images on the Peripheral Retina,' 
Psychol. Rev., 1907, XIV., p. 122. 467 
RETINAL SENSITIVITY 
izing the illumination of the visual field in all work on the color 
sensitivity of the retina, and show how it could be accomplished 
by a modification of his method of working. He already had 
at hand one of the essentials for standardizing, namely, a 
method of changing the illumination of his room. The other 
essential, a method of measurement by means of which an 
illumination could be identified with a previous illumination 
chosen as standard, might have been derived from his results. 
For example, it would seem to have been a simple matter for 
him to have chosen as standard the particular illumination at 
which the red stimulus of 2.5 mm. diameter, the blue and green 
of 12 mm. each, the yellow of 4.5 mm., and the white of 2 mm. 
were just recognizable at a distance of 5 m. Stimuli of these 
sizes, it will be seen from the tables, were just recognizable 
at this distance at the illumination called 12/15, when 15/15 
represents the illumination "bei giinstige Tagesbeleuchtung." 
Using this condition as an index of the standard illumination, 
he could at any time have adjusted the illumination of the 
room by adding to or subtracting from the layers of tissue- 
paper covering the window, until the stimuli of these sizes 
were again just recognizable at the given distance. The 
accuracy and sensitivity of this method could have been 
tested by comparing the results of a series of determinations. 
An accurate and highly sensitive method sustaining some 
similarity in principle to the method suggested here is de- 
scribed in another paper in this volume of the Review.1 
III. Experimental 
The effect of change of illumination was forced upon our 
attention early in the investigation of the factors that influence 
the color sensitivity of the retina. For example, in preliminary 
work done by the writer on a well-lighted porch on Long 
Island, changes in color-tone were observed, when certain 
colors were compared in the central and in the peripheral 
retina, that are not found at all under the more intensive 
illumination of our optics-room when neither of the curtains is 
1See footnote 2, p. I. 468 
GERTRUDE RAND 
drawn; and the peripheral limits of color were narrower by 
5° to 12°. Furthermore, on a dark day, it was found that the 
limits of stimuli exposed through an opening in a white screen 
were reduced by about 40 as compared with the limits taken 
on a bright day. The change was less considerable with black 
and gray screens. The change in color-tone was most con- 
spicuous in case of green.1 On dark days, the green stimulus 
appeared as a pale unsaturated blue before becoming colorless 
in passing from the center to the periphery of the retina. This 
zone of blue was from 70 to 230 wide, in different meridians of 
the retina, with both white and black screens, but was wider 
with the black than with the white screen. On a sunny day, 
on the other hand, with the white screen green passed into 
bluish-green, then directly into gray, except in case of the 
upper regions where it appeared blue throughout a zone of 
about 40 in width. With the black screen, the blue zone was 
found only in the upper and temporal regions of the retina. 
The transition of green to yellow in the periphery that is 
generally reported in the literature was found in these experi- 
ments only when the gray screen was used. Yellow showed a 
color change that varied in amount with the degree of the 
general illumination. On a bright d^y with the white screen, 
it appeared reddish-orange. On a cloudy day, it was seen in 
the extreme periphery as a dark saturated red. 
Working in our optics-room we found also that results 
taken on one day could not at all be duplicated on the following 
day. When the work was carried on under the most favorable 
conditions without special means of controlling illumination, 
namely, on bright days only, differences of 50 or more were 
found when the white screen was used. This necessitated a 
long series of observations if legitimate averages were to be 
obtained. Such a procedure is at best a poor makeshift and is 
besides of great disadvantage in many problems that come up 
in the work on color sensitivity. Particular instances of this 
may be found in investigations in which it is required to work 
in the region lying just within the limits of sensitivity, and in 
work on the after-images of stimuli in which no color is sensed. 
1 The green of the Hering series was used. RETINAL SENSITIVITY 
469 
In the latter case the experiment requires that the stimulus be 
exposed just outside the limits of sensitivity determined with a 
given brightness condition, and that the observer should not 
be aware of the nature of the stimulus. In order to fulfill these 
requirements the experimenter must know the limits obtaining 
with a given brightness condition. It would be impossible to 
know this when the brightness conditions were subjected to the 
influence of changing illumination unless re-determinations 
were made at the beginning of each sitting and even frequently 
during its course. This would consume a great deal of time 
and would, besides, only roughly fulfill the requirements of the 
problem. A further and still more important example of the 
disadvantage may be found in the task we had set ourselves, 
namely, to investigate from point to point the sensitivity of the 
retina to each of the principal colors for three backgrounds in 
at least sixteen different meridians. In this work it is obvious 
that unless a standard illumination were provided, all com- 
parative work would have to be done at one sitting. This is 
impossible. When time is taken between observations to guard 
against fatigue, at least three hours is required merely to out- 
line the limits of sensitivity for a given color with one back- 
ground for only one half of the retina. Even for this length 
of time there is no guarantee that the illumination has not 
altered. Thus at the outset of any extended investigation of 
color sensitivity, it is evident that, without a standard illumina- 
tion, results will be of little comparative value. 
In order better to know our factor and the ways in which 
it operates, a systematic investigation of the influence of 
changes of general illumination was carried on in our optics- 
room which is especially constructed to secure fine changes in 
illumination.1 The experimentation was conducted by means 
of a rotary campimeter, described in full by Ferree, in the 
July number of the .American Journal of Psychology.2 Three 
observers acted as subjects. Since the results of all three are 
1 A description of this optics-room was given in the Psychological Review, 
September number, 1912, pp. 367-368. 
2 Ferree, C. E., 'Description of a Rotary Campimeter,' Amer. Jown. of Psychol., 
1912, XXIII., pp. 449-453. 470 
GERTRUDE RAND 
similar in their general bearing on the problem, space will be 
taken for the results of two of them only, A and C. 
Rough preliminary experiments showed that the primary 
effect of decreasing the illumination was an increase in the 
amount of contrast induced across the stimulus by the campim- 
eter screen. With the white screen, the increased induction 
was the most pronounced and was sufficient to cause large 
changes in the limits and in the color-tone of the stimulus. 
In order to investigate this effect in detail, gradual changes of 
illumination covering a wide range were made by means of the 
curtains with which our optics-room is furnished. Attention 
was given to the following points, (i) A quantitative estimate 
was made of the influence of change of illumination upon the 
brightness induction of the campimeter screen. (2) The effect 
of this induction upon the limits of color sensitivity was deter- 
mined. (3) The limens of the colors were measured at dif- 
ferent degrees of excentricity at different illuminations. And 
(4) the influence of change of illumination upon the effect of 
the preexposure on the limens and limits of color was investi- 
gated. The degrees of illumination chosen for comparison were 
the standard illumination, the method of obtaining which was 
described in an earlier paper,1 and a decreased illumination 
which was similar to that obtaining on a cloudy afternoon. 
Measured in foot-candles by means of the Sharpe-Millar 
portable photometer, the standard illumination equalled 390 
foot-candles, the decreased 1.65 fpot-candles. 
1. Quantitative Estimate of the Influence of Change 
of Illumination upon the Induction of Brightness 
by the Surrounding Field 
The purpose of this investigation was to find out (1) how 
much the induction from white and black screens2 is affected 
by a change in the general illumination; and (2) how much 
induction is gotten at decreased illumination from the gray 
screen which matches the color at standard illumination. The 
1 Ferree and Rand, op. cit. 
* White and black screens are chosen because they represent the extreme cases 
of the effect of change of illumination. RETINAL SENSITIVITY 
471 
induction in this latter case is caused by the change in the 
brightness relation between color and screens with decrease of 
illumination.1 The campimeter screens served as inducing 
surface, grays of the brightness of the four principal colors of 
the Hering series both at standard and decreased illumination 
were used in turn as stimuli, and the amount of induction was 
estimated upon a measuring-disc, made up of adjustable 
sectors of the gray of the stimulus and white or black, according 
to the screen used. The measuring-disc was mounted on a 
motor which could be moved along the graded arm of the 
campimeter to any position from 20° to 920. The gray stimulus 
was exposed through the opening of the screen in the usual 
manner. Two preliminary precautions were observed, (a) 
Since the brightness of the gray stimulus plus the induction 
of the screen was to be estimated by means of the measuring- 
disc, and since the brightness-value of the stimulus and of the 
disc changes with the amount of light that falls upon them, 
it was necessary to make sure before each measurement that 
the same amount of light fell upon each. This precaution was 
all the more necessary because the stimulus had to be placed 
behind the screen and the measuring-disc in front. In a given 
position of the apparatus, one or the other was apt to be shaded. 
The determination was made as follows: Measuring-disc, 
campimeter screen, and gray stimulus were all given the same 
brightness-value according to determinations made under con- 
ditions about which no doubt of the equality of the illumination 
of each could be entertained. Each was then placed in position 
for the experiment, and the position of the campimeter as a 
whole and of its various parts was adjusted until stimulus, 
screen, and measuring-disc were exactly matched in brightness- 
value. When an exact match was obtained we were guaranteed 
1 This latter determination is made to show that it is impossible to standardize 
the brightness of the surrounding field against the sudden and progressive changes of 
daylight that occur during the course of a single series of observations. These changes 
alter the brightness relation between the colored stimulus and the gray used as screen; 
therefore a match made at the beginning of a series will not hold throughout its course. 
For the same reason and to an equal degree the brightness relation between preexposure 
and colored stimulus changes with change of illumination. It is, therefore, equally 
impossible to standardize the brightness of the preexposure without some means of 
securing a standard illumination. 472 
GERTRUDE RAND 
that all three were again equally illuminated, (b) The ques- 
tion arose whether brightness induction comes to its maximal 
value at once in the peripheral retina. A determination of the 
intensity curve of the contrast sensation was accordingly made 
at various points in the peripheral retina. It showed that 
contrast increases strongly for the first few seconds of stimula- 
tion. For this reason it was found to be necessary to make 
the judgment concerning the amount of induction of the 
screen, just as long after the induction had commenced as was 
done in the experiments to determine color sensitivity. In 
the color experiments an interval has to be allowed before the 
stimulus is exposed during which the observer obtains a steady 
fixation. During this interval of preexposure, the eye is being 
stimulated by the campimeter screen and by the card which 
covers the stimulus. To prevent the preexposure card from 
giving a brightness after-image which would fuse with and 
modify the stimulus, it should be chosen of a gray of the 
brightness of the color. In the same way, an interval has to 
be given in which to secure steady fixation when the amount 
of brightness induction is being measured. In order, then, 
to have the judgments made in each case the same length of 
time after induction had begun, it was necessary only to make 
the intervals of preexposure of equal duration and to require 
that the judgments of each kind be made directly at the end of 
the preexposure. In the case of the color experiments, the 
signal for the making of the judgment is the withdrawal of the 
preexposure card and the exposure of the stimulus. For the 
judgments of induction, however, in which case the stimulus 
was the gray of the brightness of the color, it is obvious that 
no preexposure card was needed, for preexposure and stimulus 
were required by the conditions of the experiment to be the 
same. In this case, a word-signal had to be given to indicate 
the termination of the preexposure interval and the instant at 
which the judgment was to be made. 
Results when White and Black Screens were Used.-Observing 
these precautions as to the equality of the illumination of 
stimulus, screen, and measuring-disc, and as to the length of 
time the induction had had in which to increase before the RETINAL SENSITIVITY 
473 
judgment was made, measurements were taken of the induction 
by white and black screens across grays of the brightness of the 
four principal colors at the illumination used. These measure- 
ments were made at various points of excentricity on the retina 
and for both standard and decreased illumination. The deter- 
mination of the equality point between the stimulus and the 
measuring-disc was made as follows: The size of the white or 
black sector of the latter was changed until a preliminary 
judgment of equality was made. Then the j. n. d. on either 
side of this point was determined both by ascending and by 
descending series and an average of the results was taken as the 
final value of the induction. Measurements were taken at 
250 and 400 on the temporal meridian, and at 550 and 700 on 
the nasal. The conditions at the nasal 550 point were very 
similar to those at 250 on the temporal side. The measure- 
ments at 700 nasal were midway in value between those at 250 
and 400 on the temporal. The 400 point is very near the 
limits of color sensitivity in this meridian, and the induction 
here is very great. For one observer, the darker stimuli 
appeared black at this point, when the white background was 
used. In such cases, the difference between the induction at 
standard and at decreased illumination is more clearly shown 
by the observations made at 25° temporal meridian and at 550 
and 700 nasal meridian than at 400 temporal. We have, 
however, chosen for two reasons to present in the following 
table only the results obtained in the temporal meridian. (1) 
The results obtained in this meridian demonstrate sufficiently 
well all the facts that need be taken into consideration. Space 
will not, therefore, be given to the results for both meridians. 
(2) The second point of our problem requires us to correlate 
the increased amount of induction caused by a given decrease 
of illumination with the change in the color limits it produces. 
The limits of color sensitivity can be more easily investigated 
in the temporal meridian because the sensitivity to some colors 
extends in the nasal region beyond the 920 point, which is the 
limit of measurement for the apparatus we used. This is 
true in particular in case of observer C as may be seen in 
Table XI. Both purposes of the investigation are, then, 474 
GERTRUDE RAND 
better satisfied by results obtained in the temporal meri- 
dian. 
The results show in general the following facts. 
I. The amount of induction from the white screen is 
greater than that from the black screen. 
2. The amount of induction increases with the distance 
from the fovea. 
3. The amount of induction increases with decrease of 
illumination.1 
4. The amount of increase under decreased illumination is 
greater in case of the white screen than in case of the black 
screen. 
5. The white and black screens induce more contrast 
across the stimuli that are farthest removed from them in 
brightness, and least across those which are most like them. 
That is, the white screen induces more black across the gray 
of the brightness of blue than across a gray of the brightness of 
yellow; and the black screen induces more white across the gray 
of the brightness of yellow than across a gray of the brightness 
of blue. 
Results are given in detail in Tables I. and II. Table I. 
gives the results for observer A taken on the temporal meridian, 
and Table II., the results for observer C for the same meridian. 
There is some difference in the amount of induction reported 
by the different observers, but since the preceding general 
statement of results is clearly borne out in every case, it is not 
deemed necessary to give space to results from all the observers 
used. In these tables, column 1 gives the degree of excen- 
tricity at which the observation was made; columns 2, 3, and 
4 show respectively the stimulus used, and the amounts of 
induction from the white and the black screens at standard 
illumination. Columns 5, 6, and 7 give the same data for 
decreased illumination. 
Results when the Gray Screen Matching the Colored Stimulus 
in Brightness at Standard Illumination is Used.-It was neces- 
1 This statement is meant to apply only to the range of illumination worked with. 
The induction was not measured when the illumination was very low, nor when it was 
very intensive. RETINAL SENSITIVITY 
475 
Table I 
A. Showing the Amount of Contrast Induced by the White and the Black Screens at Standard and Decreased Illumination 
upon the Grays of the Brightnesses of the Colored Stimuli at Standard and Decreased Illumination1 
Fixation 
Standard Illumination 
Decreased Illumination 
Stimulus (Gray of 
Brightness of Each 
of the Four Colors 
at Standard 
Illumination) 
Amt. Induction of 
White Screen 
Amt. Induction of 
Black Screen 
Stimulus (Gray of 
Brightness of Hach 
of the Four Colors 
at Decreased 
Illumination) 
Amt. Induction of 
White Screen 
Amt. Induction of 
Black Screen 
25° 
gray No. 2 
Black 1350 
White IIO0 
gray No. 2 
Black 220° 
White 1700 
gray No. 8 
Black 1550 
White 6o° 
gray No. 6 
Black 2700 
White 8o° 
gray No. 24 
Black 2300 
White 28° 
gray No. 41 
Black 3230 
White 400 
gray No. 32 
Black 2900 
White 120 
gray No. 20 
Black 3300 
White 300 
40° 
gray No. 2 
Black 2000 
White 3000 
gray No. 3 
Black 3200 
White 360° 
gray No. 8 
Black 3000 
White 13 20 
gray No. 5 
Black 360° 
White 1800 
gray No. 24 
Black o° 
White 6o° 
gray No. 50 
Black 36002 
White o°3 
gray No. 29 
Black 360" 
White 28° 
gray No. 13 
Black 360° 
White ioo° 
1 It is obvious that the method of expressing the amount of brightness induction used in this and the following tables gives an under- 
estimation. Suppose, as is shown in Table I., that No. 24 Hering gray has been darkened by induction until it matches in brightness a disc 
made up of 2300 of black and 1300 of the No. 24 gray. The amount of induction is greater than is represented by the 2300 of black because 
the induction has not lessened the amount of light coming to the eye from the grey paper while the addition of 2300 of black to the measuring- 
disc has cut off approximately 2/3 of the light coming from the gray paper. That is, in the one case enough black has been added by induction 
to reduce 360° of No. 24 gray to the given point in the brightness scale, while in the other enough black was added by direct mixing to lower 
only 1300 of No. 24 gray to this point in the scale. Moreover, the underestimation will be increased by this method of measuring in proportion 
as the amount of induction is increased because the greater the induction is the more black and the less gray will have to be used in the measur- 
ing-disc. All that can be said accurately is that a certain gray darkened or lightened by induction matches in brightness a gray made up of a 
certain amount of the given gray plus a certain amount of black or white. The exact amount of the induction can not be separated out. 
Further, just because the brightness added by contrast does not alter the amount of light coming to the eye while the brightness added in any 
method of measurement does change this amount of light, the writer knows of no way by which an exact expression can be attained. The 476 
GERTRUDE RAND 
Fixation 
Standard Illumination 
Decreased Illumination 
Stimulus (Gray of 
Brightness of Each 
of the Four Colors 
at Standard 
Illumination) 
Amt Induction of 
White Screen 
Amt. Induction of 
Black Screen 
Stimulus (Gray of 
Brightness of Each 
of the Four Colors 
at Decreased 
Illumination) 
Amt. Induction of 
White Screen 
Amt. Induction of 
Black Screen 
25° 
gray No. 2 
Black 700 
White 550 
gray No. 2 
Black 1300 
White 700 
gray No. 8 
Black 84° 
White 48° 
gray No. 6 
Black 1550 
White 590 
gray No. 24 
Black 930 
White 300 
gray No. 40 
Black 187° 
White 45° 
gray No. 32 
Black 160° 
White 150 
gray No. 17 
Black 2440 
White 22° 
40° 
gray No. 2 
Black no0 
White 2000 
gray No. 3 
Black 216° 
White 3400 
gray No. 7 
Black 1420 
White 1600 
gray No. 4 
Black 2300 
White 3 200 
gray No. 24 
Black 1800 
White 950 
gray No. 50 
Black 36004 
White o°6 
gray No. 27 
Black 2140 
White 350 
gray No. 7 
Black 3000 
White 1080 
method she has used, however, does serve as a means of comparing the amounts of induction occurring under different conditions sufficiently 
accurately for her purpose at this point. 
2 The gray No. 50 was in reality rendered blacker by the inductive action of gray No. 24 than the Hering black we used on the measuring- 
disc. A match thus could not be attained with black 360° as the table indicates. 
3 There was no brightness induction in this case because the stimulus, gray No. 50, m atches in brightness the black paper which formed 
the campimeter screen. 
4 See footnote 2 above. 
8 See footnote 3 above. 
Table II 
Observer C RETINAL SENSITIVITY 
477 
sary to perform the experiments bearing on this point at de- 
creased illumination only. For them the campimeter screens 
which matched in brightness the four principal colors of the 
Hering series at standard illumination served as inducing 
surfaces. For the contrast surfaces, grays of the brightness of 
these colors at decreased illumination were chosen. The 
methods of measuring, precautions in working, parts of the retina 
investigated, etc., were the same as in the preceding determi- 
nations. The following general statement of results may be 
made, (i) At the 250 point the brightness of yellow was 
found not to have changed at all with the decrease of illumina- 
tion produced by changing the illumination from the value 
selected as standard to the value selected for the comparison; 
the brightness of green lightened by an amount equal to the 
difference between No. 8 and No. 6 of the Hering series of 
grays; red darkened by an amount equal to the difference 
between No. 24 and No. 40; and blue lightened by an amount 
equal to the difference between No. 32 and No. 20. The 
amount of induction by the gray screen of the original bright- 
ness of the color upon the gray stimulus of the brightness of 
the color as altered by the decreased illumination, expressed 
in terms of Hering white and black, was for yellow o°, for 
green 6o° of white, for red 270 of black, and for blue 200 of 
white. (2) At the 400 point, the yellow darkened by an amount 
equal to the difference between No. 2 and No. 3 of the Hering 
grays; green lightened by an amount equal to the difference 
between No. 8 and No. 5; red darkened by an amount equal 
to the difference between No. 28 and No. 50; and blue lightened 
by an amount equal to the difference between No. 28 and No. 
13. The amount of induction produced by these changes was 
for yellow 280° of black, for green 1300 of white, for red 360° 
of black, and for blue 6o° of white. These results are shown 
in detail in Table HI. 
2. The Effect of These Amounts of Induction Upon 
the Limits of Color Sensitivity 
In order to obtain an estimate of the range of effect upon 
the limits of color sensitivity of the induction of the screens 478 
GERTRUDE RAND 
Table III 
A. Showing the Amount of Contrast Induced at Decreased Illumination on 
Grays of the Brightness of the Colors at Decreased Illumination by 
the Gray Screens Matching the Colors in Brightness at 
Standard Illumination 
Fixation 
Stimulus 
Screen 
Amount of Induction 
25° 
gray No. 2 
gray No. 2 
O° 
gray No. 6 
gray No. 8 
white 6o° 
gray No. 41 
gray No. 24 
black 270 
gray No. 20 
gray No. 32 
white 20° 
4°° 
gray No. 3 
gray No. 2 
black 280° 
grgy No. 5 
gray No. 8 
white 1300 
gray No. 50 
gray No. 24 
black 36001 
gray No. 13 
gray No. 28 
white 6o° 
at standard and at decreased illumination, the breadth of the 
color zones was determined at both illuminations (1) when 
white and black served in turn as campimeter screens; and 
(2) when a gray matching the color in brightness at standard 
illumination was used. The preexposure was in each case to 
a gray of the same brightness as the stimulus at the illumination 
used. The point at which the color lost all trace of its original 
quality was recorded as the limit of sensitivity. 
Results when White and Black Screens Were Used.-When 
the stimulus color is gotten by reflection from a pigment 
surface, two factors operate to give a change of result when 
the illumination is decreased. (1) There is a decrease in 
the amount of colored light coming to the eye. (2) There 
is an increase in the inductive action of the screen due to the 
change in the brightness relation of stimulus to screen and to 
the increased sensitivity of the eye to brightness contrast at 
decreased illumination. 
In order to find out how much of our results with the white 
and black screens should be attributed to the decrease in the 
amount of colored light coming to the eye produced by the 
decreased illumination, and how much to the increased inductive 
action of the screens, the limits of sensitivity were also deter- 
mined at both illuminations with the screens of the gray into 
1 The gray No. 50 was in reality rendered blacker by the inductive action of gray 
No. 24 than the Hering black we used on the measuring-disc. A match thus could not 
be attained with black 360° as the table indicates. RETINAL SENSITIVITY 
479 
which the color disappears in the peripheral retina. From the 
values obtained with the three screens at both illuminations, the 
amount of change due to decrease in the amount of colored 
light coming to the eye and the amount due to induction by the 
white and black screens were calculated as follows, (a) From 
the number of degrees expressing the limits for a given color 
at standard illumination with a screen of the brightness of the 
color at that illumination was subtracted the number expressing 
its limit at decreased illumination with a screen of the bright- 
ness of the color at the decreased illumination. That this gave 
the number of degrees the zone of sensitivity was narrowed 
by the decrease in the energy of the stimulus may be said 
with the following qualification. If there is any influence upon 
color sensitivity of the local brightness-adaptation of the retina 
produced by the change in the general illumination, it is, of 
course, included in this effect. But, since this influence would 
have to be brought about by previous exposure to the illumina- 
tion in question, it can be reduced to a minimum by guarding 
against an exposure to it for any considerable length of time. 
The effect of whatever adaptation there may be, however, can 
not be isolated or separated out from the above result, and the 
value expressing the amount the limit is narrowed by the 
actual decrease of the energy of colored light coming to the 
eye cannot, strictly speaking, be obtained. But it is probable 
that the adaptation effect is not sufficiently strong to influence 
the limits, since the sensitivity of the extreme peripheral retina 
falls off very abruptly from point to point. The difference, 
then, between the color limit obtained at standard illumination 
and the limit at decreased illumination, when in both cases there 
is no brightness induction from the screen, may be said to 
approximate the effect upon the limits produced by the decrease 
in the amount of colored light coming to the eye. (b) Figures 
can be obtained, however, from our results, which express the 
amount by which the zones are narrowed by the change in the 
inductive action of the white and black screens produced by 
decreasing the illumination, that are not open to theoretical 
questioning; for the influence of local brightness-adaptation, 
if there be any, is a constant for all screens at the same illu- 480 
GERTRUDE RAND 
mination. If then, the number of degrees which expresses 
the limits of sensitivity for either the white or the black screen 
at decreased illumination is subtracted from the number 
expressing the limit with a screen of the gray of the brightness 
of the color at this illumination, the result will represent the 
extent to which the limit was narrowed by the action of induc- 
tion alone. The results show in general the following facts,: 
i. At standard illumination, induction from the white screen 
narrows the limits of yellow and red; induction from the black 
screen narrows the limits of blue and green. The difference 
is in no case more than 4°. 
2. At decreased illumination, the induction from the white 
screen narrows the limits of all the colors much more consider- 
ably than does the induction from the black screen.1 
3. The values expressing the narrowing of the limits 
caused by decrease of illumination without induction are 
greatest in case of those colors which undergo maximum change 
of brightness in passing into the periphery, namely, for blue 
and red. 
We have shown by the results of the preceding section, 
that the increased induction produced by decrease of the 
general illumination is greater for the white screen than for 
the black, and, by the results of this section, that this increase 
is effective to the extent of narrowing the limits of sensitivity 
to all colors from 50 to 120 with this screen. With the black 
screen, the limits were narrowed from 30 to 6°. At standard 
illumination, the limits were narrowed only from i° to 40 with 
either the white or the black screens. 
Results in detail are given in Tables IV. and V., taken from 
the temporal meridians of the observers whose observations are 
recorded in Tables I. and II. In column 1, Tables IV. and V., 
is given the stimulus. Column 2 shows the limit of sensitivity 
to the stimulus at standard illumination with a screen of a 
1 For observer A the results for green present an exception. At the decreased 
illumination used the green stimulus appeared bluish in the central retina. The in- 
duction of the black screen caused it to appear as a pale blue at a comparatively slight 
degree of excentricity. According to our definition of color limit, this point is the 
limit of green. It is, however, obvious that the exception is due rather to the quali- 
tative than to the quantitative effect of brightness upon color. 481 
RETINAL SENSITIVITY 
gray of the brightness of the color at standard illumination; 
column 3 shows the limit with a white screen; and column 4 
with a black screen. Column 5 shows the limit at decreased 
illumination with a screen of the brightness of the color at 
decreased illumination; column 6 shows the limit with a white 
screen; and column 7 with a black screen. 
Table IV 
A. Showing the Color Limits at Standard and Decreased Illumination (a) 
with Gray Screens of the Brightnesses of the Colors at the Illumi- 
nation Used; and (b) with White and Black Screens 
Stimulus 
Standard Illumination 
Decreased Illumination 
Limit with 
Gray Screen of 
Brightness of 
Color at 
Standard 
Illumination 
Limit 
with 
White 
Screen 
Limit 
with 
Black 
Screen 
Limit with 
Gray Screen of 
Brightness of 
Color at 
Decreased 
Illumination 
Limit 
with 
White 
Screen 
Limit 
with 
Black 
Screen 
Yellow 
44° 
42° 
45° 
43° 
35° 
43° 
Green 
37° 
37° 
34° 
36° 
3i° 
27° 
Red 
43° 
42° 
44° 
40° 
3i° 
4°° 
Blue 
5° 
5°° 
49° 
48° 
36° 
43° 
Table V 
Observer C 
Stimulus 
Standard Illumination 
Decreased Illumination 
Limit with 
Gray Screen of 
Brightness of 
Color at 
Standard 
Illumination 
Limit 
with 
White 
Screen 
Limit 
with 
Black 
Screen 
Limit with 
Gray Screen of 
Brightness of 
Color at 
Decreased 
Illumination 
Limit 
with 
White 
Screen 
Limit 
with 
Black 
Screen 
Yellow 
49° 
46° 
5°° 
46° 
36° 
44° 
Green 
44° 
42° 
4°° 
41° 
28° 
33° 
Red 
45° 
4i° 
45° 
4i° 
34° 
41° 
Blue 
56° 
55 
53° 
50° 
38° 
44° 
Tables VI. and VII. have been compiled from Tables IV. 
and V. to show the following facts. 
(i) How much the decrease of illumination narrowed the 
limits of color sensitivity by causing a decrease in the energy 
of the light waves coming to the eye. This was determined 
by subtracting the value of the limit at decreased illumination 
with the screen of a gray of the brightness of the color at 
decreased illumination from its value at full illumination with 
the gray screen of the brightness of the color at full illumination. 482 
GERTRUDE RAND 
(2) How much the limits were narrowed by the action of the 
white and black screens at decreased illumination. This was 
ascertained by subtracting the values of the limit with the 
white and the black screen at decreased illumination from the 
value of the limit at decreased illumination with the gray 
screen of the brightness of the color at this illumination. (3) 
How much more the limits were narrowed by the white and 
the black screens at decreased than at full illumination. This 
was computed for the white screen, for example, as follows. 
The quantity, limit at decreased illumination for gray screen of 
brightness of color at decreased illumination, minus limit for 
white screen at decreased illumination, is subtracted from the 
Table VI 
A. Showing (i) How Much the Limits were Narrowed by Decrease in the 
Amount of Colored Light Coming to the Eye; (2) How Much They 
were Narrowed by Increased Induction of White and Black 
Screens at Decreased Illumination; and (3) How Much 
More They were Narrowed by Induction of White 
and Black Screens at Decreased Than at Full 
Illumination 
Stimulus 
How Much 
Limits were 
Narrowed by 
Decrease in 
Amount of 
Colored Light 
Coming to 
the Eye 
How Much 
Limits were 
Narrowed by 
Induction of 
White Screen 
How Much 
Limits were 
Narrowed by 
Induction of 
Black Screen 
How Much 
More Limits 
were Nar- 
rowed by 
White Screen 
at Decreased 
than at Full 
Illumination 
How Much 
More Limits 
were Nar- 
rowed by 
Black Screen 
at Decreased 
than at Full 
Illumination 
Yellow 
1° 
8° 
O° 
6° 
3° 
Green 
1° 
5° 
9° 
5° 
6° 
Red 
3° 
9° 
O° 
8° 
3° 
Blue 
2° 
12° 
5° 
12° 
4° 
Table VII 
Observer C 
Stimulus 
How Much 
Limits were 
Narrowed by 
Decrease in 
Amount of 
Colored Light 
Coming to 
the Eye 
How Much 
Limits were 
Narrowed by 
Induction of 
White Screen 
How Much 
Limits were 
Narrowed by 
Induction of 
Black Screen 
How Much 
More Limits 
were Nar- 
rowed by 
White Screen 
at Decreased 
than at Full 
Illumination 
How Much 
More Limits 
were Nar- 
rowed by 
Black Screen 
at Decreased 
than at Full 
Illumination 
Yellow 
3° 
3° 
4° 
6° 
IO° 
13° 
7° 
12° 
2° 
8° 
o° 
7° 
7° 
n° 
3° 
n° 
O 4^ 
o o o o 
Green 
Red 
Blue RETINAL SENSITIVITY 
483 
quantity, limit at full illumination for gray screen of bright- 
ness of color at full illumination minus limit for white screen 
at full illumination. A similar computation was made for the 
black screen. 
Results when a Gray Screen Matching the Color in Bright- 
ness at Standard Illumination is Used.-In these experi- 
ments a determination was made of the amount the limits 
of sensitivity are changed by the brightness induction caused 
by the alteration of the brightness relation between stimu- 
lus and screen with decrease of illumination, when a screen 
is used which matches the color in brightness at standard 
illumination. This determination was made as follows. An 
estimate was made of the amount the limits were narrowed 
by decrease of illumination when a screen of the brightness of 
the color at standard illumination is used for both standard 
and decreased illuminations. From this result was subtracted 
the amount the limits were narrowed by decrease of illumina- 
tion when the screen is made in turn of the brightness of the 
color at standard and decreased illumination. The difference 
obtained represents the value sought. It is given in Table VIII. 
Table VIII 
A. Showing How Much the Color Limits were Narrowed at Decreased Il- 
lumination by the Induction of the Screen Which Matched the 
Color in Brightness at Standard Illumination 
Stimulus 
Screen of Bright- 
ness of Color at 
Decreased 
Illumination 
Limit 
Screen of Bright- 
ness 6f Color at 
Standard 
Illumination 
Limit 
Amount Limit 
was Narrowed by 
Change in Bright- 
ness Relation Be- 
tween Stimulus 
and Screen Caused 
by Decreased 
Illumination 
Yellow.. . 
gray No. 3 
43° 
gray No. 2 
4i° 
2° 
Green . .. 
gray No. 5 
36° 
gray No. 8 
29° 
7° 
Red 
gray No. 50 
40° 
gray No. 24 
33° 
7° 
Blue 
gray No. 13 
48° 
gray No. 28 
46° 
2° 
3. The Effect of These Amounts of Induction upon the 
Limen of Color at Different Degrees of 
Excentricity 
We have shown the effect of decreasing the general illumina- 
tion upon the color sensitivity of the peripheral retina with 484 
GERTRUDE RAND 
gray, white, and black screens, by the effect on the limits of 
sensitivity. This is only an indirect means of estimating its 
influence, for the results obtained cannot be translated into 
terms of direct measurement, owing to the irregular decrease 
in sensitivity of the peripheral retina from the fovea outwards. 
In this section, we shall measure the influence of changes of 
illumination directly by the changes produced in the limen of 
sensation at various angles of excentricity. As in the previous 
section, measurement will be made of the effect upon sensitivity 
(i) of the decrease in the amount of colored light coming to the 
eye, produced by the decrease of illumination, (2) of the dif- 
ference in the inducing power of the white and black screens, 
and (3) of the change in the brightness relation of stimulus to 
background. 
To determine the first of these three points, a campimeter 
screen had to be selected that gave no brightness contrast 
with the stimulus. To provide for differences in the brightness 
of the colors at the different points observed for the two illu- 
minations at which we worked, a preliminary determination of 
the brightness of the sensation at these points was made at 
both illuminations by the flicker method. The brightness of 
the screen was chosen in each case of the brightness of the color 
according to these determinations. To eliminate the effect of 
preexposure, the stimulus previous to exposure was in every 
case covered by a gray of the brightness of the color for the 
illumination used at the point of the retina at which we were 
working. Thus no brightness after-image was carried over 
to exert an inhibitive action upon the color sensation. The 
stimulus was a disc compounded of sectors of the color and 
of the gray of the brightness of the color for the illumina- 
tion used at the point of the retina under investigation. The 
proportions of the sectors were altered until the observer gave 
the judgment of just noticeable color. The average of judg- 
ments made in ascending and descending series was chosen as 
the final value of the limen. The difference between the 
limens at standard and decreased illumination was taken as the 
measure of the loss in intensity which the stimulus had sus- 
tained by the decrease of illumination. 485 
RETINAL SENSITIVITY 
The effect upon the color limen of the increased induction 
from the white and black screens was shown by the same 
method, with the exception that the white and black screens 
were substituted for the gray of the brightness of the color. 
The stimulus was a disc composed of sectors of color and gray 
of the brightness of the color at the angle of excentricity at 
which the determination was made. 
The effect of the change in the brightness relation between 
the stimulus color and the screen produced by decrease of 
illumination was shown as follows. An estimate was made 
of the amount the limens are raised by the decrease of illumina- 
tion when a screen was used for both standard and decreased 
illumination that had a brightness value equal to the color at 
standard illumination. From these results was subtracted the 
amount the limens were raised by decreasing the illumination 
when the screens were made in turn of the brightness of the 
color at standard and at decreased illumination. The differ- 
ence obtained represents the value sought. These results are 
of particular importance because they show that the influence 
of the brightness of the surrounding field can not be eliminated 
even when a screen of the brightness of the color is used unless 
some means be had of maintaining the general illumination of 
the room constant. 
Table IX. shows how much the limens of sensitivity were 
raised at the fovea and at points 15°, 250 and 300 from the fovea 
in the horizontal meridian on the temporal side by the decrease 
in the amount of colored light coming to the eye produced by 
the decrease in the general illumination. The results of this 
table may be generalized as follows: 
1. The limen of color is higher in the periphery than in the 
center of the retina at both illuminations. 
2. The limen of color is higher at decreased illumination 
that at standard illumination. 
3. The direct effect upon the intensity of the sensation 
produced by decreasing the illumination is shown by the 
limen determinations to be inconsiderable. In the central 
retina, the difference is but 2° or 30. In the peripheral retina 
at the points considered there is a difference of from io° to 200. 486 
GERTRUDE RAND 
Table IX 
A. Showing How Much the Limens of Sensitivity were Raised at the Fovea, 
and at Points 150, 250, 300 from the Fovea in the Horizontal 
Meridian on the Temporal Side by the Decrease in the 
Amount of Colored Light Coming to the Eye 
Produced by the Decrease in the 
General Illumination 
Stimulus 
Point on Hori- 
zontal Temporal 
Meridian at 
Which I,imen 
was Taken 
Timen at Stan- 
dard Illumina- 
tion with Screen 
of Brightness of 
Color at Standard 
Illumination 
Limen at De- 
creased Illumina- 
tion with Screen 
of Brightness of 
Color at Decreased 
Illumination 
How Much Limen 
was Raised at 
Decreased 
Illumination 
Yellow 
0° 
18° 
20° 
2° 
15° 
22° 
32° 
IO° 
25° 
35° 
40° 
5° 
30° 
50° 
65° 
15° 
Green 
0° 
20° 
20° 
O° 
15° 
27° 
28° 
1° 
25° 
40° 
50° 
IO° 
Red 
o° 
11° 
2° 
15° 
13° 
4° 
25 
17° 
25° 
8° 
30° 
25° 
45° 
20° 
Blue 
o° 
9° 
10° 
IO° 
1° 
15° 
13° 
3° 
25° 
12° 
15° 
3° 
30° 
20° 
4° 
20° 
Table X 
A. Showing the Color Limens at Standard and Decreased Illuminations 
with White and With Black Screens 
Stimulus 
Point on Horizon- 
tal Temporal Me- 
ridian at Which 
Limen was Taken 
White Screen 
Black Screen 
Limen at 
Standard 
Illumination 
Limen at 
Decreased 
Illumination 
Limen at 
Standard 
Illumination 
Limen at 
Decreased 
Illumination 
Yellow .. . 
O° 
22° 
25° 
28° 
3°° 
15° 
25° 
50° 
35° 
45° 
25° 
50° 
80 
42° 
6o° 
30 
8o° 
I2S° 
6o° 
90° 
Green.... 
o° 
22° 
25° 
28° 
30° 
15° 
26° 
36° 
35° 
43° 
25° 
3°° 
75° 
75° 
220° 
Red 
o° 
13° 
20° 
10° 
14° 
15° 
19° 
35° 
13° 
19° 
25° 
3°° 
55° 
23° 
35° 
30° 
5°° 
330 
29° 
58° 
Blue 
o° 
170 
22° 
10° 
12° 
15° 
25° 
40° 
12° 
170 
25° 
35 
6o° 
18° 
25° 
30° 
40° 
9°° 
30° 
6o° RETINAL SENSITIVITY 
487 
Table X. shows the color limens at both standard and 
decreased illuminations when white and black screens are used, 
at the fovea and at points 150, 250, and 300 in the horizontal 
meridian on the temporal side. 
Table XI. has been compiled from Tables IX. and X. to 
show how much greater the limens were for white and black 
screens at decreased than at full illumination; how much of the 
effect may be ascribed to the reduction of the amount of 
colored light coming to the eye; and how much to the increased 
induction of the screens. It will be seen from the results of 
this table that the loss of the sensation in intensity due to the 
increased brightness induction is much greater than that 
caused by the reduction in the amount of colored light coming 
to the eye. 
Table XI 
A. Showing How Much Greater the Limens were with White and Black 
Screens at Decreased Than at Standard Illumination and How Much 
of This Effect may be Ascribed to the Reduction in the 
Amount of Colored Light Coming to the Eye 
and How Much to the Increased Induc- 
tive Action of the Screens 
Stimulus 
Point on 
Horizontal 
Temporal 
Meridian 
at Which 
Timen was 
Taken 
White Screen 
Black Screen 
Total 
Amount 
Greater 
Amount 
Due to De- 
crease in 
Amount 
of Colored 
Tight Com- 
ing to Eye 
Amount 
Due to 
Increased 
Induction 
of Screen 
Total 
Amount 
Greater 
Amount 
Due to De- 
crease in 
Amount 
of Colored 
Tight Com- 
ing to Eye 
Amount 
Due to 
Increased 
Induction 
of Screen 
Yellow. . . 
O° 
7° 
2° 
5° 
12° 
2° 
IO° 
15° 
28° 
IO° 
18° 
23° 
IO° 
13° 
25° 
45° 
5° 
40° 
25° 
5° 
20° 
30° 
75 
15 
6o° 
35° 
15° 
25° 
Green . .. 
o° 
5° 
0° 
5° 
10° 
o° 
IO° 
15° 
8° 
i° 
7° 
16° 
i° 
15° 
25 
35° 
10° 
25° 
i8o° 
10° 
1700 
Red 
o° 
n° 
2° 
9° 
22° 
5° 
10° 
2° 
3° 
6° 
15° 
26° 
4° 
4° 
25° 
38° 
8° 
3°° 
18° 
8° 
IO° 
3° 
305° 
20° 
285° 
33° 
20° 
13° 
Blue 
o° 
13° 
1° 
12° 
3° 
7° 
1° 
2° 
15° 
30° 
3° 
27° 
3° 
4° 
25° 
48° 
3° 
45° 
13° 
3° 
IO° 
30° 
70° 
20° 
50° 
40° 
20° 
20° 488 
GERTRUDE RAND 
It was shown in Table III. that quite a great deal of bright- 
ness induction is caused by the change in brightness relation 
between color and screen produced by decreasing the illumina- 
tion. Table VIII. shows how much this induction narrows 
the limits of sensitivity to the four colors used. Table XII. 
shows how much the limens are raised when the illumination is 
decreased by the inductive action caused by the change in the 
brightness relation between stimulus color and gray screen 
of the brightness of the color at standard illumination. 
Table XII 
A. Showing How Much the Color Limens were Raised at Decreased Illu- 
mination by the Induction of the Screens Which Matched the 
Color at Standard Illumination 
Stimulus 
Point on Hori- 
zontal Temporal 
Meridian at Which 
Umen was Taken 
Timenwith Screen 
of Brightness of 
Color at Decreased 
Illumination 
Limen with Screen 
of Brightness of 
Color at Standard 
Illumination 
Amount Limen 
was Raised by 
Change in Bright- 
ness Relation 
Between Stimulus 
and Screen Caused 
by Decrease of 
Illumination 
Yellow 
o° 
15° 
25° 
30 
o° 
15° 
25° 
0° 
15° 
25 
30° 
o° 
15° 
25 
30° 
20° 
32° 
40° 
65° 
20° 
28° 
5°° 
ii° 
13° 
25 
45° 
IO° 
13° 
15° 
40° 
20° 
32° 
40° 
1160 
20° 
40° 
190° 
11° 
24° 
48° 
I50° 
12° 
160 
23° 
55° 
O° 
O° 
o° 
51° 
0° 
12° 
I4O0 
o° 
11° 
230 
105° 
2° 
3° 
8° 
15° 
Green 
Red 
Blue 
4. The Influence of Change of Illumination upon the 
Action of the Preexposure on the Limens 
and Limits of Color 
The brightness of the preexposure exerts an influence upon 
the color observation because the eye carries over an after- 
image from the preexposure into the color observation. If, for 
example, the preexposure is to black, a white after-image is RETINAL SENSITIVITY 
489 
aroused which fuses with the succeeding color sensation and 
strongly reduces its saturation. The effect of preexposure is 
especially strong in the peripheral retina because a very strong 
brightness after-image is aroused in the peripheral retina by a 
very short period of stimulation. It is very difficult for the 
writer to predict from the data she has at hand with regard to 
the effect of change of illumination upon the sensitivity of the 
peripheral retina to the brightness after-image just what will 
be the effect of change of illumination upon the action of pre- 
exposure on the color sensitivity of the peripheral retina. But 
even though there be no change in the sensitivity of the periph- 
eral retina to the brightness after-image with change of illu- 
mination, it is obvious that there will be some effect of the 
change of illumination because of the change in the brightness 
relation of the preexposure card to the colored stimulus. In 
case the stimulus light is gotten by reflection from pigment 
surfaces, this change of brightness relation is due to the shift 
in the brightness of the colors produced by the change in the 
illumination. In case transmitted light is used as stimulus, 
the brightness of the stimulus color is independent of changes 
in illumination and will remain constant; but a change in the 
brightness relation of stimulus to preexposure will occur be- 
cause the preexposure will lighten or darken with change of 
illumination. The writer hopes to make the quantitative 
investigation of this point the subject of a future study. At 
present she can only point out that if a guarantee is wanted 
that the effect of the brightness of the preexposure is eliminated 
from the results of the observation, the preexposure must be 
to a gray of the brightness of the color and the illumination of 
the room must be kept constant. 
IV. Conclusion 
The foregoing results show how strongly the changes in the 
illumination of the visual field influence the color sensitivity 
of the peripheral retina, particularly when the stimulus is 
surrounded by a white field. They also show that the influence 
of surrounding field can not be eliminated even by means of a 
campimeter screen of the brightness of the color unless some 490 
GERTRUDE RAND 
means be had of keeping the general illumination of the room 
constant. It is obvious without further comment how im- 
portant it is that a method be devised to standardize this 
factor. The preceding experiments indicate that without this 
standardization, no experiment can be repeated from time to 
time under the same conditions relative to any one of the 
brightness factors that influence color sensitivity. Results thus 
obtained are far from comparable. As was stated earlier in 
the paper a method of standardizing was described in an earlier 
paper in this volume of the Review.1 
1See footnote, 2, p. I. 
It was the writer's intention to have had the present article precede the one in 
which the method of standardizing is described, but owing to limited space in the 
July number the Editor was compelled to use for that number the shorter article. [Reprinted from the Transactions of the Illuminating 
Engineering Society, January, 1913.] 
TESTS FOR THE EFFICIENCY OF THE EYE UNDER 
DIFFERENT SYSTEMS OF ILLUMINATION AND 
A PRELIMINARY STUDY OF THE CAUSES 
OF DISCOMFORT.* 
C. E. FERREE. 
Synopsis:-Besides outlining (I) the problem which confronts the 
investigator who would determine the effects of various lighting systems 
on the eye, this paper discusses: (II) the scale or general level of effi- 
ciency of the eye under different systems of lighting, with brief comments 
on the conventional tests for the efficiency of the eye such as, (a) color 
discrimination, (&) brightness discrimination, (c) visual acuity-the latter 
tests, modified, it is contended are adequate for the determination of the 
general level of efficiency of the unfatigued eye; (III) loss of visual 
efficiency as the result of a period of work-here it is contended that 
each of the aforementioned tests fails to show a true loss of visual effi- 
ciency, and a new test is described. The paper is concluded (IV) with a 
brief statement of some of the causes of ocular discomfort under various 
conditions, and a description of a method of making a comparative esti- 
mate of discomfort. 
I. INTRODUCTION. 
In 1911 the American Medical Association appointed a com- 
mittee to study the effect of different lighting systems on the eye. 
The writer was asked to share in the work of this committee. 
The problem presented to him was to furnish tests that would 
show the effect of different lighting systems on the eye and more 
especially to devise, if possible, a test that would show a loss of 
efficiency as the result of three or four hours of work under an 
unfavorable lighting system. It is the purpose of the following 
paper to give a preliminary report of the work that has been 
carried on by the writer in this field during the past year. 
Confronting the problem of the effect of lighting systems on 
the eye, it is obvious that the first step toward systematic work 
is to obtain some means of making a definite estimate of this 
effect. The prominent effects of bad systems of lighting are loss 
of efficiency, temporary and progressive, and eye discomfort. 
Having devised methods which after six months of testing he 
has found to be accurate and practicable, the writer has under- 
*A paper read at the sixth annual convention of the Illuminating Engineering 
Society, Niagara Falls, Ont., September 16-19, 1912. 41 
FERREE: TESTS FOR THE EFFICIENCY OF THE EYE 
taken to determine (i) the lighting conditions that give in gen- 
eral the highest level or scale of visual efficiency, (2) the condi- 
tions that give the least loss of efficiency for continued work, 
and (3) the conditions that cause the least discomfort. This 
plan of work, it is scarcely needful to remark, will involve a 
wide range of experimentation. The crux of the problem, as the 
writer conceives it, is, however, to secure reliable methods of 
estimating effect. Having these methods, the factors whatever 
they may be, intensity, quality, position of light relative to the 
eye, etc., can be varied one at a time and the effects be determined. 
From these effects it should not be difficult to ascertain what 
lighting conditions are best for the eye and what is the relative 
importance of the factors that go to make up these conditions. 
Further it should be possible on the practical side to test out and 
perfect a lighting system, so far as its effect on the eye is con- 
cerned, before we put it on the market.* 
In this report nothing more will be attempted than to indicate 
what methods may be used in the three steps of the problem as 
outlined above. 
II. THE SCALE OR GENERAL LEVEL OF EFFICIENCY OF 
THE EYE UNDER DIFFERENT SYSTEMS OF LIGHTING. 
A general survey of the field shows that at different times the 
following tests have been used in one capacity or another for de- 
termining the efficiency of the eye: brightness discrimination, col- 
or discrimination, and visual acuity. No extensive use, if any at 
all, has been made of any of these with the exception of visual 
acuity in connection with problems of the type here considered, 
but the fitness of their application in some form to such problems 
is evident at a glance. If the eye's efficiency is to change at differ- 
ent times and under different conditions of lighting, it should be 
manifested in changes in brightness discrimination, color discrimi- 
nation, or visual acuity. The first step in our work would, then, 
seem to be to devise for these points tests which are sufficiently 
sensitive for use in work of the kind we have in hand. The gen- 
eral nature of these tests is too familiar to need detailed mention 
here. A few special points may, however, be given in passing. 
(1) The threshold or limen test is the most sensitive and practical 
* This latter point was suggested to the writer by reading Dr. Ives' discussion of this 
paper (p. 57). TRANSACTIONS I. U. S.-PART II 
42 
for color sensitivity. In making this test the pre-exposure1 and 
the surrounding field2 should be of a gray of the brightness of 
the color at or near its threshold value. Further, the illumination 
of the room must be kept constant from test to test.3 If the 
colored light is to be obtained by reflection, disks of standard 
1 By pre-exposure is meant what the eye rests on immediately preceding its stimula- 
tion by color. It is obvious that there must always be some pre-exposure and, unless 
care be taken to eliminate its effect, it will influence the eye's sensitivity to color. Even 
closing the eye, as is often done before stimulating by color, is the equivalent of giving a 
black pre-exposure. All color must of cou-se be eliminated from the pre-exposure. It 
should also be of the same brightness as the color by which the eye is to be stimulated. If 
not it gives a brightness after-image which mixes with the succeeding color impression 
and reduces its saturation. This reduction of saturation takes place apparently at some 
physiological level posterior to the seat of the positive, negative, and contrast color 
processes commonly supposed to be located in the retina. (See Ferree and Rand : " The 
Fusion of Brightness with Color-The Focus of the Action," Journal of Philosophy, Psy- 
chology and Scientific Methods, VIII, 1911, p. 294.) If the pre-exposure is lighter than the 
color it adds by after-image a certain amount of black to the succeeding color impression 
and, if darker, it adds a certain amount of white. Since white inhibits color more than 
black, the effect of a dark pre-exposure is to reduce the sensitivity to color more than the 
effect of a light pre-exposure. But since both white and black as after-effect reduce the 
sensitivity to color, the eye is rendered more sensitive when no after-image is given, i. e. 
when the pre-exposure is of the same brightness as the color. The pre-exposure therefore 
should be to a gray of the brightness of the color. No brightness after-image will be 
added thereby' to the succeeding color impression to modify either its saturation or color 
tone. 
2 When the surrounding field is either lighter or darker than the color, brightness is 
induced by contrast across the colored surface. When the surrounding field is lighter 
than the color, a certain amount of black is induced, and when darker, a certain amount 
of white is induced. As stated above, the mixture of this white or black with color, 
although it does not alter the amount of colored light coming to the eye. reduces the 
saturation of the color. The effect of brightness contrast can be eliminated only' by 
making the brightness of the surrounding field a gray of the brightness of the color. 
This can be done by means of a gray screen around the color, or by a larger gray disk in 
case a color mixer is used. 
s In case the colored light used for the stimulus is obtained by reflection from a 
pigment surface, a change in the general illumination of the field of vision affects the 
results of the sensitivity tests in the following ways. (1) It changes the amount of 
colored light coming to the eye. (2) By changing its brightness adaptation it changes 
the sensitivity of that part of the retina upon which the colored light falls. (3) 'By' 
changing the sensitivity of the eye to brightness after-image and contrast, it changes the 
amount of brightness added to the color as the result of pre-exposure and surrounding 
field, and therefore changes the effect of pre-exposureand surrounding field upon the color 
impression. Moreover, the effect of pre-exposure and surrounding field cannot be elimin- 
ated even when both are made of the brightness of the color for some given illumination 
unless that illumination be kept constant throughout the test for, when it changes, the 
brightness pf the color and of the grays used as pre-exposure and surrounding field does 
not change in equal amounts ; hence, the brightness equality which is needed cannot be 
maintained. In case the colored light is not gotten by reflection from a pigment surface 
but is obtained from monochromatic sources from standard filters or from the spectrum 
only the last two of the factors stated above influence the results of the tests for color 
sensitivity. In the tests made by the writer, the general illumination was rendered 
constant by methods to be described later in the paper. 
Although for the purposes of this work the tests for color sensitivity could never be 
conducted in the dark-room, still it may be of general interest to note at this point that the 
elimination of the effect of pre-exposure and surrounding field cannot be accomplished 
in work on color sensitivity done in the dark room, because in the dark-room the pre- 
exposure and surrounding field cannot be made of the brightness of the color. They 
will always therefore exert an effect on the color impression. Moreover since the colors 
all differ in brightness, this effect will be exerted in different amounts on the different 
colors. That is, the amount of brightness added by after-image or contrast depends 
upon the amount of brightness difference, respectively, between pre-exposure and color 
and surrounding field and color. As stated above this amount, when working in the 
dark-room, will be different for the different colors. For this reason and also because 
even the same amount of brightness excitation acts with different degrees of strength 
upon the excitation set up by the different colors, it is especially important that no work 
on the comparative sensitivity of the retina to the different colors should be done in the 
dark-room. It should be done in a light room of a constant intensity of illumination and 
with pre-exposure and surrounding field in each case of the brightness of the color to be 
used. In this way alone can all the factors which influence the sensitivity of the retina 
extraneous to the source of light, be eliminated. 43 
ferrEE: tests for the efficiency of the Eye 
colored and gray papers (V. g., the papers of the Hering series) 
may be used on a color mixer.4 If, on the other hand, it is 
desirable to use the light of the spectrum or the light transmitted 
through standard filters, the colored light may be cut down to 
the threshold value by means of a sectored disk, the sectors of 
which should be covered with a gray of the brightness of the 
color at or near its threshold value. (2) For brightness discrimi- 
nation also the threshold or limen test is the most sensitive and 
practical, but when made in a well-illuminated room, it becomes 
in effect a test for a just noticeable difference. This test may be 
performed at different points in the brightness scale, e. g., when 
the standard is black, near mid-gray, or white. As before, disks 
of standard papers may be used on the color-mixer, or the light 
from a given source may be varied by means of a sectored disk.5 
(3) Visual acuity tests of the Snellen type, especially when used 
in work in which it is required to make successive tests on the same 
person, are open to the following objections, (a) The judgment 
is in terms of recognition. A letter may be recognized when it is 
not seen clearly. In any judgment based on the recognition of 
even a single letter, memory plays an important role. It is, so far 
as the writer knows, impossible to standardize this memory fac- 
tor and to obtain results strictly in terms of acuteness of vision, 
(b) The test card is made up of quite a long series of letters. 
As the test progresses the letters are memorized more and more 
completely. It is practically impossible to eliminate this progres- 
sive error when a number of successive judgments have to be 
4 In making the tests with reflected light, two sets of disks are mounted on a color 
mixer (a) an outer disk of gray of the brightness of the color to be used, and (b) an inner 
disk made up of this gray and the disk of color. To the inner gray disk, varying pre- 
portions of the color are added until the threshold value or just noticeable color is 
obtained. To facilitate the judgment of just noticeable color, the inner disk of gray plus 
color is compared with the outer disk of gray as a standard. Since both grays are of the 
brightness of the color, the addition of the colored sector to the inner disk produces no 
change of brightness either to confuse the judgment of noticeable color, or to affect the 
intensity of the color excitation actually aroused. In getting the threshold value, the 
method of ascending and descending series should be used, that is, beginning withequali- 
ty, the variation is towards noticeable difference and beginning with a difference greater 
than noticeable, the variation is towards equality. An average of the two sets of 
results is taken for the threshold value. 
5 When the test is made with reflected light two sets of disks, an outer and an inner, 
are mounted on the color mixer. Each set is made up of one white and one black disk. 
Both sets of disks are set at the point in the brightness scale from which the variation 
towards white or black is to be made. One is kept constant and the other is changed 
until the judgment different is given. In making the judgment the method of ascending 
and decending series is used and the results are averaged for the difference limen. This 
difference limen is taken as the measure of the observer's sensitivity to brightness or 
white light. TRANSACTIONS I. E. S.-PART II 
44 
made as is the case before a final result is reached in any single 
visual acuity test and as is especially the case when a number of 
successive tests have to be given to the same person, which 
happens in much of the work involved in the solution of the 
problem here proposed. It might be supposed that the memoriza- 
tion of the series could be broken up by using in each successive 
judgment in a single test or in the successive tests, as the case 
may be, cards having a different distribution of the letters in the 
series. Considerable inconvenience would, however, be involved 
in giving the tests in this way and besides no guarantee could be 
had that each judgment would present the same degree of diffi- 
culty. That is, the series is made up of similar and dissimilar 
letters. The dissimilar letters can be distinguished from each 
other with less difficulty than the similar. It is practically impos- 
sible to distribute the letters so that the individual tests may be 
equally rigorous. This objection can, of course, be eliminated 
in part by a careful selection of the test letters, but not entirely 
because a series of letters uniformly similar cannot be found, 
(c) The Snellen series contains quite a large number of letters. 
The eye is found to fatigue and vision to blur before the series 
is completed. This introduces an error which it is practically 
impossible to render constant. 
All of the above objections were eliminated in the tests finally 
adopted by us by changing the type of judgment and by making 
the test object in one case two parallel vertical lines stamped 1 
mm. apart on a white card6 and in another the letters li printed 
in small type.7 In using these cards the observer's acuity of 
6 The card is mounted on a sliding carrier which runs on a track made of two meter 
rods fastened end to end on a folding base. The base is mounted on adjustable stands 
fastened to a table. When making the test the apparatus is so adjusted that the track 
carrying the test card is just below and close to the observer's eye. In order that the 
observer's head may be held steady he is required to bite an impression of his teeth, 
previously made and hardened in wax on a mouth-board, which is rigidly fastened by a 
heavy iron rod and accessories to the table supporting the track and carrier. 
I Besides the letters li the writer would recommend the following figures as test objects. 
I 
(i) -1 -. The test is to distinguish clearly the dot at the center. A test object in the 
I 
shape of a cross has the advantage of affording a steady control of fixation. According 
to photographic records of involuntary eye-movement, where a variety of fixation 
objects has been used, the cross is found to give the best control of fixation. 
(2) | ' | or ! . In these figures also the test is to distinguish the dot clearly. The 
former figure, however, is a little too complicated. There is both a tendency to lose the 
dot and for the lines to run together on either side. A simpler criterion gives an easier 
and safer judgment. Doubtless with a little effort other figures can be found possessing 
still greater merit as test objects. 45 
FERREE : TESTS FOR THE EFFICIENCY OF THE EYE 
vision is determined by the distance at which he can just clearly 
distinguish in every detail the two test objects. The results are 
thus rendered directly in terms of clearness of vision, and there 
are no progressive errors introduced by memory and fatigue. 
We have good reason to believe that the brightness sensitivity, 
color sensitivity, and visual acuity tests rendered sensitive and 
adapted to our purpose in the manner described above will serve 
as a measure of the general level of efficiency of the unfatigued 
eye under different systems of illumination. For example, they 
show considerable difference in result when the tests are given 
under three types of lighting now in use: namely, systems of 
direct lighting, systems of indirect lighting, and daylight. In each 
of these cases, the intensity of the light falling on the test object 
measured in foot-candles is kept the same. The tests can not, 
however, be depended upon to show a loss of efficiency of the 
eye as a result of three or four hours of work even under a very 
unfavorable lighting system. 
III. LOSS OF EFFICIENCY AS THE RESULT OF A 
PERIOD OF WORK. 
We have no reason to believe that the brightness and color 
sensitivity tests have failed to show that the eye loses in efficiency 
as the result of a period of work under an unfavorable lighting 
system because of any fault in the tests. The tests used are the 
product of several years of study by the writer of the sensitivity 
of the eye to brightness and color and of the factors that in- 
fluence this sensitivity. There is doubtless very little, if any, loss 
of sensitivity during this length of time. In fact it is commonly 
believed that the brightness and color processes are compensating 
in nature. The case is quite different, however, with the conven- 
tional visual acuity test, or even with the modification of it de- 
scribed above. Although brightness and color sensitivity are fac- 
tors influencing the visual acuity test, still in every case to which 
it may be applied, it is predominantly a test of the refracting 
mechanism of the eye and its muscular control. In fact our re- 
sults for the tests of brightness and color sensitivity teach us that 
when applied to the case in hand in which there has been no 
change in the quality and intensity of the illumination or of the 
refracting mechanism from the beginning to the close of work,, TRANSACTIONS I. E. S.-PART II 
46 
the results of the visual acuity test may be ascribed practically 
entirely to changes in the muscular control of the refracting 
mechanism, or at least to changes in the muscular control of the 
eyes as a whole.8 Now the visual acuity test, when it is confined 
to a momentary judgment of clearness of vision, is not adapted 
to show a loss in muscular efficiency because, although this 
efficiency may have been lowered enormously, it may rise momen- 
tarily under the spur of the test to its usual level, or at least to 
the level obtaining at the beginning of work. Just as the runner 
may, under the spur of his will, equal in the last lap of his course 
the highest speed he has attained at any other point in the course; 
so may the flagging muscles of the eye be whipped up to their 
normal power long enough to make the judgment required by the 
visual acuity test. It was the feeling of all our observers that at 
the close of work under the system of direct lighting installed in 
our laboratory the eye had lost heavily in efficiency. A great deal 
of discomfort was felt. The test was painful and was accom- 
plished only with decided strain. Still the judgment could be 
made apparently with as much accuracy as at the beginning of 
work. But just as the runner finishing his course cannot long keep 
up his extra burst of speed, so might we expect that the eye 
cannot sustain its extra effort. This analogy led the writer 
to continue the visual acuity test through an interval of time. 
After considerable experimentation an interval of three minutes 
was chosen as best suited for our purpose. Our surmise proved 
to be correct. The fatigued eye cannot keep up its extra effort. 
The results of the test showed an enormous loss of efficiency as 
8 Before the writer would speak with full certainty, however, that the retina 
loses none of its power to function for color and brightness sensation during the above 
stated period of work, he would feel it necessary to perform another kind of test for color 
and brightness sensitivity. This test has been devised by him especially to meet the 
needs of this problem. In this test the element of time is introduced. It is possible 
that the retina may have lost in power to give color and brightness sensation as the result 
of a period of work even when the conventional test based on a momentary judgment, 
shows no loss of sensitivity. That is, it may be more susceptible to fatigue as the result 
of the preceding work. To determine this, a fatigue test should be run at the beginning 
and close of work. For color this may be done in two ways, (i) A given amount of 
colored light may be used and the time required for the eye to become completely ex- 
hausted or insensitive to this color may be determined. The difference in time required 
for this amount of fatigue to take place at the beginning and at the close of work will 
represent how much the retina has lost in its power to function for color. (2) The experi- 
ment need not be continued until complete exhaustion takes place. The amount of 
exhaustion that has taken place in a given interval of time can be measured. As before, 
this can be done at the beginning and at the close of work and the results can be compared 
to find out how much the retina has lost in power to give color sensation. 47 
FERREE: TESTS FOR THE EFFICIENCY OF THE EYE 
the consequence of three hours of work under the system of 
direct lighting, while in daylight practically no loss was shown. 
In detail the test is as follows. When the observer is required 
to look at the test card for three minutes, the test objects, even 
when the eyes are fresh, are not seen clearly for the whole time. 
The muscular effort required to keep the eyes adjusted for clear 
vision cannot be sustained steadily for that length of time. The 
test objects are seen alternately as clear and blurred. The time 
they are seen clear and blurred is recorded on a rotating drum 
upon which a line registering seconds is also run. From this 
record the ratio of the time seen clear to the time seen blurred 
is determined. This ratio may be fairly taken as a measure of 
the efficiency of the eye at the time the test is taken. In applying 
the test to our problem the record is taken at the beginning and 
at the close of work, and the ratios of the time clear and the time 
blurred are compared for the two cases to determine how much 
the eye has lost in efficiency as a result of work. Two values 
were chosen for the distance at which the test card was placed 
from the eye: (a) the maximal distance at which the test objects 
could be seen clearly in the momentary judgment, and (b) a 
distance less than this. The latter distance was chosen because 
for the maximal distance towards the close of the test, even when 
the eyes were fresh, the value of the time blurred became, it was 
thought, excessively high. Results for the two distances, there- 
fore, give probably a fairer expression of the loss in efficiency 
than for the one. 
The problem dealing with loss of efficiency as the writer has 
conceived it presents two phases. We may investigate (a) 
whether the eye shows a loss of efficiency after three or four 
hours of work under a given lighting system, and (b) whether 
there is a progressive loss of efficiency in working several 
months or years under a given lighting system. Only the first 
part of this investigation has been attempted thus far in our 
work and it has been undertaken, not so much for the purpose 
of making an exhaustive study of loss of efficiency under 
a given set of conditions, as it has been to get a sensitive and 
practical method of detecting loss of efficiency. In order to 
determine whether the method we have described is practical and 
sufficiently sensitive for our purpose, tests should be made on a TRANSACTIONS I. E. S.-PART II 
48 
large number of people under a wide range of lighting condi- 
tions. We have not as yet made tests under a wide range of 
lighting conditions. We have chosen rather to begin with three 
broad types of illumination now in general use; systems of direct 
lighting, systems of indirect lighting, and daylight. Types based 
upon the distribution of light have been selected because it has 
seemed to the writer, both from his own work and from a survey 
of the work done by others, that distribution or diffuseness of 
light is the most important factor we have yet to deal with in 
our search for conditions that give minimum loss of efficiency 
and maximum comfort in seeing. The quality of the light and 
its intensity at the source are already pretty well taken care of, 
apparently at least better taken care of in general practise, rela- 
tive to their importance to the eye, than is distribution. A 
detailed report of our results will not be given in this paper. 
The following results selected as typical from a large number 
of observations are appended, however, to show how the effi- 
ciency of the eye as measured by the above test falls off as the 
result of three hours of work under a system of direct lighting 
as compared with daylight. 
The tests were conducted in a room 30.5 feet (9.29 m.) long, 
22.3 feet (6.797 m-) wide and 9.5 feet (2.895 m-) high. The 
daylight illumination came from six windows, all on one side 
provided with thin white curtains to secure the necessary control. 
The artificial lighting9 was accomplished by means of two rows 
of fixtures of four fixtures each. Each row was 6 feet (1.828 
m.) from the side wall and the fixtures were 6 feet apart. Each 
fixture was supplied with two 16 candle-power carbon lamps 
29 inches (0.736 m.) from the ceiling with a white porcelain 
reflector 16 inches (0.406 m.) in diameter fastened directly above. 
The daylight tests were made at 9 a. m. and 12 m. Between these 
limiting times, the observer was required to read pages of type, 
uniform in size, printed upon paper of uniform texture of sur- 
face and of uniform reflecting power. The tests for the system 
9 This room gave the impression of being brilliantly lighted. The writer was amazed 
to find, however, that only 2.5 foot-candles of light were received on the test card placed 
about midway between two of the rows of lights and midway between two sets of fixtures. 
The walls and ceiling of the room were of plaster, natural finish, and the floor of dark 
tiling. Before our tests were taken, the walls and ceilings were painted white which 
nearly doubled the light received on the test card. 49 
PERREE : TESTS FOR THE EFFICIENCY OF THE EYE 
of direct lighting were taken at 7 p. m. and 10 p. m. During the 
interval intervening, the observer was required to read type of 
the same size and printed on the same paper as was used in the 
daylight work. The reading was done in each case at exactly 
the same spot in the room as at which the tests were made. The 
intensity of illumination was also in both cases made as nearly 
equal as it was possible to do by methods now available.10 The 
two tests were always given on successive days but one. In order 
to guarantee that the observer's physical and optical condition 
should be as nearly the same for the two tests as it was possible 
to obtain, he was required to rest during the day immediately 
preceding each test. Since the li test has proven to be the more 
sensitive, results will be given for it alone in the following table. 
Column 1 of this table gives the time of day at which the work 
was done and the tests were made. Column 2 gives the type of 
test. Column 3 gives the distance of the test card from the eye. 
As stated earlier in the paper, two distances were used; one the 
maximum at which the test object could be seen clearly, the other 
a distance less than this. Division A of the table gives the results 
for the former distance; division B, for the latter. Columns 4 
and 5 respectively, give the number of times the test object was 
seen clear and unclear. Column 6 gives the number of seconds 
in the three minutes that the test object was seen clear, and 
10 In order to equalize the intensity of illumination, a method of measurement is 
required. Two methods were used by us ; photometry, and a more delicate method 
based upon the sensitivity of the peripheral retina to brightness contrast. In case of the 
former, a Sharp-Millar portable photometer was used. The light falling upon the test 
card was measured in foot-candlesand was made equal for each type of lighting. Full 
details of the latter method will not be given here. As stated above it is based upon the 
extreme sensitivity of the peripheral retina to brightness contrast, especially to the 
induction by a white screen. To apply the method, some given illumination is taken 
as standard. The amount of black induced by a white campimeter screen upon a 
15 mm. area of some medium gray, (£. ,g. Hering gray No. 14) at an excentricitj' of 25 
deg. in the temporal meridian, is measured. This amount of contrast is taken as the 
index of that illumination. To duplicate the illumination at any succeeding time, the 
intensity is varied until the same amount of contrast is induced by the white screen on 
the gray at the 25 deg. point, for the same observer. This method was devised in the 
writer's laboratory and he has found by repeated trials that, although it is not so con- 
venient for many of the purposes for which the photometric method is used, it is many 
times more sensitive than the traditional photometric method. The Sharp-Millar 
photometer, like other photometers, is insensitive for the determination of the illumina- 
tion of a room by daylight. This is because the standard field illuminated by the 
tungsten lamp is deep orange in color, while the comparison field illuminated by day- 
light is clear white. This difference in color tone makes the judgment of brightness 
equality difficult to make and renders the instrument extremely insensitive for daylight 
work. TRANSACTIONS I. E. S.-PART II 
50 
column 7 the number of seconds unclear. Column 8 gives the 
ratio of the total time clear to the total time unclear. This ratio 
as stated earlier in the paper expresses the efficiency of the eye 
for clear seeing for an interval of three minutes at the time at 
which the test was taken. 
TABLE I. 
Showing How the Eye Falls Off in Efficiency as the Result of Three Hours 
of Work under a System of Direct Lighting as Compared with Daylight. 
In Division A the Test Card is Put at the Maximal Distance at Which the 
Test Object Could be Seen Clearly; in Division B, at a Distance Less than 
This.11 
Time 
of 
day 
Test 
Distance 
of card 
from eye 
cm. 
Number 
of 
times 
clear 
Number 
of 
times 
unclear 
Total 
time 
clear 
sec. 
Total 
time 
unclear 
sec. 
Total time clear 
Total time unclear 
A. 
9 
A. M. 
li 
102 
15 
J5 
105.6 
78.4 
1-4 
12 
M. 
li 
102 
15 
14 
103.1 
76.9 
1-33 
7 
P. M. 
li 
75 
18 
18 
II9.7 
60.3 
1.98 
IO 
P. M. 
li 
75 
15 
15 
55-4 
124.6 
0-44 
B. 
9 
A. M. 
li 
92 
14 
13 
136.8 
43-2 
3.16 
12' 
M. 
li 
92 
12 
12 
134-9 
45-i 
2.99 
7 
P. M. 
li 
65 
24 
23 
141.8 
38.2 
3-7 
IO 
P. M. 
li 
65 
17 
17 
75-5 
104.5 
0.72 
11 It will be noticed in the table that the ratio total time seen clear 4- total time seen 
unclear is smaller for the test both at the beginning and at the close of work in division 
A where the maximal distance at which the test object could be seen was used, than in 
division B where a distance less than this was used. This is just what should be expected 
from the nature of the test. For it may be said that, within limits, the nearer the object 
is to the eye the greater is the proportion of time it should be seen clearly ; and, con- 
versely, the farther the object is from the eye the smaller is the proportion of time it 
should be seen clearly. It will also be noticed that the ratio is slightly larger when the 
tests are made under the system of direct lighting than when made under daylight. The 
explanation of this, too, is found in terms of the distances that were chosen for the test 
object. These distances, relative to the maximal distance, were chosen shorter for the 
artificial light than for daylight. This was done because of the large falling off in the 
ratio gotten for the test at the close of work under the artificial light. Had the first ofthe 
two distances used in these tests, for example, been chosen as near to the maximal dis- 
tance for the artificial light as it was for daylight, the result of the test made at the close 
of work would have been that, after the first interval seen as clear, the test object would 
have been seen unclearly during the remainder of the test. At first glance one might 
be tempted to think that the difference in the scale of magnitude for the two ratios, is due 
to some inequality in the intensity of the illumination that was given by the two systems 
of lighting. It is obvious on reflection, however, that the intensity of illumination can 
have little or nothing to do with the scale of magnitude of these ratios. The intensity of 
the illumination influences the maximal distances at which the test object can be seen 
clearly but the scale of magnitude of the ratio, time clear to time unclear must depend 
primarily upon how near the distance chosen for the test object is to the maximal 
distance. (This principle, it is obvious, does not affect the comparison of the ratios 
obtained at the beginning and close of work under a given lighting system for the dis- 
(Continued on following page.} 51 
FFRRFF: TFSTS FOR THE) FFFICIFNCY OF THS FYF 
In order to give a typical representation in graphic form of 
the effect of three hours of work on the efficiency of the eye in 
daylight and under the system of direct lighting, estimated in 
terms of the test we have described, the results of the above table 
are given in the form of a curve. In constructing this curve the 
length of time of work is plotted along the abscissa and ratio of 
the time the test object is seen clear to the time unclear, is plotted 
along the ordinate. Each one of the large squares along the 
abscissa represents an hour of work, and along the ordinate an 
integer of the ratio. Figure I shows the result of division A and 
figure II for division B of the table. An inspection of these 
Fig. I.-Curve for division A of the table. Showing how the eye falls off in efficiency 
for three hours of work under a system of direct lighting as compared with daylight. 
curves shows that the efficiency of the eye measured by the ratio 
of the time the test objects are seen clear to the time seen unclear, 
falls off rapidly for the system of direct lighting but scarcely at 
all for daylight. 
Although it has been the purpose of this paper merely to out- 
tance. once it is chosen forthat system, is kept the same or both tests). As further 
proof that the difference in the intensity of illumination had nothing whatever to do 
with this result, the intensity of illumination was carefully determined immediately 
before and after these tests and, if the readings showed any inconstancy in the illumi- 
nation, the results were discarded and new tests were made. The above explanation 
should be borne in mind also in examining the curves plotted from the results of the table. 
The curve for division B of the table, for example, begins at a higher point on the ordi- 
nate than for division A ; and the curve for the artificial illumination starts at a higher 
point than the curve for daylight. 
It is scarcely necessary to point out that neither the scale of magnitude of ratio nor 
the point at which the curve startsis of any considerable consequence for our work. The 
important thing is not how large is the ratio at the beginning of work, but how much it 
falls off as the result of work. In fact the magnitude of ratio need not be taken into ac- 
count at all any further than that it chances to be a coincident result of a condition that 
seems to render our test more sensitive. That is, our results seem to show that the ratio 
falls off more when the distance chosen for the test object is not too near the maximal dis- 
tance. In future work, therefore, more care should be taken probably than was exercised 
in this preliminary study to choose the distances for the test object so that in case of each 
lighting system employed they shall sustain the same ratio to their corresponding 
maximal distances. TRANSACTIONS I. E. S.-PART II 
52 
line and in part to demonstrate a set of tests, a word of dis- 
cussion and interpretation of the results we have reported may 
not be out of place here. Since the visual acuity test (given 
under constant quality, intensity and distribution of light) is a 
test largely of the refractive mechanism of the eye and its mus- 
cular control and since the refractive mechanism could not have 
changed during three or four hours of work, the obvious indi- 
cation of the above result is that the loss of efficiency sustained 
by the eye in these experiments is a loss in muscular efficiency. 
This conclusion is borne out also by the fact, stated earlier in 
the paper, that the direct tests of the efficiency of the retina, 
namely, the test for brightness and color sensitivity did not show 
conclusively any loss.12 Moreover, the conlusion is in line with 
current conception. In current theory the retina is considered 
as a mechanism more or less compensating in its action, while 
Fig. II.-Curve for division B of the table. Showing how the eye falls off in efficiency 
as the result of three hours of work under a system of direct lighting as compared 
with daylight. 
the muscles of the eye are not so considered. The following 
reasons are suggested why the muscles of the eye giving both 
fixation and accommodation should be subjected to a greater 
strain by the system of direct lighting than by daylight, (i) The 
bright images of the electric bulbs falling on the peripheral retina 
which is in a perpetual state of darkness adaptation as compared 
with the central retina and is therefore extremely sensitive in its 
reaction to such intensive stimuli, sets up a reflex tendency for 
the eye to fixate them instead of the letters which the observer 
is engaged in reading. (2) Likewise, a strong reflex tendency 
to accommodate for these brilliant sources of light all at differ- 
12 This statement is also subject to the foot-note appended to the earlier statement. 53 
FERREE : TESTS FOR THE EFFICIENCY OF THE EYE 
ent distances from each other and from the lettered page, is set 
up. (3) These brilliant images falling upon a part of the retina 
that is not adapted to them causing as they do acute discomfort 
in a very short period of time,13 doubtless induce spasmodic con- 
tractions of the muscles which both disturb the clearness of 
vision and greatly accentuate the fatiguing of the muscles. The 
net result of all these causes is excessive muscular strain which 
soon shows itself as a loss in power to do work. In the illumina- 
tion of a room by daylight with a proper distribution of windows, 
the situation is quite different. The field of vision contains no 
bright sources of light to distract fixation and accommodation 
and to cause spasmodic muscular disturbances, due to the action 
of intensive light sources upon the dark adapted and sensitive 
peripheral retina. In daylight the light waves have suffered 
innumerable reflections and the light has become diffuse. The 
field of vision is uniformly illuminated. The illumination of the 
retina, therefore, falls off more or less uniformly from fovea to 
periphery as it should in order to permit of fixation and accom- 
modation for a given object with the minimum amount of strain. 
It is not our purpose to contend in this report that distribution 
is the only factor of importance in the illumination of a room. 
The intensity and the quality of the illumination must also be 
taken into consideration. To test the relative effect of these 
factors upon the working power of the eye, records would have 
to be taken when each was varied in turn and the other two 
maintained constant. In the results shown in the above tables 
the intensity alone was constant in the two cases. Both the 
quality and the distribution were different in the direct lighting 
system and the illumination by daylight. The difference in the 
results obtained will have, therefore, to be attributed both to 
13 There is no doubt in the writer's mind that the eye-discomfort experienced as the 
result of work under an unfavorable system of lighting is not by any means all muscular. 
The "sandiness" passing over into a stinging, stabbing pain which conies early in the 
experience of discomfort seems to be conjunctival. And while the retina itself is 
apparently insensitive to pain from mechanical stimulation, still when exposed to a 
source of light of a brilliancy to which it is not adapted, a painful reaction is produced 
which can scarcely be considered muscular. For example, after confinement for some 
time in a dark-room exposure to ordinary daylight is painful to the normal eye. That 
this is not entirely muscular can be shown by the fact that a similar reaction is experi- 
enced when the ciliary and iris muscles are paralyzed by atropine. The reaction is also 
experienced by aphakial subjects whose lenses have been so long removed that muscular 
atrophy must have taken place. TRANSACTIONS I. E. S.-PART II 
54 
difference in the distribution and to difference in the quality of 
the illumination. In our tests comparative of the systems of 
direct and indirect lighting, the results of which will be reported 
in a later paper, clear tungsten lamps will be used in both cases. 
The intensity will be made the same and the quality of the light 
will be approximately the same. The distribution or diffusion 
alone will be different. Whatever difference in result we get in 
these two cases can, therefore, with reasonable certainty be 
attributed to the differences in the distribution of light. 
With regard to the effect of varying the intensity of illumina- 
tion, our results show nothing; with regard to the effect of varying 
quality, nothing in isolation; and with regard to distribution, we 
have data only for such differences as are found in the three 
types of illumination now in general use. In later work, however, 
the analysis along these lines will be completed. We hope on 
the laboratory side, to make a systematic study of the effect of 
wide ranges of variation of each of the factors in turn. It will 
be comparatively easy, for example, to keep the intensity and 
distribution constant and vary the quality, or to keep the quality 
and distribution constant and vary the intensity. We hope in 
addition, to supplement this work by testing the eyes of employ- 
ees who work under a given lighting system for several hours a 
day, for evidences of a progressive loss of efficiency. 
IV. A preliminary study of the causes of 
DISCOMFORT. 
In addition to studying the conditions that give us maximal 
efficiency, it is important to determine the lighting conditions 
and eye factors that cause discomfort. In fact, it might well be 
said that our problem in lighting at present is not so much how 
to see better as it is how to see with more comfort and with less 
damage to the general health on account of eye strain. Any 
comparative study of the conditions producing discomfort neces- 
sitates a means of estimating discomfort. It is obvious that the 
core of the experience of discomfort is either a sensation or a 
complex of sensations. As such, it should have a limen or 
threshold value just as other sensations have; and just as we are 
able in general to estimate sensitivity in terms of the threshold 
value, so should we in this case be able to use the threshold 55 
FERREE : TESTS FOR THE EFFICIENCY OF THE EYE 
value in estimating the eye's sensitivity or liability to discomfort 
under a given lighting condition. Threshold values are usually 
determined by finding how much energy or intensity of a given 
stimulus applied for a short interval of time is required to arouse 
a just noticeable sensation. This form of procedure, however, 
is not adapted to the needs of our problem. It is much better to 
reverse the process and find how long the eye has to be exposed 
to a stimulus of a given intensity to arouse just noticeable dis- 
comfort. Our limen, then, becomes a time limen, and is meas- 
ured in units of time instead of in units of intensity. In order 
to determine whether the judgment of the limen of discomfort 
can be made with certainty and to test in general the feasibility 
of the method, the writer undertook to determine the compara- 
tive sensitivity of the eye to discomfort when the source of light 
was exposed in different parts of the field of vision. In order to 
carry out this investigation, a 16 candle-power lamp was attached 
to the arm of a perimeter in such a way that the end of the bulb 
was always directed towards the observer's eye. The arm of the 
perimeter could be shifted to any meridian in which it was desired 
to work, and the lamp could be moved at will along the arm. It 
was thus possible to expose the light at any point in the field of 
vision that was desired. Working in this way, we have investi- 
gated the effect of many types of variation of the distribution of 
the light in the visual field, and it is our purpose to extend the 
investigation as fast as possible to the variation of the other 
factors. Of the variations we have made in the distribution of 
the light in the field of vision, it will be necessary, however, in 
order to illustrate the general method of working, to describe 
only one, namely, the exposure of the source of light at different 
points in the field of vision for one eye when fixation and accom- 
modation were taken for a far point. 
In carrying out the investigation, the following precautions 
were observed, (a) It was found better to work in a room mod- 
erately illuminated by a source of light behind the observer and 
entirely concealed from him rather than in the dark. The inter- 
vals of dark-adaptation between exposures in the dark-room 
seemed to make the observer's eye too sensitive for our purpose. 
This was especially true for certain parts of the peripheral retina. TRANSACTIONS I. E. S.-PART II 
56 
In becoming supersensitive there was a tendency to become 
erratically sensitive, (b) It was found that blinking serves as a 
variable factor for the relief of discomfort and that the amount 
of blinking must be made constant from test to test. This was 
accomplished by having the observer blink at equal intervals dur- 
ing the exposure, timing himself by the stroke of a metronome. 
The interval most natural and suitable for this purpose was de- 
termined for each observer separately, (c) All comparisons were 
planned in series. For example, if it were desired to compare 
the sensitivity of the temporal and nasal halves of the retina in a 
given meridian, the exposure was first made at a given point in 
one half and next at the corresponding point in the other half. 
This was to guarantee that the eye should be as nearly in the 
same condition with regard to progressive fatigue, etc., as was 
possible. Further to safeguard against error in this regard ser- 
ies were compared in which the exposures were repeated in the 
reverse order, (d) An interval of recovery was allowed between 
exposures. This interval had to be determined separately for 
each observer and often had to be made different for the same 
observer on different days. It was never changed, however, 
during the course of an experiment, the results of which were to 
be compared, (e) In order that the observer's head be held 
rigidly in position during the exposure, he was required to bite 
an impression of his teeth previously made and hardened in wax 
on a mouthboard. When an exposure was to be made, the fixa- 
tion was taken, the light turned on, and a signal was given by the 
observer when a just noticeable discomfort was aroused, or, if 
it was desired, when the different stages of discomfort were 
reached. The judgment was found to present no especial diffi- 
culty, and the method, when properly applied, to provide a feas- 
ible means for comparing the sensitivity of the eye to discom- 
fort under all the conditions to which we have been able thus far 
to extend its application. In actual practise the method also 
brings out an analysis of discomfort. 
Discomfort seems to be a complex of three experiences, each of 
which develops at a different time. When the light is turned on, we 
have at once glare. This is a light sensation and though unpleas- 
ant has no painful elements. Next comes a conjunctival sensation 57 
FERREE: TESTS FOR THE EFFICIENCY OF THE EYE 
which begins with what is commonly called "sandiness" and soon 
passes over into a sharp, stinging, stabbing pain. Lastly there 
comes what is probably a muscular discomfort,-a hurting and 
aching in the ball of the eye which if the exposure is continued 
long enough seems to radiate to the socket and the surrounding 
regions of the face and head, the arch of the brow, the forehead, 
the temples, etc. Details will not be given here of the compara- 
tive sensitivity of different points of the retina to discomfort. It 
will be sufficient to say, that the periphery of the retina is more 
sensitive than the center; that the nasal half is in general more 
sensitive than the temporal half and the upper half than the 
lower half; and that in passing from the center to the periphery 
of the retina, the sensitivity is found first to increase then to de- 
crease, becoming extremely little at the limits of the field of 
vision. In the horizontal meridians both on the temporal and 
nasal sides, maximal sensitivity is found around the 45 deg. 
point. In the vertical meridians, maximal sensitivity seems to 
be near the point 15 deg. below the horizontal. In a paper soon 
to be published, a detailed statement and explanation of these 
results will be given. 
DISCUSSION. 
Dr. H. E. Ives : This paper is well worth the while of pro- 
fessional psychologists to study; and it gives to the illuminating 
engineer results which are extremely valuable. 
When the illuminating engineer has the problem of producing 
satisfactory results, he has two methods of doing so; first, the 
case system, in which he copies an illumination produced by 
nature or invention, which has proved satisfactory by experience; 
and he hopes to get the same result. But there are defects in 
this method; we are very apt to follow the example of the Chinese 
who made motors by copying the imported ones even down to 
the color of the paint on the casings and the scratches on the paint 
We may do equally foolish things by slavish copying. That is in- 
herent in the case system. Up to recently some of us have been of 
the opinion that even with the defects of this case system we could 
apply it to advantage, for instance by studying how nature pro- 
duces her lighting schemes. But in order to make any great TRANSACTIONS I. E. S.-PART II 
58 
progress we must deviate from what exists; we must experiment 
and invent. This necessitates some means of testing our results 
and this process of experiment and test constitute the second pro- 
cedure. Our Society has lately been interested in the physiological 
side of illumination, but has been sadly handicapped by the lack of 
significant tests-we have been dependent practically on labor- 
iously acquired experience. One great object in adopting a 
method of measurement is the saving of time. For instance, 
suppose our only means of measuring voltage was by the duration 
of physiological disturbances following an electric shock. In order 
to duplicate a voltage which gives a shock whose after effects 
last a day, we would require weeks or months of toil, because of 
the time necessary to wait for the results of each experiment. 
Suppose the first time we secured the desired voltage we had an 
instrument known as a "voltmeter;" it would only take a minute 
to determine that voltage. We want something for measuring 
the effect of lighting systems which will enable us to get results 
with a speed comparable with that of a voltmeter. 
Various methods of test have been proposed and Dr. Ferree 
has gone over all of these. He arrives at a conclusion which I 
think it behooves us all to observe; namely, that these tests will 
show what he calls "the general level or scale of visual efficiency," 
but they are practically useless as tests of the loss of visual 
efficiency. 
Here is a sentence which means a great deal "Just as a runner 
may, under the spur of his will, equal in the last lap of his course 
the highest speed he has attained at any other point in the course, 
so may the flagging muscles of the eye be whipped up to their 
normal power long enough to make the judgment required by the 
visual acuity test." 
Dr. Ferree here gives us the benefit of his point of view and 
experience in these matters. In this paper he has recognized the 
inefficiency of the methods now used. He realizes that we want 
a test of the loss of visual efficiency. The eye may respond 
momentarily, like the tired runner, and see the object as dis- 
tinctly as before, but we know that it is not as efficient. Dr. 
Ferree has devised a test in which is introduced a time element. 
The observer views a visual acuity test object. When the 59 
FERREE : TESTS E'OR THE EFEICIENCY OF THE EYE 
limit of visibility is found the observer is not allowed to rest, 
because he will again after an interval get just as good results 
as at first; instead he presses a key as long as the detail is clear; 
then when the tired muscles flag and the object blurs, the finger 
on the key is removed. At first it appears easy to see the detail 
clearly, but pretty soon it is not so easy and one does not distin- 
guish the chart so well. Very soon it becomes necessary to take the 
finger off the key. Intervals of clear and blurred vision alter- 
nate and at the end we have a ratio of the time the chart is dis- 
tinguishable to the time when it is not. 
Dr. Ferree has tried out daylight and a direct artificial 
lighting system and we have here for the first time the results of 
that test. They show what many of us have been sure of; 
that daylight does not decrease the efficiency nearly as much as 
artificial lighting. On the fourteenth and fifteenth pages are two 
charts showing by straight lines the falling off in efficiency which 
occurs under artificial lighting as compared to daylight. Person- 
ally, I think we should say "Eureka!" 
I hope Dr. Ferree will proceed to standardize these tests and 
tell us the best working distances and one thing and another. 
As he is not here, I have tried to bring out the most important 
points. He has given us a most valuable contribution, and I 
hope before long we will be in a position to settle these questions 
of light and dark walls by this method of test and not by 
"Kilkennycat" discussion, which brings us nowhere. 
I think we should do our best to aid Dr. Ferree to develop this 
method of test to give us what now we can get only by experience. 
I am aware that I have not done this paper justice, but I want to 
express my appreciation of his work. 
Mr. C. O. Bond : The American Medical Association is a 
body whose conclusions as to the harmful or beneficial effects of 
any types of illuminating installations will carry considerable 
weight. They have discussed time after time how they were to 
make the tests, and this paper has grown out of Dr. Ferree's 
experiments, in the hope of placing in the hands of that Com- 
mittee means of making the tests. Wei are extremely fortunate 
in having the first public report of this method. The method is 
under advisement by the Committee and I was present at one of 
their meetings when Dr. Ferree brought a set of this apparatus TRANSACTIONS I. U. S.-PART II 
60 
to Philadelphia and they made a test of it. Two or three of the 
doctors present were very much impressed with it. I think, even 
if it does not succeed as it now stands, perhaps here is the germ 
of the best possible method of test. 
Dr. C. E. Ferre® (communicated in reply) : I can express 
only great appreciation of the interest that the men who have pre- 
ceded me have taken in our work. The problem is extremely in- 
teresting to me and I hope we have here a vulnerable point of 
attack. Once we have procured a successful method of measur- 
ing the effect of different lighting systems on the eye, a broad field 
of application opens out before us. We not only can find out 
what are the favorable and what are the unfavorable features in a 
lighting system, but we can no doubt, as may be inferred from Dr. 
Ives' discussion, test out and perfect a lighting system, so far as 
its effect on the eye is concerned, before we put it on the market. 
This latter point is a good one, I think, and I thank Dr. Ives for 
the suggestion.* I feel that Dr. Ives' perspective and practical 
grasp of the situation is a distinct contribution to the paper. 
We are very much handicapped at present for funds by means 
of which to carry on this work. In the first place apparatus and 
models of lighting systems are required for the work on the 
laboratory side. Trained assistants are also needed to help out 
with the details of the work. Further, to verify and enlarge 
the work done in miniature in the laboratory, we should test the 
eyes of employees working under established lighting systems 
and in the surroundings in which these systems have to operate. 
All of this takes time and money, also entrance into commercial 
concerns. In all of these regards we need the help and influence 
of the Illuminating Engineering Society. 
This work, I suppose, could be done spontaneously and sporadi- 
cally here and there as the insight and inclination of various men 
may direct. But in the beginning, at least, I do not think 
it should be scattered. Until launched and safely moving, it 
should be done under common supervison. 
* The general idea that over and above its application to abstract investigation the 
test may have an application in the daily work of the lighting engineer has come to the 
writer by suggestion from the engineers themselves. Mr. Cravath, for example, has re- 
cently pointed out that the test should be of advantage in making the actual installation 
of a lighting system. The writer would suggest in addition that it may further be of ser- 
vice in determining the effect of different kinds of type and paper on the efficiency of the 
eye; also the effect of different kinds of desk lighting, etc. In short, it is obvious that the 
usefulness of such a test is limited along these lines only by its sensitivity. THE FACTORS THAT INFLUENCE THE 
SENSITIVITY OF THE RETINA TO COLOR: 
A QUANTITATIVE STUDY AND 
METHODS OF STANDARDIZING 
dissertation 
PRESENTED TO THE FACULTY OF BRYN MAWR COLLEGE IN PARTIAL 
FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE 
OF DOCTOR OF PHILOSOPHY 
BY 
GERTRUDE RAND 
Reprinted from 
THE PSYCHOLOGICAL MONOGRAPHS 
PRINCETON, N. J. 
March, 1913 TABLE OF CONTENTS 
I. Introduction I 
II. Historical and Critical 6 
A Factors that have been found to influence the 
sensitivity of the retina to color 6 
i. Size of stimulus 6 
2. Intensity and brightness or white value of the 
stimulus 12 
a The confusion that has arisen with regard 
to the meaning of intensity and of 
brightness, and its effect uport the 
development of methods of working. . 12 
b The effect of intensity of stimulus 20 
1) The effect upon the limens of sensitivity 20 
2) The effect upon the j. n. d. of sensitivity 22 
3) The effect upon the limits of sensitivity 24 
c The effect of brightness of the Stimulus.. 38 
3. Brightness of the field surrounding the 
stimulus 55 
4. The general illumination 57 
B Methods of standardizing these factors 63 
1. Size of the stimulus 63 
2. Intensity of the stimulus 63 
3. Brightness of the stimulus 73 
C Summary 77 
III. Experimental 79 
A Purpose of investigation 79 
B Description of optics-room and apparatus 86 
C Determination of the brightness of the colored 
stimuli employed in the investigation 91 
D The factors investigated 97 
1. Brightness of the stimulus 97 
2. Brightness of the field surrounding the 
stimulus no Table of Contents 
a The effect of the induction of the sur- 
rounding field upon the limits of color 
sensitivity 113 
b Explanation of the effect of the induction 
of the surrounding field upon the 
limits of color sensitivity 113 
1) The relative inhibitive action of black 
and white upon color in peripheral 
vision 115 
2) The rate of falling off in the sensitivity 
of the retina to color from center 
to periphery 117 
c The effect of the induction of the surround- 
ing field upon the color limens 120 
d Explanation of the effect of the induction 
of the surrounding field upon the color 
limens 121 
3. The brightness of the preexposure 126 
a Effect upon the limens of color and upon 
the limits of color sensitivity 129 
b Combined effect of surrounding field and 
preexposure upon the limits of color 
sensitivity 131 
4. The general illumination of the retina 135 
a Quantitative estimate of the influence of 
change of illumination upon the in- 
duction of brightness by the surround- 
ing field 138 
b The effect of these amounts of induction 
upon the limits of color sensitivity . . 144 
c The effect of these amounts of induction 
upon the limens of color at different 
degrees of excentricity 150 
d The influence of change of illumination 
upon the action of the preexposure on 
the limens and limits of color 156 
E Methods of standardizing these factors 158 PREFACE. 
The following study, practically as is here presented, was 
submitted to the Faculty of Bryn Mawr College in May 1911 
in partial fulfillment of the requirements for the degree of 
Doctor of Philosophy. It is the outgrowth of a series of studies 
dealing with the phenomena of color vision that was begun in 
1908 by the writer working under the direction of and in col- 
laboration with Professor C. E. Ferree of Bryn Mawr College. 
In order to show in what way the present study, which deals 
with the formulation of a technique for investigating color 
sensitivity, is the logical outcome of the initial studies in the 
series, and is, moreover, required for the completion of the later 
studies, a brief resume will be given of the work undertaken in 
the investigations preceding and following it. 
The first of these studies entitled: Colored After-image and 
Contrast Sensations from Stimuli in Which No Color Is Sensed, 
was published in the Psychological Review, 1912, XIX, pp. 195- 
239. As is shown by the title, the article deals with the con- 
ditions under which colored after-image and contrast sensations 
may be aroused from stimuli in which no color is sensed. A 
formulation of these conditions, together with allied fusion and 
limen experiments, shows the phenomenon to be a peculiarity of 
the inhibitive action of brightness upon color. Brightness fused 
with color inhibits its saturation. With the exception of the 
region just within the limits of sensitivity for two colors, the 
following may be stated roughly as a law of this action for all 
colors and all parts of the retina: white inhibits most, grays In 
the order from light to dark next, and black least. This law was 
generalized from the results of fusion and limen experiments 
in a large number of meridians of the retina. In accord with 
this law, color may be obtained in the after-image when none is 
sensed in the stimulus when an unfavorable brightness quality 
is fused with the stimulus color and a favorable one with the 
after-image color. The technique for securing these conditions VI 
PREFACE 
for the after-image sensations in central and peripheral vision, for 
the contrast sensations in central vision, and for the phenomenon 
which we have called the Purkin je-Briicke phenomenon, is des- 
cribed in detail in the paper. But the study is not quantitative. 
The results do not show, for example, that the inhibition of the 
stimulus color has no effect upon the after-image. They show 
merely that, working near the limen, the stimulus color may be 
inhibited and the complementary color still be sensed in the 
after-image. 
In order to determine as accurately as possible to what degree, 
if at all, the intensity of the after-image excitation is decreased by 
adding to the stimulus color a brightness excitation unfavorable 
to its saturation, the second study of the series was begun in 
1909. It was entitled: The Fusion of Colored with Colorless 
Light Sensation: The Physiological Level at Which the Action 
Takes Place. An abstract of this article was published in the 
Journal of Philosophy Psychology and Scientific Methods, 1911, 
VIII, pp. 294-297. The full report will shortly be published in 
the Psychological Review. The results of this study have a 
twofold bearing. (1) They make plain once for all why it is 
possible to obtain color in the after-image when none is sensed 
in the stimulus, for they show that the intensity of the after- 
image excitation is not decreased at all by adding to the stimulus 
color a brightness excitation unfavorable to its saturation. (2) 
They throw some light on the broader problem presented by 
the fusion of brightness and color. By serving to indicate the 
level at which this action takes place, they help, for example, to 
explain a number of somewhat puzzling phenomena attendant 
upon the fusion of brightness with color, in case of positive, 
after-image, and contrast sensations. This action takes place 
apparently posterior to the seat of the after-image and contrast 
processes and the cancelling action of the complementary colors. 
There are two effects of the fusion of brightness with color, both 
of which are pressed into service in drawing the above con- 
clusion : (1) it reduces the saturation of the color sensation; and 
(2) it changes the quality or tone of certain colors. This con- 
clusion is based on the following lines of argument: PREFACE 
VII 
(i) When the color of the stimulus is inhibited by the addition 
of a brightness excitation, the intensity of the after-excitation, 
judged in terms of the duration of the after-image, is not 
affected by this excitation. (2) When the tone of the color 
aroused by a given stimulus is changed by the addition of a 
brightness excitation, the color of the after-image does not 
undergo a complementary change. (3) When the saturation of 
the inducing color is inhibited by the addition of a brightness 
excitation, the saturation of the contrast color is not affected by 
the change. (4) When the tone or quality of the inducing color 
is modified by adding a brightness excitation, the tone of the con- 
trast color is not determined in the complementary direction. (5) 
When a given color is inhibited by the addition of a brightness 
excitation, its power to cancel the complementary color is not al- 
tered. (6) When the tone or quality of a color is altered by the 
addition of a brightness excitation, the tone or quality of the color 
required to cancel it is not affected by the change. (7) When the 
tone or quality of a color has been altered by the addition of a 
brightness excitation, the color component added can not be can- 
celled by mixing with the original color a color complementary to 
this component. Since, then, this fusion affects the positive and not 
the negative excitation and does not affect the cancelling action of 
the complementary colors, the conclusipn is drawn that it takes 
place at some physiological level posterior to the seat of the 
after-image and contrast processes and to the cancelling action 
of the complementary colors. This study deals, however, only 
with the measure of the effect of the fusion of brightness and 
color as it occurs in central vision. 
In order to extend the investigation to the peripheral retina, a 
third study was begun in June 1911. Its object was as follows. 
(1) It was planned to determine the effect of the fusion of 
brightness with color at a number of points from the center to 
the periphery of the retina, and to see how far the following 
points can be explained in terms of this action: (a) the influence 
upon the limits of color sensitivity of the brightness of the sur- 
rounding field and of the preexposure; and (&) the color changes 
that occur in passing from central to peripheral vision. (2) A VIII 
PREFACE 
comparative study was to have been made of the chromatic and 
achromatic phenomena of the peripheral retina. Both of these 
sets of phenomena for the peripheral retina show a number of 
striking differences from the phenomena of the central retina. 
Some of these points of difference are as follows: (a) There is 
considerable difference in the action of the achromatic qualities 
on color, (i) They inhibit or reduce the saturation of color 
much more strongly in the peripheral than in the central retina. 
(2) Of the achromatic qualities, white inhibits all colors the 
most strongly and black the least strongly in the central retina, 
while near the limits of sensitivity in the peripheral retina, black 
inhibits red and yellow the most strongly and white the least 
strongly. (3) The change in the tone or quality produced by 
adding white or black is much more pronounced in the peripheral 
retina and is often in a different direction. For example, black 
added to yellow in daylight illumination in the central retina 
turns it towards green; while in the peripheral retina the change 
is towards red. (&) Varying the brightness of the surrounding 
field has more effect on colors in the peripheral than in the 
central retina, (c) Exhaustion to color takes place more rapidly 
in the peripheral than in the central retina, that is, the change 
of saturation per unit of time is faster, (d) The colored after- 
image ,ijs of very short duration in the peripheral retina, but in 
proportion to its duration, it is much more saturated than the 
after-image of the central retina. (e) Some of the differences 
with regard to the achromatic phenomena are as follows. There 
is a very strongly increased sensitivity to contrast and to flicker; 
adaptation or exhaustion occurs very rapidly; the after-image is 
quickly aroused, is relatively very intensive, and in proportion 
to its intensity lasts a very short time; and so on. These points 
of difference raise the question how far we need go in assuming 
a different mechanism for the two parts of the retina. The 
purpose of our investigation was to have been primarily to 
determine how many of these differences are due at least in 
part to the difference in the state of brightness adaptation of the 
central and of the peripheral retina. (3) Maps were to have 
been made showing the sensitivity of the eye to the different PREFACE 
IX 
colors for three kinds of background (white, black, and gray of 
the brightness of the stimulus color). These three backgrounds 
were selected because they represent the extreme situations with 
regard to achromatic induction: maximal black induction, 
maximal white induction, and no induction; and, therefore, rep- 
resent the best conditions that can be obtained for the study of 
the effect of the brightness of the surrounding field upon the 
local sensitivity of the retina. A sufficient number of meridians 
were to have been worked over to give an accurate outline of 
the zones of sensitivity for the three kinds of background used. 
Gradients were to have been established showing the falling off 
in sensitivity from the fovea outwards. Also the changes in 
color tone were to have been determined from point to point for 
all the backgrounds. Both sets of determinations were to have 
been made by matching in central vision what is seen in peripheral 
vision. The object of this investigation was to have been to 
give a complete representation of the sensitivity of the entire 
retina, quantitative and qualitative, in terms that are more or 
less familiiar to all, namely, the sensation values of the central 
retina. It was found, however, that the large M. V. occurring in 
the work from observation to observation rendered the extended 
comparative investigation planned impossible. The original plan 
had, therefore, for the time to be abandoned, and the present 
study was undertaken. 
This study aims (i) to determine what are the factors that 
influence the sensitivity of the retina to color; (2) to make a 
quantitative examination of the factors extraneous to the stimu- 
lus; and (3) to provide methods for their standardization. For 
the sake of historical continuity, the study is preceded by an 
historical and critical resume of the analyses of factors influenc- 
ing color sensitivity that have been made up to this time and of 
the attempts to standardize. It may be stated in passing that 
with the control of factors rendered possible by this study, the 
original plan of work has been resumed and in part, completed. 
It will be published in the near future. 
There remains to be mentioned the relation of the present 
study to the final one of the series. The latter has developed PREFACE 
X 
from the historical and critical resume mentioned above of the 
investigations that have been made to determine the factors that 
influence the sensitivity of the retina to color. In the course of 
this discussion various deficiencies have been pointed out in the 
methods used by previous investigators in their attempts to 
control these factors, and ways have been devised to correct these 
deficiencies. The factors that influence the sensitivity of the 
retina to color may be divided into two classes: those pertaining 
to the stimulus, and those extraneous to the stimulus. The 
experimental part of the present study is especially directed 
toward making a quantitative estimate of the latter set of factors 
under various typical conditions obtaining in the investigation of 
color sensitivity and toward securing effective methods of control. 
No concern is had, however, to standardize the factors pertaining 
to the stimulus any farther than is necessary to accomplish this 
purpose. The more effective standardization of these factors 
will form the subject of our future work. The question of 
intensity will be taken up first. It has been shown in the historical 
part of the present study that this factor has been most inade- 
quately handled by previous investigators. In determining the 
comparative sensitivity of the retina to the different colors, for 
example, either no account has been taken of the different inten- 
sities of the colors used, or incorrect methods have been employed 
of equalizing these intensities. In no case has the determination 
been made ip terms of units that can be compared. It is the 
writer's purpose to make an exhaustive determination of the 
sensitivity of the retina to the different colors in terms of such 
units. The comparative limits of sensitivity will be determined in 
a number of meridians with stimuli equalized in energy, and the 
limens and the j.n.d's. of sensitivity at different degrees of 
intensity will be determined in terms of radiometric units at 
various points from the center to the periphery of the retina in 
the different meridians. This investigation in fact is now in 
progress. A preliminary statement of the plan of this work has 
already been published by the writer in collaboration with Pro- 
fessor Ferree (American Journal of Psychology, 1912, XXIII, 
PP- 328-332). XI 
PREFACE 
From the above discussion of the place of the present study in 
the series, it will be evident to the reader how extensively this 
study has been due to the instruction, guidance, advice, and 
assistance of Professor Ferree, and how great a debt of gratitude 
the writer owes him. In stating his share and collaboration in 
the studies preceding and following this in the series, I can indi- 
cate perhaps more fully than in any other way the share he has 
had both directly and indirectly in the production of the present 
study. I. INTRODUCTION. 
In no branch of psychological optics does one find such varied 
and contradictory results as in the work on the color sensitivity 
of the peripheral retina. This is doubtless due in minor part to 
the intrinsic difficulty of the indirect vision observation, but in 
major part it is due to the lack of adequate standardization of the 
factors that influence the local stnsitivity of the peripheral 
retina. These factors may be divided into two classes: (a) 
those pertaining to the stimulus, or the source of light; and (b) 
those extraneous to the source of light. In the former class may 
be included the size, intensity, and brightness of the stimulus; in 
the latter, the preexposure or what the eye has rested on before 
being exposed to the stimulus, the surrounding field, and the 
general illumination of the visual field.1 The work of standard- 
ization thus far has been directed largely towards the factors in 
the former class. Of the factors in the latter class, attempts 
have been made, as will be shown later in the discussion, to 
standardize only the influence of surrounding field. The recog- 
nition of the importance of this factor came relatively late in the 
development of the technique of the subject. It was at one time 
thought that the use of the perimeter and the dark-room provided 
ideal conditions for testing the local sensitivity of the peripheral 
retina, because by this means the local area alone was stimulated 
by light, hence it was thought that the influence of the surround- 
ing field was eliminated. We know now that these conditions 
were not so ideal as they seemed, that a dark- as well as a light- 
1 In case the colored light is obtained by reflection from a pigment surface, 
some exception may be taken to the above classification, for unless some 
especial device be used to illuminate the pigment surface, the intensity of the 
stimulus will depend upon the degree of the general illumination of the 
visual field, and the brightness of the stimulus will also, to a certain extent, 
be dependent upon the general illumination. In such a case, these factors 
would have to be included in both classes. If on the other hand the colored 
light is obtained by means of standard filters, or from the spectrum, the 
illumination of the visual field will exercise its influence entirely indepen- 
dently of any effect on the stimulus. GERTRUDE RAND 
2 
adapted retina influences by contrast the sensitivity of the area 
stimulated. The use of the perimeter and dark-room accom- 
plishes, then, but a very small part of the purpose for which it 
was intended. Instead of eliminating altogether the influence of 
the brightness of the surrounding field, it makes only one phase 
of it constant. It standardizes by giving us one state of 
brightness-adaptation alone, namely, the adaptation of the dark- 
room.2 The campimeter was devised especially to correct this 
deficiency. Its purpose is to control and standardize the influence 
of the brightness of the surrounding field when one is working 
with a light-adapted retina. But the campimeter, like the peri- 
meter, has accomplished only in part the purpose for which it was 
intended. It standardizes the influence of the surrounding field 
for one degree of illumination only, because the influence of the 
campimeter screen changes markedly with chang'es in the illumin- 
ation of the visual field. There are two reasons for this, (a) A 
brightness match between the colored stimulus and the gray of 
the surrounding field made at one illumination will not hold at 
another. And (&) the sensitivity of the retina to brightness 
induction changes markedly with changes in the general illumina- 
tion. This latter point is especially true in the peripheral retina 
where changes which are too small to be detectable by any current 
photometric device produce quite a noticeable change in the 
amount of induction between two surfaces of different bright- 
ness. The campimeter, then, is almost useless as an instrument 
of precision, unless the general illumination can be rendered 
constant or some means can be devised for standardizing the 
observation for changes of illumination. No satisfactory method 
has as yet been obtained for keeping the illumination of a room 
by daylight constant. To keep it constant presupposes what has 
not as yet been provided, namely, a sensitive means of measure- 
ment. Constancy may be approximated by artificial illumination, 
2 As will be shown later in the paper, neither the influence of surrounding 
field nor of preexposure can be eliminated when the observation is made in 
the dark-room. The influence of these two factors can be eliminated only 
by working in a light-room of constant intensity of illumination, and by using 
a preexposure and a surrounding field of the brightness of the color used 
for the stimulus. INTRODUCTION 
3 
but no artificial source has yet been devised which gives a light 
that approaches average daylight3 sufficiently closely in composi- 
tion to warrant its use in color work. Of the various sources of 
light the Moore Tube comes nearest to doing this, but spectro- 
photometric and colorimetric determinations show that the light 
from it contains an excess of blue4 and, therefore, although it has 
been adopted by various textile concerns for use in color matching, 
its substitution for daylight can scarcely be recommended for the 
more exact requirements of color optics. Ives and Luckiesh5 
attack the problem of producing artificial daylight from another 
side. By their subtraction method they claim to have gotten the 
closest approximation to average daylight yet attained. They 
aim to cut out by absorbing screens the excess of red and yellow 
in artificial light due to the comparatively low temperature of 
artificial illuminants. Tungsten lamps are used by them as the 
source of light, and two kinds of commercial glass approximating 
in their absorptive action cobalt blue and signal green are used 
as screens. In order to correct for the pronounced band of 
yellow-green transmitted by the cobalt blue, a film of gelatine 
dyed with rozazeine is also used. Although according to com- 
parative measurements made by Ives and Luckiesh the light thus 
gotten is the closest approximation to average daylight yet ob- 
tained, still it shows a deficiency of 15% in the green and about 
25% in the blue. Moreover, the spectrum of this light does not 
show the brightness distribution of the spectrum of daylight. 
Since the absorbing screens cut down the light emitted by the 
tungsten lamp to 15% of its original intensity, the spectrum of 
8 For results of measurements of the color values of average daylight, see 
Nichols, E. L. Transactions of the Illuminating Engineering Society, 1908, 
HI, p. 301. Ives, H. E. The Daylight Efficiency of Artificial Illuminants. 
Transactions of the Illuminating Engineering Society, 1909, IV, pp. 434-442; 
Color Measurements of Illuminants. Transactions of the Illuminating En- 
gineering Society, 1910, V, pp. 189-207. 
4 See Ives, H. E. Color Measurements of Illuminants. Transactions of 
the Illuminating Engineering Society, 1910, V, p. 206; and Rosa, E. B., quoted 
by Moore, D. McF. A Standard for Color Values. Transactions of the 
Illuminating Engineering Society, 191°) IV, p. 224. 
B Ives, H. E. and Luckiesh, M. Subtractive Production of Artificial Day- 
light. Electrical World, 1911, LVII., pp. 1092-1094. 4 
GERTRUDE RAND 
the light finally given out shows the brightness distribution char- 
acteristic of lights of low intensity, unless the original light-source 
is of extremely high candle-power. 
We seem thus compelled either to give up the investigation of 
the sensitivity of the retina for daylight illumination, or to devise 
some means of keeping this illumination constant. At an early stage 
in the study of the color phenomena of the peripheral retina begun 
four years ago and still in progress in the Bryn Mawr Labora- 
tory, the writer was compelled to take into account the influence 
of the changes in the illumination of the visual field upon the 
color observation. The changes of illumination that took place 
from day to day, the progressive changes during the day, and 
the many sudden changes even in the course of an hour, rendered 
any constancy, or close reproduction of results entirely out of 
the question. The consideration of this factor led in turn to a 
general study of the conditions that influence the color observa- 
tion. It is the purpose of this paper to report the results of 
that study. The report will take the following form. 
(i) A resume and criticism will be given of previous studies 
of factors, and of attempts to standardize. (2) The color obser- 
vation will be reexamined for the factors that influence its 
results, and a study of these factors will be made with the fol- 
lowing points in view: (a) Their influence will be measured 
under various typical conditions obtaining in the work on color 
sensitivity. (&) Their effect on the limen of color at different 
points in the retina and on the limits of color sensitivity will be 
determined, (c) An explanation based on the conclusions drawn 
from (a) and (&) will be made of the results of other experi- 
menters and of the contradictions found in these results. (3) 
From this study of the influence of the factors, it will be deter- 
mined what factors need to be standardized in the various kinds 
of work on color sensitivity and methods will be devised for 
their standardization. 
In the latter part of the work especial attention will be given 
to the effect of general illumination and of local preexposure. The 
writer finds these to be the two most important factors extraneous 
to the source of light that influence the results of the color obser- INTRODUCTION 
5 
vation, and yet, so far as she is able to determine, up to this time 
no attempt worthy of more than passing consideration has been 
made to standardize either factor in investigations of color sensi- 
tivity. In fact, it can scarcely be said that either has been included 
in the list of factors by any previous writer. The effect of the 
general illumination has received only casual mention by Ole 
Bull and a few others, and the brightness of the preexposure has 
not been clearly recognized as exerting any influence whatever. II. HISTORICAL AND CRITICAL. 
A. FACTORS THAT HAVE BEEN FOUND TO INFLUENCE THE SEN- 
SITIVITY OF THE RETINA TO COLOR. 
1. Size of the Stimulus. 
An increase in the size of the stimulus is generally considered 
to be equivalent in some proportion to an increase in intensity.1 
It is but natural, then, to think that an increase in the size of 
the stimulus would both lower the limen of sensitivity and 
extend the limit of the zone within which a given color can 
be sensed. The question with regard to the limits of sensitivity 
is, however, not so simple as it seems. In the first place, the 
limit of the zone may not be extended, because the retinal sen- 
sitivity may fall off so rapidly at the point worked upon, that 
the increase of stimulation is not sufficient to overweigh the 
loss. In the second place, the effect of the increase of area may 
depend to some extent upon the area of the original stimulus. 
For example, fatigue is set up so easily with very small stimuli 
that an increase up to a certain point is advantageous, while, on 
the other hand, the outer margin of large stimuli may extend 
so far into the zone of relative insensitivity that a further in- 
crease of area becomes ineffective. In the third place, the effect 
may vary with the meridian of the retina investigated. Two 
reasons may be assigned for this variation, (a) We should 
expect the effect to be in some measure proportional to the 
rapidity with which the retina falls off in sensitivity from the 
fovea to the periphery. For example, in the temporal and lower 
1 Raehlmann, E. Ueber Farbenempfindung in den peripherischen 
Netzhautpartien in Bezug auf normale und pathologische Brechungs- 
zustande. Inaug. Diss., Halle, 1872. 
While the work of Raehlmann and others shows in general the truth of the 
above statement, no systematic determination of the exact relation of change 
of area to change of intensity has yet been made. This determination for 
the sensations aroused both by white and colored light is now in progress in 
the Bryn Mawr laboratory. HISTORICAL AND CRITICAL 
7 
meridians, where the sensitivity falls off sharply, we should 
expect little if any effect; while in the nasal and upper regions, 
where the decrease is much more gradual, we should expect 
considerable effect. (&) In the nasal and upper meridians, the 
limits of sensitivity extend much farther toward the periphery 
than in the temporal and lower meridians. There is in these 
meridians, then, as the limits of sensitivity are approached, a 
relatively greater shrinkage in one dimension of the stimulus, 
owing to the greater angle of excentricity, than occurs in the 
temporal and lower regions. In proportion as this shrinkage 
causes a shortening of one dimension of the stimulus, it adds 
to the range of areas over which an increase is of advantage for 
extending the limit of sensitivity. 
A survey of the literature on peripheral vision shows that the 
size of the stimulus was early recognized as one of the factors 
influencing the sensitivity of the peripheral retina. In fact, the 
first investigation of peripheral sensitivity was made to deter- 
mine the effect of the size of the stimulus. This work done by 
Hueck2 in 1840, may be considered as pioneer, for although 
Troxler3 and Purkinje4 had previously mentioned the phenomena 
of peripheral vision, they had made no systematic attempt to 
investigate these phenomena. Hueck's object was primarily to 
study the effect of increase in the size of the stimulus upon the 
limits of the field of vision. Using gray paper stimuli of very 
small area, he observed the effect on the limit of vision (a) 
when their objective size was increased, and (&) when their 
apparent size was altered by a decrease in their distance from 
the observer, that is, by enlargement of the visual angle. He 
found that an increase in size produced in either of these ways 
caused a widening of the field of vision for that quality of 
stimulus. The investigation was also extended to color. Pig- 
1 Hueck, A. Von den Grenzen des Sehensvermogens. Muller's Archiv, 
1840, p. 95. 
3 Troxler, D. Ueber das Verschiwinden gegebener Gegenstande innerhalb 
unseres Gesichtskreises. Ophthal. Bibliothek herausgegeben von Himly u. 
Schmidt, Jena, 1804, I, 2. pp. 1-53- 
''Purkinje, J. Beitrage zur Kenntniss des Sehens. 1823, I, p. 76; 1825, 
II, p. 14. 8 
GERTRUDE RAND 
ment papers were used. This investigation showed that the 
limits of color sensitivity also are influenced by the size of the 
stimulus. An interesting table was compiled which shows that 
by altering either the size of the stimulus or the visual angle 
it subtends, the limit of sensitivity can be made to vary by 
amounts equal to 10 over a wide range of the retina. Hueck's con- 
clusion, that the limits of color vision are influenced by the 
size of stimulus, was confirmed by Aubert5 in 1865. Five 
years later it was contradicted by Woinow.6 Woinow worked 
in the dark-room, using for stimuli colored glasses illuminated 
by a shaft of sunlight of variable extent. He claimed that "die 
Grenze immer dieselbe ist, ohne Riicksicht auf die Grosse der 
Pigmentflache, wenn die Gesichtswinkel nicht von der Mitte 
sondern von dem dem Auge zugekehrten Rande der Pigment- 
flache berechnet werden." No information is given as to the 
size of stimuli employed. Kriikow,7 repeating Woinow's pre- 
caution of measuring the angular distance to the inner edge of 
the stimulus rather than to the middle, confirmed the conclusion 
that the boundaries of the color zones are absolute, within cer- 
tain limits of size of stimulus. His stimuli were 3, 6, and 9 
mm. square. Aubert8 in 1876, repeated the observations recorded 
in his earlier work. Colored squares with sides varying from 1 
mm. to 32 mm. placed at a distance of 20 cm. from the eye were 
used as stimuli. The results he obtained led him to believe that 
the size of the stimulus is a factor in determining the limits of 
sensitivity. He writes: "Die Grosse des farbigen Objectes 
massgebend ist fiir die Entfernung vom Centrum, in welcher 
es noch farbig empfunden wird. Die gegentheilige Behauptung 
Woinow's . . . muss ich nach vielfacher, wiederholter Unter- 
suchung fiir falsch erklaren." He mentions the precaution used by 
Woinow as to the measurement of the angular distance, but does 
not definitely state that he himself took this precaution. Raehl- 
'Aubert, H. Physiologic der Netzhaut. Breslau, 1865, p. 121. 
'Woinow, M. Zur Farbenempfindung. A. f. O., 1870, XVI, p. 219. 
1 Kriikow. Objective Farbenempfindung auf den peripherischen Theilen 
der Nezhaut. A. f. O., 1874, XX,„ pp. 255-296. 
'Aubert, H. Physiologische Optik. Leipzig, 1876, pp. 541-544. HISTORICAL AND CRITICAL 
9 
mann,9 Schon,10 Schirmer,11 and Briesewitz,12 all agree with 
Aubert; but they also have not mentioned their method of meas- 
urement. 
In Tschermak,13 however, we find an investigator who has 
observed Woinow's precaution and yet has obtained results 
that are contradictory to Woinow's. He investigated the factors 
which condition the colorless vision of the peripheral retina, and 
showed that neither the red-green nor the totally color-blind 
zones of the normal retina are invariable in extent. Size of 
stimulus was found by him to be one of the factors that deter- 
mine the breadth of these zones. This he demonstrated on the 
Hering apparatus for investigating the color sensitivity of the 
peripheral retina, an apparatus consisting of a campimeter screen 
with an opening behind which the stimulus is placed. Tscher- 
mak's screen was of gray paper. The size of the stimulus-open- 
ing was regulated by means of two gray slides which widened the 
opening either on the side within the visual field and toward the 
fovea, or on the other side toward the periphery. Using a small 
stimulus-opening, he determined the degree of excentricity at 
which Urgriln and Urroth appeared colorless. He then widened 
the stimulus-opening toward the fovea and found that the 
color was sensed. No conclusion, however, can be drawn from 
this because he had extended the inner margin of the stimulus 
into the region sensitive to color, hence the sensation aroused 
may have been due to that cause rather than to the increase' 
made in the area of the stimulus. He next widened the original 
stimulus-opening toward the periphery. This caused an increase 
in the area of the retina stimulated, without extending the margin 
of the stimulus into the field sensitive to color. Since in this 
case also the color was sensed, Tschermak concludes that the 
8 Raehlmann, E. loc. cit. 
10 Schon, W. Ueber die Grenzen der Farbenempfindung in pathologischen 
Fallen. Klinische Monatsblatter, 1873, p. 171. 
11 Schirmer, R. Ueber erworbene and angeborene Anomalien des Farben- 
sinns. A. f. O., 1873, XIX, p. 194. 
12 Briesewitz. Ueber das Farbensehen bei normalem and atropishem Nervus 
Opticus. Inaug. Diss., Greifswald, 1873. 
13 Tschermak, A. Beobachtungen iiber die relative Farbenblindheit' in 
indirectem Sehen. Pfliiger's Archiv, 1890, LXXXII, pp. 559-560. 10 
GERTRUDE RAND 
limits of color sensitivity are influenced by the area of the stim- 
ulus. It is in the second method of increasing the area of the 
stimulus that Tschermak took the precaution mentioned by 
Woinow relative to the measurement of the angle of excentricity. 
But it is obvious that the method of measurement need not have 
been the only cause of variable results in work of this kind. The 
meridian tested may have been, as we have already suggested, a 
second cause. That certain regions of the retina react differently 
to an increase in the area of the stimulus, is noted in a brief 
paragraph by Kirschmann.14 Using stimuli of 28, 40, and 58 
mm. in diameter, he found that the color sensitivity of the 
peripheral retina is dependent to different degrees in different 
meridians upon the size of the stimulus. In the lower and 
temporal meridians, the zone sensitive to each color was widened 
very slightly by increasing the area of the stimulus, and never 
beyond certain limits. On the upper and nasal parts of the 
retina, however, the possibility of widening the zones by this 
means seemed to be, he says, unlimited. Now we know that 
Woinow and Kriikow obtained their results on the temporal 
meridian. Tschermak, however, does not state what region he 
investigated. If he worked in the nasal region, his conclusions 
may be reconciled with those of Woinow in the light of Kirsch- 
mann's work. These variations in the effect of area in different 
regions of the retina are no doubt due largely to the difference 
in the rapidity with which sensitivity falls off from the center to 
the periphery of the retina along the several meridians. Where 
the decrease is gradual, as is the case in meridians that have wide 
limits of sensitivity, more effect might be expected than where 
the sensitivity decreases rapidly. 
A third cause of the variable results recorded may have been 
the range of size of stimuli employed. The retina fatigues 
easily to very small stimuli; hence an increase in size up to a 
certain point is advantageous. On the other hand, the margins of 
very large stimuli may extend so far into the zone of insensi- 
tivity that a further increase is ineffective. Kriikow no doubt 
u Kirschmann. A. Die Farbenempfindung bei indirectem Sehen. Philos. 
Studien, 1893, VIII, p. 612, 613. HISTORICAL AND CRITICAL 
11 
referred to this fact when he said that color sensation is indepen-1 
dent of the size of the stimulus, but only within certain limits. 
An interesting conclusion reached by Abney15 may also be men- 
tioned in this connection. He wished to determine at what 
intensity the different colored spectral lights were brought below 
the limen of sensation both in central and at various points in 
peripheral vision. He found, however, that the intensity of the 
stimulus is not the only factor to be considered. A stimulus 2 
inches in diameter, for example, was seen at a lesser intensity 
than a stimulus inch in diameter. He further found that it 
is not the area, but the shortest dimension of the stimulus, ver- 
tical or horizontal, which determines the intensity required to 
render the stimulus subliminal. 
The following table has been compiled from the results of 
his work in central vision. 
Stimulus Relative Intensity Value of Light 
disc .95 in. diameter 234 97.4 
square .84 in. x .84 in. 216 
rectangle 1.68 in. x .42 152 495-2 
square -84 in. x .42 154 478.4 
The areas of the disc, the square, and the first rectangle are 
equal, but the rectangle which has the shortest dimension, re- 
quires 400 units more of light intensity than does the disc in 
order to be made just subliminal. The fourth stimulus has half 
the area of the third, but their shortest dimensions are equal, and 
accordingly the same amount of light is required to render them 
both just subliminal. The experiments were extended by Abney 
to the peripheral retina, and the conclusion was again reached 
that the shortest dimension of the stimulus and not its area 
determines the reduction in intensity necessary to render the 
stimulus just subliminal (p. 183). Since Abney found further 
that "there is a simple connection [relation] between the intensity 
of the stimulus color and the extent of the color field," we may 
infer that he would have us conclude that the extent of the color 
field is also influenced by the shortest dimension of the stimulus.' 
The present status of this point may be summarized as follows: 
18 Abney W. de W. The Sensitiveness of the Retina to Light and Colour. 
Philos. Trans., 1897, CXC, Ser. A, pp. 169-171. GERTRUDE RAND 
12 
Within a certain range of dimensions and particularly for certain 
regions of the retina, the size of the stimulus is an important 
factor in determining the extent of the color zones. And with 
regard to size, the shortest dimension of the stimulus and not 
its area is, according to Abney, the determining factor. Care 
must be taken, therefore, to measure accurately the size of the 
stimulus used in peripheral investigation, also its distance from 
the observer, and to keep these measurements uniform through- 
out the investigation. 
2. Intensity and Brightness or White-Value of the Stimulus. 
(a) The confusion that has arisen with regard to the meaning 
of intensity and of brightness, and its effect upon the development 
of methods of working. 
Before we attempt to discuss the influence of the intensity 
and the brightness of the stimulus upon the limits of color sensi- 
tivity, some attention should be given to a definition of terms. 
The need for this will be shown by a brief examination of the 
literature on these subjects. A great deal of confusion as to 
terminology seems to exist, and not a little misinterpretation of 
fact seems traceable to this confusion. The greater part of the 
confusion arises from the use of the word intensity. This term 
has been employed at various times to indicate (a) the energy of 
a beam of spectral light homogeneous as to color; (&) the 
white-value of a color; (c) the saturation of a color; and (J) 
the energy of light-waves reflected from a pigment surface as 
conditioned by the general illumination of the visual field. This 
equivocal use of the term has now and then apparently led to a 
wrong interpretation of results, and this in turn to the modifica- 
tion of experimental technique. An example of this is found in 
the work done by Baird in the Cornell laboratory on "The Color 
Sensitivity of the Peripheral Retina."16 In his review of the 
literature, Baird finds data that lead him to assume that an 
equation of the white-values of the stimuli employed is essential 
for a determination of the relative extent of the retina's sensi- 
tivity to the different colors. Apparently these data are derived 
18 Baird, J. W. The Color Sensitivity of the Peripheral Retina. Carnegie 
Institution of Washington, 1905. HISTORICAL AND CRITICAL 
13 
mainly from three sources: (a) from a study of the color sensi- 
tivity of the peripheral retina made by Aubert; (&) from a study 
by Abney of the effect of changes in the energy or intensity of 
spectral light upon sensation; and (c) from a study of the limits 
of color sensitivity made by Landolt.17 An examination of the 
investigations made by these men shows, however, that Baird's 
conclusion is apparently based upon a loose construction put upon 
the meaning of certain terms. The most striking example of 
this, as we shall see, results from the interpretation given by 
Baird to the term intensity. Baird uses intensity to indicate 
luminosity and, as we shall show, he also uses luminosity inter- 
changeably with brightness or white-value. Landolt and Abney, 
on the other hand, from the results of whose investigations of 
the effect of intensity on color sensitivity Baird largely draws his 
conclusions as to the need of equating the white-value of his 
stimuli, clearly use the term intensity to mean the energy of the 
light-waves coming to the eye. 
According to Baird, the first mention of the need to equate in 
brightness was made by Aubert. Baird writes: "His [Aubert's] 
results may be summarized as follows: 
"1. The brightness of the background has a most pronounced 
influence upon the extension both of the color sensitivity and 
of the brightness sensitivity. 
"2. The extension of the color zones increases with increase of 
area of stimulus. 
"3. The color sensitivity decreases at very different rates upon 
different retinal meridians. 
17 Baird claims to derive authority also from the work of Raehlmann, Klug, 
Chodin, Bull, Hess, and Hegg. We have considered that this authority is 
derived mainly from Aubert, Landolt, and Abney, however, because Baird 
discusses the results of these three men and to some extent their methods 
of working, giving several sentences to show that their work points out the 
need for equation of the white-values of the stimuli employed to investigate 
color limits. To the other men from whom he claims to derive authority, 
Baird devotes merely a sentence to each which states, in case of 
Raehlmann and Klug, that they had found that the color limits vary with 
changing brightness of stimulus; in case of Chodin, that he believed that 
brightness equation was necessary; in case of Bull, Hess, and Hegg, that they 
had equated the white-values of their stimuli. The discussion of these cases 
will be taken up later in the paper (see pp. 51-53). 14 
GERTRUDE RAND 
"4. The transitions of color tone are as follows: Red passes 
through reddish-yellow and yellowish-gray to gray; green be- 
comes yellowish, while yellow and blue undergo no change of 
tone, but decrease in saturation and finally appear gray. 
"5. The relative extension of the color zones can not be deter- 
mined with any degree of accuracy. Since the width of the 
color zone is a function of the luminosity of the stimulus, the 
color-stimuli employed in the determination of comparative reti- 
nal limits must all be equated in brightness. 
"6. There is a close analogy between the functioning of the 
central and peripheral parts of the retina."18 
Baird seems to derive his authority for the need to equate 
in brightness, so far as Aubert is concerned, from the fifth of 
these points of summary. From the wording of the text it is 
impossible to state the exact source of Baird's quotation since 
he bears himself out in his summary only by a general reference 
to a long list of Aubert's articles on vision. 
But the organization of this summary is so closely akin to 
that given in the Physiologische Optik,-the only difference being 
in the omission by Baird of the third item in the Optik summary, 
-that one seems justified in asserting that this work contains 
the source of the statement quoted above. In no other of 
Aubert's articles are all of the points mentioned by Baird touched 
upon. The earlier articles are narrower in scope than the Optik 
and treat of fewer factors. 
Aubert's statement of results in the Physiologische Optik is 
as follows: "Durch meine Versuche wurde festgestellt 
"1. der grosse Einfluss welchen die Umgebung der Pigmente 
auf die Farbenempfindung auch beim indirecten Sehen hat. 
"2. der Umstand, dass die Grosse des farbigen Objectes mass- 
gebend ist fiir die Entfernung vom Centrum, in welcher es noch 
farbig empfunden wird. 
"3. dass Pigmente verschiedener Farbentone unter sonst glei- 
chen Umstanden verschiedene Grenzzonen fiir die Erkennbarkeit 
der Farbe zeigen. 
"4. dass in die verschiedenen Meridianen der Netzhaut die 
18 Baird, J. op. cit., pp. 12-13. HISTORICAL AND CRITICAL 
15 
Grenzzonen fiir die Farben sehr verschieden weit von dem Fixa- 
tionspunkte liegen. 
"5. Schon Purkinje hat verschiedene Uebergange durch Far- 
bentone und Farbennuancen beobachtet, und zwar geht auf 
schwarzem Grunde nach Aubert: Roth durch Rothgelb und Gelb- 
grau zu Grau, Blau durch immer weisslichere Nuancen zu Grau, 
Grim durch Graugelb zu Grau, Gelb durch Graugelb zu Grau. 
"6. Bonders und Landolt haben nachgewiesen, dass die Far- 
benempfindung auf den peripherischen Netzhautzonen eine dem 
Centrum gleiche bleibt, wenn die Intensitat der Beleuchtung 
gesteigert wird: also auch beim indirecten Sehen sind Gesichts- 
winkel und Helligkeit massgebend fiir die Farbenperception. 
"7. dass die peripherischen Theile der Netzhaut fiir die Far- 
benempfindung viel schneller ermiiden, also die centralen."19 
To Baird's fifth point of summary, the closest approximation 
that the writer is able to find anywhere in Aubert's works is 
the sixth conclusion quoted above from the Optik. This is: 
"Bonders and Landolt haben nachgewiesen, dass die Farben- 
empfindung auf die peripherischen Netzhautzonen eine dem Cen- 
trum gleiche bleibt, wenn die Intensitat der Beleuchtung gesteigert 
wird: also auch beim indirecten Sehen sind Gesichtswinkel und 
Helligkeit massgebend fur die Farbenperception. Nagel bestatigt 
Landolt's Angabe."20 The question here is whether Helligkeit in 
the above quotation means brightness of stimulus which, if our 
assumption with regard to the source of his authority is correct, 
Baird has apparently interpreted it to mean. The following points 
may be cited to show that such an interpretation is very strongly 
open to question, (a) Aubert himself does not use Helligkeit 
in connection with a qualifying phrase, for example, Helligkeit 
der Farben or its equivalent, while so far as the writer is able 
to determine, he never uses Helligkeit as referring to the bright- 
ness of color without the qualifying phrase. For example, in 
his Physiologische Optik p. 527 he uses Helligkeit der Farben; 
p. 528, Helligkeiten der verschiedenen Abtheilungen des Spec- 
trums; p. 529, Helligkeiten der Farbentone des Sonnenspectrums; 
** Aubert, H. Physiologische Optik, Leipzig, 1876, pp. 541-545. 
30 Aubert, H. Physiologische Optik, Leipzig, 1876, p. 545. 16 
GERTRUDE RAND 
same page, Helligkeiten der Farben; p. 530, Helligkeiten der 
Abtheilungen des Spectrums. In his Physiologic der Netzhaut p. 
109 he uses, Helligkeit der Pigmente. The same expression is 
used again on pp. m and 112. (b) When using Helligkeit 
without a 'qualifying phrase, Aubert commonly refers to the 
intensity or brightness of the general illumination. For examples 
of this usage, see Physiologic der Netzhaut, pp. 109, no, 124; 
Physiologische Optik, p. 532, and other places, (c) No evidence 
can be obtained from Landolt from whom the citation is made 
that brightness of color is referred to. In fact the evidence is 
strongly against this interpretation. As will be shown in detail 
(pp. 24-25), Landolt worked with colors of varying energy. 
He used as stimuli very intense spectral light and pigment papers. 
The energy of the former was varied directly, of the latter by 
changing the general illumination which altered the amount of 
colored light reflected to the eye. In the general statement of 
his problem Landolt is not concerned with the effect of the 
brightness of his colors; nor are his results couched in terms of 
the effect of the brightness of colors. His sole interest was to 
find the effect on sensation of using colors of great energy or 
intensity. In the above quotation from Aubert, then, we may 
conclude that it is strongly open to question whether Helligkeit 
is not also used here in the sense in which we have shown that 
Aubert most frequently uses the term: the brightness of the 
general illumination (see (&) above). If so, what he really 
does claim in this statement, therefore, is that when one is 
working with pigment colors, the degree or brightness of the 
general illumination with all of its influences, namely, its effect 
on the intensity of color, on the brightness of color, on the bright- 
ness of the surrounding field, the preexposure, etc., is one of the 
factors that determine the extent of the color field; not simply 
one of these influences, the brightness of color, as he is interpreted 
to claim by Baird.21 Moreover, the writer is compelled to say 
that in a careful reading of all the articles by Aubert contained 
" For Aubert's own statement of his opinion on the question of the influence 
upon color sensitivity exerted by the brightness of the stimulus see this 
article, pp. 40-44. HISTORICAL AND CRITICAL 
17 
in the long list to which Baird refers, she is unable to find a 
single statement that would justify the conclusion that Baird has 
drawn in his fifth point of summary. 
Towards Landolt and Abney, Baird takes a slightly different 
attitude. He does not hold that they have mentioned a need to 
equate in brightness. In their results, however, he finds justifi- 
cation for equating the white-values of his own stimuli from the 
construction he puts upon their use of the words luminosity and 
intensity. Abney,22 a physicist, had defined luminosity as equiva- 
lent to intensity. In his experiments he had increased or dimin- 
ished the luminosity or, as he also says, the intensity of a beam 
of light by interposing in its path a wedge graduated in thickness. 
The wedge was made of gelatine in which were scattered black 
opaque particles. The energy of the beam of light was dimin- 
ished by amounts depending on the thickness of that part of 
the wedge through which it was made to pass. This "obstruction 
method" resulted not only in a decrease of the energy of the 
light, but in a darkening of the stimulus. But Abney was not 
at all concerned with the effect of the lightness or darkness aspect 
of the stimulus:-in other words, with the relative inhibitive 
action of white and black and its influence on the limits of color 
sensitivity. His purpose was to vary the energy of the light- 
waves coming to the eye by "obstructing" them by known 
amounts, and to ascertain the effect of this change upon the 
color limits. It is true that the lightness or darkness of the 
sensation quality was altered incidentally, but he apparently had 
no thought of saying that this was in any sense responsible for 
the effect obtained, nor is it a necessary inference from his 
work. Hence Baird is not justified in stating (p. 31) that Abney 
makes the brightness of the stimulus a factor in determining the 
color limit, if we take our clue as to what Baird means by 
brightness from the following passage. "No determination of 
the relative extensions of the various color zones can ever yield 
comparative results unless it be accomplished by means of stimuli 
of equal brightness, or, more correctly speaking, of equal white- 
value" (see p. 37). For Abney assuredly, does not mean by lum- 
22 Abney, W. op. cit., 155-195- 18 
GERTRUDE RAND 
inosity what Baird calls brightness or white-value. Baird falls 
into a similar error in his treatment of Landolt's results. He 
says: "An important feature of Landolt's paper is his insistence 
that no investigation of color vision is complete unless it takes 
into account the relative luminosity of the stimuli employed" 
(p. 17). Now when we read in a footnote (see Baird p. 34) that 
"Under brightness [of stimulus] is included both absolute and 
relative luminosity of stimulus, i.e., its own brightness and its 
brightness contrast with its background," we see that the relative 
luminosity referred to by Landolt means for Baird relative 
brightness or white-value, and that the work of Landolt is 
brought forward as as evidence for the necessity of equating the 
white values of the stimuli. Now Landolt worked with both 
spectral and pigment stimuli. In case of the first, his method was 
to increase the energy of spectral light; and in case of the 
second, to increase the amount of light coming to the eye from 
a pigment surface by increasing the general illumination of the 
room. Here, as in the case of Abney, we have a change in the 
amount or energy of the colored light coming to the eye, and 
incidentally a change in the white-value of the sensation aroused. 
No separation is made, however, of the two factors: (a) the 
altered energy of the colored light, and (&) the change in the 
white-value of the sensation aroused. Yet Baird finds reason to 
conclude from Landolt's results that the white-value of the color 
influences the limits of sensitivity. It is obvious that Landolt's 
results do not show this at all. All that they do show is that 
when the amount of colored light given to the eye is increased or 
decreased, the extent of the zones of sensitivity is altered.23 
Whether or not the white-value of the stimulus can be con- 
sidered to any degree responsible for changes in the color limits 
" These three cases taken from Baird's discussion of the work of Aubert, 
Abney, and Landolt are examples of the cases referred to earlier in the chapter 
in which a confusion as to terminology has led to a wrong interpretation of 
results which in turn has been the cause of changes in technique. Baird 
was led by this confusion, in part intrinsic and in part due to his own 
misinterpretations, to think that these three men considered that the white- 
value of the colored stimulus affects the extent of the retina's sensitivity to 
it, and was influenced thereby to equate the white-values of his stimuli 
without further investigation. HISTORICAL AND CRITICAL 
19 
will be considered by the writer in the experimental section of 
this paper. What we wish to point out here is that without the 
isolation and the separate investigation of this factor, Baird 
concludes from the work of previous investigators in which inten- 
sity or energy changes have been made in the stimulus, that the 
white-value of the stimulus influences the boundaries of the color 
zones and that, therefore, stimuli should be equated in white- 
value in all work on the limits of color sensitivity. He says: 
"It has been been established in hosts of instances24 that change 
of luminosity is, within limits, invariably attended by a corres- 
ponding change in the extension of the retinal zone within which 
the color of the stimulus is recognized. Its significance for the 
problem is self-evident. No determination of the relative exten- 
sion of the various color zones can ever yield really comparative 
results unless it be accomplished by means of stimuli of equal 
brighness, or, speaking more correctly, of equal white-value" (p. 
37). While we grant the significance of changes of luminosity 
or intensity in the sense in which Abney and Landolt use the 
terms, we do not admit that this aspect of the stimulus can be 
standarized in terms of white-value; nor do we grant that any 
definite evidence whatsoever can be gathered from the results we 
have quoted above, to show that changes in the white-value of a 
stimulus affect the limits of the retina's sensitivity to its color, 
provided the amount of colored light coming to the eye remains 
unaltered. 
To sum up: (a) The white-value of a stimulus may be varied 
without altering the amount of colored light coming to the eye. 
This factor, then, may be isolated and its effect on the limits of 
sensitivity determined apart from any change in the physical 
intensity of the stimulus. (&) Unless this separation is made, we 
have no right to conclude that the white-value of the stimulus 
affects the limits of sensitivity. Baird, for example, drew this 
conclusion from work in which the separation was not made, 
(c) The confusion that exists with regard to color terminology 
"Beside the three instances mentioned here, Baird cites in support of 
his position the work of Raehlmann, Chodin, Klug, Bull, Hess, and Hegg. 
How far their work can justly be cited in support of his position will be 
shown on pp. 51-53- GERTRUDE RAND 
20 
has been, we believe, in no small measure responsible for Baird's 
conclusion. 
The terminology which we propose to use in this report may be 
outlined as follows: Intensity of stimulus will be used to indicate 
the energy of light-waves coming to the eye. Intensity of sen- 
sation, or apparent intensity, will be used as its correlative sub- 
jective term. So used, it will signify merely energy or volumin- 
ousness of sensation and will have no reference whatever to the 
white-value of a color. Saturation of the stimulus will be used 
to indicate the proportion of colored to colorless light coming to 
the eye. Saturation of color or saturation of the sensation will 
be used as its correlative term and will refer to the proportion of 
chromatic to achromatic quality in the sensation. The achro- 
matic sensations will be designated by the terms white, black, and 
gray; and the terms brightness and white-value will be used inter- 
changeably to indicate the lightness or darkness of a color. 
(&) The effect of intensity of stimulus. 
A dependence of color sensation upon the intensity of the 
stimulus has been recognized since the observations of Purkinje. 
Purkinje noted also that a color stimulus gave a less intense 
sensation in the peripheral retina than in the central retina. Since 
that time, it has been claimed (a) that with stimuli of minimal 
intensity, no color sensation is aroused; (&) that the light-waves 
arousing the different monochromatic sensations must be of dif- 
ferent intensities to give liminal color sensations; that is, the eye 
is not equally sensitive to waves of different lengths; (c) that, 
progressively, greater intensity of stimulus is required to give 
sensation as the stimulus is moved from the fovea to the periph- 
ery; {d} that the extent of the color fields is determined within 
certain limits by the intensity of the stimulus. 
The influence of. changes in the intensity of the stimulus upon 
the sensation of color has been investigated by three methods: 
(i) by determining the effect upon the limens of sensitivity; (2) 
by determining the effect upon the j.n.d. of sensitivity; (3) by 
determining the effect upon the limits of sensitivity. 
The effect upon the limens of sensitivity. When working by 
this method, the investigator started with a stimulus that was HISTORICAL AND CRITICAL 
21 
below the threshold of sensitivity, and increased its intensity 
until the sensation of color was just noticeable. This increase in 
intensity was accomplished in three ways: (a) by increasing the 
illumination of a pigment surface, and consequently the amount 
of colored light reflected to the eye; (&) by increasing the inten- 
sity of the light used to give a spectrum; (c) by increasing the 
proportion of color in a mixture of colored and gray pigment 
stimuli. 
The first method of increasing the intensity of the stimulus was 
used in central vision by Purkinje25 and Aubert.26 Purkinje 
observed a representation of the spectrum in pigment colors while 
daylight advanced. He found that blue was the first color to be 
seen in its true color tone, green next, yellow next, and red last. 
He made no measurements, however, of the amount of light 
required to give a just noticeable sensation. Aubert illuminated 
a pigment surface 10 mm. square by daylight admitted into a 
dark-room through an adjustable opening in a window. He 
found that with an opening ^4, or 1 cm. square, no sensation 
of color was obtained. His results are given in the following 
table. 
Opening in window 
Stimulus 
Sensation 
%-%-! cm' 
all 
no color 
1%-!^ 
orange 
red 
2 
O,Y,R, rose 
O',Y,R, rose 
3 
blue 
blue 
3 
light green 
brown 
3^ 
light green 
light green 
5 
green 
blue 
8 
green 
green 
Aubert found that the eye was most sensitive in order to orange, 
red and yellow, blue, and least sensitive to green. 
The second method was used by Raehlmann and Butz. They 
both used the Bunsen spectroscopic apparatus, which provides for 
changes in the intensity of the stimulus by means of the Nichol's 
prism. Raehlmann27 determined the limens of sensitivity to the 
23 Purkinje, J. op. cit., 1825, II, p. 109. 
26 Aubert, H. Untersuchungen liber die Sinnesthatigkeit der Netzhaut. 
Pogg. Annal.,. 1862, CXV, pp. 87-116. 
27 Raehlmann, E. Ueber Schwellenwerte der verschiedenen Spectralfarben 
an verschiedenen Stellen der Netzhaut. A. f. O., 1874, XX(1 pp. 232-254. 22 
GERTRUDE RAND 
different spectral colors at the center of the retina, and at 30° 
and 6o° in the horizontal nasal meridian. He found that the 
center was most sensitive in order to green, yellow, blue, violet, 
and red; and the periphery to yellow, blue, green, violet, and red. 
Butz's28 procedure was as follows: He first determined the lim- 
inal value for each of his colors at the center. Starting with 
this value as unit, he determined how much this value had to be 
altered to give liminal sensation at 300 and at 6o° in the horizon- 
tal nasal meridian. He found (a) that the sensitivity to each 
color increases from o° to 300 and decreases from 300 to 6o° ; 
(&) that the amount of increase in sensitivity from o° to 300 
and the amount of decrease from 300 to 6o° is different for the 
different colors, that is, the ratio of liminal sensitivity to any two 
colors is not the same from center to periphery; and (<?) that 
the amount of the increase is greatest, and of the decrease is less 
in order for violet, yellow, blue, green, and red. 
The third method was used by Aubert and Chodin, both of 
whom employed the Masson disc to find the limen of color sen- 
sitivity. Aubert29 found that at the fovea, the eye is more sen- 
sitive to orange and yellow , than to red and blue.. Chodin30 
found the sensitivity in the central retina to be greatest in order 
for orange, yellow, green, and least for blue. In the periphery, 
he found that the retina is more sensitive to blue and yellow 
than to red and green. 
(2) The effect upon the j. n. d. of sensitivity. The effect of 
intensity upon the j. n. d. of sensitivity has been investigated by 
Lamansky and Dobrowolsky. Lamansky31 used polarized spec- 
tral light, and worked in central vision. He found (a) that the 
j. n. cl. of intensity for the different colors increases or, in other 
words, the sensitivity decreases, as the intensity of the stimulus is 
28 Butz, R. Vorlaufige Mittheilungen uber Untersuchungen der physiolo- 
gischen Functionen der Peripherie der Netzhaut. Archiv fur Anatomie 
und Physiologic, 1881, pp. 437-445. 
29 Aubert, H., Physiologic der Netzhaut. p. 136. 
30 Chodin, A. Ueber die Empfindlichkeit fiir Farben in der Peripherie der 
Netzhaut. A. f. O., 1877, XXIII,2, pp. 177-208. 
31 Lamansky, S. Ueber die Granzen der Empfindlichkeit des Auges fiir 
Spectralfarben. Pogg. Annal., 1871, CXLIII., pp. 633-643. HISTORICAL AND CRITICAL 
23 
decreased, and (b) that the j. n. d. of intensity is smallest, or the 
sensitivity is greatest in order for yellow and green, blue, and red. 
Dobrowolsky32 in 1876 worked with the colors of the spectrum 
in central vision and at various points in peripheral vision. Em- 
ploying standard and comparison fields, he altered the intensity 
of the comparison by rotating a Nichol's prism before the light- 
source until its intensity was just noticeably different from that of 
the standard. Considering that sensitivity varies inversely as the 
magnitude of the j. n. d., he found (a) that the sensitivity to the 
different colors decreases with increase of excentricity; and (&) 
that the comparative sensitivity for the different colors is the same 
in the center and in the periphery, that is, the order of sensitivity 
from greatest to least is in each case, blue, green, red. Later in 
188133 he worked at seventeen different points in the intensity 
scale. He found again that the j. n. d. for blue is smallest, for 
green next, and for red largest. In addition he found that the 
ratio of sensitivity between any two colors as measured by the 
j. n. d. is not the same for different points in the intensity scale. 
In regard to the methods used in the investigations reported 
above, it may be noted that in no one of the cases have the inten- 
sities of the stimuli used been measured and standardized. Tests 
of the comparative sensitivity of different parts of the retina to 
the same color and to different colors may be made with propriety 
by such methods, but not of a given part of the retina to the dif- 
ferent colors. Conclusions can not, then, be drawn by the preced- 
ing investigators with regard to the comparative sensitivity either 
of the center or of the periphery of the retina to the different col- 
ors. They can, however, show that the center has not the same 
comparative sensitivity to the different colors that the periphery 
has. This conclusion may in fact be drawn from the results of all 
except Dobrowolsky. Dobrowolsky's results alone show that the 
center and periphery have the same relative sensitivity for all of 
32 Dorowolsky, W. Ueber die Empfindlichkeit des Auges gegen die Lichtin- 
tensitat der Farben im Centrum und auf der Peripherie der Netzhaut. 
Pfliiger's Archiv., 1876, XII, pp. 441-471. 
33 Dobrowolsky, W. Ueber die Veranderung der Empfindlichkeit des Augen 
gegen Spectralfarben bei wechselnder Lichtstairke dierselben. Pfliiger's' 
Archiv., 1881, pp. 189-202. 24 
GERTRUDE RAND 
the colors with which he has worked. A fair test of the compara- 
tive sensitivity of the eye to the different colors demands either 
that stimuli of equal energy be used or that the sensitivity be 
estimated in terms of units that can be compared. So far as the 
writer knows, no test of the sensitivity of the retina to color has 
ever been made with stimuli representing equal amounts of energy. 
Langley (1889) worked with stimuli of equal energy, but his test 
was for visual acuity.34 Until stimuli of equal energy are used, 
it will remain an open question whether or not the retina, either in 
the center or in the periphery, possesses a different degree of sen- 
sitivity to each of the colors. 
(5) The effect upon the limits of sensitivity. That change in 
the intensity of the stimulus has an effect upon the limits of sen- 
sitivity, has been shown by Abney35 and others. Abney carried 
on an elaborate series of experiments with spectral light to show 
the effect of changes of intensity upon the extent of the color 
fields. He decreased the intensity of the stimulus by placing be- 
fore it a gelatine wedge in the form of an annulus or ring. This 
annular wedge was one inch broad. It was graduated in thickness 
and its transparency was further regulated by black opaque parti- 
cles which had been mixed with the gelatine in its semi-fluid state. 
The value of light admitted at O or at the thinnest part of the ring, 
was 10,000 units;36 that admitted at 360° was 8 units. By in- 
terposing the annular wedge in a plane perpendicular to the path 
of the light and producing the proper amount of rotation, the 
intensity of the stimulus was reduced by graded amounts. Abney 
concluded that there is a simple relation between the intensity of 
the stimulus and the size of the color field. 
The extreme position with regard to the effect of intensity upon 
the extent of the color field is taken by Landolt in the following 
passage. "In ein absolut dunkles Zimmer fiel nur durch eine 
kleine Offung im Fensterladen directes Sonnenlicht. Dieses wurde 
auf das ausserste Ende des Perimeterbogens gelenkt. Wahrend 
wir unser Auge ins Centrum des Bogens setzen, bracht man in die 
34 For discussion of Langley's work, see this paper p. 71. 
35 Abney, W. op. cit., pp. 155-195. 
36 No statement of the value of these units is made by Abney. HISTORICAL AND CRITICAL 
25 
kleine, intensive beleuchtete Stelle farbige Papiere von moglich- 
ster Intensitat der Farbung. Nun bewegtet sich das Auge lang- 
sam vom entgegengesetzen Ende des Bogens nach Scheitelpunkte 
zu und es zeigte sich dabei, dass wenigstens mit der innern Netz- 
hautpartie alle Farben schon bei 900 erkannt wurden. Die 
Grosse des Objectes betrug weniger als 1 cm2. 
"Ais dieselben Priifungen auch mit Spectralfarben zu machen, 
entwarfen wir ein Sonnenspectrum im sonst dunkeln Zimmer und 
liessen es durch eine achromatische Linse auf einen Ende des 
Perimeters befindlichen Schirm fallen. Dieser hatte eine verand- 
liche Spalte, mittelst welcher man die einzelnen Farben aus dem 
Spectrum isolieren konnte. Wahrend wir nun wiederum nach 
langer Adaptation, und bei verbundenem zweiten Auge das eine 
Ende des Bogens fixierten, wiirde von einem Assistenten irgend- 
eine Farbe des Spectrums auf die Spalte gelenkt, und wir drehten 
nun, unter stehter Fixation unserer Fingerspitze, welche sich auf 
dem Bogen bewegte, das Auge allmahlig der Farbe entgegen. 
Es zeigte sich auch hier wiederum dass alle Farbe schon bei 900 
erkannt werden, wenn sie intensiv genug sind."37 
The first to recognize the need for making any sort of intensity 
equation of the stimuli used to investigate the relative sensitivity 
of the retina to the different colors was Ole Bull.38 His purpose 
was to find an accurate method for investigating and measuring 
the sensitivity of the retina to the different colors. The first 
essential condition of the method, he considered, must be to equate 
the colors in saturation and brightness. Briefly, his method of 
equating in saturation consisted of using complementary colors 
of such relative intensity that they cancelled each other in a 1:1 
37 Landolt und Snellen. Ophthalmometroldgie. Handbuch der ges. Augen- 
heilk. von Graefe und Saemische, 1874, III., p. 70. The above quotation is 
given in full in the original in order to confirm (a) the statement made p. 16 
concerning the interpretation of Aubert's statement of Landolt's results, and 
(b) the writer's interpretation, as opposed to Baird's, of Landolt's results, 
stated p. 18. 
A brief summary of the above work is given by Landolt in Klinische 
Monatsblatter ffir Augenheilkunde, 1873, XL, pp. 376-3771 and in Annales 
d'Oculistique, 1874, LXXL, pp.44-46. 
38 Bull, O. Studien fiber Lichtsinn und Farbensinn. A. f. O. 1881, XXVII, 
pp. 54-154. 26 
GERTRUDE RAND 
ratio; that is, he used colors whose color-cancelling or color- 
quenching power was equal. It will be shown later in the paper 
(pp. 63-69), that this method is an anomoly, and, so far as the 
writer knows, is not justified in any investigation of color sensi- 
tivity that has yet been proposed. It certainly does not warrant 
conclusions concerning the relative limits of color nor the relative 
sensitivity of the retina to the different colors. Using this method 
of equating Bull concludes, however, that the retina, central and 
peripheral, is most sensitive to blue, then to yellow, then to red 
and green. 
The second to recognize this essential condition was Hess39 
who made an exhaustive "Priifung des Farbensinnes auf der peri- 
pherischen Netzhaut." Hess's object was primarily to furnish 
Hering with experimental evidence that would enable him to 
refute the Young-Helmholtz theory as modified by Fick to ex- 
plain color-blindness. Young's view that congenital color- 
blindness is due to the absence of one of the three kinds of nerve 
fibres conceived by him to exist in the retina was adopted by 
Maxwell and Helmholtz, and extended by the latter40 to explain 
the so-called peripheral color-blindness of the normal eye. Helm- 
holtz believed that the peripheral retina is red-blind, and that this 
fact could be explained by assuming the absence of the red- 
sensing fibre. In 1873 Fick41 and Leber42 independently pointed 
out that this explanation of the peripheral color-blindness is 
inconsistent with the fundamental assumptions of the Young- 
Helmholtz theory. Fick declared that according to this theory the 
sensation of white can be aroused only by the stimulation of all 
three fibres in equal amounts, or to express it in another way, in 
balanced proportions. Now, if one fibre were inactive in the 
39 Hess, C. Ueber den Farbensinn bei indirectem Sehen. A. f. O., 1889, 
XXXV, pp. 1-62. 
10 Helmholtz, H. Handbuch der physiologischen Optik. 1st ed.,1867, pp. 
301, 845. 
41 Fick, A. Zur Theorie der Farbenblindheit. Arbeiten aus dem physiol. 
Laborat. der Wiirzburger Hochschule, pp. 213-217. 
43 Leber, T. Ueber die Theorie des Farbenblindheit und liber die Art und 
Weise, wie gewisse, der Untersuchung von Farbenblinden entnommene 
Einwande gegen die Young-Helmholtz'sche Theorie such mit derselben 
vereinigen lassen. Klin. Monatsblatter f. Augenheilk., 1873, XI, pp. 467-473. HISTORICAL AND CRITICAL 
27 
peripheral retina, the sensation of white could never be 
produced in that part of the retina; but instead, the sensation 
proper to the combined action of the other two fibres would 
be aroused by white light. If, for example, in the middle 
region of the retina, the red-sensing fibre, and in the 
more peripheral regions both the red- and the green-sensing 
fibres were lacking, the sensation produced by white light in the 
former case would be blue-green, in the latter blue.43 Further, 
if one or more fibres were absent in the outer zones of the retina, 
all the color sensations in these regions would be more saturated 
than those in the more central regions. For when all three fibres 
are present, as in the central retina, red, green, or violet light, 
for example, will stimulate not only its own proper fibre strongly, 
but will also stimulate the other two weakly. A certain amount 
of this total stimulation will be in the proportion to give the 
sensation of white, or more properly speaking, the sensation of 
gray, and the effect of this colorless component will be to reduce 
the saturation of the color sensation aroused. If, then, in the 
peripheral retina only one or two fibres are present, the colorless 
component will not be present to reduce the saturation of the color 
sensation, hence, other things being equal, the sensation of color 
should be more saturated here than that given in the central retina. 
Fick proposed the following modification of the theory to account 
for the color phenomena of the peripheral retina. He assumed 
that from the middle toward the periphery of the retina the 
relative excitability of the three nerve fibres to lights of the 
various wave-lengths constantly alters in such a way that at a 
certain distance from the fovea, namely, in the zone called by 
Helmholtz red-blind, the red-sensing fibres possess the same excit- 
ability as the green-sensing fibres toward lights of all wave-lengths; 
"While in the opinion of the writer, Fick is correct in saying that white 
could not be produced in the extreme periphery by the action of two of the 
retinal fibres he is not right in saying what would be produced. For by a 
literal interpretation of the curves of excitation drawn by Helmholtz to 
represent his theory, none of the color experiences can be produced by the 
action of two fibres, excepting those in a small region of the spectrum in the 
violet. All other color experiences are produced by the combined excitation 
of all three fibres in some proportion. 28 
GERTRUDE RAND 
and that further toward the extreme periphery, all difference be- 
tween the relative excitability of the three fibres diminishes and 
finally disappears. In the red-blind zone, then, the intensity- 
curves for the red- and green-sensing fibres coincide, and in the 
totally color-blind zone, the curves for all three coincide. Curves 
drawn in accord with these assumptions will, it is contended by 
Fick, explain the types of color-blindness found in the peripheral 
retina without violating any of the fundamental principles of 
the Young-Helmholtz theory. Helmholtz accepts the essential 
points of this modification and incorporates them in his theory in 
his later edition of the Physiologische Optik.44 
In order to disprove Fick's assumption that the relative in- 
tensity of response of the three fibres varies from center to peri- 
phery of the retina, and thus discredit the Helmholtz theory, 
Hess advanced three lines of argument. These arguments are 
as follows. In the first place he claims that there are three 
colors of the spectrum, a yellow, a green, and a blue, and a mixed 
color, a bluish-red, which are all invariable in tone from the 
center to the periphery of the retina.45 In the second place, he 
shows that the proportions in which the complementary colors 
combine to produce white do not change for the different parts 
of the extramacular retina. And in the third place he attempts 
to show that a constant ratio of sensitivity to the members of 
each pair of complementary colors obtains throughout the retina. 
In order to make his third point he attempted to obtain red and 
green stimuli such that the red Valenz of the one, or, as he defines 
Valenz, its capacity to arouse red sensation,46 should equal the 
44 Helmholtz, H. Handbuch der Physiologischen Optik. 2nd ed., 1896, p. 373. 
45 For a refutation of this point see footnote, p. 85. 
40 Hess defines the Valenz of his stimuli as their power to arouse color 
sensation. The writer would question this use of the term. According to its 
accepted chemical usage, the term might be applied with some degree of 
propriety to the power which, in terms of the Hering theory, a color 
possesses to combine with or cancel its complementary color, but scarcely 
to its power to arouse sensation. Hess, we presume, applies this term to the 
power of a color to arouse sensation because he assumes that this power is 
the same or at least equivalent to its power to cancel the complementary 
color. But since, as we shall show later (see p. 65), this assumption is far 
from correct, we strongly question the propriety of calling the power of a 
color to arouse sensation its Valenz. HISTORICAL AND CRITICAL 
29 
green Valenz of the other. He describes his method of doing 
this as follows: "Ais das Nachstliegende erscheint es nun, einen 
roth- und einen griin-wirkenden Pigmente den gleichen Roth- 
und Griinwerth dann zuzuschreiben wenn dieselben zu gleiche 
Theilen, z.B. auf dem Kreisel gemischt eine farblose Mischung 
geben" (p. 39). The same procedure was used in obtaining his 
blue and yellow stimuli. Using as stimuli, then, a red equal in 
cancelling power to a green, and a blue to a yellow, he determines 
the limits of sensitivity to- these four colors. He finds that the 
limits for his red stimulus coincide with the limits for his green, 
and the limits for his blue with the limits for his yellow. The 
limits for blue and yellow, however, fall further out from the 
fovea than they do for red and green. From these results he 
concludes (a) the sensitivity of the retina from center to periph- 
ery falls off with the same rapidity for red as for green, and for 
blue as for yellow: and (&) the sensitivity for both red and green 
falls off more rapidly than for blue and yellow. Among his 
conclusions one also finds: "Bei den von uns mitgetheilten Unter- 
suchungen ist die Priifung des Farbensinnes auf der peripheren 
Netzhaut zum ersten Male mit genauer Beriicksichtigung aller 
jener Bedingungen vorgenommen worden, welche unverlasslich 
sind wenn die mit verschiedenen Farben gewonnenen Resultate 
untereinander vergleichen werden sollen" (p. 56). These con- 
clusions are open to the following criticisms: (1) His results do 
not warrant the statement that the sensitivity for red falls off as 
rapidly as for green, and for blue as for yellow. For this state- 
ment is based on the assumption that if in passing from center 
to periphery, sensitivity ends at the same point on the retina for 
two stimuli which have equal power to arouse sensation at the 
center, they must still have equal power to arouse sensation at 
the periphery, that is, sensitivity has fallen off as rapidly for 
one as it has for the other. Now in the first place this assump- 
tion begins with a fallacy. For his stimuli were not equated in 
power to arouse sensation but in cancelling power, and, as we 
shall show later, the power of a color to arouse sensation and 
its power to cancel its complementary color are not at all equiva- 
lent. And in the second place the assumption is itself incorrect, 30 
GERTRUDE RAND 
for, because of the abrupt decrease in sensitivity as the limits are 
approached, the relative sensitivity to the two colors may have 
changed greatly and still their limits have coincided. We have 
found for example that working with colors of normal satura- 
tion under good illumination, it takes, varying with the color, 
a difference of 900 to 1200 of color to make a difference of 
i° in the limits. Hess should have determined the limens of 
color at various points from the center to the periphery of the 
retina, and have found out whether the ratios of the liminal 
values of his pairs of stimuli were equal at all of these points.47 
If so, the sensitivity to each member of the pair must have fallen 
off with equal rapidity from point to point, otherwise a change of 
ratio would have occurred. This method would have had the 
following advantages, (a) Account would have been taken of 
sensitivity at a large number of points from fovea to limits. (&) 
Much smaller changes in sensitivity would affect the limens than 
would affect the limits, especially until very near the limits, 
(c) No equation of stimuli with its attendant disadvantages would 
have been needed as long as the liminal values for each color 
were obtained in terms of the same stimulus all the way out. 
(2) But even if an equation made in terms of cancelling power 
were the equivalent of an equation made in terms of the power 
to arouse sensation, he would not have been justified in conclud- 
ing, so far as his method of working is concerned, that the red- 
green sense decreases more rapidly than the blue-yellow sense, 
because this method afforded him no means of equating the 
intensities of the members of one of the pairs with those of the 
other pair. (3) Nor is he justified in his claim that he is the 
first to investigate the color sensitivity of the peripheral retina 
who has paid due regard to all the conditions which are essential 
if the results obtained for this sensitivity are to be compared with 
one another. One scarcely knows where to begin to refute a 
43 While in the opinion of the writer, Fick is correct in saying that white 
just noticeable difference of sensation could have been determined at each of 
these points in the retina for various points in the intensity scale. An 
account could thus have been had of the retina's sensitivity to the members 
of the pairs of colors at as many degrees of intensity of sensation as was 
desired. HISTORICAL AND CRITICAL 
31 
statement so broadly overdrawn as this. The inadequacy of his 
treatment of the factors: general illumination, brightness of the 
stimulus, brightness of the surrounding field and preexposure, is 
obvious to anyone who knows the effect of these factors. This 
inadequacy, however, will be noted at various other points in 
the paper. It will be sufficient at this point to consider only his 
handling of the factor, intensity, and that only briefly, for it will 
also be discussed in some detail at another point in the paper. 
Hess's problem was narrow and happens to furnish one of the 
very rare cases in which a subjective equation of the intensity of 
the stimuli used is justified. But his conclusion with regard to 
his method of handling the intensity factor is broad and specific- 
ally refers to all investigations of sensitivity in which results are 
to be compared. Such a conclusion is most assuredly not justi- 
fied. In fact one might almost say that the converse of his 
conclusion is true. In no investigation of the comparative sensi- 
tivity of the retina to the different colors where absolute values 
are wanted is a subjective equation permissible. Such an equation 
begs the question at the outset. In such an investigation when 
an equation is needed, it should be made in terms of a common 
objective unit, for example, the unit of energy, and when an 
equation is not needed, the sensitivity should be estimated and 
expressed in terms of this common objective unit, or some other 
unit in terms of which results can be compared. In no case, so 
far as the writer is at present able to outline the field, is a sub- 
jective equation justified in an investigation of sensitivity except 
in certain problems relating to existing color theories or assump- 
tions made for systematic purposes. And not even in these 
problems, so far as the writer is familiar with them, is an equa- 
tion made in terms of cancelling power justified in an investiga- 
tion of sensitivity. 
Hegg48 was the third investigator to attempt intensity equa- 
tion. He posits three conditions to be fulfilled "fur die Unter- 
suchung des peripheren Farbensinnes." (a) The colors must be 
physiologically pure, that is, each color must be sensed similarly 
in all parts of the retina sensitive to it. (&) The colors must be 
48Hegg, E. Zur Farbenperimetrie. A.f.O., 1892. XXXVIII., pp. 145-168. 32 
GERTRUDE RAND 
of equal value in regard to brightness, (c) They must be equal 
in regard to color content (farbigen Gehalt). Concerning the 
third condition, Hegg writes: "Wenn es sich nun, was von 
theoretischen und praktischen Gesichtspunkten aus betrachtet von 
gleich grosser Wichtigkeit ist, darum handelt, die verschiedenen 
Farbempfindungen mit einander zu vergleichen, die physiolo- 
gische Erregbarkeit entsprechender Nervenelemente nach einem 
gemeinsamen Massstab zu messen, so ist es selbstverstandlich 
unumganglich nothwendig, mit gleichgemessenen Reizen die 
Versuche anzustellen und unsere definirbaren, invariablen Far- 
ben anzupassen" (pp. 148-149). Hegg's "gleichgemessene Rei- 
zen" were obtained according to the method first used by Bull and 
endorsed by Hess. But this method of equating can not warrant 
any conclusion concerning the relative limits of peripheral sensi- 
tivity to the different colors. Hegg is then not justified in con- 
cluding "dass die Grenzen fur Roth und Grun zusammenfallen. 
Die Grenze fur Gelb sind durchwegs ca. i° enger als fur Blau, 
vielleicht wegen der starkerem Brechung der blauen Strahlen?" 
(p. 166). 
Baird49 was the next to use this method of equating. He 
determined the relative extension of the different color zones, 
employing as stimuli light transmitted through gelatine filters. 
Like his predecessors he also concludes beyond what is justified 
by his method of working. He expresses his results as follows: 
"The zone of stable red is coextensive with that of stable green; 
the zone of stable yellow is coextensive with that of stable blue; 
and the yellow-blue zone is much more widely extended in all 
directions than is the red-green zone" (p. 61 ).50 
48 Baird, loc. cit. 
"Fernald (Psychol. Rev. Monog., 1909, X, pp. 60-67) as a minor point 
in her study of the color sensitivity of the peripheral retina, made a hurried 
investigation in the nasal and temporal meridians of the limits of an Urroth 
and an Urgriin, an Urgelb and an Urblau that she claimed were of equal 
saturation. Her method of equating these stimuli was the same as that of 
Bull, Hess, Hegg, and Baird. The limits of these stimuli were, she states, 
determined hurriedly, most of the observations being made at intervals of 
5° on the peripheral retina; for example, she states that Urgelb and Urblau 
were both seen at 85°, and not seen at 900. In spite of the looseness of 
these determinations, she concludes that "the limits for the Urgriin are 
practically coextensive with those for the Urroth, and the fields for the HISTORICAL AND CRITICAL 
33 
Bull, Hess, Hegg, and Baird all claim, then, that when the 
investigation is made with stimuli equated in brightness and in 
terms of cancelling power, the limits for Urroth and Ur grim, 
for Urgelb and Urblau coincide. It may be inferred from their 
work that they believe that at least one of the reasons for the 
non-coincidence of limits obtained by previous investigators is 
that the colors used were not equated in intensity. Evidence, 
however, can be derived from the work of Kirschmann,51 1893, 
that this can not at least be considered the sole reason. In 
Kirschmann's case, in fact, it apparently can not be considered as 
having any influence at all in producing the non-coincidence 
obtained. Although his pairs of colors were not equated in 
intensity, it is obvious that the deviations he obtained from coinci- 
dence of limits can not be explained as due to differences in 
intensity between his stimuli. Kirschmann mapped many merid- 
ians of the retina for the limits of sensitivity with both spectral 
and pigment stimuli. The pairs of colors were not equated in 
intensity. His color maps show that the outline of the field for 
each color is irregular in the different meridians, and is different 
from that of any other color. In general the field for blue is 
wider than that for yellow, but in certain meridians this order 
is reversed. The red field is generally wider than the green, but 
in some meridians the green is the same or wider than the red. 
Further, the difference between the limits of the colors in some 
meridians is considerably greater than in others. It is evident 
Urgelb with those for the Urblau" (p. 65). In addition to applying to 
this work the criticisms passed above concerning the more careful work of 
Bull, Hegg, and Baird, all of whom drew conclusions similar to Fernald's, 
one may express surprise that work so sketchy should be considered as 
warranting any conclusion whatever. Fernald states, however, that dif- 
ferences of 5° in limits are too small to be of any significance (p. 66). She 
makes this statement apparently because with the varying conditions of 
illumination under which she worked, 50 seem to be a normal variation in 
limits. (It may be too that she considers that this gives her warrant for 
working only at intervals of 5°.) But as will be shown in the experimental 
section of this paper, a difference of 5° in limits represents a difference in 
sensitivity sufficient to raise the limen of sensitivity 2000. In more careful 
work, then, limits which varied 5° would hardly be considered as "practically 
coextensive." 
B1 Kirschmann, A. Die Farbenempfindung bei indirectem Sehen. Philos. 
Studien, 1893, VIII., pp. 562-614. 34 
GERTRUDE RAND 
that irregularities of this kind can not be due to a difference in 
the intensity of the stimuli employed, for if blue and yellow, for 
example, have the same limit when equated in intensity, their 
limits should retain the same general outline when the colors are 
of unequal intensity, and should differ only in their distance from 
the fovea. The zone for the more intense color should in every 
meridian be regularly some degrees wider than that for the less 
intense. Relative to the issue between Kirschmann and Bull, 
Hess, Hegg, and Baird, it is interesting to find that Bull, Hess, 
and Baird, who all claimed to find coincident limits in all parts 
of the retina for the paired colors obtained results which, when 
examined in detail, show the same deviations from coincidence as 
those which Kirschmann found. Baird, for example, determined 
the limits of the four principal colors in eight meridians, and 
concludes: "The results show that the zone of stable-red is 
coextensive with that of stable-green; that the zone of stable- 
yellow is coextensive with that of stable-blue."52 An inspection 
of his table of results shows however, that this coincidence is 
extremely rough. In the results for every observer it is found 
that in some meridians the green field is wider than the red by 
1°, 2°, or 30; in other meridians, there is coincidence of limits; 
and in still other meridians, the green field is narrower than the 
red by i°, 2°, or 30. The same thing is true of blue and yellow. 
In general, but not in every meridian, yellow seems to be wider 
than blue on the nasal retina, blue wider than yellow on the 
temporal. Hess's and Bull's results show similar variations which 
are in some cases of even greater extent. It is evident, however, 
from their conclusion concerning the coincidence of limits that 
they regard these variations as insignificant, probably no more 
than their normal M.V. for the rough conditions under which 
they worked. But it should be borne in mind that 2° or 30 of 
difference in limits is not insignificant when conclusions are to 
be drawn from the results with regard to the relative sensitivity 
of the peripheral retina to the members of the pairs of comple- 
mentary colors. Because of the abrupt falling off in sensitivity 
just before the limit is reached (see p. 117 ff.), a difference of 
52 Baird, J. o-p. cit., p. 61. HISTORICAL AND CRITICAL 
35 
2° or 30 in the limits represents quite a large difference in sensi- 
tivity. For example, according to our results a difference of 20 
in limits represents a difference in sensitivity sufficient to raise 
the limen for yellow 1200, for green ioo°, for red 1600, for 
blue 1700; and a difference of 30 represents sufficient to raise the 
limen for yellow 2100, for green 2150, for red 2100, for blue 
2150. Obviously the point is too important to be passed over 
without re-examination by better methods of working. The 
limits should be re-determined under conditions that do not give 
so large an M.V. This has been done by the writer with such a 
control of all the factors that cause variable results that her M.V. 
from observation to observation is less than i°. The results 
obtained show that this difference in limits is real, and not an 
error due to any inaccuracy in the method of working. In some 
meridians the limits coincide, in others they diverge. In short, 
when the zones are outlined by lines connecting the points rep- 
resenting the limits in the different meridians, these lines for 
the pairs of colors do not coincide, but criss-cross in a very 
irregular manner. Therefore, on the basis of our own results as 
well as those of Kirschmann, the inference can not be drawn from 
the work of the men who have equated their stimuli in terms of 
cancelling power that coincidence of limits had not been obtained 
up to that time because stimuli so equated had not been used. 
Nor can the conclusion be drawn that even if the stimuli had been 
properly equated, coincidence of limits would have been obtained. 
In fact the converse of this conclusion can apparently be drawn.53 
58 The writer used as stimuli in these experiments the standard yellow, 
red, green, and blue of the Hering series. She did not use colois stable in 
tone as Bull, Hess, Hegg, and Baird claimed to use, because an exhaustive 
study of this question showed her that stability of tone for all meridians 
of the retina can be obtained for blue alone of the four principal colors. 
For a further statement in support of this point, see footnote p. 85. In any 
event, she is unable to see how the fact that the red and green of this 
series appear yellow in a small region of the peripheral retina could have 
any effect on the coincidence of limits for red and green, unless the limit 
for red, for example, is taken as the point at which all sensation of color 
disappears, regardless of whether this color is red or yellow. (The writer 
considered in her own experiments that the limit for a color was the point 
at which the color sensation lost all trace of its original quality.) However, 
even if the point at which all sensation of color whatever disappeared be 36 
GERTRUDE RAND 
Therefore, the argument based upon it against Fick's modifica- 
tion of the Helmholtz theory to explain the color-blindness of 
the peripheral retina can also be refuted insofar as the results on 
the coincidence of limits can be considered as furnishing argu- 
ment. This argument is, it will be remembered, that the sensi- 
tivity of the retina from fovea to periphery falls off as rapidly to 
one of the members of the pairs of complementary colors as to 
the other. It should be borne in mind, however, that the results 
needed in order to make this argument can not be obtained by a 
determination of the comparative limits of sensitivity alone. 
considered the limit, the following consideration shows that the small 
component of yellow present in the peripheral retina in the sensations 
aroused by the red and the green of this series could not have caused the 
criss-crossing of limits any more than a difference in intensity could have 
caused it. In fact the point in reality reduces to a question of intensity. If, 
for example, one of the colors were stronger than the other, this color would 
have relatively more power to arouse this yellow component, than it would if 
the colors were of equal intensity. That is, the zone through which the yellow 
component would be sensed would be broader for this color than it would 
have been had both colors been of the intensity of the weaker color. And 
relative to the zone in which a yellow component was sensed, it would not be 
irregularly broader in some meridians for the stronger color and in other 
meridians for the weaker color, for in regions in which the yellow sense is 
relatively weak the zone would narrow for both stimuli. In short, while the 
influence of the yellow component might cause the zone of sensitivity to 
broaden for one member of the pairs of stimuli as compared with the other 
member, it could not cause it to become alternately broader and narrower, in 
other words, to criss-cross. It is just as obvious that difference in brightness 
can not be offered as an explanation of this criss-crossing, even if difference 
in brightness could be shown to have an effect on the limits of sensitivity to 
the pairs of colors in question. But, as we will show in the experimental sec- 
tion, difference in brightness can be considered as having no effect whatever on 
the limits of sensitivity. Hence on both counts, difference in brightness can 
be ruled out of consideration. It would seem, then, that we can conclude 
that the criss-crossing of limits represents a real relation of sensitivity to 
the members of the pairs of colors in the far periphery of the retina, even 
before the investigation is made with stimuli properly equated in intensity. 
Moreover, when we take into account the fact that a difference of 2° to 30 in 
the limits represents a large difference in sensitivity we have considerable 
reason for believing that the ratio of the sensitivity to one member of a pair 
of complementary colors to the sensitivity to the other member is not the same 
in all parts of the retina. The point will be definitely determined in the near 
future by the writer by a careful determination of limens from point to point 
in many meridians of the retina, with stimuli which consist of lights of 
spectral purity, measured in terms of a common unit of intensity. HISTORICAL AND CRITICAL 
37 
Knowledge is needed of the comparative sensitivity all the way 
out. Moreover, this knowledge must be based on actual determin- 
ations at points not widely separated from each other. The 
limit is only one of the points at which determinations should be 
made. In fact, it can scarcely be said to sustain any more 
important relation to the problem than any other point far re- 
moved from the fovea. Furthermore, conclusions can not be drawn 
from work done by the method of limits with regard to the 
comparative falling off in sensitivity from fovea to periphery 
unless it has previously been inferred from a comparison of the 
results at the center with the results at the limits what the com- 
parative sensitivity should be at the points intervening, as was 
done by Hess. But this is wholly unjustifiable, for if there is 
one principle above another that the determinations of sensitivity 
in the peripheral retina bring out, it is that inferences can not be 
drawn about sensitivity between points at all widely separated. 
In fact, practically no conclusions of systematic importance can 
be justified at all from results obtained by the method of limits, 
although conclusions of great importance to theory have fre- 
quently been drawn from such results. The method of limens 
should be used instead, for by means of it accurate account can 
be taken of sensitivity at any point that is desired.54 
A brief survey of our discussion of intensity shows the fol- 
lowing facts: 
1. The intensity of the stimulus is a factor influencing both the 
limens and the limits of color sensitivity. It should, therefore, 
be carefully standardized in all determinations of sensitivity. 
2. The conclusions that have been drawn up to this time con- 
cerning the comparative sensitivity of the retina to the different 
colors have not been justified because of the methods of working 
that were used in the investigations from the results of which 
they were drawn; for (a) either no standardization of the inten- 
sity of the stimuli used had been made; or (&) this standardization 
had been made by an improper method. From the results of 
81 As stated in footnote p. 83, if a more exhaustive study of the point 
is wanted, the results of the method of limens should be supplemented 
by a determination of the just noticeable difference in sensation for 
various points in the intensity scale at each point of the retina investigated. 38 
GERTRUDE RAND 
the work as it was done the following conclusions alone can be 
drawn: (a) The comparative sensitivity to the different colors is 
not the same at the center as it is in the periphery of the retina. 
This conclusion may be drawn from the results of Chodin and 
Raehlmann and Butz for the limens of sensitivity. And (b) it 
is not the same for the members of the pairs of complementary 
colors at all points even in the small region of the peripheral 
retina that has been examined, namely in the region compre- 
hended by the criss-crossing limits for these colors. This con- 
clusion may be drawn from Kirschmann's results, from our own, 
also from those of Bull, Hess, and Baird as shown in their 
tables giving the limits of their stimuli in different meridians. 
3. Broader conclusions than this are justified only when the 
comparative sensitivity of the retina to the different colors is 
determined with stimuli properly standardized in intensity. As 
has been briefly pointed out, the comparative limits of sensitivity 
can be determined only when the intensities of the stimuli used 
have been standardized in terms of a common objective unit, for 
example, the unit of energy; and the comparative limens only 
when the intensities of the stimuli used have been estimated in 
terms of this common objective unit, or some other unit in 
terms of which results can be compared.55 This point will be 
discussed more fully later in the paper (see p. 63 ff). 
(c) The effect of brightness of the stimulus. 
There has been very little discussion by previous investigators 
whether the brightness or white-value of a color affects the retinal 
limits of sensitivity to that color, and whether this aspect of the 
stimulus need then be taken into account in determining the rela- 
tive extent of the color zones. We have already shown that in 
such work as there has been, the brightness factor has been 
obscured by and confounded with the intensity factor. 
The confusion of the intensity and brightness aspects of color was noted 
by Langley in an article entitled Energy and Vision, Philos. Mag., 1889, 
XXVII, gth ser., p. 1. Langley writes: "While it is quite a familiar fact that 
the luminosity of any spectral ray increases proportionately to the heat in 
55 It is scarcely needful to point out that the same kind of standardization 
is required in order to determine the comparative j. n. d's. for the different 
colors. HISTORICAL AND CRITICAL 
39 
this ray, and indeed is but another manifestation of the same energy, I have 
recently had occasion to notice that there is, on the part of some physicists, a 
failure to recognize how totally different optical effects may be produced by 
one and the same amount of energy according to the wave-length in which 
this energy is exhibited. I should not perhaps have thought it advisable to 
make this last remark, were it not that there has appeared in a recent 
number of Wiedemann's Annalen a paper by H. F. Weber on "The Emission 
of Light", in which he tacitly makes the assumption that the luminosity of 
a color is proportionate to the energy which produces it, an assumption 
which it is surprising to find in a paper of such general merit and interest." 
This confusion, insofar as it has not been due to a misinterpre- 
tation of terms, seems to have arisen because the usual method of 
varying the one factor has necessitated an accompanying variation 
of the other. Further, the fact that when the intensity or energy 
of spectral light is increased to a maximum, the color is lightened 
until white is produced,56 has been responsible for the view that 
light colors are more intense than dark colors, and has led to the 
custom of determining the energy or intensity of colored light by 
photometric methods. The photometric method can not, how- 
ever, be used directly for estimating the intensity of colored light 
for two reasons, (a) Direct radiometric measurements of energy 
show that the relative values of the colors of the spectrum as 
determined by the two methods do not at all coincide. The photo- 
metric curve, for example, of the spectra of all light sources of 
normal intensity is highest in the yellow-green and lowest in the 
blue and red. The radiometric curve of the visible spectrum of 
sunlight of the same intensity is, on the other hand, according 
to Langley,57 highest in the red near the C line and lowest in 
the violet; while the radiometric curves of the visible spectra of 
most of the artificial sources of light, such as the Nernst, tungsten, 
and arc lights, are highest in the extreme red and lowest in the 
violet. (&) The relative photometric values of the colors of all 
spectra differ widely for different intensities of the same light- 
'* This statement is true only if the original color has maximal saturation. 
When one increases the intensity of colors whose original intensity is 
slight, red and yellow become lighter, blue and green darker. These 
brightness changes are known as the Purkinje phenomenon. 
67 Langley, Energy and Vision. Amer. Journ. of Science, 1888, XXXVI, 
3rd Ser., pp. 359-379; also Philos. Mag., 1889, XXVII, 5th Ser., p. 1; and 
Invisible Solar and Lunar Spectra, Philos. Mag., 1888, XXVI., 5th Ser., pp. 
505-520. 40 
GERTRUDE RAND 
source. For medium intensities, for example, the curve is highest 
in the yellow-green and lowest in the blue. But as the intensity is 
decreased the curve levels, while its maximum height shifts to 
the green and its minimum to the red. In short, the photometric 
value of a color is not a constant but a variable function of its 
intensity. From the above consideration, it is obvious, then, (a) 
that the photometric method can not be used to estimate the 
relative intensities of the colors of the spectrum even for a 
single intensity of light-source unless for each point of the spec- 
trum considered a factor be determined which will transform the 
photometric into the radiometric value; and (&) that it can not 
be used over a wide range of intensities of light-source unless 
this calibration be previously made for each degree of intensity 
used. Furthermore, the brightness factor is not inseparably 
bound up with the intensity factor. It can be isolated. When, for 
example, one uses a constant amount of colored light, spectral or 
pigment, and mixes with it a constant amount of white, black, or 
gray, one obtains stimuli which contain an equal amount of col- 
ored light but which have different brightnesses. Up to the 
present this method of isolating the brightness factor to test its 
influence has been employed only by Hess, and in his work, as 
we shall see, it has been used very inadequately. 
The discussion whether the brightness of the stimulus affects 
color sensitivity was raised by Aubert. Aubert was unable to 
reach positive conclusions concerning its influence. He was led 
to a consideration of the question by the outcome of an investiga- 
tion planned to determine the influence of visual angle upon the 
perception of color in central vision. His results showed that the 
liminal visual angle was different in case of the different colors, 
and further, that it differed when the colors were viewed upon 
white and black grounds. In the discussion of these results he 
writes: "Zum Theil beruhen diese Verschiedenheiten wohl auf 
einer verschieden starken Affection des Farbensinnes, zum gros- 
seren Theil aber wohl auf Helligkeitsdifferenzen. Wir haben 
dabei drei Momente zu beriicksichtigen, namlich die Farbennii- 
ance, die Farbenintensitat, und die Helligkeitsdifferenz oder den 
Contrast der Pigmente."58 These Momente Aubert defines as 
8S Aubert, H. Physiologic der Netzhaut. Breslau, 1865, p. in. HISTORICAL AND CRITICAL 
41 
follows: Farbennuance is the sensation given when a color is 
mixed with white, black or gray. A light blue, for example, con- 
tains more white and less blue light than a saturated blue.59 Far- 
benintensitat is defined as "the impression which is dependent on 
the intensity of colors: in case of spectral colors, on the ampli- 
tude of vibration; in case of pigments, on the intensity of the 
illumination." "Maxwell," he says, "calls this 'shade' \ one color 
may be lighter or darker than another."60 Whether or not Aubert 
also uses the term intensity of color as synonymous with white- 
value is open to question. There is evidence in his discussion, 
however, that he uses it synonymously with brightness. For 
example, when discussing FarbenintensitaF1 he claims that the 
intensity factor cannot be made standard because it is impossible 
to determine which of two colors is the brighter, and because the 
photometric values of the spectral colors is unknown, the results 
of Melloni, Dove, and Helmholtz on this point differing widely. 
But whether or not brightness means for him also white-value 
depends upon what he thinks is measured by the photometric 
method. Helligkeitsdifferenz is the brightness relation between 
a color and its background.62 The three factors, then, that Aubert 
believes we have to consider are (a) the amount of colorless 
light mixed with the color; (b} the intensity or brightness of 
the color; and (c) its contrast with the background. The first of 
these factors, he contends, influences color perception. He finds 
that the more colorless light is mixed with a pigment color, the 
the greater must be the illumination at which the color can be 
liminally sensed; in other words, when the pigment surface re- 
flects a small amount of colored light, the intensity of its illumi- 
nation must be proportionately greater to give liminal sensa- 
tion.63 The third of these factors he also considers very 
important. Colors have different limens and limits of visibility 
on white and on black grounds. Now the brightness of the 
°* Aubert, H. loc. cit. 
80 Aubert, H. op. cit., p. 108. 
"Aubert, H. op. cit., p. m. 
62 Aubert, H. op. cit., p. 112. 
88 Aubert, H., Untersuchungen uber die Sinnesthatigkeit der Netzhaut. 
Pogg. Ann. d. Physik und Chemie, 1862, CXV, p. in, 114. 42 
GERTRUDE RAND 
color determines how great the difference is in the two cases. 
Blue, for example, is very dark. It is in great contrast to the 
white field and in much less contrast to the black. Its limit in 
each case is respectively 150 and 36°,-a difference of 21 °. Red 
is less dark. Its limits with the white field is 160, with the 
black 300,-a difference of 140. Green is a lighter color and 
more nearly of mean brightness between white and black. In 
this case the difference in its limits with white and black fields 
is only 40. Aubert contends "dass die specifische Farbenwahrneh- 
mung an der Seitentheilen der Netzhaut um so eher in eine 
blosse Wahrnehmung von Hell und Dunkel iibergeht, je starker 
dieselbe mit der Umgebung contrastiert."64 This is stated again 
in the Physiologic der Netzhaut: "Contrast und Helligkeit der 
Farben sind von grossem Einfluss auf die qualitative Farben- 
empfindung, so wie auf die Grosse der Netzhautparthie, inner- 
halb welcher die Farben empfunden werden sein."65 In this re- 
gard, then, the difference in the brightness of the colors has an 
effect on color sensitivity only when the brightness of the sur- 
rounding field is made the same for all of the colors. But 
since it need not and should not be made the same for all of the 
colors in any investigation of sensitivity, unless the purpose of 
the investigation is to test the effect of surrounding field on the 
different colors, this case may be ruled out of consideration. 
The second Moment, the intensity or brightness of the colors, 
Aubert finds himself unable to isolate. He works with pigment 
stimuli and can not alter the intensity without at the same time 
changing the proportion of colored to colorless light reflected 
from his stimuli; or as he calls it, the nuance of the color. That 
is, in order to change the intensity of the color, he is compelled 
to change the amount of white, black, or gray mixed with it.66 
Further, he thinks it is often impossible to determine which of 
two colors is the brighter.67 However, his experiments to de- 
termine the relative intensity of illumination necessary for the 
64 Aubert, H. Ueber die Grenzen der Farbetiwahrnehmung auf die seitlichen 
Theilen der Netzhaut. A. f. O., 1857, III,2. p. 54. 
65 Aubert, H. Physiologic der Netzhaut. Breslau, 1865, p. 122. 
08 Aubert, H. op. cit., p. 153. 
67 Aubert, H. op. cit., p. ill. HISTORICAL AND CRITICAL 
43 
liminal visibility of the different colors lead him to think that the 
brightness of color is not in itself a factor. In these experiments 
he finds that when the colors are arranged in order according to 
the intensity of illumination necessary for their perception, they 
are from least to greatest: orange, yellow, red, light blue, light 
green, blue, green. But when the illumination is decreased so that 
all the stimuli are seen as grays, they differ in brightness from 
light to dark in the sequence: yellow, light blue and green, blue, 
green, orange, red. Since the order differs in the two cases, he 
concludes that the differences in the perceptibility of the colors 
can not be due to their difference in brightness. He says: 
"Vergleicht man diese Ordnung der Farben nach ihrer Helligkeit 
mit ihrer Reihenfolge hinsichtlich ihrer Erkennbarkeit bei be- 
schranktem Lichtzutritt, so sieht man, dass die Helligkeit der Pig- 
mente nicht die Ursache seyn kann, dass diese oder jenes Far- 
benquadrat bei einer geringeren Lichtintensitat farbig er- 
scheint."68 His final conclusion from this discussion of the three 
Momente is that the colors are not equally perceptible; but it can 
not be determined in how far this depends upon color tone, color 
intensity, or color nuance.69 
In clearing the ground by the discussion of these three Mo- 
mente, Aubert, then, finds himself utterly unable to answer the 
question: Does the brightness difference between two colors 
affect the sensitivity of the retina to these colors ? because he can 
not isolate the factors involved for investigation. He, however, 
recognizes the possibility that brightness difference may affect 
color sensitivity and for this reason is inclined to think that his 
own may be only a rough determination. He writes: "To ob- 
tain a fine estimation, the influence of brightnesses must be 
eliminated; and pigments of equal intensity and nuance must be 
observed upon a background of the same brightness as the pig- 
ment. But we do not possess such pigments; and since the 
photometric value of prismatic colors is not known, an exact 
estimation of the liminal visual angle at which the colors can 
be sensed seems impracticable."70 • As will be shown later, Bull 
63 Aubert, H. Pogg. Ann. d. Physik und Chemie, 1862, CXV, p. 105. 
89 Aubert, H. Physiologie der Netzhaut. Breslau, 1865, p. 123. 
70 Aubert, H. op. cit, p. 112. 44 
GERTRUDE RAND 
and Hegg take this statement as authority to equate colors in 
brightness for investigations of the peripheral limit of color 
sensitivity. It is obvious, however, that the statement contains 
no real authority for the equation of colors to be used even in 
investigations of color sensitivity in central vision, because 
Aubert confesses that he is unable to demonstrate whether or not 
the brightness of a color affects the retina's sensitivity to that 
color. He has merely expressed a belief that in fine determina- 
tions the equation should be made. Baird, as we have seen, also 
claims Aubert as authority for equation in brightness, but ap- 
parently he has not even this much basis upon which to rest his 
claim, for in Baird's long list of references to Aubert, none is 
found to the section containing the statement quoted above. 
This statement occurs in Aubert's discussion of the influence of 
visual angle on the perception of colors in direct vision (Physi- 
ologic der Netzhaut, pp. 108-115). Baird's references to the 
Physiologic der Netzhaut are pp. 89-105; 116-124. As has been 
shown on p. 13, Baird apparently gets his authority from a 
discussion in which Aubert is clearly concerned not with the 
white-value of color but with the total effect of changes in the 
general illumination. Furthermore, so far as the writer has 
been able to ascertain, Aubert nowhere else gives as much 
authority for equating in brightness as is contained in the state- 
ment selected by Bull and Hegg upon which to base their claim. 
Aubert's opinion that the effect of the brightness of a color 
upon the retina's sensitivity to that color can not be determined 
was also held by Chodin who was unable to isolate brightness 
and intensity from each other. 
Chodin writes :71 "Es bleibt nur iibrig die Farben bei gleichcr 
Sattigung und bei mittlerer Lichtintensitat zu vergleichen, und 
da sie unter dieser Bedingung von verschiedener Intensitat sind 
(vielleicht sind die idealen Farben von gleicher Helligkeit, wie 
Hering sich vorstellt aber wir wissen dies nicht und konnen es 
nicht wissen) so liegt es sehr nahe anzunehmen, dass diese ver- 
schiedene Helligkeit eine constante Eigenschaft der Farbe selbst 
sei, welche, wenigstens in merklichen Grade, weder vermehrt 
71 Chodin, A. op. cit., p. 178. HISTORICAL AND CRITICAL 
45 
noch vermindert werden kann ohne eine Veranderung des Char- 
akters der Farben selbst herbeizufiihren." Chodin goes on to 
point out one of the difficulties of attempting to obtain stimuli of 
equal brightness. In case of some of the colors, both spectral 
and pigment, a change of quality or tone takes place, when, in 
their state of greatest saturation, their brightness is altered. For 
example, spectral yellow when darkened gives a reddish-yellow or 
brown appearance, and the yellow of pigment paper, a decided 
olive-green appearance. Blue, when altered in brightness, ap-. 
pears reddish. These color changes have been recorded by 
Chodin, Briicke,72 Hegg,73 Rood,74 and others. They are marked 
and are particularly troublesome in that the blue and yellow 
stimuli, which undergo the greatest changes in quality, are the 
stimuli that must be altered the most in order to be equated in 
brightness. 
Ole Bull73 was the first to mention the problem under discus- 
sion in its specific relation to the determination of peripheral 
color limits. He made no direct test of the influence of the 
brightness of the stimulus upon sensitivity, but quoted as author- 
ity for equating his own stimuli Aubert's statement that for fine 
determinations of color sensitivity brightness differences must 
be eliminated. Like Aubert, he was unable to show that the 
brightness of the stimulus constitutes a factor, but unlike Aubert 
he does not recognize the need for demonstrating this point. 
He thus equated his stimuli apparently without ever even having 
realized the need to investigate whether or not this equation 
should be made. 
Hess76 seems to have been the only investigator who has made 
any attempt whatever to determine whether a light and a dark 
color of equal intensity have the same or different limits of visi- 
bility in the peripheral retina. His test was simple and was ap- 
plied only to one color. Using two stimuli, the one composed 
73 Briicke. Wiener Berichte, 1865, LI, p. 10. 
73 Hogg, E. loc. cit. 
u Rood. On the Effects Produced by Mixing White with Colored Light. 
Phil. Mag., 1880, X., p. 209. 
75 Bull, O. op. cit., p. 93. 
73 Hess., C. op. cit., p. 42. 48 
GERTRUDE RAND 
is obvious, then, that if the intensity of the colored stimulus is 
sufficient to cause its limit to fall within this zone of rapid de- 
crease in sensitivity, the difference in saturation between a light 
color and a dark color of equal physical intensity would not be 
sufficient to cause a widening of the limits. On the other hand, 
if a smaller intensity of color were used so that the limit of the 
color occurred within the retinal region where color sensitivity 
alters but slightly from point to point, the greater apparent satur- 
ation of the dark color would be sufficient to allow the color to 
be visible at a point where the light color of less apparent satur- 
ation though of equal physical intensity is not sensed. Now the 
limen for the different colors when mixed with white at the 
point where the rapid decrease in sensitivity to each begins, is as 
follows: 120° of yellow; 1300 of green; 1350 of red; and 
1450 of blue. These values, then, represent the amount of color 
that would have to be had in the stimulus to which white had 
been added to make the limits of sensitivity to them fall just at 
the boundary between the zones of gradual and abrupt decrease 
in sensitivity. And since black inhibits color less than white, 
even less color would be required in the stimuli to which black 
had been added to make the limits fall at this point. For any 
amount of color greater than these values, the limits would fall 
within the zone of rapid decrease. One could, then, add white to 
his stimuli by amounts varying from o° to 2I5°-24O° for the 
different colors, and black in still greater amounts, and still work 
within the zone of abrupt decrease. And as long as one works 
within the region of abrupt decrease, the limits for the colors 
lightened and darkened by the above amounts will coincide. It 
need hardly be said that 215°-24O° multiplied by two to express 
the variation in both directions, represents a much greater bright- 
ness difference than exists between the standard colors. On the 
other hand, for any amount less than the above values, the limits 
would fall in the zone of gradual decrease. Now Hess's stimuli 
contained less than this amount of color because his Urgriin con- 
tained only 244 0 of green and 1160 of blue. 1800 of this 
Urgriin would, then, contain but 1220 of green and 58° of blue. 
The intensity of this stimulus was, therefore, so small that ac- HISTORICAL AND CRITICAL 
49 
cording to our results the limits for both component colors would 
fall in the zone of gradual decrease of sensitivity. In this 
zone the greater saturation of the color when mixed 
with black would cause it to be sensed farther out than the 
color when mixed with an equal amount of white; that is, the 
limits for the darkened color would fall further from the fovea 
than the limits for the light color, as Hess found them to do. 
In this regard, too, then, it may be said that if in Hess's test, 
a brightness change was made greater than the brightness dif- 
ference between the standard colors, and consequently a stimulus 
was used having smaller intensity than it would have had, if the 
proper brightness change had been made, the test would have 
been unfair because the degree of intensity of the stimulus is a 
factor determining whether or not there is a different limit for 
light and dark colors of equal physical intensities. 
While it is sufficient for our point to show in what respects 
Hess is open to criticism for arbitrarily selecting 1800 of bright- 
ness variation without making any determination of how much 
need be made, a word may be added to show that 1800 of 
brightness variation is much more than was needed. For, the 
table in which Hess78 states the values of the stimuli determined 
by different observers to be equal in cancelling-action and bright- 
ness, shows that in no case was a variation of more than 1330 
made in one direction, that is, toward either white or black. It 
is often difficult to see just how much variation was made, be- 
cause in the majority of cases both black and white were added, 
hence the net variation was less than the value of either. For 
example, 1050 of yellow, 6o° of white, and 1950 of black were 
found by one observer to be equal in cancelling action and bright- 
ness to 85° of blue, 84° of white, and 1910 of black; for an- 
other observer, 2000 of red, 48° of blue, 66° of black, and 46° 
of white were equal in cancelling action and brightness to 1880 
of green, 390 of blue, and 1330 of black. But in no case is the 
variation in one direction more than 1330, or less than 91 °. We 
may conclude, then, that by his choice of 1800 of brightness 
variation by means of which to compare the limits of a light and 
" Hess, op. cit., pp. 45-46. 50 
GERTRUDE RAND 
a dark color of equal physical intensity, he has, roughly speaking, 
made his brightness variation one and one-half times greater 
and his intensity only four-fifths of what it should have been, 
judging from the greatest variation he had to make for his 
brightness equation. Either one of these items: the overestima- 
tion of brightness change or the underestimation of intensity is 
sufficient both to render his test unfair and, moreover, to account 
for a very considerable difference in the limits for color mixed 
with white and for the same color mixed with black. 
The second point of criticism of Hess's test is in regard to 
his method of controlling the brightness of the surrounding field 
and of the preexposure. He did not use the proper conditions 
of brightness of screen and of preexposure card. He used a 
screen and a preexposure which were of the same brightness as 
the dark stimulus, both in the tests with the dark and with the 
light color. There was, then, a considerable amount of white 
added to the light stimulus by contrast from the dark background 
and by after-image from the dark preexposure over and above 
the white that was added by objective mixing. Thus while he 
aimed to add equal amounts of white and black to his pairs of 
stimuli, he really added much more white than black. Now, we 
have found that white subjectively aroused apparently has as 
much effect on the color excitation as white objectively aroused. 
His test was thus again rendered unfair by his lack of proper 
conditions with regard to the brightness of the surrounding field 
and of the preexposure. The surrounding field and the pre- 
exposure should have been made in each case equal in brightness 
to the stimulus. 
Hess does not appear to have realized the importance either of 
the brightness of the surrounding field or of preexposure as 
factors influencing color limits. He generally equalized the 
brightness relation between stimulus and background because he 
seemed to think that accuracy of judgment was fostered there- 
by,-that one can more readily tell when the color has disap- 
peared from a given stimulus when there is no brightness 
difference between the stimulus and the background to confuse 
the judgment.79 The preexposure he made of the same quality 
79 See Hess, op.cit., p. 25. HISTORICAL AND CRITICAL 
51 
and brightness as the background. No reason was given for 
this procedure.80 Thus in the above test, he overlooked the 
diminution in the intensity of the sensation aroused by the 
lighter stimulus due to the contrast from the surrounding field 
and to the after-image from the preexposure,-a diminution 
which is also in itself sufficient to account for quite a large dif- 
ference in the limits for a light and for a dark stimulus of the 
slight degree of intensity used by Hess. Hess's results, then, 
assuredly can not be considered as showing that the brightness 
difference between the normal colors affects their limits of sensi- 
tivity. 
Hegg,81 like Bull, made no test to determine whether the 
brightness of the stimulus constituted a factor in the determina- 
tion of color limits. He gives no reason for the attempt he 
makes to equate in brightness, other than the fact that Aubert 
and Bull had mentioned the necessity for this procedure. 
Baird was the next to state that the stimuli used to investigate 
the peripheral color sense must be of equal brightness. His 
reasons for making this statement have already been discussed 
in part (pp. 12-19) in order to show how the prevailing confusion 
with regard to terms has led to misinterpretation. To connect 
this preceding discussion with what is to follow, a few words to 
resume will probably be of service here. It will be remembered 
that like Bull and Hegg, Baird made no attempt to determine 
whether or not the brightness or white-value of a color exerts 
an influence on the limits of sensitivity to that color. Unlike 
Bull and Hegg, however, he claims to be able to derive authority 
for equating stimuli in white-value from the work of many in- 
vestigators,-primarily from that of Aubert,82 Landolt,83 and 
Abney,84 but also from Raehlmann,85 Chodin,86 Klug,87 Bull,88 
80 See Hess, c. op. cit., p. 44. 
81 Hegg, E. op. cit., p. 146. 
82 Baird, J. op. cit., p. 12. 
83 ibid, p. 16. 
84 ibid, p. 31. 
85 ibid, p. 17. 
86 ibid, p. 20. 
87 ibid, p. 20. 
88 ibid, p. 22. 52 
GERTRUDE RAND 
Hess,89 and Hegg.90 With regard to these sources of authority, 
it has been shown (a) that Abney and Landolt do not even claim 
that brightness difference affects the sensitivity of the retina to 
color, and that Aubert does not in the references given by Baird; 
(&) that Bull and Hegg equated their stimuli in brightness 
merely because Aubert had expressed the belief that such pro- 
cedure is necessary in making fine determinations of color sensi- 
tivity, but since Aubert was unable to demonstrate this necessity, 
their reason for making the equation has no value;91 and (c) 
that Hess's test, upon the results of which he bases his conclusion 
with regard to the need to equate, was both incomplete and 
wrongly devised. We have yet to show, then, that no authority 
can be derived by Baird from the work of Raehlmann, Klug, 
and Chodin. A claim to authority was derived from the work 
of Raehlmann and Klug by misinterpretations similar to those 
made in the cases of Landolt and Abney. Raehlmann and Klug 
both worked with spectral light and sought to find the effect 
of decreasing the intensity of the colored light upon the limits of 
sensitivity. Raehlmann92 decreased the intensity of his stimuli as 
follows: Light reflected from a heliostat was passed through a 
prism and the spectrum from this source was thrown upon a 
screen, which may be called screen^ the distance of which from 
the light-source was kept constant. This screen contained an 
opening for the transmission of the colored light. The amount 
of light transmitted could be regulated by the size of this open- 
ing, and the quality could be regulated by shifting the position 
of the opening along the spectrum. The colored light fell upon 
a second screen, screen2, so arranged that its distance from 
screen! could be varied. In two ways, then, could diminution of 
89 Baird, J. op. cit., p. 27. 
80 ibid, p. 29. 
91 As has been shown, pp. 40-44, Baird, as well as Bull and Hegg, might have 
had some justification in citing Aubert's authority on the question of 
brightness equation, from the latter's statement that all brightness differences 
must be eliminated from stimuli used to make fine determinations of color 
sensitivity. But this statement of Aubert's is not included in the references 
to Aubert from which Baird drew his authority. 
93 Raehlmann, E. Ueber Verhaltnisse der Farbenempfindung bei indirectem 
und directem Sehen. A. f O., 1874, XX,1( p. 18. HISTORICAL AND CRITICAL 
53 
intensity be produced: (a) by decreasing the size of the opening 
in screen-t, and (b) by increasing the distance between screen! 
and screen2. Both methods were used by Raehlmann for pro- 
ducing what he terms die Abnahme der Lichtstarke. A change 
in brightness may have been incidentally produced, but this as- 
pect of the stimulus was of no concern to him. He found that 
a decrease in intensity decreased the zone in which a color was 
sensed in its characteristic tone; he most assuredly does not claim, 
however, as Baird says he does that "the color limits were found 
to vary with changing brightness of stimulus" (see Baird, p. 17) 
in the sense in which Baird uses brightness, namely, as white- 
value. Klug93 used a method somewhat similar to Abney's. He 
weakened a beam of light by interposing respectively one, two, 
and three thicknesses of ground glass, and found that the color 
limits were narrowed in each successive case. Thus he also made 
no attempt to isolate the effect of the brightness of the stimulus, 
and his work can not be cited as having any bearing on that 
problem. In Chodin's work, as we have already seen, the ad- 
visability of equating in brightness was discussed and decided 
against because of lack of evidence for the need of equating and 
because of the changes in color tone produced by changing the 
brightness of the colors. In giving Chodin as one of his authori- 
ties for equating, Baird refers to the passage in Chodin's article 
quoted in the original in this paper, pp. 44-45. 
Baird writes: "Chodin remarks in his introduction: Tt is 
self-evident that in comparing the retinal sensitivity to different 
colors, the color stimuli employed must be of equal brightness 
and of equal saturation.' But this very essential condition was 
not fulfilled in his own experiments" (see Baird, p. 20). Baird 
has here again made a misinterpretation. The rather free trans- 
lation of Chodin's statement: "Es bleibt nur iibrig die Farben 
bei gleicher Sattigung und bei mittlerer Lichtintensitat zu ver- 
gleichen"94 and the failure to read carefully the discussion fol- 
lowing it, are responsible, we presume, for the misinterpretation. 
It is obvious from the foregoing resume that the factor, bright- 
83 Klug, F. Ueber Farbenemfindung bei indirectem Sehen. A. f. O., 1875, 
XXI,pp. 274-278. 
91 Chodin, A. op. cit., p. 178. 54 
GERTRUDE RAND 
ness of the stimulus, has been very inadequately treated in the 
literature. The specific question has never been answered, in 
fact has never really been investigated: Does the amount of 
brightness difference existing between the colors influence their 
limits of sensitivity in the peripheral retina? Aubert95 and 
Chodin96 and others have shown that the sensation limen of 
color when mixed with white is higher than the limen when 
mixed with black. This fact may be explained as due to the 
superior inhibitive power of white. But within what limits this 
greater inhibitive action of white is sufficient to cause the peri- 
pheral limit of a color mixed with white to be narrower than 
that of an equally intense color mixed with black has not been 
determined. And certainly it has never been shown that the 
brightness difference that exists between the standard colors 
at full saturation exerts an inhibitive action sufficiently strong 
to cause a change in the peripheral limits. It has never been 
claimed, for example, that a light color in its state of maximal 
saturation is more inhibited for sensation than a dark color in 
its state of maximal saturation by the brightness component in- 
herent in each; in other words, that a saturated yellow is more 
inhibited by its brightness component than is a saturated blue 
by its brightness component. In fact, in strange contradiction 
to this, it has often been held that the colors which have the 
stronger white component are the more intense. Yellow, for 
example, has been frequently called a more intense color than 
blue just because of its proximity to white. 
We must conclude, then, that the assumption that color limits, 
must be investigated with stimuli of equal brightness is prob- 
ably based upon the belief that stimuli differing in brightness 
differ also in intensity. This belief has doubtless arisen from the 
fact that as stimuli are ordinarily varied, a change of brightness 
is accompanied by a change of intensity, and conversely a change 
of intensity is accompanied by a change of brightness. But 
brightness and intensity are not inseparable variants. Conclu- 
sions should not be drawn, therefore, until the influence of bright- 
95 Aubert, H. Physiologische Optik, p. 532. 
98 Chodin, A. op. cit., p. 183. HISTORICAL AND CRITICAL 
55 
ness change has been investigated in separation from intensity 
change. Since this investigation has not been made, we are 
forced to consider that the influence of the brightness of the 
stimulus upon the limits of color sensitivity is at present an open 
question, despite the verdict to the contrary by Bull, Hegg, 
Hess, and Baird. We have, therefore, included it in our own 
work as one of the points to be investigated. The results of 
this investigation are reported in the experimental section of 
this paper. 
3. Brightness of the Field Surrounding the Stimulus. 
The recognition of the influence exerted by the field surround- 
ing the stimulus upon the limits of the color zones, has led to the 
substitution of the campimeter for the perimeter in investigations 
of the color sensitivity of the peripheral retina. The campimeter 
provides a means of readily changing the brightness of the field 
which surrounds the stimulus, so that the effects of these changes 
may be studied both upon the limens and limits of color and upon 
the quality changes that appear as the stimulus is carried from 
the fovea to the periphery. 
The influence of the field surrounding the stimulus is two- 
fold. In the first place, it directly modifies the stimulus by 
contrast induction, provided there is brightness opposition. This 
effect was observed and to some extent investigated by Aubert 
and Woinow before the campimeter came into use. In the 
second place, the campimeter screen, when of sufficient size, 
stimulates the entire retina uniformly and guarantees an equal 
brightness-adaptation of every portion. It was the recognition 
by Kriikow97 that former methods had allowed the retina to be- 
come unequally fatigued to chance objects in the surounding 
room, that led directly to the first use of the campimetrical 
method of working. Kriikow did not, however, study the effect 
of different backgrounds. He used a uniform gray field to stimu- 
late the surrounding retina, and mounted stimuli on cards of 
equal quality. In subsequent investigations, a black background 
was used almost exclusively. So far as induction is concerned, 
" Kriikow, loc. cit. 56 
GERTRUDE RAND 
this screen gives conditions with the light-adapted retina some- 
what similar to those existing in dark-room work; that is, in 
each case the stimuli are lightened by contrast from the sur- 
rounding dark field. 
Woinow and Aubert worked only with small areas of back- 
ground and thus secured the given brightness stimulation over 
but a small zone surrounding the part of the retina stimulated 
to color. Woinow98 placed a disc made of black and white 
sectors behind the stimulus to be investigated. He found that, 
when the sectors were so adjusted that when rotated they formed 
a dark gray, the color limits were the same as when the sectors 
were arranged to give a light gray sensation. From these re- 
sults, he concludes that the color zones are not influenced by 
the brightness of the field surrounding the stimulus. Aubert99 
fastened colored paper stimuli on white and on black cards. He 
found that the black card gave relatively wider limits for red; 
and that the white card gave relatively wider limits for yellow, 
green, and blue, except for very small stimuli. 
The first campimeter described was apparently what is now 
called the Hering color-mixer.100 Hess101 was the first to employ 
it for an investigation of peripheral color sensitivity. He and 
later Tschermak102 tested by means of it the influence of the 
brightness of the surrounding field upon the color limits. With 
this apparatus pigment stimuli are observed through an opening in 
a large gray screen, placed in the horizontal, which can be turned 
toward or away from the source of light, and in this way a sur- 
rounding field can be obtained that is lighter, darker, or equal in 
brightness to the stimulus at its point of disappearance as color. 
Hess and Tschermak both found that the limits of sensitivity to 
98 Woinow, M. loc. cit. 
99 Aubert, H. Physiologische Optik. pp. 541-543. 
190Titchener in Experimental Psychology, Instructor's Manual, Qualitative, 
1901, p. 20, ascribes the description of this apparatus to Hering, giving as 
reference A.f.O., 1889, XXV4, p. 63. The writer is unable to find any men- 
tion of this apparatus in this or any other of Hering's articles. It is, how- 
ever, described in some detail on p. 25 of the paper by Hess which just 
precedes and accompanies the Hering article to which Titchener refers. 
101 Hess, C. loc cit. 
102 Tschermak, A. op. cit., p. 561. HISTORICAL AND CRITICAL 
57 
color were widest when the surrounding field was equal in bright- 
ness to the stimulus. If the stimulus appeared lighter or darker 
than the surrounding field, the limits were narrowed proportion- 
ately to the loss of saturation of the stimulus color due to the ac- 
tion upon it of the brightness quality induced by the background. 
Fernald103 used a vertical campimeter. She summarizes her 
results with white and black screens as follows: "All the colors 
except the reds are perceived at a greater angle of eccentricity 
with the dark than with the light backgrounds." 
The only quantitative estimates of the effect of different 
backgrounds reported by these experimenters is given in terms 
of the effect upon the color limits. In no case has the amount 
of white or black induced by a given screen been determined, 
nor has the effect of the induction upon color sensitivity ever 
been tested in any part of the retina by the most direct means 
available, namely, the determination of the limen or threshold of 
sensation. Neither has any attempt been made to isolate the 
influence of the background from the influence of the brightness 
of whatever stimulates the retina immediately before the exposure 
of the stimulus. This factor, which we shall discuss under the 
name of preexposure, is effective through the intensive brightness 
after-image that is set up on the retina and is superimposed upon 
the colored stimulus when it is exposed. Its importance has 
never been recognized by previous investigators, nor has its 
effect ever been studied in isolation from the effect of the bright- 
ness of the background. In short, in surveying the literature, 
one can scarcely help but feel that the study of the influence of the 
surrounding field has been neither analytic nor systematic. 
4. The General Illumination. 
The effect of the general illumination of the retina on color 
sensitivity has been recognized since the time of Purkinje and 
Aubert. It has been studied in some detail by a number of 
experimenters, among whom may be mentioned Kramer and 
Wolff berg. Both have shown that the sensation aroused by the 
colored stimulus is weakened by a reduction of the general 
103 Fernald, G. M. The Effect of Achromatic Conditions on the Color 
Phenomena of Peripheral Vision. Psychol. Rev. Monog., 1909, X, No. 42. 58 
GERTRUDE RAND 
illumination, but neither, it may be mentioned, has given a method 
of keeping the general illumination constant. Kramer's104 pur- 
pose was to determine the sensitivity of the eye under different 
intensities of daylight and artificial illumination. His method 
was as follows. Stimuli, 4 mm. square, of blue, yellow, red, and 
green paper on a black background were used. The distance at 
which the stimulus had to be placed from the observer to be 
just recognized as colored, was tested by sunlight and when 
the sky was obscured by clouds and for three intensities of each 
of the following sources of artificial illumination: candle light, 
gas, petroleum, sodium, potassium, strontium, and calcium lights. 
His results may be summarized as follows: (1) Red is seen 
at the greatest distance in all lights except calcium, in which 
case green is seen when farther away than red. The other 
colors are recognized in the order green, yellow, blue. (2) All 
the colors are recognized at a greater distance when seen by sun- 
light than when illuminated by artificial light or the dull light 
from a clouded sky. (3) As the intensity of the artificial 
illumination is decreased, the colors must be placed nearer the 
eye to be recognized. Kramer's method of working, however, 
may be criticized because he ignored the white contrast which 
the black background induced across the stimuli. The induction 
across the stimuli whose sizes were only 4 mm. square must 
have been considerable. It was, moreover, of different amounts 
in each case; because brightness contrast is greatest when there is 
maximal brightness opposition. The modification of the light 
colors, as a result of contrast induction, must, therefore, have 
been greater than that of the dark colors. 
Wolffberg's105 interest was in the influence of gradual alter- 
ations of the general illumination on the light and the color 
sensitivity of the central and of the peripheral retina. His 
room was illuminated by daylight entering through a window. 
Fifteen different degrees of illumination were produced by fasten- 
ing from one to fifteen thicknesses of tissue-paper over the win- 
1M Kramer, J. Untersuchungen liber die Abhangigkeit der Farbenempfindung 
von der Art und dem Grade der Beleuchtung. Inaug. Diss., Marburg, 1882. 
105 Wolffberg. Ueber die Priifung des Lichtsinnes. A. f. O., 1887, XXXI., 
pp. 1-78. HISTORICAL AND CRITICAL 
59 
dow. The illumination obtained when the window was uncovered 
was called 15/15; when covered with one thickness of tissue- 
paper, 14/15, etc. His method of determining the effect of varia- 
tions of illumination upon the central retina was as follows: Pig- 
ment stimuli were placed at a standard distance of 5 meters from 
the observer, and the size of stimulus necessary to render it just 
visible in its true color was determined. In the peripheral 
retina, he investigated to what extent the limits of white and 
of colored stimuli were altered by reducing the illumination. In 
all his experiments, the stimuli were fastened on a black back- 
ground. Wolffberg's results for the central retina are shown 
in the following table. The stimuli were circular in shape and of 
diameters given in columns 2, 3, 4, 5, and 6. 
Illumination 
Size of 
Red Stimulus 
Size of Blue 
Size of Green 
Size of Yellow 
Size of White 
15/15 
.5 mm. 
3 mm. 
3 mm. 
1.5 mm. 
.2 mm. 
I4/I5 
1-5 
5 
4 
2 
•5 
13/15 
2 
6 
6 
4 
1. 
12/15 
2-5 
12 
12 
4-5 
2 
11/15 
3 
20 
20 
5 
2-5 
5/15 
10 
5° 
5° 
10 
6 
These results show that in the central retina a decrease of illu- 
mination has a greater effect upon the sensation of color than 
upon the sensation of white. Wolff berg next tested the effect 
of a gradual decrease of illumination upon the limits of sensitivity 
to white and to the colors. He found that the extent of the 
visual field was not narrowed for white when the illumination 
was decreased to 1/15. The color limits, however, narrowed 
gradually when the illumination was decreased from 15/15 to 
3/15. The narrowing was in no case more than 150. The 
relative extents of the fields remained unaltered, that is, the 
order of size was in every case blue, red, and green. 
Although special investigations have been conducted by Kra- 
mer, Wolffberg and others to show the effect of changes in the 
general illumination upon color sensitivity, in general little if 
any precautions have been taken by earlier experimenters to 
prevent such changes when investigating color sensitivity. 
Either the experimenter has not considered the influence of the 60 
GERTRUDE RAND 
general illumination, or he has been satisfied to take the rough 
precaution to work only on bright days at stated hours. Ole 
Bull,106 for example, commented at length on the factor of 
general illumination, but suggested no method for its standardiza- 
tion. He writes: "The amount and nature of the general illumi- 
nation are of more significance in perimetrical observations than 
one is accustomed to consider. It must always be noted whether 
the sky is clear or cloudy, whether it rains or snows. The 
extreme limits of the visual field for mixed light undergo such 
wide fluctuations that it is of little value to establish an average 
limit on the basis of a number of measurements. Changing 
illumination, conditioned by the time of day and of year during 
which the work is carried on, as well as the locality in which 
it is undertaken, produce variations in the same stimulus large 
enough to cause differences of from io° to 200 [in the limit of 
sensitivity]. Especially in the nasal parts of the retina does 
the illumination influence the color limits, while their position 
remains more constant in the temporal retina." Fernald,107 
however, did make some attempt to obtain a standard illumina- 
tion. She arranged white curtains at the windows of her optics- 
room which could be lowered on bright days and drawn on dark 
days. This rather crude method was used also by Thompson 
and Gordon.108 It is scarcely necessary to point out that the 
method lacks the first essential of standardization, namely, a 
means of measuring. 
It is surprising that Wolff berg as the logical corollary of his 
work, did not draw attention to the importance of standardizing 
the illumination of the visual field in all work on the color sensi- 
tivity of the retina, and show how it could be accomplished by a 
modification of his method of working. He already had at hand 
one of the essentials for standardizing, namely, a method of 
changing the illumination of his room. The other essential, a 
method of measurement by means of which an illumination could 
be identified with a previous illumination chosen as standard, 
100 Ole Bull. Perimetrie. Bonn, 1895, p. 8. 
107 Fernald, G. M. Psychol. Rev., 1905, XII, p. 392. 
108 Thompson and Gordon. A Study of After-images on the Peripheral 
Retina. Psychol. Rev., 1907, XIV, p. 122. HISTORICAL AND CRITICAL 
61 
might have been derived from his results. For example, it 
would seem to have been a simple matter for him to have chosen 
as standard the particular illumination at which the red stimulus 
of 2.5 mm. diameter, the blue and green of 12 mm. each, the yel- 
low of 4.5 mm., and the white of 2 mm. were just recognizable 
at a distance of 5 m. Stimuli of those sizes, it will be seen from 
the tables, were just recognizable at this distance at the illumina- 
tion called 12/15, when 15/15 represents the illumination "bei 
giinstige Tagesbeleuchtung." Using this condition as an index 
of the standard illumination, he could at any time have adjusted 
the illumination of the room by adding to or subtracting from 
the layers of tissue-paper covering the window, until the stimuli 
of these sizes were again just recognizable at the given distance. 
The accuracy and sensitivity of this method could have been 
tested by comparing the results of a series of determinations. 
An accurate and highly sensitive method sustaining some simi- 
larity in principle to the method suggested here, will be described 
by the writer in the experimental part of this paper. 
The influence of changes in the intensity of the general illumination 
upon visual acuity has received somle attention from physiologists and 
oculists. Although their work has no direct bearing on the influence of 
change of illumination upon color sensitivity, it may be of interest to note 
briefly their methods of dealing with these changes. 
Schweigger109 in 1876, using the Snellen series of optotypes and the 
N 
formula V = in which n represents the distance of the test-object from 
n 
the eye of the observer, and N the number of the series of the smallest of 
optotype series that can be recognized at that distance, found that on 
a clear day his visual acuity equalled 20/15, on a cloudy day it equalled 
30/15. To correct for the errors in visual acuity introduced by changes 
in the illumination he first found the number of the series of the smallest 
optotypes that he himself could read at a given distance, then he de- 
termined this Value for the patient at the same distance. Using his own 
N . . . . 
results V = - as standard, he determined the ratio of the patient s results 
n 
N1 N1 
V -■ to his own. This ratio -, he considers the expression of what the 
n N 
patient's visual acuity would be at standard illumination. 
Cohn's and von Hoffman's interests lay mainly in testing the eyes of 
schoolchildren and in determining what was the lowest intensity of il- 
lumination of the schoolroom suitable for work. Cohn110 in 1867 and 1883 
109 Schweigger, E. Sehproben. Berlin, 1876, Preface, pp. III-IV. 
110 Cohn, H. Untersuchungen der Augen von 10060 Schulkindern. Leipzig, 
1876, p. 101; Hygiene of the Eye in Schools, translated by Turnbull, 1883, p. 131. 62 
GERTRUDE RAND 
claims that as there is no photometer available for the measurement of the 
intensity of daylight, the eye must be its own photometer. Later in 1892111 
he states that L. Weber has made a daylight photometer, but as this 
apparatus is difficult of access, he would recommend apparently that the 
changes in visual acuity experienced by the eye with changes of illumination 
be used as a means of identifying a given degree of illumination. He 
endorses von Hoffmian's112 method of accomplishing this. According to this 
method, Type No. 30 of the Snellen optotypes is placed in the schoolroom 15 
feet from the eyes of a child whose visual acuity is 15/15. If the child recog- 
nizes the letters of the test, the room is sufficiently well-lighted. Work in 
the room is to be suspended as soon as the child can no longer recognize 
the letters of the test. This provided a practical method, not for measuring 
the illumination of a room, but for detecting when a room has insufficient 
light for purposes of schoolwork. 
Nicati113 tested the influence of change of the intensity of artificial il- 
lumination upon visual acuity. His work was purely quantitative. He pro- 
poses a unit of measure by means of which to study this effect. This 
unit he calls a photo. A photo is the smallest intensity of light which when 
placed 1 meter from a test-object printed in black on a white card gives to 
nornnal monocular vision a normal acuity. The method of measuring the 
intensity of an illumination in photos is as follows. A source of light is 
brought towaids the test-object until the observer has normal acuity. The 
intensity of the source then equals as many photos as the square of the 
distance of the light-source from the test-object, measured in meters. Nicati 
finds that there is an absolute logarithmic relation between visual acuity 
and intensity of illumination. As visual acuity is decreased in arithmetical 
series, intensity of illumination decreases in geometrical series. His table 
showing this relation is as follows: 
Visual Acuity 
i 1 
•9 
.8 
•7 
.6 
•5 
• 4 
Distance of source 
iM | 
Intensity in photos. 
1 
7 
1 
1 
J 
1 
1 
A 
Since this relation exists, either the intensity of illumination can be con- 
considered a measure of visual acuity, or visual acuity can be considered a 
measure of the intensity of illumination. That is, the scale of visual acuity 
is a photometric scale and can be used as such. To measure the illumination 
of a room in photos, then, the visual acuity should be determined in different 
portions of the room, and the average of the photos corresponding to these 
values in the acuity scale be taken as the measure of the illumination of the 
room in photos. The method as formulated is apparently serviceable as a 
means of estimating the illumination of a room chiefly, if not entirely, when 
it is below what is needed for normal acuity of vision. 
111 Cohn, H. Lehrbuch der Hygiene des Auges. Wien und Leipzig, 1892, 
PP- 34-35- 
112 von Hoffman, H. Augenuntersuchungen in vier Wiesbadener Schulen. 
Klin. Monatsbl. f. Augenheilk., 1873, pp. 289-290. 
113 Nicati, W. Physiologic Oculaire humaine et comparee normale et 
pathologique. 1909, p. i63ff. HISTORICAL AND CRITICAL 
63 
B. METHODS OF STANDARDIZING THESE FACTORS. 
i. Size of the Stimulus. 
No special method of standardizing the size of the stimulus is 
required. Each experimenter who has recognized it as a factor 
has chosen what he considered the most favorable area to work 
with, and has used that area in all of his comparative determina- 
tions. No limitations as to area have been prescribed other 
than that it must not be too large nor too small. 
2. Intensity of the Stimulus. 
As a general introduction to our discussion of the methods 
that have been used to standardize the influence of intensity, we 
wish to call attention once more to the fact that in no one of 
these investigations have the stimuli employed been equated 
with regard to the energy of the light given to the eye, nor have 
they been standardized in terms of any fixed unit of intensity 
that can be compared; and that until this is done, we have no 
proper means of determining the comparative limits of sensitivity 
to the different colors; nor of determining and expressing the 
comparative limen or j. n. d. of sensitivity. For example, 
attempts to determine the relative sensitivity of the retina to the 
four principal colors have been made among others by Aubert, 
Chodin, Raehlmann, Butz, Lam ansky, and Dobrowolsky: also 
by Bull, Hess, Hegg, and Baird. As stated above, however, 
these experimenters did not equate their stimuli with regard to 
the energy of light given to the eye for the investigations of the 
limits of sensitivity nor estimate sensitivity in terms of a com- 
mon unit for the work on limens and j. n. d.'s of sensitivity. 
The last four, however, did attempt to equate in intensity, but 
the equation was made in terms of a subjective measure arbitra- 
rily selected, namely, the proportion in which the pairs of an- 
tagonistic colors must be combined to produce gray, or, in terms 
of the Hering theory, to cancel each other. 
At this point, two general criticisms may be passed on this 
method of equating, (a) The stimuli should not be equated 
in terms of any subjective measure if one is to test the compara- 
tive sensitivity of the retina to the different colors. This begs 64 
GERTRUDE RAND 
the question at the outset. If, for example, a direct judgment of 
the intensity of the sensations aroused by two stimuli either at 
the limen or higher in the intensity scale be taken as the criterion 
for determining their equality, the method begins by making the 
stimuli of such physical intensity that they are sensed equally. 
In no fashion could the comparative sensitivity of the retina 
to the colors in question be determined by such stimuli. Suppose, 
for example, the limits of sensitivity of the peripheral retina 
were to be investigated, and for this, the stimuli which had been 
made subjectively equal for the central retina were used, the 
results obtained would not at all express the comparative limits 
for the colors in question. If these limits should be found to 
coincide, the conclusion could not be drawn either that the 
sensitivity of the retina to these colors extended only to this 
point, or that there was equal sensitivity at this point to the 
colors used. At most no more could be said than that approxi- 
mately the same ratio of sensitivity to the two colors obtained 
in this region that was present at the point in the central retina 
for which they were equated; but this ratio may not be a 1:1 
ratio. In fact, the investigator who gets his limits to coincide 
with stimuli so equated finds himself in the somewhat ludicrous 
position of having made his conditions such that the limits 
could not help but coincide, regardless of whether they actually 
ought to do so or not. He is not working with real limits, but 
with limits arbitrarily established, and the coincidence he finds 
is not a fact, at least not so far as he is able to determine by 
his method of working, but an artifact. To illustrate, the retina 
might be much more sensitive to red than to green; but if the 
red stimulus were reduced in physical intensity until the sensation 
it aroused was equal in intensity to the sensation aroused by the 
green, it is obvious that the comparative sensitivity to these colors 
could not be directly tested at any point. The limits of the red 
zone so determined might, for example, coincide with the limits 
of the green zone, although the extent of the red zone would 
have been much wider had stimuli of equal physical intensity 
been used. So much may be said for this as a type of subjective 
measure of equality, and what is said in criticism of its use for HISTORICAL AND CRITICAL 
65 
investigating the comparative sensitivity of the retina has a 
general application to all subjective measures. 
(b) The stimuli, especially should not be equated in terms 
of the cancelling power of antagonistic colors.114 This method 
is analomous. One scarcely knows what it does accomplish. On 
the one hand, stimuli so equated are in no way equated in physical 
intensity; and on the other hand, it would be the merest assump- 
tion to say that they have equal power to arouse sensation. To 
demonstrate this, let us compare certain colors with regard to 
their comparative power to arouse sensation and to cancel each 
other. There are three ways by which we may judge the power 
of a color to arouse sensation:- (a) the value of its limen when 
mixed with a gray of equal brightness; (&) the value of its 
j. n. d. at different points in the intensity scale; and (c) the 
direct or introspective judgment of its intensity. Estimated 
in all of these ways, spectral red, as prepared in pigment colors 
by Maxwell after Helmholtz, has greater power to arouse sen- 
sation than the green of the series. And yet on the color-mixer, 
it requires 240 0 of this red to cancel or • obliterate all trace of 
1200 of the green. And if sufficient blue be added to the mixture 
to give gray, the proportions of red and green are 165° and 
1150 respectively. Thus, whether the result of the combina- 
tion is yellow or gray, the red stimulus, although it has the 
greater power to arouse sensation, is required to be in consider- 
able excess of the green. We regret that we can not make a 
similar comparison for the blue and yellow of the Maxwell 
series, because we were unable to procure them in season for 
this work. But using the blue and the yellow of the Hering 
pigment papers, we find that this blue has a lower limen in a 
gray of its own brightness than the yellow has in a gray of its 
brightness, and that introspectively it is judged much more 
saturated than the yellow. And still at the general illumination 
we have chosen as our standard, 2000 of blue are required to 
cancel 1600 of yellow.115 
114 There are problems in the optics of color in which subjective equations 
of intensity may be desired. For a description of what the writer considers 
a proper method of making the subjective equation, see this paper p. 83, 
footnote. 
115 If the illumination is decreased, we find that a different proportion 66 
GERTRUDE RAND 
In a later paper, we shall show that if the complementary 
colors can be assumed to cancel each other, because a certain 
amount of the one when combined with a certain amount of the 
other, kills it for sensation, then, by the same token, the non- 
complementary colors may be assumed to exercise a degree of this 
cancelling action. For a definite and considerable amount of 
each must be mixed with any other before it is sensed, the 
amount required varying over a wide range when one of them 
is combined with each of the others in turn. We may draw 
upon the non-complementary colors, then, for a further demon- 
stration of our thesis that the cancelling power of colors upon 
each other is no measure of their power to arouse sensation. 
A notable instance of this is the combination of red and yellow 
to give orange. Working with the Hering pigments, we find 
that the standard orange of the series, judged as sensation, 
seems to be equally red and yellow. The standard red of the 
series, however, which is chosen to form the orange, has greater 
power to arouse sensation than the yellow by all of the tests 
mentioned above. Still 295 ° of red are required to be combined 
with 65° of the yellow to give the orange which as we have 
said is equally red and yellow as sensation. Orange furnishes 
us with only one instance of the non-equivalence of cancelling 
power to sensation-arousing power that may be found among 
the combinations of non-complementary colors.116 
We must, then, conclude that even if one were to err so pro- 
foundly as to choose a subjective measure for equating the 
intensity of his stimuli for an investigation of the comparative 
is required to produce cancellation. This difference depends upon the fact 
that colors do not lose their saturation with equal rapidity with decrease of 
illumination. The dependence of cancelling proportions upon the general 
illumination may be pointed out as a minor source of error in the use that 
has been made of this method by investigators who have conducted their 
experiments in daylight, for they did not work at an invariable or standard 
illumination. 
u" In the above demonstration that cancelling power can not be taken as 
the equivalent of sensation-arousing power, we have assumed that we have 
not left our work open to criticism in the use of pigment instead of spectral 
colors, or even in passing from one series of pigment colors to another, 
because in each case, our test of the cancelling power and of the sensation- 
arousing power was made with the same stimuli. HISTORICAL AND CRITICAL 
67 
sensitivity of the retina to the different colors, this measure should 
not be selected in preference to the power to arouse sensation 
as determined either by limen tests, by direct introspective judg- 
ments of intensities, nor by the j. n. d. method described on p. 83 
footnote. One can only surmise the following reasons for 
the selection, (a) The limen and the just noticeable difference 
tests, either of which is a proper method of estimating the power 
of a stimulus to' arouse sensation, were probably not taken into 
consideration at all. The alternative test, the introspective judg- 
ment of equal intensities, is difficult to make. And the method 
given on p. 83 footnote, which is preferable to any method 
known to the writer for subjectively equating stimuli of all de- 
grees of intensity, has never been suggested even in principle 
prior to the publication of this paper. It may have been thought 
that equality based on cancelling power could be substituted for 
the equality as determined by the other methods. 
The reason for selection of equation in terms of cancelling power is 
not stated by any one of the four men who has made this selection. Con- 
sidering the statement of each in turn, we find that Bull who wished to obtain 
color stimuli of equal saturation and brightness, merely states that he 
does this by establishing pairs of colors such that equal amounts of each 
cancel and give a gray that conforms to the standard gray he has chosen 
(A. f. O., 1881, XXVII., p. 95). Hess, aiming to obtain a red light such that 
its red Valens, or, as he says, its capacity to arouse red sensation, is 
equal to the green Valens of a green light, writes: "Ais Nachstliegende 
erscheint es nun, einem roth- und einem griin-wirkenden Pigmente den 
gleichen Roth und Griinwerthe dann zuzuschreiben, wenn dieselben zu 
gleichen Theilen eine farblose Mischung geben" (A. f O., 1889, XXXV., 
p. 39). Hegg wishing to equalize the color-values of his stimuli, claims 
that this is possible for only the two members of each pair of antagonistic 
colors. Red can be made equal in color-value to green, he says, or yellow to 
blue, but green and blue can not be equalized. He continues: "Wir 
betrachten ein Roth und ein Grun als chromatisch aquivalent, wenn sie auf 
der Rotationssoheibe zu gleichen Theilen gemischt sich gegenseitig total 
aufheben, so d'ass eine Mischung entsteht, welche weder ins rothliche 
noch griinliche sticht." (A. f. O., 1892, XXXVIII., p. 149). Baird offers no 
reason whatever for selecting the method of cancelling power by means of 
which to equate color-values. He apparently takes for granted that this 
method is the only one, and does no more than describe how the method 
was applied to his particular stimuli. 
The search for a reason for this selection may, however, be pushed back 
a little further. We find that in 1880 before any of the work mentioned 
above was published, Hering (Lotos. Jahrbuch fur Naturwissenschaft, 68 
GERTRUDE RAND 
1880, I, pp. 76-107) had made statements from which we can conclude that 
in his opinion two antagonistic colors of equal sensation-arousing power 
cancel. Hering's statements leading to this conclusion are as follows: 
(a) "Die Vermogen der Lichtstrahlen, die weisse Empfindung zu fordern, 
will ich die weisse Valenz der Lichtstrahlen nenne." (p. 79.) (While this 
definition is not specifically repeated for colored light, still it is obvious from 
the text that it applies to colored as well as to white light).117 
117 The following quotations are appended to show the use of Valenz by 
Hering and Hess: 
Hering (Zur Erklarung der Farbenblindheit aus der Theorie der Gegen- 
farben. Lotos-Jahrbuch fiir Naturwissenschaft, 1880, I., p. 76-107) writes 
(p. 79) : "Die Vermogen der Lichtstrahlen, die weisse Empfindung zu 
fordern, will ich die weisse Valenz der Lichtstrahlen nennen. 
"Die Grosse dieser Valenz ist offenbar von zwei Factoren abhangig: ersten 
von der objectiven Intensitat oder lebendigen Kraft, mit welchen die 
Strahlen verschiedener Wellenlange bis zur empfindlichen Netzhautschicht 
gelangen, und zweitens von dem, was wir die specifische weisse Erregbarkeit 
des Sehorgans gegeniiber den Strahlen verschiedenen Wellenlange nennen, 
d.i. das Vermogen dieses Organs, unter dem Einflusse jener Strahlen die 
Weissempfindung deutlicher werden lassen. 
"Ausser der weissen Valenz, welche alien Lichtstrahlen gemeinsam ist, 
kommen nun den einzelnen Strahlenarten verschiedene farbige Valenzen zu. 
Allen Strahlen vom aussersten Roth oder vom Anfange des Spectrums bis 
zu jenem im Tone reinen Grim, welches eine Griindfarbe ist und welches wir 
das Urgriin nennen wollen, haben eine gelbe, alien Strahlen vom Urgriin bis 
zum violetten Ende des Spectrums eine blaue Valenz." 
Hess (Ueber den Farbensinn bei indirectem Sehen, A. f. O., 1889, XXXV., 
pp. 1-60.) writes (p. 30) : "Unter weisser Valenz eines farbigen homogenen 
oder zusammengesetzten Lichtes versteht Hering den Helligkeitswerth 
desselben fiir eine Netzhautstelle, welche das farbige Licht wegen mangelhaften 
Farbensinnes oder aus anderen Griinden farblos sieht" (p. 39) : "Um uber 
das gegenseitige Verhaltniss der Abnahme der Empfindungsvermogens fur 
Roth und Grun, resp. Blau und Gelb uberhaupt Untersuchungen anstellen 
zu konnen, ist es zunachst erfbrderlich, fiir beide Arten des Empfindungs- 
vermogens ein gemeinsames Maass zu finden. Verschiedene griine Lichter 
besitzten die Fahigkeit, griine Empfindung zu erzeugen, in sehr verschiedenem 
Maasse, sie sind, urn es kurz zu bezeichnen, sehr verschieden griinwirkend. 
Das mehr oder minder grosse Vermogen eines Lichtes, griin zu wirken, 
bezeichnen wir mit Hering als die griine Valenz oder den Griinwerth des 
beziiglichen Lichtes". (p. 40) : "Bestimmen und messen lasst sich derselbe 
nur in Bezug auf ein als Normalgriin gewahltes Pigment, welches unter 
genau denselben Beleuchtungsverhaltnissen wie das zu untersuchende gesehen 
wird. Ganz analoges gilt von dem Roth-, Gelb-, und Blauwerthe eines 
Pigmentes. 
"Fiir die vorliegende Frage handelt es sich aber nicht bloss darum die 
Griinwerthe oder die Rothwerthe verschiedenen Pigmente je unter sich 
zu vergleichen, sondern den Griinwerthe eines griinwirkenden mit dem 
Rothwerthe eines rothwirkenden Pigmentes. 
"Als das Nachstliegende erscheint es nun, einem roth- und einem griinwirken- HISTORICAL AND CRITICAL 
69 
(b) "Zwei homogenen Lichter, nun, von welchen das eine ebenso gelb (oder 
roth) wirkt, und das andere blau (oder griin) so dass beide Valenzen sich 
aufheben, nenne ich gegenfarbig aquivalent" (pp. 83-84). In the first of 
these statements he directly calls the capacity of a color to arouse sensation 
its Valenz. And from the second it may readily be derived that when the 
yellow-sense, for example, is affected as strongly by yellow light as the blue- 
sense is affected by blue light, complete cancellation will ensue,-that is, 
equality in cancelling power may be considered as the equivalent of equality 
in capacity to arouse sensation. In making this deduction we have of course 
assumed that wirkt refers to sensation-arousing action and not to cancelling 
action. We have no doubt that this assumption is correct, still it may be 
wrorth while to bring forward direct evidence in support of this point from 
a statement made by Hess while working under Hering's direction. Hess 
writes: "Das mehr oder minder grosse Vermogen eines Lichtes griin zu 
wirken, bezeichnen wir mit Hering als die grime Valenz oder den Griin- 
werth des beziiglichen Lichtes" (op. cit., p. 39). Here Vermogen griin zu 
wirken is made the equivalent of Valenz and Valenz by definition is the 
capacity of a color to arouse sensation. Hence we have little hesitation in 
assuming that in the case in question wirkt also refers to the sensation- 
arousing action of the colored light and not to its cancelling action, and in 
concluding, therefore, that Hering believed that antagonistic colors of equal 
power to arouse sensation would also have equal power to cancel each other. 
Since this is true, it is probable that the followers of Hering (Hess and Hegg) 
assumed the equivalence of power to arouse sensation and power to cancel 
and equated their stimuli accordingly. That Hess was actuated by some 
such reason is shown by a statement made by him in his discussion of this 
point. He Writes: Die von Herrn Professor Hering angegebene, oben 
geschriebene Vntersuchungsweise gestattet mit grosse Genauigkeit den zu 
vergleichenden Pigmenten gleich grosse farbige und gleich grosse weisse 
Valenz zu geben, sie ermoglicht es, fur die Werthigkeit der Farben einen 
genauen numerischen Ausdruck zu winnen und in die Rechnung einzufuhren" 
(op. cit., p. 58). Hegg also seems to refer back to Hering, for he uses the 
Hering terminology in discussing the equation of his stimuli. 
Or (b} since cancellation is the corollary to the assumption of 
an assimilation-dissimilation mechanism, it may have been con- 
sidered for some reason, not readily understood by the writer, 
that an equation based upon it is the proper one to make. 
Having said this much about the impropriety of selecting a 
subjective measure for the intensity equation of stimuli, let us 
den Pigmente den gleichen Roth-und Grunwerth dann zuzuschreiben, wenn 
dieselben zu gleichen Theilen, z.B. auf dem Kreisel gemischt eine farblose 
Mischung geben, im Faile sie dazu aber in einem anderen Verhaltnisse gemischt 
werden miissen, anzunehmen, dass sich der Rothwerthe des einen zum Griin- 
werthe des anderen umgekehrt verhalt wie die Grosse der beiden zur Her- 
stellung einer farbosen Mischung nothigen Sectoren." 70 
GERTRUDE RAND 
pass to a resume of the attempts that have been made to apply 
this measure by Bull, Hess, Hegg, and Baird. Hegg selected 
four stimuli that suffered no' alteration of color tone in passing 
from the center to the periphery.118 These were a bluish-red, a 
bluish-green, a blue, and a yellow. They were equated in pairs, 
the bluish-red to the bluish-green, and the blue to the yellow, as 
follows. It was determined in what proportions the members 
of each pair had to be combined to produce gray, and from these 
proportions, values of the sectors of the stimulus disc were cal- 
culated for each color. The procedures of Bull and Hess were 
essentially similar.. 
Baird, employing the light transmitted by gelatines, prepared 
blue, yellow, red, and green stimuli as follows. A lantern con- 
taining an incandescent lamp of 16 candle-power was used as 
source of light. The stimulus light was emitted from the lan- 
tern through a circular aperture, 15 mm. in diameter. Gelatines 
were placed over the aperture in combinations which gave the 
four stable colors, and their spectral values were obtained. A 
disc in which two windows of equal size had been cut, was rotated 
on a motor in front of the lantern. The combination of gela- 
tines to give the red stimulus was fastened across one of the 
windows, while the green combination was used to cover the 
other window. As the windows were of equal size, the rotation 
of the disc gave a mixture which contained equal proportions of 
both stimuli. The gelatine combinations were changed by adding, 
subtracting, or substituting until the mixture showed no trace 
of color. Similar equations were obtained for the blue and the 
yellow stimuli. 
It will be seen from the work of these men that even if their 
methods had been based upon a proper principle of equating, 
they would not be adequate for all that is involved in the problem 
118 Only one meridian was used for determining this invariability of color 
tone. It is obvious that a conclusion should not be drawn from such a 
scant investigation of the sensitivity of the retina. For example, working with 
the red, green, blue, and yellow of the Hering standard papers, the writer 
has found that with a careful standardization of factors, an investigation 
in any considerable number of meridians shows that stability of tone is 
possessed by the blue alone. HISTORICAL AND CRITICAL 
71 
of determining the comparative sensitivity of the retina to the 
different colors. For not only is the comparative sensitivity to 
the complementary colors desired, but to the non-complementary 
colors as well. The method offers no possibility, for example, of 
equating red and green to blue and yellow. One can only con- 
jecture how much of our present conception of the comparative 
extent of the different zones of color sensitivity is an artifact 
due to the use of stimuli that have not been equated with refer- 
ence to the energy of the light-waves they give to the eye. In 
addition, then, to the objection that the methods that have been 
used thus far to equate the color stimuli in intensity are found 
to be essentially wrong in principle, the further criticism may 
be offered that they are not adequate in scope. An energy equa- 
tion of the light-waves by means of some radiometric device, 
for example, the thermopile, the bolometer, the selenium cell, or 
what not, alone seems adequate to the requirements set by the 
problem of determining the comparative sensitivity of the retina 
to the different colors, or the comparative limits of the zones of 
sensitivity. 
Energy equations in terms of radiometric units have been 
made by Langley and Pfund, but up to this time no investigation 
of color sensitivity has been made with colors equalized in energy. 
Langley119 invented the bolometer and determined by means of 
it the relative distribution of energy in the normal spectrum. In 
order to equalize the energy of the different colors, he states that 
one may vary the width of the collimator-slit until equal radio- 
metric readings are obtained. In his own experiments on visual 
acuity, he does not, however, proceed in this way. Tables of 
logarithms were illuminated in a dark-room by monochromatic 
light representing known amounts of energy. The greatest 
distance at which the figures could be read was determined for 
each of the colors, and corrections were applied for inequalities 
in the energy of the different lights. The corrections were made 
in terms of the distribution obtaining in the following table. 
Pfund used the first method suggested by Langley. In an 
"'Langley. Energy and Vision. Amer. Journ of Science, 1888, XXXVI., 
3rd Ser., pp. 359-379- 72 
GERTRUDE RAND 
investigation of the changes in the resistance of selenium to 
lights of varying wave-length, he employed differently colored 
Wave-length 
Heat 
M -35 
1.8 
M .38 
5-3 
M -45 
11.9 
M. 50 
17-3 
-55 
20.7 
.60 
21.9 
M .65 
22.2 
M .70 
21.4 
M -75 
20.7 
.768 
20.2 
beams of equal intensity. The intensity equations were made as 
follows. Using first a Rubens thermopile120 and later a radio- 
micrometer,121 Pfund determined which wave-length gave the 
least galvanometer deflection. He then reduced the more intense 
beams by interposing a smoked wedge of the proper thickness 
until every portion of the spectrum produced the same deflection. 
In this way he obtained colored lights of known and constant 
energy. 
Psychological investigators have been slow to recognize the 
importance of standardization of intensity in radiometric terms 
of the colors which are to be used for the investigation of sensi- 
tivity. The only equations of intensity have been made in sub- 
jective terms, a procedure which if done by a proper method may 
be legitimate for work on certain points relative to existing color 
theories, but which is not adequate (see this paper, pp. 64-65) to 
meet the requirements of the problems which deal with the 
comparative sensitivity of the retina to the different colors. 
Note.-Since the completion of this paper, the report of Watson and 
Yerkes concerning methods of studying vision in animals has been published 
(Behavior Monographs, 1911, I, pp. 1-89). For the measurement of the 
intensity of the stimulus they find two methods available, photometry and 
radiometry. They write: "The method of photometry in all its forms is 
dependent upon the visual capacity, training, and the special skill of the 
observer who attempts to use it. For this reason, and others only less 
important, it is usually desirable to supplement photometric measurements 
of photic stimuli by measurements of their value in terms of energy. Hence 
the pertinence of physical measurements. Determination of the value 
of photic stimuli in terms of heat units by radiometric procedure has 
proved feasible. Radiometry yields a measurement which is relatively inde- 
pendent of the visual peculiarities of the observer, and it therefore supple- 
ments in an invaluable manner the results of photometry" (p. 11). 
The authors in question then decide in favor of radiometric measurement 
ia> Pfund, A. A Study of the Selenium Cell. Philos. Mag., 1904, VII, 
Ser. 6, p. 26. 
121 Pfund, A. The Electrical and Optical Properties of Metallic Selenium. 
Phys. Rev., 1909, XXVIII, p. 326. HISTORICAL AND CRITICAL 
73 
and control of the stimuli to be used in determining the animal's color sense. 
Their reasons for this decision are not, however, those stated in the above 
criticism of subjective methods of equation either by cancelling power or by 
sensation-arousing power, namely, that these methods are essentially wrong 
in principle for tests for the comparative sensitivity of the eye to different 
colors. That they do not consider them wrong in principle for work of 
this kind is shown in fact by their recommendation of the Hegg colored 
papers. The colors of the Hegg papers are equated in intensity in terms 
of the cancelling power of the complementary colors, the worst of the 
subjective methods discussed. They write: "These [the Hegg papers] 
are mixtures of oils on paper yielding the hues red, yellow, green, and 
blue. These hues are claimed to be equal in intensity and saturation for 
the human eye. The set is useful as a means of ascertaining, in a 
preliminary survey, whether an animal readily discriminates two hues which 
for us are of nearly the same intensity and saturation" (p. 32). It is 
obvious also that they do not consider the photometric method of equating 
intensities (also a subjective method) wrong in principle. The method 
is not recommended merely because it depends upon the visual capacity, 
training, and special skill of the observer. But the fact that they endorse 
this method to supplement the radiometric procedure or rather, as quoted 
above, the radiometric to supplement the photometric shows that they do 
not realize the absolute diversity of the photometric and of the radio- 
metric curve. Their conclusion, then, in favor of the method of radiometry 
for measuring the intensity of the stimulus is based upon very different 
arguments from those which have governed the similar decision reached 
in the above discussion. They do not seem to entertain any criticism of 
the subjective method of equating, either the method which measures 
cancelling power, or the method of photometry, nor do' they recommend 
that either be discarded. Their choice of energy measurement is due 
largely to the fact that they wish a method which is as free as possible from 
subjective errors. 
3. Brightness of the Stimulus 
The same investigators who sought to obtain stimuli of equal 
intensity, attempted also to equate these stimuli in brightness. 
This may be done in two ways: the white-values of the colors 
as they appear in direct vision may be equated, or the white- 
values as they appear in indirect vision may be equalized. The 
first method was used by Bull who made direct comparison judg- 
ments of the relative brightness of the colors, facilitating 
his comparisons by the use of intermediate color-tones. 
For example, a blue was changed in brightness until it 
appeared as light as a given blue-green. Green was then made 
equal to the blue-green; yellow-green to the green; yellow, to 
the yellow-green, etc. Hess, Hegg, and Baird employed the second GERTRUDE RAND 
74 
method. The stimuli were carried to a point in the field of the 
peripheral retina at which they appeared colorless and their 
brightness values were altered until the gray sensations obtained 
from all the stimuli were equal. 
Hegg, who used pigment colors, observed the stimulus through 
an opening in a gray screen, whose brightness could be altered by 
turning it toward or away from the source of light. He adjusted 
the screen so that its brightness was the same as that of the 
gray sensation aroused by the green stimulus in the peripheral 
retina. Retaining this setting of the screen, he replaced the 
green by the red stimulus, the intensity of which he had previ- 
ously equated to the intensity of the green by the method de- 
scribed and criticised in the preceding section. The red stimulus, 
which was composed of 216° of red, 550 of blue, 89° of white, 
when observed in the extreme periphery, was seen as a gray that 
was lighter than the screen. To make the stimulus and the screen 
of equal brightness, 5° of the white sector had to be replaced by 
black. A complication arose when blue and yellow were equated 
to this brightness, resulting from the changes in color-tone which 
took place. Hegg found that when he added white to lighten 
the blue stimulus, a sensation of reddish-blue was aroused. 
(Chodin, it will be remembered (see p. 45), saw in this fact 
an argument against the possibility of equating the brightness of 
colors for investigations of this kind.) To cancel this effect, he 
added green. The addition of black to yellow, which was neces- 
sary in order to equate the brightness of yellow to green, resulted 
in a greenish-yellow sensation. To this he added red in a suf- 
ficient amount to cancel the greenish appearance of the fusion.122 
1X1 In connection with a study (done in cooperation with Dr. C. E. 
Ferree) to determine the physiological level at which the fusion of colored 
with colorless light sensation takes place, the writer attempted to add suf- 
ficient red to cancel the green in a mixture of yellow and black. A 
curious paradox was observed. Starting with 550 of yellow, and 3050 
of black, and keeping these proportions relatively constant while red was 
being added to the mixture, it was reported by a number of observers 
that, after the addition of about io° of red, it was seen in the mixture with 
the green. As more was added, the green and red continued together in 
varying proportions, until, with about 450 of red in the mixture, it 
dominated the fusion, which was seen as a dark brownish-orange. Our ob- HISTORICAL AND CRITICAL 
75 
Hegg does not give the proportions of the final white-green-blue 
Urblau and the black-red-yellow Urgelb which, he claims, 
were equal in brightness to the Urroth and Urgrun. 
Baird also used the method of indirect vision comparison. The 
two stimuli to be equated were placed one above the other at a 
point at which both appeared colorless in the periphery of the 
retina. The brightness of blue was chosen as standard, and the 
red, green, and yellow stimuli were darkened to equal it by rotat- 
ing an episcotister in front of each of them in turn. The sectors 
of the episcotister were adjusted so that each stimulus was 
darkened as much as necessary to cause the colorless sensation 
aroused by it in the periphery to be the same as was aroused by 
the blue stimulus. Baird does not say that his work was compli- 
cated by changes in color-tone. His method would at first glance 
seem to be more simple than that of Hegg. When, however, we 
remember that the equation of brightness and apparent intensity 
had to be carried on hand in hand, we see some of the difficulties 
he must have encountered. His problem, was to bring the com- 
plementary colors to such intensity, that 1800 of one cancelled 
1800 of the other; and at the same time, to maintain them all of 
the brightness of the blue. But it is apparent that by his method 
of equating in brightness, an alteration in the amount of colored 
light coming to the eye is produced every time a change in 
brightness is made. And as the brightness of the several stimuli 
had to be changed by unequal amounts to bring them all to the 
brightness of blue, the amount of colored light coming to the 
eye was also changed by unequal amounts. This much of the 
procedure is sufficient to show the difficulty that confronts the 
experimenter. To equate either for brightness or cancelling 
power, disturbs the equation established for the other; that is, 
when the stimuli are brought to equal brightness, their cancelling 
power will no longer be equal, and vice versa. It is obvious that 
servation has been verified too many times and by too many observers 
for us to question its validity. It stands, then, in direct contradic- 
tion to Hegg's claim that a change in color-tone produced by altering the 
white-value of a color can be remedied by adding the complementary color 
to the stimulus. The difficulty then, seems insurmountable, and stands as 
one of the objections to the attempt to equate colors in brightness. 76 
GERTRUDE RAND 
the goal desired, if it can be attained at all, must be reached by a 
series of approximations; and that in the end the experimenter 
will have very much altered stimuli. Since to equate for both at 
once, involves making much more radical changes in the stimuli 
than to equate for one alone, it is plain that in doing both, we 
but add to the objections we have already made when each is 
done alone. 
It is to be deplored that Baird does not tell us just how he 
worked in this most difficult part of his technique. The defect is 
serious, for as the report of his method stands, one can neither 
pass judgment on its adequacy, nor be sufficiently guided by it, 
should one attempt to repeat the work. Of the technique that is 
described, however, the following criticism may be offered. In 
the equation of the brightness of two stimuli, Baird carried them 
to an angle of excentricity, at which both appeared colorless. 
Now, it is seen from his tables, that the limit of blue is some 150 
wider than that of green. He has, then, either to show that the 
brightness of green is the same at its limit as it is 150 peripheral- 
wards, or to equate the brightness of the colors at their individual 
limits by some means, such as the flicker method. 
Since Bull equated his stimuli by the direct vision method, and 
Hess, Hegg, and Baird by the indirect vision method, a 
word may be said in concluding this topic with regard to which 
is the proper method. Obviously, the decision rests upon whether 
or not the colors have different relative white-values at center and 
periphery. That they do has been reported among others by 
Tschermak,123 and we have been able to confirm this statement. 
The equation should, therefore, be made in the peripheral retina. 
As we shall show in the experimental section of our paper, how- 
ever, no equation should have been made by any of these men 
for the work they were doing, because unless the stimuli used 
are extremely weak in saturation, to equate in brightness for the 
investigation of the limits of sensitivity not only is unnecessary, 
but results in positive harm. If, however, in other work in the 
peripheral retina the need for equating should arise, the writer 
would urge not only that the equation should be made in the 
123 Tschermak, A. op. cit., pp. 564-575. HISTORICAL AND CRITICAL 
77 
peripheral retina, but that it should be obtained at the point at 
which the investigation is to be made. 
c. Summary. 
With regard to the attempts that have been made to standard- 
ize, the results of our historical survey are found tO' be largely 
destructive in character. They show, however, that a decided 
need for standardizing has been recognized. This in itself was 
a first step in the right direction. The following factors have 
been discussed: (a) The size of the stimulus. This factor has 
been the most adequately treated by previous investigators. Its 
influence as a factor has been shown, and with it the need of 
careful measurement of the actual size of the stimulus and of its 
apparent size as determined by its distance from the eye of the 
observer. There is still need, however, for further work. While 
it has been generally held that an increase in the area of the 
stimulus functions in some degree as the equivalent of an increase 
in intensity, and thus influences the limits and limens of color 
sensitivity, no quantitative estimate has been made of the degree 
of this equivalence. Exact knowledge of this point is not only 
of general interest in psychological optics, but it is needed in turn 
in certain problems of standardizing. For example, it is often 
required that the size of the stimulus be varied and its intensity 
for sensation be kept constant. This can be done only when the 
ratio of equivalence is known. As stated on p. 6, this ratio is 
now being worked out in this laboratory. The results will be 
reported later. 
(&) The intensity of the stimulus. The influence of the inten- 
sity of the stimulus upon color limens and color limits has been 
pointed out, but no adequate standard of measure has been 
employed. In dealing with the comparative sensitivity of the 
retina to the different colors, estimated in terms of the limits, it 
is obvious that equal amounts of light should be used. Estimated 
in terms of the limens, the amounts used should be determined 
in terms of units that can be compared. The problem of the 
measurement of these amounts of light is wholly physical, hence 
the standard of the physicist should be adopted. The determina- 
tion should be in terms of energy as measured by the bolometer, 78 
GERTRUDE RAND 
the thermopile, or other radiometric device. Only in this way 
so far as we know, can the retina's sensitivity to the different 
colored lights be obtained in terms of units that can be compared. 
(c) The brightness of the stimulus. Brightness and intensity 
have been much confused in the literature of the subject. The 
effect obtained by varying both factors has often been attributed 
to change in brightness alone. The effect of change in brightness 
has never been investigated in isolation. This factor, then, 
occupies the novel position of having been standardized for work 
on the limits of color sensitivity before the need for such control 
has been shown. 
{d} The preexposure. Only in a very general way has the 
effect of the brightness of the preexposure been recognized, and 
the precise reason for its influence has been very little understood. 
No quantitative estimate of the effect has been made, and no 
attempt at standardization has been undertaken which has shown 
any comprehensive knowledge of how the factor works. 
(?) The field surrounding the stimulus. Considerable atten- 
tion has been given to this factor. A small amount of qualitative 
work has been done, and some attempts have been made to 
secure control of the factor. More detailed knowledge, however, 
is needed of its influence, quantitative and qualitative, over a 
wider range of the retina. Especially should its relation to 
general illumination be studied. Until this relation is under- 
stood and some means is taken to render the general illumina- 
tion constant, no effective estimation of the influence of the 
brightness of the surrounding field can be obtained, nor can it 
be eliminated as a factor from the color observation. 
(/) The general illumination. The influence of the illumina- 
tion of the visual field on color sensitivity has been recognized and 
rough attempts have been made to determine the amount of this 
influence. The different ways in which changes in general illum- 
ination affect color sensitivity have not, however, been deter- 
mined, and the relative importance of each has not been estimated. 
Very little attempt at standardization has been made because the 
first essential of standardization, namely, a sensitive means of 
measurement, has not been had. III. EXPERIMENTAL. 
A. Purpose of Investigation. 
The purpose of this investigation includes the following points. 
(i) The color observation will be analyzed for the brightness 
factors that influence its results. (2) A systematic study will be 
made of these factors with special reference to the determination 
of their effect upon the color sensitivity of the retina and upon 
the limits of sensitivity to different colors. (3) It will be ascer- 
tained whether the effect of these factors can not be explained in 
terms of the action of brightness upon color in the peripheral 
retina and of the rapidity with which the sensitivity to color 
decreases from the fovea outwards. (4) Methods will be devised 
to standardize these factors in so far as our results show the 
need of standardization. No attempt will be made at this point 
to study the factors that pertain to the source of light with the 
following exception. Brightness will be isolated from intensity 
and the effect on the limits of sensitivity of changes in the bright- 
ness of the stimulus, made without altering the amount of colored 
light coming to the eye, will be determined in order to find out 
whether or not colors should be equated in brightness when the 
limits of sensitivity are investigated. Moreover, since our 
problem is concerned only with the brightness factors that influ- 
ence the action of the colored stimulus upon the retina, the 
writer will not feel obliged to concern herself with the standard- 
ization of her stimulus with regard to either quality or intensity 
any further than is needed to show the effect of the brightness 
factors upon the retina's response to these stimuli. All the 
standardization that is needed will be accomplished by using the 
same stimulus for all observations the results of which are to be 
compared; that is, no comparisons will be made except of the 
effect of the different brightness factors upon the same stimulus. 
For obtaining results so purely comparative the standardization 
afforded by pigment papers should be adequate, provided a 
standard illumination can be obtained so that the amount of 80 
GERTRUDE RAND 
colored light reflected from the pigments will be constant from 
test to test. Since we were able to secure a highly sensitive 
means of duplicating our illumination from observation to obser- 
vation, the standardization of the stimulus afforded by the Hering 
pigment papers has been considered adequate. More especially 
has this degree of standardization been considered adequate be- 
cause the results are to be used primarily merely as a guide in 
the formulation of a method of working. Having secured a 
method of working, however, that will permit of a close duplica- 
tion of results from observation to observation with the pigment 
papers, the writer will attempt to adapt the method to work in 
which the colors of the spectrum are used. In order to do this, 
the following requirements will have to be met. (1) A spectro- 
scope will have to be devised by means of which the retina can 
be stimulated at any degree of excentricity in any meridian that 
is desired, for example, a spectroscope that can be used in con- 
junction with the rotary campimeter1 in all its adjustrhents. Such 
a spectroscope having all the freedom of movements of its parts 
needed for use with the rotary campimeter has been devised in 
this laboratory and is now under construction. (2) In order 
that the stimulus-opening in our campimeter be filled with light 
sufficiently homogeneous for our purpose, a prism of high disper- 
sive power will have to be procured for use in our spectroscope. 
A compound prism of the Cassie type2 seems adequate for this 
requirement. Such a prism constructed to our special order is 
now being made for us in Germany. (3) In order that the light 
may undergo high dispersion and still be sufficiently intense for 
work in a room lighted to the degree that some phases of our 
problem demand, a source of light of high intrinsic brilliancy 
is needed. Constancy in candle-power should also be had. A 
high voltage Nernst filament seasoned for 100 hours or more 
and operated on a steady circuit will give, the writer believes, 
the intensity and constancy required. 
Having completed our work of standardizing the factors 
extraneous to the source of light, an attempt will next be made 
1 For a description of the rotary campimeter see this paper p. 87 ff. 
2Cassie. Philos. Mag., 1902, III. Ser 6., p. 449. EXPERIMENTAL 
81 
to secure a better control of the source. Standardization up to 
the present can be considered successful only with regard to the 
quality of light. No adequate work has been done on the 
standardization of the quantity of light for work on color 
sensitivity. As stated earlier in the paper, the writer believes 
that this can be done only by means of energy determinations. 
She expects to do her radiometric work by means of a surface 
thermopile (Coblentz model)3 and a DuBois-Rubens Panzer gal- 
vanometer, unless future results show that some other combin- 
ation of radiometer and galvanometer is more satisfactory. 
Finally from the work of standardization it is our hope to 
return to the investigation of the problems which we were in the 
beginning forced to abandon because the work could not be 
satisfactorily done by the methods now in use in the optics of 
color. A brief statement of the plan of our future work has 
already been given in an article published in conjunction with 
Dr. Ferree in the American Journal of Psychology.4 In order 
that the scope of this work be known at this point, and that the 
importance of the present investigation be understood in relation 
to this work, the statement is appended here. 
"About a year ago5 the writers undertook to determine the 
retina's sensitivity, relative and absolute, to colored light in 
terms of units that can be compared. Since several years will be 
required to complete this work, they have thought it best to pub- 
lish a preliminary note showing briefly the purpose and scope 
of the investigation. The following points will serve to indicate 
what is being attempted in this study. 
"(1) All measurements of sensitivity will be made in radio- 
metric terms. This will give an expression of the sensitivity of 
the retina in units which are directly comparable with one another. 
At present we have no direct estimate of the comparative sensi- 
tivity of the retina to the different colors further than is ex- 
3 Coblentz, W. W. Instruments and Methods Used in Radiometry. Re- 
print No. 188, Bulletin of the Bureau of Standards, 1911, IX., pp. 22-23. 
4 Ferree, C. E. and Rand, G. A note on the Determination of the 
Retina's Sensitivity to Colored Light in Terms of Radiometric Units. 
Amer. Journ. of Psychol., 1912, XXIII, pp. 328-332. 
5 The first public statement of our intention to use radiometric units 
in the investigation of the retina's sensitivity to color was made to 
the committee in charge of the Sarah Berliner Research Fellowship, 
February 1, 1911. 82 
GERTRUDE RAND 
pressed, for example, by the relative width of the collimator-slit 
that has to be used to arouse color sensation when a light-source 
of a given candle-power is used. This kind of comparison is 
obviously unfair because such different amounts of energy are 
represented from point to point in the spectrum that a given width 
of slit would admit many times the amount of energy at one part 
of the spectrum that it would at another. In short, no adequate 
estimation and expression of the retina's sensitivity to color, 
comparative or absolute, can be made by means of the methods 
now in common use.6 
"(2) Comparisons of results on many other points with such 
disparate stimuli seem equally inadequate: the relative time 
Required for the different color sensations to attain their 
sTwo criticisms have been received from private sources which it 
may be well to take account of here. In one the possibility of a point 
of view is implied, in the other a point of view is stated. The point of 
view, the possibilty of which is implied in the first criticism, is that it 
is not proper to estimate the sensitivity of the retina in terms of physi- 
cal units, because it is generally conceded by modern investigators of 
color vision that the retinal processes which transform the physical 
energy of the color stimulus into nervous energy is essentially chemical 
in its nature; and one can not assume that a certain amount of physical 
energy arouses an equal amount of chemical energy in the retina, nor 
that equal amounts of physical energy arouse equal amounts of chem- 
ical energy. In answer to this, the writers would point out that these 
chemical substances are a part of the retina and their respective iner- 
tiae constitute one set of factors that determine the sensitivity of the 
retina to the different colored lights. It is not necessary to assume, 
therefore, that a given amount of physical energy arouses an, equal 
amount of chemical energy, etc., in order to make our determinations 
of the comparative sensitivity of the retina to the different colors in 
terms of physical units. That would be necessary only if we were try- 
ing to separate out the nerve filaments, and to measure or compare 
their sensitivity to the different colors in terms of physical units. But 
even in chemical theories when speaking of the comparative sensitivity 
of the retina to the different colors, we do not mean the comparative 
sensitivity of the nerve filaments alone. We include the reaction of 
the chemical substances as well. Our contention, then, is that if the 
determination of the comparative sensitivity of the retina to the differ- 
ent colors is a proper problem, the determination should be made in 
terms of quantities that can be compared. This can be done either 
(a) by using lights equalized in energy and determining by means of a 
sectored disc the relative amounts of these lights that are required to 
arouse sensation; or (b) by using lights representing different amounts 
of energy and measuring directly in terms of radiometric units the amounts 
required to arouse sensation. We scarcely need point out that in speaking of 
the comparative sensitivity of the retina to the different colors we are not 
raising a new problem, but are merely recognizing a very old one. 
The second criticism is in substance that a quantitative comparison 
of the effect of the different wave-lengths on the retina is improper 
because the different wave-lengths constitute stimuli too different in kind to 
permit such comparison. This criticism we leave open, because we do not 
wish to discuss in this paper the propriety of the problem of comparing 
sensitivities. EXPERIMENTAL 
83 
maximum of intensity, or retinal inertia; the relative rate of 
fatigue to the different colors; after-image and contrast sensi- 
tivity, etc.7 In fact there is not a quantitative problem dealing 
7 It is conceivable that two points of view may be held with regard 
to what is meant by after-image and contrast sensitivity. (i) After- 
image and contrast sensitivity may express a relation between the amount 
of light required to arouse after-image and contrast sensations and the 
unit of light used. (2) It may express a relation between the amount of 
light required to arouse the after-image and contrast sensations and the 
amount required to arouse positive sensation. If the former view should be 
held it will be convenient to start with stimuli equalized in energy, and to 
determine the relative amounts of light required to arouse the after-image 
or contrast sensation by means of a sectored disc. If the second view 
should be held, the energy of the lights used may first be rendered pro- 
portional to the sensitivity of the eye to the colors in question; and the 
liminal values m)ay then be determined by means of the sectored disc. 
In each case the relative sensitivity may be expressed by the inverse 
ratio of the open to the closed sectors. 
Similarly two views may be held with regard to the determination of 
the comparative rates of fatigue, and of the development-time of sen- 
sation. (1) Lights equalized in energy may be used. (2) The energy 
of the lights may be made inversely proportional to the sensitivity of the 
eye to the different colors. 
The need in both the above cases is equally great for a method of 
regulating and determining the amounts of light to be used in terms 
of a common unit of measurement. For example, in the second case two 
ways might be conceived of making the amounts of the colored lights 
proportional to the eye's sensitivity to these lights. (1) The limens might 
be determined and the intensity of the lights always be kept directly 
proportional to these liminal values. But the ratio needed to maintain 
this proportion could not be established unless some means were available 
of measuring the limen-values in terms of a common unit. And if this 
were established, we have no right to assume that it expresses the relative 
sensitivity of the eye to the colors in question when greater amounts of 
light are used. To make this assumption, we would have to maintain (a) 
not only that Weber's law holds for colored as well as for white light, 
but also that the ratio of increase which gives the just noticeable change 
in intensity is the same for all colors. We do not even know that there 
is a constant ratio over any considerable range of the intensity scale 
for even a single color, (b) We would have to maintain that this ratio is 
the same at the limen as at greater intensities, in other words, that Weber's 
law holds down to the limen. The consensus of opinion amiong investi- 
gators is that this is not true. (2) A curve may be constructed for the 
particular observer in which just noticeable changes in sensation are 
plotted along one coordinate and the energy changes required to give 
these changes in sensation are plotted along the other coordinate;. The 
subjective equation, then, would be made by choosing points on the curves 
for each of the colors all representing the same number of just noticeable 
changes in intensity of sensation from the limen. The amounts of light 
required to give these equally intensive sensations could then readily be 
read off from; the energy coordinate of the curve. The energy measure- 
ments required to construct such a curve would be comparatively simple, 
for once the limen-value was measured in terms of energy units, the 
remainder of the values could be determined by means of the sectored disc, 
that is, the energy change required to produce a just noticeable change in 
sensation is directly proportional to the ratio of change of open to closed 
sectors in the disc. 84 
GERTRUDE RAND 
with the comparative functioning- of the retina to the different 
colors in which there does not seem to be a need for the regulation 
and estimation of the stimulus in terms of a common unit of 
measurement. It is the purpose of the writers to extend the work 
as fast as possible into these related fields. 
"(3) We wish to make a careful study of the sensitivity of 
the peripheral retina, quantitative8 and qualitative, in a large 
number of meridians. In general too much uniformity has been 
assumed with regard to the sensitivity of the peripheral retina. 
8 The following are two of the points we wish to take up: (i) A de- 
termination will be made of the ratio of sensitivity of peripheral to central 
retina from point to point for a single color in several meridians. This 
will show at what rate the retina falls off in sensitivity in a single meridian, 
and how uniform this decrease is in the different meridians. We have 
found in a preliminary study that this knowledge is greatly needed in 
explaining certain phenomena of th'e peripheral retina. Futhermore, 
when this determiination is made for each of the colors with which we wish 
to work, the ratios of sensitivity for these colors at all the points can be 
calculated and a definite answer can be given to the question whether or not 
uniformity of ratio obtains throughout the retina. This question has been 
given considerable importance in the discussion of color theories. (2) The 
limits of sensitivity will be investigated. In general two problems are 
involved here. (a) The limits may be considered in relation to the 
comparative sensitivity of the retina to the different colors. (b) They 
may be considered in relation to existing color theories. In the first 
of these problems the limits should be obtained with stimuli equalized in 
energy. So obtained the results will constitute merely another expression 
of the comparative sensitivity of the retina to the different colors. The 
second problem is more complicated and will later be made the subject of 
a separate paper. A word indicating its relation to our present plan of 
work may, however, not be out of place here. It may be logically assumed, 
for example, that the Hering theory demands that wherever the blue-sensing 
substance is found, the yellow-sensing substance must also be found. We 
have no means of knowing where these substances are except by the sen- 
sation aroused. Speaking in terms of the theory, then, we have a right to 
assume that wherever the blue sensation can be aroused the yellow sensation 
should be able to be aroused also, provided a sufficiently intensive stimlulus 
be used. If, therefore, in passing towards the periphery of the retina, a point 
be found where blue can be aroused and yellow can not, the evidence will 
be strongly in favor of the conclusion that no yellow substance is present, 
unless it can be shown that elsewhere in the retina so much greater energy 
of yellow light than of blue is required to arouse sensation that the 
amount needed for this far peripheral point is greater than can be obtained. 
To establish this point the comparative sensitivity to these colors would 
have to be obtained at various points in the retina. This would involve the 
determination of a ratio based upon the amounts of blue and yellow light 
required to arouse sensation. Two methods of measurement may be used, 
(a) The amounts needed may be measured directly by means of a thermo- 
pile of the type we use, or other sensitive radiometer. In a determination 
of limens the number of readings required would render this method 
tedious, (b) The energy of the two lights may be made equal by means of 
a thermopile and the final amounts required to arouse sensation may be 
secured by means of a sectored disc. From the ratio of open to closed 
sectors the amount the light is cut down in each case may be calculated and 
the ratios of energy may be determined from these amounts. EXPERIMENTAL 
85 
Generalizations of great importance to color theory have fre- 
quently been based upon the results of work in which careful 
investigation was made in only one or two meridians. The con- 
ception of stable colors, and its application in support of the 
Hering Urfarben may be taken as a fair example of a sweeping 
conclusion which is based upon work too limited in its range. 
With a careful standardization of factors, an investigation in any 
considerable number of meridians shows that stable colors do 
not exist.9 Many other points of interest have come out in our 
more detailed study of the peripheral retina. For example, we 
find in the periphery of the normal retina small areas which are 
exact replicas of the Schumann case of color-blindness. 
"(4) We wish to conduct our investigation in full daylight 
instead of in the dark-room. This is to eliminate the influence 
of the field surrounding the colored stimulus and of the pre- 
exposure. When the surrounding field is black, white is induced 
by contrast across the stimulus color. Since the colors all differ 
in brightness, the induction takes place in different amounts for 
the different colors. This white, in proportion to its amount, 
reduces the action of the colors on the retina. Further, a given 
amount of white affects to different degrees the action of the 
different colors on the retina. To eliminate this twofold unequal 
action, the surrounding field should be made in each case of the 
brightness of the color to be used. This can be done by working 
in a light-room of constant intensity of illumination and making 
the surrounding field of a gray paper of the brightness of the 
stimulus color. In order to accomplish this, and at the same 
time be able to work upon any meridian of the retina we choose, 
we have constructed a special piece of apparatus which we call a 
rotary campimeter. The influence of preexposure is even more 
important than of surrounding field. If the preexposure is too 
black, white is added as after-image to the stimulus color. The 
effect of a black preexposure upon the stimulus color is greater 
than the effect of a surrounding field of black, because more 
'The following points are offered in support of the above statement, (i) 
A red and green cannot be obtained which in every meridian of the peri- 
pheral retina will pass into gray without an intermediate change into yellow 
or blue. (2) The amount of blue that has to be added to a mixture of red 
and green to produce gray varies from point to point in a given mieridian 
even where the extramacular region alone is considered. Further, a 
series of determinations made for a given meridian will not hold for the 
remaining meridians. (3) A red, green, and yellow can not be obtained 
which will not change in color-tone in passing from the center to the 
periphery of the retina in a single meridian. 
Blue alone of the four principal colors is stable in tone for all parts of 
the retina. 86 
GERTRUDE RAND 
white is added as after-image of preexposure than is induced by 
contrast from the surrounding field. This effect also can be 
eliminated only by working in a light-room of constant intensity 
of illumination and by choosing as preexposure a gray of the 
brightness of the color to be used." 
B. Description of Optics-Room and Apparatus. 
The work was carried on in a well-lighted optics-room, 
12^2 x io ft. The room is situated on the upper floor of an 
isolated building and is lighted by a skylight, 8 x 7J/2 ft. Be- 
neath the skylight, two diffusion-sashes, 4 x 7^2 ft. are swung on 
hinges so that they can be raised or lowered as desired. The 
framework of these sashes is made of a light-weight iron. For 
convenience of local control of illumination, if needed, each sash 
is divided into four units by means of cross-pieces. The sashes 
are filled with double-strength glass ground on one side, so 
adjusted to the frame that they can be removed easily for cleaning 
or for the substitution of some other kind of glass in case that 
is desired. This glass diffuses the light so effectively that local 
shadows cast by the cross-pieces in the framework of the skylight 
are completely eliminated, while the sudden changes of illumina- 
tion produced by the passage of the sun behind a cloud are 
reduced to a minimum. This diffusion seems to have the further 
advantage of reducing the yellowness of direct sunlight below 
the limen of sensation. At least, when working under the sash, 
the observer never judged a gray exposed through the campi- 
meter-opening as yellow under any local conditions, as frequently 
happened when working under direct sunlight. 
The room is planned also so that small changes of illumination 
can be produced, ranging from the intensive illumination of a 
south-exposure skylight to the blackness of a moderately good 
dark-room. Two provisions are made for this. (a) The 
diffusion-sashes are made so that any or all of the panes of 
ground glass can be quickly and easily taken from the sash, and 
anything can be substituted that is desired; or the illumination 
can be varied by placing layers of tissue-paper above the glass. 
(&) The room is provided with two curtains mounted on heavy 
spring rollers. One is a white curtain made of thin muslin; the EXPERIMENTAL 
87 
other is a black light-proof curtain so mounted that, when drawn, 
its edges are deeply enclosed in light-proof boxing extending 
along the four walls of the room. One or both of these curtains 
can be drawn any distance that is desired, and the illumination 
can thus be changed gradually from a very intense brightness to 
a fairly good blackness. To aid in getting dark-room effects, 
the doors of the room are carefully boxed and curtained. One 
requirement of a perfect dark-room, however, is lacking, namely, 
the walls and floor of the room are painted white. This is because 
it is of advantage in the light-room work, and because complete 
blackness is not needed in the type of work for which the room is 
devised. 
The apparatus used in the investigation consists of a rotary 
campimeter devised to meet the requirements of the task in hand 
by Dr. C. E. Ferree10 of Bryn Mawr College. The object of 
this apparatus is to add to the vertical campimeter the rotary 
features of the perimeter and thus to allow investigation of every 
possible meridian of the retina with as much ease and precision 
as was possible with the old form of campimeter in the nasal 
meridian only, or at most, in the nasal and temporal meridians. 
The apparatus consists of two parts with proper supports and 
accessories; a stimulus screen, and a campimeter screen which 
rotates on a collar around a circular support. The stimulus is 
exposed through an opening in the center of the campimeter 
screen. One arm of the framework of this screen carries the 
fixation-points, and also a right-angled extension which allows 
fixation to be given at an excentricity of 92 °. This arm may be 
rotated to any position desired, and thus any meridian of the 
retina may be explored. In order that the sensation received in 
the peripheral retina may be accurately expressed in terms of 
color- and brightness-values of the central retina, the fixation- 
arm of the screen is further provided with a small detachable 
motor upon which may be rotated the proper combination of discs 
for matching peripheral sensation. This increases greatly the 
definiteness of work on the sensitivity of the peripheral retina. 
10 For the orginal description of this apparatus, see C. E. Ferree. Descrip- 
tion of a Rotary Campimeter. Amer. Journ. of Psychol., 1912, XXIII, pp. 
449-453- 88 
GERTRUDE RAND 
The feature was added to the apparatus so that complete maps 
might be made of the changes in the sensitivity of the retina 
from center to periphery and from one meridian to another, 
with tables showing the value of the changes from point to point. 
Photographs of the skeleton apparatus and of the front and 
back views of the campimeter in readiness for use are appended. 
Figure I shows the skeleton apparatus. It consists of the 
following parts: supporting base, frame for campimeter screen, 
and frame for the stimulus card. The supporting base consists 
of a horizontal steel bar, 83 cm. long, supported by two iron 
tripod rests (B and B'). To this bar are clamped two uprights 
(C and C'), which are adjustable along its length. The anterior 
upright (C) supports the frame on which the background of card- 
board and the campimeter screen (D) are fastened. The poster- 
ior upright (C') supports the stimulus frame (E). The height 
from the table of each of these frameworks is adjustable by 
means of set-screws (F and F'). The framework for the 
campimeter screen consists of central support and radiating arms. 
The central support consists of a stationary brass ring,11 about 
which rotates a larger brass collar (H), 20 cm. in diameter. The 
back surface of collar (H) is graduated from o° to 360°. To 
this collar are fastened the radiating arms. There are eight of 
these arms, one for each 450 mark of the graduated collar. They 
are made of steel and are 2 cm. broad and 40 cm. long. The 
eighth arm (I-F) differs from the other seven. It forms a 
right angle, one side of which is in the plane of the background 
and the other in front of this plane. The part in the plane of the 
background is 30 cm. long, and the part at right angles to this 
u This ring was made large in diameter for two reasons, (a) The ring 
had to be made very thick in order to give sufficient rigidity to support 
the campimeter screen and to furnish proper attachment for the rotary 
collar. Had the circumferance been small, the effect of the ring would 
have been that of a short tube. If the stimulus were viewed through a short 
tube, an induction factor would have been involved which would have been 
difficult, if not impossible, to standardize. The opening in the ring was, there- 
fore, made considerably larger than any stimulus we wished to use in order 
to avoid the introduction of this factor, (b) The large circumference of the 
ring makes the apparatus available for investigating the effect upon sensitivity 
of varying the size of stimulus. EXPERIMENTAL 
89 
plane is 28 cm. long. The arm is graduated from 180 to 570 
along the section that lies in the plane of the background and 
from 570 to 920 along the section at right angles. The gradua- 
tions are based on the arc of a circle of 25 cm. radius. The arm 
is also split lengthwise to form two narrow arms, each 1 cm. 
wide, so separated that there is an opening (J) 0.8 cm. in width 
between them to admit the shank of the motor for rotating the 
discs needed to match the peripheral sensation. The opening to 
admit the shank of the motor may be clearly seen in all the 
pictures of the campimeter. The motor is shown at K on the 
right of Figure 1 and more clearly on the left of Figure 3. It 
has a shank 4 cm. long and 0.3 cm. in diameter, which can readily 
be thrust through the opening (J). The weight of the motor 
is so great that it can not be clamped to the arm (I-I') and thus 
be shifted with the arm as the retina is tested in different 
meridians. It has then to be supported so that it can readily and 
quickly be moved to any point in any meridian to which the 
arm (I-I') may be rotated. This is accomplished by the use of 
two rods-one vertical (L) and the other horizontal (M). The 
vertical rod (L) may be clamped to the table or other support on 
either side of the campimeter, and M is clamped to L. The 
vertical adjustment for any setting of the motor can thus be 
made along L and the horizontal adjustment along M. Holes 
are punched in each of the eight arms at six or more places to 
allow the insertion of small metal fasteners to hold the back- 
ground screen to the frarpe. The stimulus frame may be 
seen at E. It is 20 cm. square and carries a groove for the 
insertion of the stimulus card. The stimulus card may be made 
of whatever colored paper the experimenter desires to use. 
Figure 2 shows the front view of the campimeter in readiness 
for use; and Figure 3, the back view. A cardboard background 
has been fastened to the steel arms by means of paper-fasteners. 
Since the background is fastened to the arms attached to the 
brass collar (H), a circular gap is left at its center. This gap 
is filled by a disc (N), shown in Figure 3, which has been fast- 
ened to the arms just outside of the collar (H). The disc is 27 
cm. in diameter and contains the stimulus-opening (O), the size GERTRUDE RAND 
90 
of which may be varied to accord with the purpose of the investi- 
gation. In the experiments reported in this paper, it was 15 mm. 
in diameter throughout. In order to complete the graduations 
on the fixation-arm to the stimulus-opening, disc (N) is grad- 
uated from o° to 18°. A background 40 cm. in height is 
fastened to the extension arm (I). In the picture a paper screen 
made of No. 7 of the Hering series of grays has been attached .by 
thumb tacks to the cardboard background. A strip of paper of 
the same quality as the background is placed along the opening 
(J), and the graduations from o° to 920 are pricked on this 
strip as indicated by the markings on the back of disc (N) and 
arm (I-I')- These constitute the fixation-points. The card in 
the stimulus frame (E) is seen through opening (O). A disc 
(P) composed of black and white sectors has been placed on 
the motor (K). 
The method of using the apparatus is as follows: The observer 
is seated in front of the campimeter screen with his head held 
in a rigid position by means of a mouthboard bearing the impres- 
sion of the teeth in sealing wax. Since the graduations of the 
fixation-arm are based on the arc of a circle of 25 cm. radius, the 
distance of the eye from the stimulus-opening is chosen as 25 
cm. The position of the eye in the observing plane may be 
obtained according to the method described by Fernald.12 In 
order to facilitate excentric fixation in the nasal and temporal 
meridians, the head should be turned 450 nasalwards or temporal- 
wards, as the case may be. With the head so placed, the eye can 
swing easily from the stimulus-opening to a fixation-point whose 
excentricity exceeds 900. The unused eye is closed and covered 
by a bandage. The arm (I-I') is placed in the meridian desired, 
the position being indicated by the graduations on the collar (H). 
The experimenter covers the stimulus in the stimulus frame 
with a card, which we shall call the preexposure card, while the 
observer takes the fixation required. At a signal given by the 
observer, the preexposure card is withdrawn, the stimulus is 
exposed for three seconds, and the preexposure card is replaced 
12 Fernald, G. M. The Effect of Achromatic Conditions on the Color 
Phenomena of Peripheral Vision. Psychol. Rev., Monograph Supplements, 
1909, X., p. 18. EXPERIMENTAL 
91 
over the stimulus. The observer is required to rest the eye 
after each observation. Further provisions against fatigue are 
made by periods of rest after each fifteen minutes of observation. 
When it is desired to measure the stimulus as seen in the peri- 
pheral retina in terms of brightness- and color-values of the 
central retina, the motor shown at K in Figures i and 3 is used. 
The method of making the measurement is as follows: If a 
direct vision judgment, for example, of the appearance of yel- 
low at 250 in the temporal meridian is wanted, the cord (R) 
carrying a movable fixation-point, seen in Figure 2, is fastened 
in front of the 250 point on the graduated background. The 
observer, in position, fixates the 250 point and brings the movable 
point in line with the eye and the 250 point. This point then 
serves as the new fixation-point, and the graduated strip cover- 
ing the opening (J) is removed. The required discs are placed 
on the motor immediately behind the new fixation-point, and 
their proportions are changed until the observer judges that the 
sensation aroused in the periphery is matched by that aroused in 
the center by the measuring-disc on the motor. In making this 
judgment, the method of ascending and descending series was 
used. 
In this investigation, stimuli of blue, yellow, red, and green 
pigment papers of the Hering series were employed. White, 
black, and gray papers of the Hering series served to make the 
backgrounds. Results were obtained from three observers: 
Miss Campbell, C, graduate student in Bryn Mawr College, who 
had no knowledge of the problem in hand, Dr. Ferre, B, and the 
writer, A. 
C. Determination of the Brightness of the Colored 
Stimuli Employed in the Investigation. 
At every turn in our problem, it was necessary to know the 
black-white-values of the colored papers that formed our stimuli, 
as they appeared in the central and peripheral retina at full and 
decreased illumination. It was thought best, therefore, to devote 
a separate chapter of our report to a discussion of the methods 
used in determining these values. The method of flicker photo- 
metry was used throughout except at the limit of peripheral color 
vision, where it was possible to use the method of direct com- 92 
GERTRUDE RAND 
parison. The black-white-values of the colors were determined 
for the central retina by means of the Schenck Flimmer Photo- 
meter. As this apparatus is not adapted to indirect vision 
work, it was necessary to devise a means by which the 
brightness of the stimuli at any point in the peripheral retina 
could be determined by the flicker method. The conditions 
of our experiment made it essential that these determinations be 
made not only in terms of black-white-value but also of colorless 
pigment paper, the brightness of which would approximately be 
the same as that of the colored stimulus. In order to make the 
latter determination possible, a series of gray papers varying 
in brightness by very small amounts was required. The Hering 
papers, ranging in number from I to 50, were found to furnish 
a series which varied in brightness by amounts sufficiently small 
to serve our purpose. 
The use of the flicker method in photometry is based on the 
fact that two surfaces are considered equal in brightness when 
upon their alternation one with the other at a certain favorable 
rate of speed, no experience of flicker results. Obviously a very 
important point in the method is to determine what this rate of 
alternation should be. It should be determined empirically for 
each observer in a preliminary experiment. To make the deter- 
mination we must be able to produce known brightness differences 
in different parts of the scale and to try the effect on flicker 
of different rates of speed for these brightness differences. 
This can be accomplished by making the preliminary experi- 
ment with colorless surfaces, for very small differences in 
brightness between two colorless surfaces can be estimated by 
the method of direct comparison. (This could not be done if one 
or both of the surfaces were colored.) Working then with color- 
less surfaces by the aid of the method of direct comparison, not 
only do we know at every stage of the experiment how much 
brightness difference is produced, but we standardize the flicker 
determination in terms of the method of direct comparison to 
which all indirect methods of determining brightness equality 
must conform if their results are to be of any value. In making 
our preliminary determination, then, colorless surfaces should be 
used and that rate of alternation, equal to or in excess of the Figure I 
Figure II EXPERIMENTAL 
93 
fusion rate for the color in question, should be selected at which 
the smallest difference in brightness between the two surfaces pro- 
duces flicker. This speed may be considered as giving the maxi- 
mum sensitivity to the method for the given observer and may 
be used for that observer in the color work. Applying this 
method to the determination of the brightness of a colored paper, 
we may consider that a colored and a gray paper are of equal 
brightness when no flicker is produced by the rotation of equal 
sectors of each at the chosen rate of speed. In order to prevent 
induction from the surrounding field, the rotating disc should 
be viewed through an opening in a screen of the same brightness- 
value as the disc itself. This requires that the gray sector, the 
colored sector, and the screen all be of equal brightness-value. 
The final value of this brightness must, however, be reached by a 
series of approximations. That is, the gray sector and the screen 
must at each trial be chosen of the same brightness, and both 
must be changed alike until a gray is finally obtained which does 
not produce flicker when it is rotated with the colored sector at 
the chosen rate of speed. Owing to the great number of steps in 
the series of approximations needed to reach the gray-value of 
the color, it was impracticable to change the large campimeter 
screen at each step. Moreover, to prevent brightness induction 
over the stimulus, it was not necessary to have so large a part of 
the surrounding field, as was comprised in the entire screen, of 
the same brightness as the stimulus. Squares 30 cm. on a side 
were found to be quite adequate for the purpose and to have the 
practical advantage that they could be quickly and easily changed. 
Since the sensitivity of the retina to flicker varies from point to 
point as we pass from the fovea outwards, it is found to be 
important that the opening in the screen be made small so that 
the area of the retina stimulated shall be uniformly sensitive to 
flicker. And, further, since the area of the retina influenced by 
a given stimulus decreases as we pass from the fovea to the 
periphery, because of decrease in the visual angle, it is necessary 
that the stimulus-opening be proportionately increased in the 
peripheral observations, in order to maintain the size of the 
stimulated area of the retina constant. Both of these conditions 
were met sufficiently accurately for our purpose, by using two 94 
GERTRUDE RAND 
openings of different size, the smaller to be employed at all 
points from the fovea to 200 peripheralwards, and the larger to 
be used from the 200 point to the extreme periphery. 
The method used to determine the brightness in terms of gray 
paper of a given color in central vision was as follows. From the 
series of squares of gray papers having the smaller of the two 
sizes of stimulus-openings (3 mm. x 1 mm.) one was selected 
which was judged by the method of direct comparison roughly to 
approximate the brightness of the color in question. This square 
was fastened upon the campimeter screen so that the stimulus- 
opening passed vertically through the center of the opening in 
the original screen. A disc compounded of 1800 of the given 
color and 1800 of the gray of the brightness of the square, was 
rotated behind the stimulus-opening and the observation made for 
flicker. Lighter and darker grays were in turn substituted in 
disc and screen and the observation was repeated. The gray 
which produced no flicker at the chosen rate of speed is, in terms 
of the method, the gray of the brightness of the color. The 
determinations were not at all difficult nor uncertain. Flicker was 
readily discernible in the gray lighter and darker than the one 
which was chosen, at the speed of rotation at which the one 
chosen showed no flicker. Our determinations showed that at 
the standard illumination13 used throughout the work, the method 
of obtaining which will be discussed later, the brightness of Her- 
ing blue for central vision equalled that of Hering gray No. 41; 
of red equalled gray No. 24; of green equalled gray No. 8; and 
of yellow equalled gray No. 2. These values were the same for 
all of our observers. 
For the determination of the brightness of the colored stimuli 
in peripheral vision, the same method was used with the exception 
that at points in the periphery beyond 200, the screens having 
the large stimulus-openings, 15 mm. in diameter, were used. The 
results of the peripheral experiments differed from the central at 
standard illumination only in case of blue. Blue was found to 
lighten in the periphery so much that gray No. 21 was determined 
13 Measured in foot-candles by means of the Sharpe-Millar portable photo- 
meter, the standard illumination equalled 390 foot-candles. EXPERIMENTAL 
95 
by one observer to equal it in brightness. The determinations at 
the peripheral limit of sensitivity to color were made by the 
method of direct comparison. On the campimeter was mounted a 
gray screen for each of the colors in turn the brightness of which 
was such that when the stimulus was observed beyond the limits 
of color sensitivity, the color in each case changed into the 
gray of the brightness of the screen. If a gray lighter or darker 
was used, the stimulus appeared either darker or lighter than the 
screen. 
The black-white-values of the colors for peripheral vision can 
not be determined directly because for the direct determination 
the Schenck Flimmer Photometer with its graduated sectors of 
black and white or some similar device must be used. This 
photometer is not adapted for peripheral vision work. A deter- 
mination, then, had to be made with the photometer in the central 
retina for the grays which had been found by the method 
described in the preceding paragraph to equal the colors in bright- 
ness in the peripheral retina. 
In one section of the investigation, it was found necessary to 
work at decreased illumination, and to know the brightness of 
the colored stimuli under these conditions. The decreased illum- 
ination was obtained by drawing the black curtains, until the 
illumination was slightly less than that of a cloudy afternoon.14 
The colors appeared a little less saturated, and slightly altered in 
color tone. The green was a trifle bluish, the yellow was changed 
toward orange, and the blue appeared slightly reddish. Deter- 
minations were made at this illumination according to the method 
described above, of the brightness of the colors in central and 
in peripheral retina. The results are stated in Tables I and II. 
The first column shows the white-black-values of the stimuli at 
standard illumination in central vision; the second shows the 
same values in terms of the Hering gray papers; the third repre- 
sents the brightness-values in peripheral vision at the limit of 
sensitivity; and the fourth and fifth columns show the brightness 
of each stimulus at the center and at the periphery under condi- 
tions of decreased illumination. 
14 Measured in foot-candles by means of the Sharpe-Millar portable photo- 
meter, this decreased illumination equalled 1.65 foot-candles. 96 
GERTRUDE RAND 
Table I records the results of Observer A; Table II those of 
Observer C. 
Table I. 
A. Showing the brightness-values of the Hering principal colors 
at standard and decreased illumination. 
Stimulus 
Brightness at Standard Illumination 
Brightness at Decreased 
Illumination 
At center 
At limit of 
sensitivity in 
periphery 
At center 
At limit of 
sensitivity in 
periphery 
Yellow 
white 236° 
black 1240 
gray no. 2 
gray no. 2 
gray no. 2 
gray no. 3 
Green 
white ioo° 
black 260° 
gray no. 8 
gray no. 8 
gray no. 8 
gray no. 5 
Red 
white 410 
black 3190 
gray no. 24 
gray no. 24 
gray no. 24 
gray no. 50 
Blue 
white 150 
black 3450 
gray no. 41 
gray no. 28 
gray no. 35 
gray no. 13 
Table II. 
Observer C. 
Yellow 
white 236° 
black 1240 
gray no. 2. 
gray no. 2 
gray no. 2 
gray no. 3 
Green 
white ioo° 
black 260° 
gray no. 8 
gray no. 7 
gray no. 7 
gray no. 4 
Red 
white 410 
black 3190 
gray no. 24 
gray no. 24 
gray no. 24 
gray no. 50 
Blue 
white 150 
black 3450 
gray no. 41 
gray no. 21 
gray no. 35 
gray no. 7 
Whether or not the peripheral retina functions differently from 
the central retina and must, therefore, be assumed to possess a 
different sensory mechanism, is a question of considerable 
importance to theories of color vision. Upon this question 
the results shown in Tables I and II have a direct systematic 
bearing. But since the comparative functioning of central and 
peripheral retina will be made the subject of a later report of 
work already completed by the writer, the significance of these 
results need not detain us here. We need only note in passing, EXPERIMENTAL 
97 
that the brightness changes that occur when a stimulus is carried 
from central to peripheral vision are similar to those that obtain 
in central vision when the illumination is decreased. With regard 
to this point our tables show (a) that in the peripheral retina at 
standard illumination, the colors have very nearly the same 
brightness relations that they have in the center at the decreased 
illumination we used; and (&) that the brightness relations of the 
colors seen in the peripheral retina at decreased illumination 
approximate those of the colors in the center when the illumina- 
tion is further decreased. These latter changes known as the 
Purkinje phenomenon are in the following directions: blue and 
green relatively lighten; red and yellow relatively darken. 
In Tables I and II, showing the comparative brightnesses of 
the color sensations at center and periphery, the lightening of 
blue and green and the darkening of red in the periphery are 
sufficiently pronounced to need no comment. But measured by 
these results, the change in yellow seems to be insignificant. If, 
however, yellow is observed in the periphery at decreased illumin- 
ation and is compared with gray No. 2, that is, the brightness of 
yellow both in center and periphery at standard illumination and 
in center at decreased, it appears to be much darker than the gray 
screen. Contrast from the screen exaggerates this darkening 
to some extent but the change in the brightness of the sensation 
due to peripheral stimulation alone is considerable. 
D. The Factors Investigated. 
The factors we have investigated with regard to their influence 
upon the color observation are: (1) the brightness of the stimu- 
lus; (2) the brightness of the field surrounding the stimulus; (3) 
the brightness of the preexposure; and (4) the general illumina- 
tion of the retina. 
1. Brightness of the Stimulus. 
It will be remembered from the historical discussion (p. 45 ff.) 
that the four men,-Bull, Hegg, Hess, and Baird,-who recog- 
nized the need of equating the intensity of colored stimuli for a 
determination of the relative limits of color sensitivity, equated 
them also in brightness. They apparently assumed the need of 98 
GERTRUDE RAND 
this equation without having investigated the influence of the 
brightness difference between the colors upon the breadth of the 
color zones. As a result of a careful investigation of this ques- 
tion, we are able to show that not only is no advantage gained 
by equating the brightness of colors when determining their 
limits of sensitivity, but a positive disadvantage is suffered. The 
following reasons may be cited in support of the latter statement, 
(a) The quality of certain colors is changed when their bright- 
ness is altered. This disadvantage, first mentioned by Chodin, 
was a source of great difficulty to Hegg, as we have seen (p. 
74).13 (b) Colored stimuli which have been equated in bright- 
ness are necessarily reduced in intensity. For this reason, no true 
nor comprehensive estimate of the color sensitivity of the retina 
can be obtained with stimuli equated in brightness, (c) The 
technique involved is extremely cumbersome. 
Our investigation of the influence which the brightness dif- 
ference between the four principal colors exerts upon their limits 
covers four points, (a) The work on the first took its start from 
Hegg.16 Hegg apparently assumed that the difference in the 
brightness of the four colors,-red, green, blue, and yellow,'-at 
full saturation is sufficient to affect their limits, and, therefore, 
that they must be equated in brightness before a determination 
of their relative limits is made possible. With the view of testing 
the validity of this assumption, we sought to ascertain whether 
a brightness change in any one of the four colors, equal to the 
maximal brightness alteration made in Hegg's equation, affects 
the limits of sensitivity to that color, provided the alteration is 
produced without changing the amount of colored light coming 
to the eye. The test was made doubly strict by varying the colors 
both toward white and toward black, thus covering a variation 
whose range was twice as great as was required. That is, both 
white and black in turn were added to each of the colors in 
16 Mrs. Franklin remarks concerning Hegg's oil papers: "Of the 'normal' 
colors prepared by Hegg, the red and the yellow would not strike the 
plain man as at all deserving of the name" (Psychol. Rev., 1897, IV., p. 96). 
16 A start was taken from Hegg, in preference to the other three investi- 
gators, because he is the only one who gives adequate numerical data con- 
cerning the extent of brightness alteration he made in obtaining this equation. EXPERIMENTAL 
99 
amounts equal to Hegg's maximal change, and the limits of the 
stimuli thus obtained were compared with each other, and with 
the limit of a stimulus equal to them in physical intensity and to 
the original color in brightness. No effect whatever on the limits 
was found as a result of these brightness alterations. (&) We 
next sought to ascertain whether a brightness equation is neces- 
sary when working with the standard pigment papers of the Her- 
ing series. A determination of the maximal brightness difference 
between the colors was made at the limit of sensitivity to each, 
and the above experiments were repeated using the maximal 
value obtained in these determinations, as the amount of varia- 
tion. In no case were the limits affected, (c) Following Hegg's 
plan of equating the four colors to one of mid-brightness, we 
next determined the maximal amount of change required to 
equate the standard Hering colors to the brightness of green. 
Since the amount of change was obviously much less than the 
variations used in (d), it was not necessary to repeat the experi- 
ments on the limits of sensitivity with this amount of variation. 
In two ways, then, we shall have shown that it is unnecessary to 
equate the Hering colors for brightness when determining their 
relative limits of sensitivity, since neither the maximal amount 
of change required to bring them all to a medium brightness, 
nor the maximal amount of brightness difference between the 
colors, has any effect upon the color limit when this change is 
applied as a variant in the direction of either white or black, 
provided that the amount of colored light coming to the eye re- 
mains unaltered throughout. These results may seem contra- 
dictory to the statement made by certain other writers and by 
ourselves17 that dark colors appear more saturated than light 
colors of equal physical intensity, that is, white exerts a greater 
inhibitive action than black upon color. This brings us to our 
fourth point, (d) We have to explain why these brightness 
changes which are known in general to affect the sensitivity of 
17 Ferree, C. E. and Rand, G. Colored After-image and Contrast Sensations 
from Stimuli in Which No Color Is Sensed. Psychol. Rev., 1912, XIX., 
pp. 215; An Experimental Study of the Fusion of Colored and Colorless 
Light Sensation: The Locus of the Action. Journ. of Philos. Psychol, and 
Scientific Methods, 1911, VIII., pp. 294-297. 100 
GERTRUDE RAND 
the retina to color do not change the limits of color. The ex- 
planation, as will be shown on p. 104, is found in the extreme 
rapidity with which color sensitivity of the retina falls off near 
the limits. The amounts of brightness dealt with in the above 
cases do not produce a sufficient change in the saturations of the 
light and the dark color to cause their limits to differ by even 
i°, because the stimuli reduced in intensity by this amount of 
brightness are still sufficiently intensive to cause the color limits to 
occur within the zone of rapid decrease in sensitivity. If the stim- 
uli had given a very small amount of colored light to the eye, and 
the limit of sensitivity had consequently occurred nearer the center 
of the retina where the sensitivity falls off more gradually, the 
difference in saturation between a dark and a light color of equal 
physical intensity, might have been sufficient to cause the latter to 
have a narrower extent of visibility. But since the amount of 
intensity at which this exception occurs is much less than is ever 
likely to be used in investigating the color limits, the exception 
can scarcely be entitled to more than theoretical consideration. 
The writer regrets to report that she has not carried on this in- 
vestigation with spectral light. While she has no reason for 
believing that the results would in general be different, still for 
the sake of knowing the exact values of the brightness quantities 
obtaining in case of spectral colors, she hopes to make the inves- 
tigation in the near future. 
To investigate these points, it was necessary to devise a method 
whereby the brightness of the color can be altered without 
changing the amount of colored light coming to the eye. When 
one is working with pigment papers, the brightness of the stimuli 
can easily be varied without changing the intensity of the stimu- 
lus. For example, discs can be compounded of 260° of yellow 
and ioo° of white, 260° of yellow and ioo° of black, 260° of 
yellow and ioo° of the gray of the brightness of yellow. In 
these cases we have, in order, a tint of yellow, a shade of yellow, 
and a yellow reduced in saturation but not changed in brightness, 
-all giving the same amount of yellow light to the eye. If it 
should be desired to make the tints darker and the shades lighter, 
the brightness sectors can be chosen of white or black in any 
proportion that is required. EXPERIMENTAL 
101 
In determining the size of the brightness sector to be used for 
the first point in this investigation, we are obliged to proceed 
largely by inference from Hegg's rather meager report of his 
work. He had equated the peripheral brightness-values of his 
four stimuli to green. In doing this, 84° of white and 5° of 
black were added to red. No statement whatever is made by him 
with regard to the amount of white and black added to blue and 
yellow. These amounts have, therefore, to be inferred. In do- 
ing so, care was taken to make the amount sufficiently large to 
give our test due rigor. We have mentioned that blue lightens 
in the periphery until its brightness is much like that of red, 
For Observer A it is slightly darker than red, for C slightly 
lighter. It is fair, then, to assume that ioo° represents the 
maximal brightness difference that Hegg found to obtain be- 
tween the colors in their peripheral values. Observations were 
taken from sets of discs composed of sectors of 260° of each 
of the four principal colors and ioo° in turn of white, of black, 
and of the gray of the brightness of the color. The surrounding 
field and preexposure in each case were of the gray into which 
the stimulus color changed at the limit of sensitivity. Several 
observers were used and several meridians explored, but in no 
case could a difference in the limits of sensitivity be detected 
for the color mixed with white for the color mixed with black, 
and for the color mixed with the gray of the brightness of the 
color. Space will not be taken here to record the results for all 
observers in all meridians investigated. The results obtained 
for Observer A for the temporal and nasal meridians are selected 
as typical. They are shown in Table III. 
We have seen that the alteration made by Hegg in his bright- 
ness equations, when applied to stimuli of Hering standard 
papers, does not affect their limits of sensitivity. We now pass 
to our second point, namely, whether the full brightness dif- 
ference in these colored papers should be considered as in any 
way affecting their limits. This is somewhat different from the 
preceding point in which we were concerned merely to find out 
whether Hegg's attempt to reduce all the colors to a mid-bright- 
ness could be considered as having any effect upon their limits. 102 
GERTRUDE RAND 
Table III. 
A. Showing the limits of sensitivity18 when the colors are mixed 
in turn with ioo° of gray of the brightness of color, ioo° oft 
white, and ioo° of black without altering the amount of colored 
light coming to the eye. 
Stimulus 
Meridian 
Limit of stimu- 
lus when mixed 
with ioo° gray 
Limit of stimu- 
lus when mixed 
with ioo° white 
Limit of stimu- 
lus when mixed 
with ioo° black 
Yellow 
Temporal 
42° 
42° 
42° 
Green 
35° 
35° 
35° 
Red 
4i° 
4i° 
4i° 
Blue 
5i° 
5i° 
51° 
Yellow 
Nasal 
88° 
88° 
88° 
Green 
59° 
59° 
59° 
Red 
85° 
85° 
85° 
Blue 
9i° 
9i° 
9i° 
We wish here to find out whether the actual difference in bright- 
ness between the extreme members of the series, blue and 
yellow, affects the limits of any one of the series. To do this, 
our method was to determine the difference between blue and 
yellow at their limits of sensitivity, to vary each color toward 
both white and black by the amount of this difference, and to 
find out whether the limits of the light and the dark stimulus 
differ from each other or from that of a stimulus of equal 
intensity which has retained the original brightness of the color. 
This amount of variation was greater than was needed in case 
of red and green, because they do not differ from any member 
of the series by so great an amount. We have used this maximal 
amount, however, because we have not wished to leave room for 
any question as to the rigor of our test. 
In order to ascertain the difference between the white-values 
of the colors seen in the extreme periphery, the Hering gray that 
represented the peripheral brightness of each stimulus as deter- 
mined by the method of direct comparison at the limit of sensi- 
tivity (see Tables I and II), was mounted on the Schenck 
18 The point at which color loses all trace of its original quality is recorded! 
as the limit of sensitivity. EXPERIMENTAL 
103 
Flimmer Photometer, and its white-value determined. For 
Observer A, blue was the darkest color. Its brightness 
was equal to white 370, black 3230; that of red was 
equal to white 41 °, black 3190; that of green was equal to white 
ioo°, black 260°; that of yellow, the lightest color, was equal to 
white 236°, black 1240. The maximal brightness difference, 
then, was between blue and yellow, and was equal to 1990. To 
ascertain whether this brightness difference is sufficiently great to 
influence the breadth of the color zones, the limits of stimuli 
composed of 1610 of each color and 1990 of black, and of 1610 
of color and 1990 of white were compared with each other and 
with the limit of a stimulus composed of 1610 of color and 1990 
of gray of the brightness of the color. The first two stimuli, it 
will be observed, were composed of the color altered in brightness 
toward black and toward white by an amount equal to the dif- 
erence in white-value between blue and yellow; the third stimulus 
retained the original brightness of the color while it sent the 
same amount of colored light to the eye as the other two. In 
every case, on either the nasal or the temporal meridian, the limit 
of color visibility was the same whether the stimulus was the 
color in its original brightness or whether its brightness was 
changed in either direction, toward black or toward white, by 
an amount equal to the maximal difference between the white- 
values of the colors as seen in peripheral vision. 
As we have said, our test is unnecessarily severe. Not only 
have we lightened blue and darkened yellow by an amount equal 
to the difference in their white-values, but we have also darkened 
blue and lightened yellow by the same amount. If a brightness 
equation were found to be necessary, the variation would by no 
means be as wide as the one we have made. It would be necessary 
merely to darken some colors and to lighten others to a medium 
brightness.19 We feel confident then in stating the following 
19 The change we have made in one direction is no greater than had to be 
made by Baird who equated all of his colors to the brightness of blue. Baird, 
it will be remembered, was forced to employ this brightness as standard 
because his equation of brightness was made by interposing an episcotister 
between the stimulus and the eye of the observer. This method permitted 
change only in one direction, towards black. The defects of his method 104 
GERTRUDE RAND 
two points:-(a) The amount of change required to equate in 
brightness the colors, red, green, blue, and yellow, has no effect 
upon their color limits and the precaution of equating is, there- 
fore, superfluous. (&) The actual brightness difference in the 
colors at standard saturation has no effect upon their relative 
limits. 
While we have shown that variations of brightness in the above 
amounts do not affect the limits when there is no alteration in in- 
tensity of the colored light, we do not claim that there might 
not be a change sufficiently large to influence the limits. This 
would be a broader thesis than we wish to maintain. We have 
merely been concerned with showing that brightness alterations 
as great as the difference between the white-values of the Hering 
standard papers do not affect the limits. Strictly speaking, this 
is as far as our criticism of previous attempts to standardize 
brightness need carry us. But it is a matter of fact that a color 
mixed with black gives us a sensation that is more intensive than 
that produced by a color of equal physical intensity which is 
mixed with white, and that the limen of color is much lower 
when the color is mixed with black than when mixed 
with white. Brightness change, then, does affect the 
retina's sensitivity to color, and,, within limits, the breadth of 
the zones of sensitivity. We have, therefore, extended our in- 
vestigation to explain why changes of the order given above do 
not affect the color limits, and to determine roughly to what ex- 
tent brightness change may be made without affecting them. As 
already indicated we must look for the explanation of our results 
to the rapidity with which the sensitivity of the retina falls off 
from point to point from center to periphery. If, for example, 
it be found that sensitivity falls off gradually from the fovea 
to near the limits (as determined with stimuli of full intensity,) 
and from that point on, it falls off abruptly, we might expect 
that light and dark colors of equal physical intensity will have 
different limits up to the point on the retina at which the abrupt 
have already been pointed out. With a spectroscopic mixer as the ideal 
apparatus for investigations with the light of the spectrum, the brightness 
changes can be readily made in both directions, as they can with pigment 
paper stimuli. EXPERIMENTAL 
105 
change in sensitivity begins, and the same limits from that point 
on. It is obvious that in either case, whether or not there is a 
difference in limit, depends upon whether the difference in the 
inhibitive action of white and black upon the color is equal 
to the amount of change of intensity required to affect the limit. 
If sensitivity falls off gradually, a relatively small change in 
intensity is sufficient to widen the limit, and, if abruptly, a rela- 
tively large amount of change is required. 
By way of explanation, it is our purpose to show (a) that 
the sensitivity of the retina falls off gradually to a point within 
5° of the limit and from that point to the limit, it falls off very 
abruptly; (&) that the white and black sectors added to the 
colored stimuli in the foregoing tests did not weaken the stimuli 
sufficiently to narrow their limits more than 30; and. (c) that 
within the zone 30 from the limit, the difference between the 
apparent saturations of our light and dark stimuli was not suffi- 
cient to affect their limits. 
An inspection of the results given in Table VIII and discussed 
in the next section (p. 117 ff.) will show the rate at which the 
sensitivity of the retina falls off from the fovea to the periphery 
and will establish our first point. The decrease is gradual from 
the center to within 50 of the limit, beyond which point it grows 
progressively more abrupt, becoming extremely abrupt from a 
point 30 from the limit to the limit. For example, when the 
screen and preexposure of the gray of the brightness of color 
are used, the limen of yellow at the fovea is 180 ; at 390 from 
the fovea in the temporal meridian, that is, 50 from the limit 
of yellow, it is ioo°. Thus over a space of 390, the limen has 
increased only 82°, and average of little more than 2° of in- 
crease per degree of retina traversed. At 41 °, however, it has 
reached a value of 1500, an average of 340 of increase per 
degree of retina traversed; at 42°, a value of 2400, an average of 
900 of increase per degree of retina traversed; at 430, a value of 
3300, an average also of 900 per degree of retina traversed. 
With regard to the second point, it will be remembered that the 
extreme amount of white or black we added to our colors was 
1990. This left 1610 of color in the stimulus discs. Table VIII 106 
GERTRUDE RAND 
(page 119) which gives the values of the color limens at dif- 
ferent points near the limit, shows that this amount of color is 
above the limen for each color at 30 from the limit. In our 
tests, then, we were working well within the 50 limit bounding 
the zone of abrupt decrease in sensitivity, as our explanation 
required us to show. With regard to the third point, it will 
be seen from the same table that, when working at the point 30 
within the limit, in order to extend the limit i°, an increase of 
the colored sector by amounts ranging from 65° in the case of 
blue to 1150 in the case of green, is required. It scarcely need 
be pointed out that the apparent saturation of a stimulus com- 
posed of 1610 of color and 1990 of black is not greater than the 
apparent saturation of a stimulus composed of 1610 of color 
and 1990 of white by an amount equivalent to from 65° to 1150 
of color. 
Having explained why brightness differences equal to those 
found in red, green, blue, and yellow papers of standard satura- 
tion have no effect upon the limits of color sensitivity, we turn 
next to a determination of the range within which brightness 
change may be made without affecting the limits of sensitivity. 
Two ways occur to us by means of which a rough estimate of 
this range may be obtained, (a) Stimulus colors at full satura- 
tion may be used and the brightness excitation be added as 
after-image or contrast or both. In this way the amount of 
colored light coming to the eye is not altered by the brightness 
added, that is, the physical intensity of the color in the stimulus 
is not affected. If we wish to use the contrast and after-image 
effects, the card which covers the stimulus before exposure can 
be adjusted so that an intensive after-image is superimposed upon 
the stimulus when the card is removed. By a proper regulation 
of this card and of the campimeter screen, which causes contrast 
induction across the stimulus, varying amounts of white and 
black can be added to the stimulus, care being taken to measure 
these amounts and to keep them equal, each to each. Since, 
according to our measurements in this region of the retina, the 
after-image and the contrast excitations from white are more 
intensive than those from black, the quality of the screen and EXPERIMENTAL 
107 
preexposure designed to give dark contrast must be regulated 
until the brightness excitation aroused is found to be equal in 
amount to the white given by the black screen and preexposure. 
A series of these changes can be made until a point is reached 
where the sensations are reduced in intensity sufficiently to allow 
the more saturated dark color to be seen farther out than the 
light color. The sum, then, of the amounts of white and black 
added in turn to the stimulus, will give the range of brightness 
change that may be made in a stimulus of full intensity without 
causing the difference in brightness to be a factor influencing 
color limits. (&) Equal sectors of white and black may be 
added to the stimulus color until a point is reached where the 
darkened color is seen farther out than the lightened. This 
method has the disadvantage that with each addition to the 
brightness sector, there is a corresponding subtraction from the 
color sector. On the other hand, however, it may have a possible 
advantage over the former method in that the brightness excita- 
tion that is added to the color is aroused by light-waves, as is 
the case with the standard colors whose brightness differences 
gave rise to our problem; hence any theoretical questioning is 
obviated as to the quantitative equivalence of the action of a 
brightness excitation objectively aroused to an excitation aroused 
as after-image or contrast. But since we can not work with 
colors at full saturation, the disadvantage is probably much in 
excess of the advantage. We can doubtless come much closer 
to the value we are seeking by the first method. As the work by 
this method is not completed, its report will be deferred until a 
later paper. The results obtained by the second method are 
given in Table IV. In this table we have shown how much the 
colored sector may be reduced by the addition of black and white, 
without changing the limits for the darkened and the lightened 
color. If a further reduction is made, the darkened color will be 
seen at a greater excentricity than the lightened color. The results 
show that 2400 of black, white, or gray of the brightness of 
color may be added to yellow and the limits for the three shades 
of color so formed will still coincide; 2250 to red; 2150 to blue; 
and 2300 to green. Since, roughly speaking, the amount of 
inhibition will be inversely proportional to the amount of color 108 
GERTRUDE RAND 
present, it is obvious that if the colors could have been maintained 
at full intensity, as they usually are in the investigation of sensi- 
tivity, a still greater brightness change would have been possible. 
While we may not have determined by this method just how much 
brightness difference there may be between colors at full satura- 
tion without affecting the limits, we have shown beyond doubt 
that there may be much more than is found between the standard 
pigment colors. 
Table IV gives some of the results of this investigation for 
Observer A in the temporal meridian. Each observation was 
taken with screen and preexposure card of a gray of the bright- 
ness of the color. Since the results in the nasal meridians are 
very similar to these, space will not be taken to report them. As 
the sensitivity of the retina falls off gradually in all directions 
until within 5° of the limit, the limen at this 50 point is almost 
identical, whatever the meridian. 
Table IV. 
A. Showing how much white, black or gray of the brightness 
of the color we may add to a colored stimulus and still have a 
coincidence of limits for the three shades of color, providing the 
amount of colored light coming to the eye is kept constant. 
Stimulus 
Value of 
colored 
sector 
Value of 
brightness 
sector (gray 
of brightness 
of colon 
white, or 
black) 
Limit of 
sensitivity 
when color 
is mixed 
with gray of 
brightness of 
color 
Limit of 
sensitivity 
when color 
is mixed 
with black 
Limit of 
sensitivity 
when color 
is mixed 
with white 
Yellow 
260° 
ioo° 
42° 
42° 
42° 
1800 
1800 
40° 
40° 
40° 
90° 
2700 
37° 
38° 
37° 
120° 
2400 
40° 
40° 
40° 
1050 
255° 
39° 
40° 
39° 
Green 
120° 
2400 
30° 
3i° 
30° 
130° 
2300 
3i° 
3i° 
31° 
Red 
120° 
2400 
37° 
38° 
35° 
135° 
225° 
39° 
39° 
39" 
Blue 
135° 
225° 
48° 
48° 
43° 20 
145° 
215° 
49° 
49° 
49° 
20 The decided narrowing of the limit of the blue stimulus in this case 
stimulus in this case EXPERIMENTAL 
109 
But there is more than one kind of problem which deals with 
peripheral color sensitivity. To avoid any possible misunder- 
standing of our position, a word may be added to show when it 
is of advantage and when of disadvantage to equate stimuli in 
brightness, (a) When investigating the limits of color sensitivity 
and when the brightness of the surrounding field is the same as 
the brightness of the stimulus color, a brightness equation of the 
different colors, within the limits we have just determined, is not 
only unnecessary, but a positive harm. This, moreover, is the 
proper regulation of the brightness of the surrounding field for 
all investigations of the relative and absolute limits of sensitivity 
and of the limens of color at different points on the retina. (&) 
When, however, the brightness of the surrounding field is differ- 
ent from that of the color, the factor of the induction of the 
screen must be taken into account. Since brightness contrast 
follows the law that maximal contrast occurs when there is a 
maximal brightness opposition, different amounts of contrast 
will be induced across colors of different brightnesses. But under 
these conditions, only one legitimate problem can arise, namely, 
to test the effect of the screen. There are two points to this 
problem, (f) Knowledge of the effect of different screens upon 
the same color may be desired. In this case, the problem of 
is due to the following cause. For Observer A there is a small spot in the 
horizontal temporal region of the right eye that is totally insensitive to blue 
light. This miniature spot of blue-blindness extends from 430 to 470 in the 
horizontal temporal meridian. Now since the apparent intensity of the sen- 
sation aroused by the stimulus composed of 1350 of blue and 2350 of white 
was not sufficient to allow the color to be seen on the peripheral side of this 
blue-blind spot, its limit occurred on the foveal limit of the spot, at 430. 
It may be added that spots of this type are not unusual. The writer has found 
in every eye she has tested one or more spots that are partially or totally 
insensitive to one color alone. Relative to these blind spots, the following 
interesting features may be noted, (a) Although totally blind to a given 
color, they have normal sensitivity to its complementary color, (b) They give 
a fully saturated complementary-colored after-image of this color to which 
they are blind, (c) They show the usual cancelling action between the color 
to which they are blind and its atagonistic color. In short, they seem 
to be exact replicas in the periphery of the normal eye of the unique type 
of color-blindness described by Schumann (see Schumann, F. Ein 
ungewdhnlicher Fall von Farbenblindheit. Bericht uber die 1. und 2. Kon- 
gress fiir experimentelle Psychologic, 1904, pp. 10-13. 110 
GERTRUDE RAND 
brightness equation would not arise, (w) Knowledge of the 
effect of the same screen on different colors, or of the compara- 
tive effect of different screens on more than one color may be 
desired. In this case the colors may or may not be equated in 
brightness:-the question depending upon the requirements of 
the problem. If they are not equated in brightness, there will be 
different amounts of induction with each screen for each color. 
If they are, the colors will be altered in intensity and often in 
color tone. No general rule can be laid down as to equation or 
non-equation in these cases. Each has to be settled on its own 
merits and in accord with the requirements of the problem in 
hand. What we wish to emphasize more than anything else at 
this point is that, while at different times in color work, one may 
need to make legitimate use of a surrounding field which differs 
in brightness from the stimulus color, it should never be done 
in any investigation of the relative or absolute limits or limens of 
color sensitivity. The use of the perimeter and the dark-room is 
a notable instance of the violation of this precaution. The sur- 
rounding field of intensive blackness induces a different amount 
of white over each of the colors unless they are of the same 
brightness. And if they are equated in brightness, all the dis- 
advantages which, as pointed out earlier in the paper, result from 
this equation, are suffered in the investigation. Moreover, to equate 
the stimuli in brightness is not to get rid of the induction of the 
surrounding field. We still have, after equating, a large amount 
of brightness induction which operates against a determination 
of absolute limits by tending to narrow the limits of sensitivity 
for all colors; and against a determination of relative limits by 
narrowing the different colors unequally, depending upon the 
difference in the inhibitive action of the same amount of white 
upon them. 
2. Brightness of the Field Surrounding the Stimulus. 
When a small color stimulus is surrounded by a large field of 
white or black, a sensation is given which consists of the color 
mixed with black or white, due to contrast induction from the 
surrounding field. The influence of the brightness of the sur- 
rounding field upon color sensitivity resolves itself, then, into the EXPERIMENTAL 
111 
question of the fusion of colored with colorless light sensation in 
central or peripheral vision, according to the part of the retina that 
is stimulated. The details of this fusion in central vision have 
been taken up by the writer working in collaboration with Dr. 
C. E. Ferree,21 in which work it was shown that the effect of 
fusing a colored sensation with white, black, or gray is twofold, 
(n) There is a quantitative effect due to the inhibition of chro- 
matic excitation by achromatic. White inhibits color most, the 
grays in order from light to dark next, and black the least. The 
records of all the observers used in this investigation show that 
the achromatic series inhibits red and yellow considerably less 
than blue and green. (&) There is also a qualitative effect. The 
tone of certain colors is changed by the action of the achromatic 
excitation. The change is greatest when the stimuli are blue and 
yellow.22 Yellow, when mixed with black, gives a sensation of 
olive-green; and blue when mixed with white, black, or gray gives 
a sensation of reddish-blue. 
As a factor influencing the limits and limens of the sensitivity 
of the retina to color, the inhibitive, or quantitative effect of the 
fusion concerns us more than the qualitative. As we have stated, 
a white surrounding field, for example, a white campimeter 
screen, induces black across the stimulus which fuses with and 
modifies the resulting sensation; while a black screen induces 
white. For an estimate of the amount of brightness contrast that 
is induced by white and black screens across yellow, green, red, 
and blue stimuli, the reader is referred to the section: Quantita- 
tive Estimate of the Influence of the Change of Illumination upon 
the Induction of Brightness by the Surrounding Field (p. 138). 
The question is considered in detail in that section rather than in 
the present one, because it will be necessary at that point to 
compare the amounts of brightness induced by the white and 
11 Ferree and Rand. An Experimental Study of the Fusion of Colored 
and Colorless Light Sensation: The Locus of the Action. Journ. of Philos. 
Psychol, and Scientific Methods, 1911, VIII., pp. 294-297. This is only a 
brief preliminary report of the work. A full report will be published later. 
23 How far the qualitative effects of the fusion of colored with colorless 
light sensation in central vision are paralleled in peripheral vision, will form 
the discussion of a later chapter of this investigation, not reported in this 
paper. 112 
GERTRUDE RAND 
black screens at standard and decreased illumination. In that 
section is shown also in what way the amounts of brightness 
induced by the screens were estimated, and within what limits 
the values obtained can be said to represent these amounts. Tables 
XII and XIII (pp. 142-143), columns 1, 2 and 3, give the amount 
of contrast that is induced by the white and black screens at 
standard illumination across the grays of the brightness of the 
colored stimuli at 250 and 400 in the horizontal temporal meridian 
for Observers A and C. The results of these tables may be 
summarized as follows: 
1. The amount of induction from the white and black screens 
increases with the distance from the fovea. 
2. The amount of induction from the white screen is greater 
than that from the black screen.23 
3. The white and black screens induce most across the stimuli 
that are farthest removed from them in brightness, and least 
across those which are nearest to them in brightness. That is, 
the white screen induces more black across the gray of the 
brightness of blue than across the gray of the brightness of 
yellow; the black screen induces more white across the gray of 
the brightness of yellow than across the gray of the brightness 
of blue. 
The effect of this induction of the surrounding field may be 
shown by two methods: (a) by its effect on the limits of color 
sensitivity; and (&) by its effect upon the limens of color sensi- 
tivity.23a Up to this time, so far as the writer knows, the effect 
of the surrounding field has been estimated only by the first of 
these two methods, by its effect on the color limits. This method, 
however, estimates the effect of the surrounding field upon the 
color sensitivity of the extreme peripheral retina alone. By the 
23 See footnote p. 141. 
2Sa Since sensitivity to color is measured by determining both the limen 
and j. n. d. of color, it might be thought that the effect of surrounding 
field could be measured in both of these ways. The determination of 
the j. n. d. would, however, show very little, because the induction of 
the surrounding field would affect both the standard and comparison surfaces. 
This will be true also of the effect of the brightness of the preexposure, 
and of changes in the general illumination. In none of these cases has the 
writer considered it worth while to make the determination of the j. n. d. EXPERIMENTAL 
113 
second method, on the other hand, this effect can be measured in 
the central and paracentral regions, as well as in all parts of the 
peripheral retina. In order to make a complete study of the 
effect of the brightness of the surrounding field on color sensi- 
tivity, we have used both of these methods. The report of the 
work done by them is as follows: 
a. The effect of the induction of the surrounding field upon 
the limit of color sensitivity. 
Assuming that the law of brightness inhibition of color for the 
central retina holds for the peripheral retina, we should expect to 
find that, since colors have a lower limen in black than in gray 
or white, a white screen, which causes black induction across the 
stimulus, would be more advantageous to color vision than would 
a black screen, which causes white induction. Further, we should 
expect to find that a gray screen of the brightness of the stimulus, 
which causes no induction whatever, would be the most favorable. 
An investigation of the color limits with screens of white, black, 
and gray of the brightness of the color, shows, however, the 
following facts: 
i. Blue and green have widest limits with the gray screen, 
slightly narrower with the white, and narrowest with the black. 
2. Red and yellow have widest limits with the black screen, 
slightly narrower with the gray, and narrowest with the white. 
The color limits of Observers A and C, taken on the temporal 
and on the nasal meridian, are given in Table V and VI. 
b. Explanation of the effect of the induction of the surrounding 
field on the limits of color sensitivity. 
Turning to the explanation of these results, we shall here 
endeavor to account for the results obtained with the white and 
black screens. We have the following points to explain: (a) Blue 
and green have wider limits with the white screen than with the 
black, but the difference is comparatively small. According to 
the law of the action of white and black on colors, formulated 
from the results of work in the central retina, we should expect 
to find wider limits with the white screen, which induces black, 
than with the black screen, which induces white. Thus far, then, 
the results are in accord with the law, but the difference found 114 
GERTRUDE RAND 
Table V. 
A. Showing the limits of color sensitivity with screens of 
white, black and gray of the brightness of the color. 
Stimulus 
Limit with gray 
screen of the 
brightness of 
the color 
Limit with 
white screen 
Limit with 
black screen 
Meridian 
Yellow 
44° 
42° 
45° 
900 Temporal 
Green 
37° 
36° 
34° 
Red 
43° 
42° 
44° 
Blue 
53° 
50° 
49° 
Yellow 
90° 
88° 
92° 
900 Nasal 
Green 
64° 
62° 
6o° 
Red 
89° 
87° 
89° 
Blue 
92° 
92° 
92° 
Table VI. 
Observer C. 
Yellow 
Green 
Red 
Blue 
49° 
44° 
45° 
56° 
46° 
42° 
41° 
55° 
50° 
40° 
45° 
53° 
90° 
Temporal 
Yellow 
92° 
92° 
92° 
90° 
Nasal 
Green 
87° 
84° 
53° 21 
Red 
92° 
92° 
92° 
Blue 
92° 
92° 
92° 
between the inhibitive action of white and of black in the central 
retina would lead us to expect a greater effect on the limits. (&) 
Yellow and red have wider limits with the black screen than with 
the white. This is in direct contradiction to the law of fusion 
formulated for the central retina. With regard to explanation, 
two points must be considered. (I) The relative inhibitive action 
of black and white upon color must be investigated in peripheral 
vision; and (2) the rate of falling off in sensitivity of the periph- 
eral retina must be ascertained. 
24 In this case, the qualitative change of green to blue caused the decided 
narrowing of the limit. EXPERIMENTAL 
115 
(i) The relative inhibitive action of black and white upon color 
in peripheral vision. The relative inhibitive action of white and. 
black upon the colors must be investigated at all points from the 
fovea to the limits of sensitivity to see whether the law estab- 
lished for the central retina holds for all degrees of excentricity. If 
we find that the difference between their inhibitive actions lessens 
as we go towards the limits, we have a reason for the small 
widening of the zones of blue and green by the white screen. 
And if we find just within the limits of sensitivity for red and 
yellow that black inhibits these colors more than white does, we 
have a reason for the relative widening of the zones of sensitivity 
for these colors with the white screens, provided we can show 
that the effect of neither screen will carry the limits farther 
towards the fovea than the inner margin of this zone within 
which the exception is found. 
To test the relative inhibitive power of black and white in the 
peripheral retina, the limen of color in black and white had to be 
determined. Two methods of procedure were possible with the 
apparatus used. By the first method, the stimulus was a disc with 
sectors of color and white or black which could be adjusted so that 
a liminal sensation of color was produced. In order to prevent 
brightness induction the screen had to be of a gray of the bright- 
ness of the stimulus used. With each addition of color to the 
stimulus, a change of the brightness was produced. The screen 
then had to be altered in brightness by an equal amount. Of the two 
methods of determining the brightness of the stimulus, described 
p. 91, the method of comparing the brightness of the colorless 
peripheral sensation with the surrounding field was obviously 
better adapted to the requirements of this observation than was 
the more cumbersome flicker method because the brightness of 
the stimulus was being continually altered. For the present case, 
the gray squares were used for surrounding field that had served 
a similar purpose for the determinations of the brightness of the 
stimuli in the periphery at standard and decreased illumination. 
The observer first made a preliminary judgment of just notice- 
able color, and then determined the gray that was equal in 
brightness to the stimulus. A square of this gray was then 116 
GERTRUDE RAND 
mounted on the campimeter and the final determination of the 
limen was made. By the second method, the screens were re- 
moved and the skeleton apparatus alone was used. A disc com- 
posed of white and black sectors was placed on the motor so 
that it just filled the large circular ring at the center. This gave 
a surrounding field whose brightness could be adjusted at will. 
A small disc, 2 cm. in diameter, composed of sectors of the color 
to be investigated and black or white, was placed over the large 
black and white disc. The method of procedure was as follows. 
The observer took the required fixation, and observed the small 
disc to find the smallest amount of color that could be sensed 
when fused with white or with black, as the case happened to be. 
Before each determination, the experimenter adjusted the black 
and white sectors of the large disc, so that they equalled the 
brightness of the inner disc. This brightness was readily calcu- 
lated from the following quantities:-the number of degrees in 
the colored sector, its black-white-value, and the number of 
degrees of white or black in the remainder of the disc. 
Since the point in question was of considerable importance, 
both of these methods were used, the one as a check on the other. 
The first had the advantage of greater ease of manipulation and of 
employing a stimulus which was the same size as that used in 
the sensitivity experiments. The second had a possible advantage 
in the adjustment of the brightness of the surrounding field, but 
it was of disadvantage because the surrounding field could not be 
made so wide as by the former method and because a stimulus 
larger than that usually employed had to be used. 
Results from both of these methods show the following facts: 
(1) As the fixation becomes more excentric, the difference in the 
inhibitive action of white and black decreases. (2) From center 
to periphery, the limens of green and blue are greater when mixed 
with white than when mixed with black; that is, the law of the 
greater inhibitive power of white holds for these colors in the 
periphery as well as in the center. (3) An exception to this law 
is found for yellow and red near the limits of sensitivity. From 
the center to within about 5° of the limit of sensitivity,25 white 
15 By the limit of sensitivity is meant the widest limit of color determined 
at standard illumination. EXPERIMENTAL 
117 
has a greater inhibitive power than black over these two colors. 
But from this point to the limit, the reverse relation obtains, 
and red and yellow in this region have a greater limen in black 
than in white. How much this apparent exception to the law of 
fusion as it obtains in central vision is due to the natural darken- 
ing of red and yellow as they pass into the peripheral field of 
vision, we are not at this time prepared to state. Because of this 
darkening, there is more black fused with red and yellow than the 
results of Table VII express. These results represent the values 
of the colored and black sectors in the stimulus discs only and 
not the actual proportions of color and black excitations aroused. 
Results are shown in detail in Table VII. They are taken from 
the records of Observer A, on the temporal meridian by the first 
method described. Column 1 indicates the stimulus used; column 
2, the fixation at which the liminal determination was made, and 
columns 3 and 4, the limens of color mixed with white and black. 
Table VII. 
A. Showing the inhibitive action of white and black upon 
color in peripheral vision. 
Stimulus 
Fixation 
Limen of color 
in white 
Limen of color 
in black 
Yellow 
o° 
40° 
3° 
35° 
85° 
65° 
38° 
95° 
85° 
40° 
120° 
ii5° 
42° 
290° 
320° 
Green 
0° 
45° 
5° 
25° 
8o° 
50° 
31° 
130° 
100° 
33° 
200° 
175° 
Red 
0° 
30° 
3° 
35° 
8o° 
65° 
38° 
120° 
110° 
39° 
135° 
135° 
40° 
155° 
1700 
42° 
2900 
3io° 
Blue 
o° 
6o° 
10° 
35° 
1250 
65° 
4i° 
145° 
1400 
42° 
1800 
1700 
5i° 
3000 
280° 
(2) The rate of falling off in the sensitivity of the retina to 118 
GERTRUDE RAND 
color from center to periphery. To determine the falling off in 
sensitivity of the retina, the limen of color must be known at 
several points of excentricity. For this determination, the re- 
sults given in Table VIII, which shows the limens of color when 
the brightness influence of the screen has been eliminated, best 
serve our purpose. They show that at 50 from the limit, the 
limen has been increased from three to tenfold as compared with 
the limen at 25 °, or six to fourteenfold as compared with the 
limen at the center. The distance between the point 250 from 
the center, and the point 50 inwards from the limit averages for 
all colors about io°. It is readily seen that the sensitivity falls- 
off much faster from the point 250 from the center to 50 from 
the limit than it does from the center to the 250 point. At the 
point 30 inwards from the limit, the limen ranges from 1450 
of color, in the case of blue and green, to 1500 of color, in the 
case of yellow and red. It is from this point that the sensitivity 
falls off with extreme rapidity. As was mentioned earlier in 
the discussion (see p. 105), a change in the fixation of i° 
peripheralwards causes an increase in the limen of 65° or more, 
an increase that represents a greater lessening of sensitivity in 
i° of excentricity than there was in the first 250 from the fovea. 
Values of the limen for all colors with gray screens of the 
brightness of the color at o° and 250 from the center, and 50, 
30, 20, and i° from the limit are shown in Table VIII. They 
were determined in the temporal meridian of Observer A and 
are selected as typical. An equal zone of rapidly decreasing 
sensitivity was found on the nasal meridian also in every case 
where the limit of color sensitivity occurred within the range of 
our apparatus. 
In Tables V and VI, it was shown that the limits of color are 
not changed more than 50 with the white and black screens 
from their values with screens of the brightness of the color 
used. The results of Table VIII show why this is so. The 
comparatively large amounts of induction by the white and the 
black screens narrow the limits so little because of the extreme 
rapidity with which sensitivity^ falls off in this zone. To narrow 
the limits even 30, enough brightness must be induced, roughly 
speaking, to completely inhibit more than 2000 of color. EXPERIMENTAL 
119 
Table VIII. 
A. Showing the rapid falling off in sensitivity of the extreme 
peripheral retina. 
Stimulus 
Li men 
at o° 
Limen 
at 250 
Limen 
5° from 
limit 
Limen 
3° from 
limit 
Limen 
2° from 
limit 
Limen 
i° from 
limit 
Limit 
Yellow 
i8° 
35° 
IOO° 
150° 
2400 
330° 
44° 
Green 
20° 
40° 
130° 
145° 
260° 
345° 
37° 
Red 
9° 
17° 
132° 
150° 
200° 
320° 
43° 
Blue 
9° 
12° 
130° 
145° 
200° 
3io° 
53° 
The following points, then, needed in our explanation of the 
influence of the white and black screens on the limits of color 
sensitivity have been established. (<z) The white screen, which 
induces black, narrows the limits of sensitivity to red and yellow 
more than the black screen, which induces white, because neither 
screen narrows the limit more than 50, and within this zone of 
50, red and yellow are inhibited by black more than by white. 
(b) The limits of blue and green are narrowed by the black 
screen more than by the white screen, because within this zone 
of 50, as at the center, these colors are inhibited more by white 
than by black. But they are narrowed less by the black screen 
than might be expected from the inhibitive action of white found 
to obtain at the center, because as we go towards the periphery, 
the difference between the inhibitive actions of white and black 
decreases. And (c) neither screen narrows the limits for any 
color more than 50, because within the zone 50 from the limits, 
the sensitivity falls off so abruptly from point to point that more 
brightness action is required to change the limits beyond this 
amount than either the white or the black screen induces. 
We have explained the limits of sensitivity to the four colors 
when black and white screens are used. We have still to explain 
the results obtained with the gray screen. Since it causes no 
brightness induction, we might expect our widest limits to occur 
with this screen. Table V and VI, however, show that while this 
is true to some extent for blue and green, it is not true for red 
and yellow. The limits for red and yellow with the gray screen 120 
GERTRUDE RAND 
of the brightness of the color are in each case slightly narrower 
than with the black screen and wider than with the white. As 
we are still working on this point, we do not at present feel 
justified in saying anything final by way of explanation. We 
may point out, though, that red and yellow darken in passing 
into the peripheral field of vision. The black screen tends to 
lessen this effect by contrast, and the white screen to augment it. 
It seems reasonable to expect, then, that the black screen, which 
lessens, by means of the white contrast, the amount of black 
fused with these colors in darkening, would widen their limits; 
and that the white screen, which increases it by means of black 
contrast, would narrow their limits, as compared with the gray 
screen, which exerts no effect at all. We can speak only tenta- 
tively, however, until the amounts of brightness dealt with in 
each case can be more accurately ascertained. 
C. The Effect of the Induction of the Surrounding Field upon 
the Color Limens. 
In order to estimate the effect of the induction of the sur- 
rounding field upon the limen of sensitivity to the different 
colors, the limens of color were determined at the center, and at 
15°, 25 °, and 30° of excentricity in the peripheral retina (a) 
when the surrounding field was of the gray of the brightness 
of the color; (&) when it was white; and (c) when it was 
black. 
The preexposure was in each case to the gray of the bright- 
ness of color. The limen was determined as follows: The 
stimulus composed of sectors of the color and the gray of the 
brightness of the color at the excentricity for which the limen 
was to be determined, was placed on the motor behind the 
campimeter screen. The proportions of the sectors were changed 
until the observer made the judgment of just noticeable color. 
Judgments were taken in ascending and descending series, and 
the average was taken as the value of the limen. 
The results show that the influence of the brightness of the 
surrounding field upon the color limen is as follows: 
1. The limen is lowest when the surrounding field is of the 
gray of the brightness of the color. EXPERIMENTAL 
121 
2. The difference in the effect of the white and black screens 
upon the limen increases from the fovea outwards. 
3. For yellow and green the limen is highest when the field 
is black and the induction white, and lower when the field is white 
and the induction black. 
4. For red and blue, the limen is highest when the field is 
white and the induction black, and lower when the field is black 
and the induction white. 
5. The difference in the effect of white and black screens on 
the limens is not so great as one at first thought might be led to 
expect from the results obtained by the objective mixing of 
white and black with color in the central retina. 
The results for Observer A are given in detail in Table IX. 
D. Explanation of the Effect of the Induction of the Surround- 
ing Field upon the Color Limens. 
We have, then, the following facts to explain: (1) The limen 
of sensitivity to color is lowest when the surrounding field is of 
the gray of the brightness of the color. This is what should 
be expected, because in case of this screen there is no induction 
present to fuse with the color sensation, and to affect the limen 
of sensitivity. (2) The difference in the effect of the white 
and black screens increases from the fovea outwards. This is 
because the sensitivity of the retina to brightness contrast in- 
creases from the fovea outwards, as the table for the amounts 
of induction shows. More white and black, then, are induced, 
and as our results with objective mixing show, the greater are 
the amounts of white and black mixed with color, the greater is 
the difference between the inhibit!ve actions of equal amounts of 
each.26 (3) The limen of sensitivity to yellow and green is high- 
"A rough demonstration of this can be easily made as follows. Set up 
two discs, of blue for example, side by side on color-mixers. Add a small 
sector of white to the one and an equal sector of black to the other, and 
observe the apparent saturations of each. Repeat the observation several 
times, each time increasing the sectors of black and white by equal amounts. 
It will be observed that the difference in the apparent saturations of the 
equally saturated discs becomes greater and greater, until at 1800 the disc 
to which white was added appears almost colorless while the disc to which 
black was added is still a well-saturated dark blue. 122 
GERTRUDE RAND 
est when the surrounding field is black, and lower when the 
surrounding field is white. This is in accord with the general 
law of the inhibitive action of white and black on color. That is, 
since color is inhibited less by black than by white, we should 
expect in terms of the law that the limen of color would be lower 
with the white screen which induces black than with the black 
screen which induces white. The limens obtained for yellow and 
green present no exception to this law. (4) The limens for red 
and blue are highest when the surrounding field is white, and 
lower when the surrounding field is black. But this is in ap- 
parent contradiction to our general law of the relative inhibitive 
action of white and black upon the colors. An explanation of 
why we have this apparent contradition in case of red and blue 
and not in case of yellow and green may be readily found, how- 
ever, in the relative amounts of contrast induced by the white 
and black screens across these colors. Table XII (p. 142) shows 
the amount of contrast that is induced by the white and the 
black screens across the grays of the brightness of the colors. 
As we have already mentioned, the white screen induces more 
black across the grays of the brightness of red and blue, than 
of yellow and green; the black screen induces more white across 
the grays of the brightness of yellow and green, than of red and 
blue. For example, Observer A estimated the amount of black 
induced by the white screen at 250 in the horizontal temporal 
meridian as 1350 for yellow, and 1550 for green; and the amount 
of white induced by the black screen as no° for yellow, and 6o° 
for green. There is, then, less white induced across these two 
colors by the black screen than there is black induced by the 
white screen. In spite of this, however, the greater inhibitive 
power of this smaller amount of white is sufficient to raise the 
limen of sensitivity to yellow and green slightly higher than it 
is raised by the less inhibitive power of the larger amount of 
black. For red and blue, on the other hand, the black induced 
by the white screen is estimated as 230° for red, and 2900 for 
blue; while the white induced by the black screen is estimated as 
only 28° for red and only 120 for blue. In these cases there is 
a very much greater amount of black induced than of white. EXPERIMENTAL 
123 
And this very much greater amount of black is sufficient to raise 
the limen of sensitivity to the colors with which it is fused 
higher than it is raised by the very small amount of white, in 
spite of the fact that when equal amounts of black and white 
are mixed with a color, its saturation is inhibited much more by 
white than by black. (5) The difference in the effect of white 
and black screens on the limens is not so great as one at first 
thought might be led to expect from the results obtained by the 
objective mixing of white and black with color in the central 
retina. This may be explained as follows. (1) The relative 
amounts of white and black induced upon the different colors by 
the screens vary greatly. We have thus not a simple case of a 
difference in the inhibitive action of equal amounts of black 
and white. In case of yellow and green, for example, there is 
so much more black induced than white that the white raises the 
limen very little more than the black. And in case of red and 
blue, the amount of black induced is so very much in excess of 
the white that the limen is raised even more by the black than by 
the white. It is not raised much more, however, (even less than 
the excess for white in case of yellow and green), because (a) 
the excess of black induction is not sufficiently large greatly to 
overweigh the superior inhibitive power of white; and (&) the 
difference between the inhibitive powers of white and black is 
high for red and blue, especially for blue. (2) The difference in 
the inhibitive power of white and black on colors decreases from 
the center to the periphery of the retina. Thus not so great a 
difference is found in the limens for white and black screens 
in the peripheral retina as one might be led to expect from the 
amounts of induction present. An inspection of the table shows 
that the difference in the limens for the white and black screens 
increases from the center towards the periphery, but this increase 
caused by the greatly increased amounts of induction27 is not so 
great as it would have been, were there no decrease in the differ- 
ence in the inhibitive power of white and black on the different 
colors. 
27 It has already been shown, footnote, p. 121, that the greater are the equal 
amounts of white and black added to color, the greater will be the differ- 
ence in the inhibitive actions exerted by these equal amounts. 124 
GERTRUDE RAND 
Table IX 
A. Showing the limens of color sensitivity with screens of 
white, black, and gray of the brightness of color. 
Stimulus 
Point on hori- 
zontal temporal 
meridian at 
which limem 
was taken 
Limen with 
screen of gray 
of brightness of 
color 
Limen with 
white screen 
Limen with 
black screen 
Yellow 
0° 
i8° 
22° 
28° 
15° 
22° 
25° 
35° 
25° 
35° 
50° 
65° 
30° 
50° 
8o° 
95° 
Green 
0° 
20° 
22° 
28° 
15° 
27° 
30° 
35° 
25° 
40° 
50° 
75° 
Red 
o° 
9° 
13° 
io° 
15° 
9° 
19° 
15° 
25° 
17° 
30° 
23° 
30° 
25° 
50° 
29° 
Blue 
o° 
9° 
17° 
10° 
15° 
10° 
25° 
12° 
25° 
12° 
35° 
i8° 
30° 
20° 
40° 
30° 
We have explained the effect of the induction of the surround- 
ing field on the limits of color sensitivity, and on the limens of 
sensitivity. As we have said, the effect on the limit takes place 
in the extreme peripheral retina; the effect on the limen has been 
measured in the more central regions of the retina,-at o°, 150, 
25 °, and 300 of excentricity. We have remaining to compare 
the effect of the induction of the surrounding field in the extreme 
peripheral retina, as estimated by the limit, with its effect in the 
more central regions, as estimated by the limen, and in turn to 
determine how both sets of effects harmonize with our law of 
the inhibitive action of brightness on the colors. The comparison 
of the results obtained by the two methods for each of the colors 
is as follows: 
1. For yellow, the limen of sensitivity was lower with the 
white screen but its limit was wider with the black screen. The 
effect of the screen upon the limen for this color is in accord with 
our general law of the relative inhibitive action of white and 
black upon the colors; and the effect of the screen on the limit EXPERIMENTAL 
125 
is in accord with its exception formulated for the extreme peri- 
pheral retina; that is, that in the region 50 from the limit for 
yellow, black inhibits yellow more than white does. 
2. For green, the limen was lower and the limit was wider 
with the white than with the black screen. The effect of the in- 
duction of the screens, then, on both limens and limits, is in ac- 
cord with our general law. 
3. For red, the limen was lower and the limit was wider 
with the black than with the white screen. The effect of the 
induction of the screens on the limen is not in accord with our 
general law that white inhibits color more than black; however, 
the exception is readily explained by the much greater amount 
of black than of white that is fused with the color sensation by 
the induction of the screens. The effect on the limit is in accord 
with our exception to this law formulated for the extreme 
peripheral retina; that is, that within the region 50 from the 
limit of sensitivity to red, black inhibits red more than white 
does. 
4. For blue, the limen was lower with the black screen, but 
the limit was wider with the white screen. The effect of the 
induction of the screen on the limen is not in accord with our 
general law, but the exception may be explained, as in case of 
red, in terms of the very much greater amount of black induced 
by the white screen than of white induced by the black screen. 
We have here, however, an apparent paradox with regard to 
the limits. That is, since the law of the relative inhibitive action 
of white and black is the same at the limit for blue as it is at 
the center, we might expect that if the black induction was 
sufficiently in excess of the white to make the limens higher for 
the white screen than they were for the black, it would also 
correspondingly make the limits narrower for the white screen 
than for the black. The reverse, however, it will be remembered, 
was true. The reason for this lies in the fact often mentioned 
previously that blue lightens in the periphery, so that near its 
limit of sensitivity it is not in so much greater contrast to the 
white screen than to the black screen as it is in the center. For 
example, for Observer A the brightness of blue in the periphery 126 
GERTRUDE RAND 
equalled gray No. 28. In the periphery, then, the amount of 
white induced by the black screen is sufficient to inhibit the blue 
sensation more than it is inhibited by the amount of black 
induced by the white screen. It may be mentioned, however, 
that the difference between the limits for blue with the white 
screen and with the black screen is smaller for all observers used 
than is the difference between the limits with these screens for 
any other color with which we worked (see Tables V and VI, 
p. 114). 
3. The Brightness of the Preexposure. 
When making the color observation in the peripheral retina, 
the observer is given a short period of preparation before the 
stimulus is exposed, in which to obtain and hold a steady and 
accurate fixation. This introduces the factor of preexposure, for 
during this period of preparation, the area which is to be stimu- 
lated by color receives a previous stimulation. It seems strange 
to the writer that this factor, which exerts a greater influence 
over the extent of color sensitivity than any we are examining, 
with the possible exception of large changes in the general illu- 
mination, should have been so generally overlooked in the work 
of earlier investigators. It has always been considered a suffi- 
cient precaution to eliminate all color from the preexposure. 
This, however, is not enough. It should also be of the same 
brightness as the color by which the eye is to be stimulated. If 
not, it gives an after-image which mixes with the succeeding 
color sensation and both reduces its saturation and modifies its 
color tone.28 If the preexposure is lighter than the stimulus 
color, it adds by after-image a certain amount of black to the 
succeeding color impression; if darker, it adds a certain amount 
of white. Since white inhibits color more than black, the effect 
of a dark preexposure is to reduce the sensitivity to color more 
23 This action takes place apparently at some physiological level posterior 
to the seat of the positive, negative, and contrast color processes commonly 
supposed to be located ini the retina. (See Ferree and Rand. An Experi- 
mental Study of the Fusion of Colored and Colorless Light Sensation: The 
Locus of the Action. Journ. of Philos. Psychol, and Scientific Methods, 
1911, VIII., pp. 294-297.) EXPERIMENTAL 
127 
than the effect of a light preexposure.29 But since both white 
and black as after-effect reduce the sensitivity to color, the eye 
is rendered more sensitive when no after-image is given, that 
is, when the preexposure is of the same brightness as the color. 
The preexposure should, therefore, be to a gray of the brightness 
of the color. No brightness after-image will be added to the 
succeeding color impression to modify either its saturation or its 
color tone. Even closing the eye, as is frequently done before 
stimulating, is equivalent to giving a black preexposure. 
No thought apparently was given by previous experimenters to 
the intense after-effect which follows the exposure of the eye to 
a brightness quality differing from that of the stimulus. Hess,30 
Fernald,31 and Thompson and Gordon,32 it is true, covered the 
stimulus before exposure with a card matching in quality the 
campimeter screen, but since the campimeter screen was not 
always of the same brightness as the color used for the stimu- 
lus, this by no means ruled out the effect of preexposure. The 
motive of each of these experimenters seems to have been to 
standardize the observation for the effect of preexposure, but no 
notion of its action sufficiently clear to guide them in formulating 
their technique seems to have been entertained. Since the action 
of the preexposure is by way of arousing a brightness after- 
image, it is obvious that the preexposure card should, as stated 
above, be matched in brightness to the stimulus color rather than 
to the screen. 
In the articles, "Colored After-Image and Contrast Sensa- 
29 A very striking demonstration of the effect of preexposure upon the 
sensitivity of the retina to color can be made for class or lecture room 
purposes as follows. Mount a sheet of the blue paper of the Hering series 
on cardboard. Cover one-half of another sheet of cardboard of the same 
size with white of the Hering series of papers, the other half with velvet 
black. Place this card immediately in front of the first card and fixate its 
center for io or 15 seconds. Remove and observe the comparative effect 
of the white and black preexposures thus obtained upon the color impression 
gotten from the blue surface. 
39 Hess, C. loc. cit. 
81 Fernald, G. M. Psychol. Rev., 1905, XII., p. 394; Psychol. Rev. Monog. 
Sup., 1909, X., No. 42, p. 17. 
32 Thompson and Gordon. A Study of After-images on the Peripheral 
Retina. Psychol. Rev., 1907, XIV., p. 123. 128 
GERTRUDE RAND 
tions from Stimuli in Which No Color Is Sensed/'33 and "The 
Fusion of Colored with Colorless Light Sensation.-The Physi- 
ological Level at Which the Action Takes Place,"3* the effect of 
the after-image due to previous brightness exposure upon color 
sensitivity has already been shown for both central and peripheral 
retina. The general fact need not further be dwelt on here. We 
do, however, need to show why in the peripheral retina the short 
preexposure which takes place while the eye is obtaining a steady 
fixation has so much effect upon the color stimulation immediately 
following. Two reasons are found for this, (a) The peri- 
pheral retina is extremely sensitive to short stimulation. While 
some slight variation is found at different angles of excentricity, 
the peripheral after-image reaches in general its maximal in- 
tensity with two or three seconds stimulation. This amount of 
time is usually consumed in obtaining fixation, hence in each 
observation there is fused with the color sensation about as 
strong a brightness after-image as can be aroused. For this 
reason alone, it is readily seen why the brightness of the pre- 
exposure is of so much greater consequence in the peripheral 
retina than it is in the central retina, where the maximal strength 
of the after-image is obtained with from forty to sixty seconds 
stimulation. (&) There is apparently no latent period in case 
of the peripheral after-image. It flashes out at full intensity im- 
mediately upon the cessation of the stimulus. Thus, there is no 
possibility of escaping the full effect of the brightness after- 
image upon the stimulus color, as might happen in the central 
retina, where the latent period obtains, if there were a very 
short exposure to the stimulus color. 
If when working with the campimeter, for example, a black 
card is used to cover the stimulus-opening during the period of 
preparation, an intensive white after-image is aroused which 
33 Ferree and Rand. Psychol. Rev., 1912, XIX., pp. 195-239. 
31 For abstract of the article, see Journ. of Philos. Psychol, and Scientific 
Methods, 1911, VIII., pp. 294-297. The article will soon be published in 
full. See also Ferree, C. E. Tests for the Efficiency of the Eye Under 
Different Systems of Illumination and a Preliminary Study of the Causes 
of Discomfort. Transactions of the Illuminating Engineering Society, 1913, 
VIII., pp. 40-60. EXPERIMENTAL 
129 
fuses with the succeeding color sensation, strongly reducing its 
saturation. If, on the other hand, a white card is used, a black 
after-image is obtained, which, according to our law of the 
action of the achromatic sensation upon color, has less effect 
than the white after-image upon blue and green, and also upon 
red and yellow if the after-image is sufficiently strong to narrow 
the limits to red and yellow more than 50. In each case, the 
intensity of the after-image will in part depend on the brightness 
of the subsequent color exposure, the projection field. The 
after-image due to preexposure to white will be more intensive 
when blue than when yellow forms its projection field. The 
after-image from black will be more intensive when projected on 
yellow or green than on blue or red. If, however, a gray card 
of the brightness of the stimulus color be used as preexposure, 
there will be no after-image to modify the color sensation. The 
only brightness change acting upon it will be due to the slight 
adaptation to this gray during the short time of preexposure. 
The method, then, of eliminating the effect of preexposure 
consists in making it of the brightness of the color to be used 
as stimulus. And in case the brightness of the color alters in 
passing from the center to the periphery of the retina, the bright- 
ness of the preexposure must be correspondingly altered. For 
example, at standard illumination, Hering Gray No. 41, or its 
equivalent should be used for preexposure to blue when the 
central retina is investigated. But for the peripheral retina, 
a much lighter gray should be used because in this region blue 
lightens by an amount depending upon the excentricity of the 
stimulation and in part upon individual variation. 
A. Effect upon the Limens of Color and upon the Limits bf 
Color Sensitivity. 
To test the importance of preexposure, two methods of 
measurement were employed. In the first, the limen of color 
was obtained at o° and at 350 on the temporal meridian, when 
the screen was of the gray of the brightness of color, and the 
preexposure was in turn to the same gray, to white, and to 
black. In the second method, the limits of color sensitivity 
were investigated under the same conditions of campimeter 130 
GERTRUDE RAND 
screen and preexposure card. The results for Observer A are 
recorded in Table X and in rows 1, 4, 7, 10 of Table XI. They 
show in every case, (a) that the limen is raised and the limit 
is considerably narrowed when the preexposure is not to the 
gray of the brightness of the color, that is, when it gives a 
brightness after-image; and (&) that the limen is higher and 
the limit narrower when the preexposure is to black and its 
after-image is white, than when the preexposure is to white and 
its after-image black.35 This is in accord with the law of the 
35 There is one exception to this statement. The limen for blue at 0° 
for both white and black preexposures is 130. The following reasons may- 
be given for this, (a) It is one of the fundamental laws of brightness 
after-images that the intensity of the after-image depends in part upon the 
brightness relation of stimulus to projection field. When the brightness 
difference between the stimulus and the projection field is small, a weak 
after-effect is obtained; when it is greater, a more intensive after-effect 
is obtained. Blue, as aroused in the central retina, is very near to black 
in brightness, and very far removed from white. We should then expect 
a very much more intensive after-effect of the exposure to white than 
to black when the projection! field is blue, (b) The brightness relation of 
the surrounding field to the preexposure also exerts an effect on the intensity 
of the after-mage given by the preexposure. When the surrounding field 
differs in brightness from the preexposure, contrast is induced. This con- 
trast quality in turn also gives an after-image which mixes with and modi- 
fies the after-image given by the preexposure. When the surrounding field 
is of the gray of the brightness of blue, for example, and the preexposures 
are in turn white and black, the influence of the surrounding field is to 
make the after-image of white stronger than that of black. That is, when 
the preexposure is white, this white is strongly intensified by contrast with 
the surrounding field of dark gray, and in consequence the black after-image 
is strongly intensified. But when the preexposure is black, little intensifica- 
tion results by contrast with the surrounding field and little effect is had 
on the after-image. Both of these influences, then, tend to cause much 
more black to be added to a blue stimulus in the central retina as a result 
of preexposure to white than white to be added as a result of preexposure 
to black. The effect of this excess of black, as is shown by the table, is to 
raise the limen for blue for the white preexposure as much as it is raised 
by the black preexposure; in other words, to make the limens equal. At 35° 
in the periphery, however, the limen of blue is seen in the table to be 
higher when the preexposure is to black and the after-image white, than 
when the preexposure is to white and the after-image black. We have here 
a different case. The difference in the effect of white and black preex- 
posures at 0° and at 350 is due to the fact that at 0° blue is very dark 
while at 350 in the periphery it has lightened until its brightness 
equals No. 35 gray. In the latter case (a) there is less difference than 
there was at 0° between the brightness relations of the stimulus to EXPERIMENTAL 
131 
action of brightness upon color and holds for all colors used. 
It might be expected from the exception to this law, mentioned 
on page 116, that the limits for red and yellow would be narrowed 
more by the black after-image than by the white. This is not 
found to be true because the effect of each after-image is suf- 
ficiently strong to narrow the limits of red and yellow more 
than 5°, and thus to carry the limits for these colors outside of 
the zone in which they are inhibited more by black than by white. 
B. Combined Effect of Surrounding Field and Preexposure 
upon the Limits of Color Sensitivity. 
The effect of preexposure upon color limits was also investi- 
gated when white and black screens were used. The results 
obtained are of interest in two regards, (a) They show under 
these typical conditions in what way and to what extent the in- 
ductive action of the screen combines with the effect of preex- 
posure to modify the limits of color, (b) They help to explain 
some of the conflicting results obtained by previous investigators 
who did not carefully standardize their observations with refer- 
ence to the effect of preexposure and surrounding field. A few 
points may be noted in advance of the tables showing the 
combined action of preexposure and surrounding field upon the 
extent of color sensitivity. It is obvious that when the 
campimeter screen is either white or black and the preexposure 
is to the same brightness quality, there will be no inductive 
the white preexposure and of the stimulus to the black preexposure, and 
consequently less difference between the intensities of the after-images 
from these preexposures; and (b) there is less difference than there was 
at 0° between the brightness relations of the surrounding field, which is 
of the brightness of blue at the point at which we are working, to the 
white preexposure and of the surrounding field to the black preexposure, 
and for this reason also, less difference between the intensities of the 
after-effect of the contrast induced by the surrounding field upon these 
preexposures. For both reasons, therefore, the after-image from the white 
preexposure at 35°, when projected on blue, is not so much more intense than 
the after-image from the black preexposure, as it was at 0°. At 35°, then, 
in these experiments the white after-effect due to preexposure to black is 
sufficiently intensive to raise the limen of blue higher by virtue of its greater 
inhibitive power than it is raised by the black after-image due to pre- 
exposure to white. 132 
GERTRUDE RAND 
action by the screen upon the preexposure either to intensify or 
to weaken it. In this case both preexposure and screen will 
add the same brightness quality to the stimulus color, the former 
by contrast, and the latter by after-image. The effect of this 
action upon the color limits is shown in Table XI, Column 4, 
rows 2, 5, 8, 11; and column 5, rows 3, 6, 9, 12. If, however, 
the preexposure and the screen are different in quality, their 
action may be either antagonistic or supplementary, depending 
upon the brightness relations between the screen and the pre- 
exposure, on the one hand, and between the screen and the 
color sensation fused with the after-image of preexposure on the 
other. For example, if a black preexposure and a white campi- 
meter screen are used, the white screen will intensify the 
blackness of the preexposure by contrast but will tend to darken 
the fusion of color sensation and white after-image, and thus 
will lessen the action of the latter upon the color. This effect 
is shown in Table XI, in column 4, rows 3, 6, 9, 12; and 
column 5, rows 2, 5, 8, 11. If, however, the preexposure is black, 
and the campimeter screen is the gray of the stimulus color, the 
Table X. 
A. Showing the effect upon the color limens of preexposure 
of the gray of the brightness of the color used, of white, and of 
black. 
Stimulus 
Surrounding 
field 
Preexposure 
Limen at o° 
Limen at 35° 
Yellow 
gray no. 2 
gray no. 2 
. 180 
83° 
white 
35° 
105° 
black 
45° 
125 0 
Green 
gray no. 8 
gray no. 8 
20° 
IOO° 38 
white 
33° 
155° 
black 
40° 
1800 
Red 
gray no. 24 
gray no. 24 
9° 
90° 
black 
I3V 
135° 
white 
20° 
148° 
Blue 
gray no. 41 at o° 
gray 
9° 
30° 
no. 35 "35° 
white 
13° 
40° 
black 
13° 
43° 
38 As green has a narrow zone of visibility, the limen was taken at 30°. EXPERIMENTAL 
133 
screen will intensify the blackness of the preexposure by contrast, 
but will lighten the fusion of stimulus color and white after- 
image, thus will add to the action of the latter upon the color. 
This effect is shown in Table XI, columns 4 and 5, rows 1, 4, 
7, 10. 
Table XI. 
A. Showing the combined effect of campimeter screen and 
preexposure upon the limits of color sensitivity. 
Stimulus 
Campimeter 
screen 
Limit with gray 
preexposure of 
the; brightness 
of the color 
Limit with 
white 
preexposure 
Limit with 
black 
preexposure 
Yellow 
gray no. 2 
44° 
40° 
38° 
white 
42° 
42° 
43° 
black 
45° 
43° 
40° 
Green 
gray no. 8 
37° 
35° 
3i° 
white 
36° 
30° 
25° 
black 
34° 
36° 
25° 
Red 
gray no. 24 
43° 
38° 
37° 
white 
42° 
38° 
39° 
black 
44° 
43° 
38° 
Blue 
gray no. 28 
53° 
420 87 
420 87 
white 
50° 
420 37 
48° 
black 
49° 
51° 
420 37 
Having seen from Tables X and XI the effect of preexposure 
and the combined effect of preexposure and surrounding field 
upon the limen and the limits of color, we may turn our attention 
to a third point of interest; namely, the explanation of some 
of the differences between our results and those obtained by 
earlier investigators in terms of the different conditions of pre- 
exposure used. As Hess and Fernald alone have stated their 
conditions of preexposure, we shall have to limit our discussion 
to them. Hess wished to find the comparative limits of color 
sensitivity. He attempted to standardize the intensity and the 
brightness of the stimuli used, and worked with white, black, 
and gray campimeter screens. The eye of the observer followed 
37 For the explanation of this decided narrowing of the limit for blue, see 
footnote, p. 108. 134 
GERTRUDE RAND 
a moving fixation-point while the stimulus was exposed from 
time to time. In every case, the stimulus was covered by a card 
of the same brightness and quality as the screen. He turned 
the gray screen toward or away from the source of light until 
its brightness was such that the color disappeared in the peri- 
phery into a gray of the brightness of the screens. He thus 
worked with screens of the gray of the brightness of the color, 
of white, and of black, and with preexposures of the same 
brightness as the screens. In only one of these cases, namely, 
the first, has he eliminated the effect of preexposure. He found 
that for every color, the limit of sensitivity was widest with the 
gray screen, narrower with the black, and narrowest with the 
white. His results are in contradiction to ours with regard to 
the influence of these screens, when the factor of preexposure 
is eliminated (see p. 113). They are, however, very nearly 
confirmed by the results we obtained when the preexposure was 
of the same brightness as the screen (See Table XI). Had 
our illumination been slightly less, and consequently the induction 
from the white screen greater, our results would have been very 
similar to those of Hess. 
Fernald, working with a white and a black background and 
using preexposures to match, obtained results that are similar 
to those of Hess for all colors except red, and even in this case 
the exception is only apparent. She says: "All the colors 
except the reds are perceived at a greater degree of eccentricity 
with the dark than with the light background. Red is seen as 
red to about the same degree of eccentricity with the dark as 
with the light background, but is seen as yellow or orange with 
the dark background at the same points at which it is seen as 
colorless with the light background."38 The latter part of the 
quotation shows that the exception to Hess in case of red was 
due rather to a difference in the method of measurement than 
to a difference in result. She recorded as the limit of color the 
point at which the sensation took on any trace of a foreign 
quality. Although the red stimulus appeared as red-yellow in 
the periphery with the black screen, there was some red in the 
38 Fernald, G. Psychol. Rev. Monog., 1909, X, p. 23. EXPERIMENTAL 
135 
sensation received from it at a greater angle of excentricity 
with the dark than with the light screen. 
As was said above, the conclusion of Hess and of Fernald 
that the colors have wider limits with the black than with the 
white screen, is confirmed by our work at certain illuminations 
when white and black preexposures are used. But, since we do 
not obtain similar results when the influence of preexposure is 
eliminated and the effect of the brightness of the field determined 
in isolation from this factor, we maintain that their results were 
due to brightness conditions not connected with the surrounding 
field, but with the preexposure. Our contention is that many of 
the conflicting conclusions concerning the effect of background 
upon color limits have resulted from ignorance of the important 
factor of preexposure. Comparative color limits can be obtained 
with any screen only when the preexposure is to the gray of 
the brightness of the stimulus. This principle was used by Hess 
in his work with the gray screen, but apparently without a 
definite purpose and without knowledge of its importance. 
4. The General Illumination of the Retina. 
The effect of change of illumination was forced upon our 
attention early in the investigation of the factors that influence 
the color sensitivity of the retina. For example, in preliminary 
work done by the writer on a well-lighted porch on Long Island, 
changes in color tone were reported, when certain colors were 
compared in the central and in the peripheral retina, that are 
not found at all under the more intensive illumination of our 
optics-room, when neither of the curtains is drawn; and the 
peripheral limits of color were narrower by 50 to 120. Further- 
more, on a dark day, it was found that the limits of stimuli 
exposed through an opening in a white screen were reduced by 
about 40 as compared with the limits taken on a bright day 
The change was less considerable with black and gray screens. 
The change in color tone was most conspicuous in case of green.39 
On dark days the green stimulus appeared as a pale unsaturated 
blue before becoming colorless in passing from the center to the 
periphery of the retina. This zone of blue was from 70 to 230 
89 The green of the Hering series was used. 136 
GERTRUDE RAND 
wide in different meridians of the retina with both white and 
black screens, but was wider with the black than with the white 
screen. On a sunny day, on the other hand, with the white 
screen green passed into bluish-green, then directly into gray, 
except in case of the upper regions where it appeared blue 
throughout a zone of about 40 in width. With the black screen, 
the blue zone was found only in the upper and temporal regions of 
the retina. The transition of green to yellow in the periphery that 
is generally reported in the literature was found in these experi- 
ments only when the gray screen was used. Yellow showed a color 
change that varied in amount with the degree of the general 
illumination. On a bright day, it appeared reddish-orange with 
the white screen. On a cloudy day, it was seen in the extreme 
periphery as a dark saturated red. 
Working in our optics-room we found also that results taken 
on one day could not at all be duplicated on the following day. 
When the work was carried on under the most favorable con- 
ditions without special means of controlling illumination, namely, 
on bright days only, differences of 50 or more were found when 
the white screen was used. This necessitated a long series of 
observations if legitimate averages were to be obtained. Such 
a procedure is at best a poor makeshift and is besides of great 
disadvantage in many problems that come up in the work on 
color sensitivity. Particular instances of this may be found in 
investigations in which it is required to work in the region lying 
just within the limits of sensitivity, and in work on the after- 
images of stimuli in which no color is sensed. In the latter case 
the experiment requires that the stimulus be exposed just outside 
the limits of sensitivity determined with a given brightness 
condition, and that the observer should not be aware of the 
nature of the stimulus. In order to fulfill these requirements 
the experimenter must know the limits obtaining with a given 
brightness condition. It would be impossible to know this when 
the brightness conditions were subjected to the influence of 
changing illumination unless re-determinations were made at the 
beginning of each sitting and even frequently during its course. 
This would consume a great deal of time and would, besides, EXPERIMENTAL 
137 
only roughly fulfill the requirements of the problem. A further 
and still more important example of the disadvantage may be 
found in the task we had set ourselves, namely, to investigate 
from point to point the sensitivity of the retina to each of the 
principal colors for three backgrounds in at least sixteen dif- 
ferent meridians. In this work it is obvious that unless a 
standard illumination were provided, all comparative work would 
have to be done at one sitting. This is impossible. When time 
is taken between observations to guard against fatigue, at least 
three hours is required merely to outline the limits of sensi- 
tivity for a given color with one background for only one-half 
of the retina. Even for this length of time there is no guarantee 
that the illumination has not altered. Thus at the outset of any 
extended investigation of color sensitivity, it is evident that 
without a standard illumination, results will be of little com- 
parative value. 
In order better to know our factor and the ways in which 
it operates, a systematic investigation of the influence of changes 
of general illumination was carried on in our optics-room, which 
is especially constructed to secure fine changes in illumination. 
Rough preliminary experiments showed that the primary effect 
of decreasing the illumination was an alteration of the amount 
of contrast induced across the stimulus by the campimeter screen. 
With the white screen, the increased induction was the most 
pronounced and was sufficient to cause large changes in the 
limits and in the color tone of the stimulus. In order to investi- 
gate this effect in detail, gradual changes of illumination cover- 
ing a wide range were made by means of the curtains with which 
our optics-room is furnished, described on p. 86. Attention was 
given to the following points, (a) A quantitative estimate was 
made of the influence of change of illumination upon the bright- 
ness induction of the campimeter screen. (&) The effect of this 
induction upon the limits of color sensitivity was determined, (c) 
The limens of the colors were measured at different degrees of 
excentricity at different illuminations. And (d) the influence 
of change of illumination upon the effect of the preexposure on 
the limens and limits of color was investigated. The degrees 138 
GERTRUDE RAND 
of illumination chosen for comparison were the standard illu- 
mination, the method of obtaining which will be described later, 
and the decreased illumination mentioned above (see p. 95 and 
Tables I and II) for which the white-values of the stimuli were 
determined. Measured in foot-candles by means of the Sharpe- 
Millar portable photometer, the standard illumination equalled 
390 foot-candles, the decreased 1.65 foot-candles. 
A. Quantitative Estimate of the Influence of Change of Illumin- 
ation upon the Induction of Brightness by the Surrounding 
Field. 
The purpose of this investigation was to find out how much 
the induction from white and black screens40 is affected by a 
change in the general illumination; and (b} how much induction 
is gotten at decreased illumination from the gray screen which 
matches the color in brightness at standard illumination. The 
induction in this latter case is caused by the change in the bright- 
ness relation between color and screens with decrease of illu- 
mination.41 The campimeter screens served as inducing surface, 
grays of the brightness of the four principal colors of the Hering 
series both at standard and decreased illumination were used 
in turn as stimuli, and the amount of induction was estimated 
upon a measuring-disc, made up of adjustable sectors of the 
gray of the stimulus and white or black, according to the screen 
used. The measuring-disc was mounted on a motor (see p. 87 
and Fig. I) which could be moved along the graded arm of the 
campimeter to any position from 200 to 92 °. The gray stimulus 
was exposed through the opening of the screen in the usual man- 
ner. Two preliminary precautions were observed, (a) Since 
40 White and black screens are chosen because they represent the extreme 
cases of the effect of change of illumination. 
"• This latter determination is made to show that it is impossible to stan- 
dardize the brightness of the surrounding field against the sudden and 
progressive changes of daylight that occur during the course of a single 
series of observations. These changes alter the brightness relation be- 
tween the colored stimulus and the gray used as screen; therefore a 
match made at the beginning of a series will not hold throughout its course. 
For the same reason and to an equal degree the brightness relation between 
preexposure and colored stimulus changes with change of illumination. It is, 
therefore, equally impossible to standardize the brightness of the pre- 
exposure without some means of securing a standard illumination. EXPERIMENTAL 
139 
the brightness of the gray stimulus plus the induction of the 
screen was to be estimated by means of the measuring-disc, and 
since the brightness-value of the stimulus and of the disc changes 
with the amount of light that falls upon them, it was necessary 
to make sure before each measurement that the same amount of 
light fell upon each. This precaution was all the more nec- 
essary because the stimulus had to be placed behind the screen 
and the measuring-disc in front. In a given position of the 
apparatus, one or the other was apt to be shaded. The determina- 
tion was made as follows: Measuring-disc, campimeter screen, 
and gray stimulus were all given the same brightness-value ac- 
cording to determinations made under conditions about which no 
doubt of the equality of the illumination of each could be enter- 
tained. Each was then placed in position for the experiment, 
and the position of the campimeter as a whole and of its various 
parts was adjusted until stimulus, screen, and measuring-disc 
were exactly matched in brightness-value. When an exact match 
was obtained we were guaranteed that all three were again 
equally illuminated. This precaution was particularly necessary 
in the investigation we are discussing in this section. It was 
carefully observed, however, throughout the entire work. (&) 
The question arose whether brightness induction comes to its 
maximal value at once in the peripheral retina. A determination 
of the intensity curve of the contrast sensation was accordingly 
made at various points in the peripheral retina. It showed that 
contrast increases strongly for the first few seconds of stimula- 
tion. For this reason it was found to be necessary to make the 
judgment concerning the amount of induction of the screen, just 
as long after the induction had commenced as was done in the 
experiments to determine color sensitivity. In the color experi- 
ments an interval has to be allowed before the stimulus is 
exposed during which the observer obtains a steady fixation. 
During this interval of preexposure, the eye is being stimulated 
by the campimeter screen and by the card which covers the stimu- 
lus. To prevent the preexposure card from giving a brightness 
after-image which would fuse with and modify the sensation 
immediately following, it should be chosen of a gray of the GERTRUDE RAND 
140 
brightness of the color. In the same way, an interval had to be 
given in which to secure steady fixation when the amount of 
brightness induction was being measured. In order, then, to 
have the judgments made in each case the same length of time 
after induction had begun it was necessary only to make the 
intervals of preexposure of equal duration and to require that 
the judgments of each kind be made directly at the end of the 
preexposure. In the case of the color experiments, the signal 
for the making of the judgment is the withdrawal of the pre- 
exposure card and the exposure of the stimulus. For the judg- 
ments of induction, however, in which case the stimulus was the 
gray of the brightness of the color, it is obvious that no preex- 
posure card was needed, for preexposure and stimulus were 
required by the conditions of the experiment to be the same. In 
this case, a word-signal had to be given to indicate the termina- 
tion of the preexposure interval and the instant at which the 
judgment was to be made. 
Results 'when white and black screens were used.- 
Observing these precautions as to the equality of the 
illumination of stimulus, screen, and measuring-disc, and as 
to the length of time the induction had had in which to increase 
before the judgment was made, measurements were taken of 
the induction by white and black screens across grays of the 
brightness of the four principal colors at the illumination used. 
These measurements were made at various points of excentricity 
on the retina, and for both standard and decreased illuminations. 
The determination of the equality point between the stimulus and 
the measuring-disc was made as follows: The size of the white 
or black sector of the latter was changed until a preliminary 
judgment of equality was made. Then the j. n. d. on either side 
of this point was determined both by ascending and by descending 
series and an average of the results was taken as the value of the 
induction. Measurements were taken at 250 and at 400 on the 
temporal meridian, and at 550 and 700 on the nasal. The con- 
ditions at the nasal 550 were very similar to those at 250 on the 
temporal side. The measurements at 700 nasal were midway 
in value between those at 250 and at 400 on the temporal. The 
400 point is very near the limits of color sensitivity in this 
meridian, and the induction here is very great. For one ob- EXPERIMENTAL 
141 
server, the darker stimuli appeared black at this point, when the 
white background was used. In such cases, the difference be- 
tween the induction at standard and at decreased illumination 
is more clearly shown by the observations made at 250 temporal 
meridian and at 55 0 and 70 0 nasal meridian than at 40 0 temporal. 
We have, however, chosen for two reasons to present in the 
following table only the results obtained in the temporal meridian, 
(a) The results obtained in this meridian demonstrate sufficiently 
well all the facts that need be taken into consideration. Space 
will not, therefore, be given to the results for both meridians. 
(&) The second point of our problem requires us to correlate the 
increased amount of induction caused by a given decrease of 
illumination with the change in the color limits it produces. The 
limits of color sensitivity can be more easily investigated in the 
temporal meridian because the sensitivity to some colors extends 
in the nasal region beyond the 920 point, which is the limit of 
measurement for the apparatus we used. This is true in particular 
in case of Observer C as may be seen in Table VI. Both 
purposes of the investigation are, then, better satisfied by results 
obtained in the temporal meridian. 
The results show in general the following facts. 
(i)The amount of induction increases with the distance from 
the fovea. 
(2) The amount of induction increases with decrease of 
illumination.42 
(3) The amount of induction from the white screen is greater 
than that from the black screen.4211 
(4) The amount of increase of induction at decreased illumi- 
nation is greater in case of the white screen than in case of the 
black screen. 
(5) The white and black screens induce most across the 
stimuli that are farthest removed from them in brightness, and 
least across those which are nearest to them in brightness. That 
42 This statement is meant to apply only to the range of illumination worked 
with. The induction was not measured when the illumination was very low, 
nor when it was very intensive. 
42a An exception to this statement of result occurs in case of gray No. 2 
at 400. This stimulus is so near to white in brightness that the induction 
across it, according to the principle stated in (5) above, is greater for 
the black screen than for the white. 142 
GERTRUDE RAND 
is, the white screen induces more black across the gray of the 
brightness of blue than across the gray of the brightness of 
yellow; the black screen induces more white across the gray 
of the brightness of yellow than across the gray of the brightness 
of blue. 
Results are given in detail in Tables XII and XIII. Table 
XII gives the results for observer A taken on the temporal 
meridian, and Table XIII, the results for Observer C for the 
same meridian. There is some difference in the amount of induc- 
tion reported by the different observers, but since the preceding 
general statement of results is clearly borne out in every case, it 
is not deemed necessary to give space to results from all the 
observers used. In these tables, column 1 gives the degree of 
excentricity at which the observation was made; columns 2, 3, 
and 4, show respectively the stimulus used, and the amounts of 
induction from the white and from the black screens at standard 
illumination. Columns 5, 6, and 7, give the same data for 
decreased illumination. 
Table XII. 
A. Showing the amount of contrast induced by the white 
and the black screens at standard and decreased illumination 
upon the grays of the brightness of the colored stimuli at standard 
and at decreased illumination.^ 
Standard illumination 
Decreased illumination 
Stimulus 
(gray of 
Stimulus 
(gray of 
c 
arightness of 
Amt. indue- 
Amt. indue- 
brightness of 
Amt. indue- 
Amt. indue- 
4-> 
each of the 
tion of white 
tion of black 
each of the 
tion of white 
tion of black 
£ 
four colors 
at standard 
illumination) 
screen 
screen 
four colors 
at decreased 
illumination) 
screen 
screen 
25° 
gray no. 2 
Black 1350 
iWhite 1 io° 
gray no. 2 
Black 2200 
White 1700 
gray no. 8 
" 155° 
" 6o° 
gray no. 6 
" 2700 
" 8o° 
gray no. 24 
" 2300 
" 28° 
gray no. 41 
" 320° 
" 400 
gray no. 37 
" 2900 
" 12° 
gray no. 20 
" 330° 
" 30° 
40° 
gray no. 2 
" 200° 
" 300° 
gray no. 3 
" 3200 
" 360° 
gray no. 8 
" 3000 
" 1320 
gray no. 5 
" 360° 
" 1800 
gray no. 24 
" 360° 
" 6o° 
gray no. 50 
" 360°" 
" 0° 
gray no. 29 
" 360° 
" 28° 
gray no. 13 
" 360° 
" 100° 
43 It is obvious that the method used in this and the following tables of 
expressing the amount of brightness induction gives an underestimation. EXPERIMENTAL 
143 
Table XIII. 
Observer C. 
25° 
gray no. 2 
Black 700 
White 55° 
gray no. 2 
Black 1300 
White 700 
gray no. 8 
" 84° 
" 48° 
gray no. 6 
" 155° 
" 59° 
gray no. 24 
" 93° 
" 30° 
gray no. 40 
" 187° 
" 45° 
gray no. 37 
" 1600 
" 15° 
gray no. 17 
" 244° 
" 22° 
40° 
gray no. 2 
" 110° 
" 200° 
gray no. 3 
" 216° 
" 340° 
gray nd. 7 
" 1420 
" 160° 
gray no. 4 
" 2300 
" 320° 
gray no. 24 
" 1800 
" 95° 
gray no. 50 
" 360°" 
" o° 
gray no. 29 
" 214° 
" 35° 
gray no. 7 
" 3000 
" 1080 
Results when the gray screen matching the colored stimulus in 
brightness at standard illumination is used. It was necessary to 
perform the experiments bearing on this point at decreased 
illumination only. For them the campimeter screens which 
matched in brightness the four principal colors of the Hering 
series at standard illumination served as inducing surfaces. For 
Suppose, as is shown in Table XII, that No. 24 Hering gray has been 
darkened by induction until it matches in brightness a disc made up of 230° 
of black and 130° of the No. 24 gray. The amount of induction is greater 
than is represented by the 2300 of black because the induction has not 
lessened the amount of light coming to the eye from the gray paper while 
the addition of 2300 of black to the measuring-disc has cut off approximately 
% of the light coming from the gray paper. That is, in the one case 
enough black has been added by induction to reduce 360° of No. 24 gray 
to the given point in the brightness scale, while in the other enough black 
was added by direct mixing to lower only 1300 of No. 24 gray to this point 
in the scale. Moreover, the underestimation will be increased by this method 
of measuring in proportion as the amount of induction is increased because 
the greater the induction is the more black and the less gray will have to be 
used in the measuring-disc. All that can be said accurately is that a certain 
gray darkened or lightened by induction matches in brightness a gray made 
up of a certain amount of the given gray plus a certain amount of black or 
white. The exact amount of the induction can not be separated out. Further 
just because the brightness added by contrast does not alter the amount of 
light coming to the eye while the brightness added in any method of 
measurement does change this amount of light, the writer knows of 
no way by which an exact expression can be obtained. The method she has 
used, however, does serve as a means of comparing the amounts of induction 
occurring under different conditions sufficiently accurately for her purpose 
at this point. 
44 The gray No. 50 was in reality rendered blacker by the inductive action 
of gray No. 24 than the Hering black we used on the measuring-disc. A 
match thus could not be attained with black 360° as the table indicates. 144 
GERTRUDE RAND 
the contrast surfaces, grays of the brightness of these colors at 
decreased illumination were chosen. The methods of measuring, 
precautions in working, parts of the retina investigated, etc., 
were the same as in the preceding determinations. The following 
general statement of results may be made. 
i. At the 250 point the brightness of yellow was found not 
to have changed at all with the decrease of illumination produced 
by changing the illumination from the value selected as standard 
to the value selected for the comparison; the brightness of 
green lightened by an amount equal to the difference between 
No. 8 and No. 6 of the Hering series of grays; red darkened 
by an amount equal to the difference between No. 24 and No. 40; 
and blue lightened by an amount equal to the difference between 
No. 32 and No. 20. The amount of induction by the gray 
screen of the original brightness of the color upon the gray 
stimulus of the brightness of the color as altered by the de- 
creased illumination, expressed in terms of Hering white and 
black, was for yellow o°, for green 6o° of white, for red 270 of 
black, and for blue 200 of white. 
2. At the 400 point, the yellow darkened by an amount 
equal to the difference between No. 2 and No. 3 of the Hering 
grays; green lightened by an amount equal to the difference 
between No. 8 and No. 5; red darkened by an amount equal to 
the difference between No. 28 and No. 50; and blue lightened 
by an amount equal to the difference between No. 28 and No. 13. 
The amount of induction produced by these changes was for yel- 
low 280° of black, for green 1300 of white, for red 360° of 
black, and for blue 6o° of white. These results are shown in 
detail in Table XIV. 
(B.) The Effect of These Amounts of Induction upon the Limits 
of Color Sensitivity. 
In order to obtain an estimate of the range of effect upon the 
limits of color sensitivity of the induction of the screens at stan- 
dard and at decreased illumination, the breadth of the color zones 
was determined at both illuminations (a) when white and black 
served in turn as campimeter screens; and (&) when a gray 
matching the color in brightness at standard illumination was EXPERIMENTAL 
145 
Table XIV. 
A. Showing the amount of contrast induced at decreased 
illumination on grays of the brightness of the colors at decreased 
illumination by the gray screens matching the colors in brightness 
at standard illumination. 
Fixation 
Stimulus 
Screen 
Amount of Induction 
25° 
gray no. 2 
gray no. 2 
0 
gray no. 6 
gray no. 8 
white 6o° 
gray no. 41 
gray no. 24 
black 270 
gray no. 20 
gray no. 37 
white 200 
40° 
gray no. 3 
gray no. 2 
black 280° 
gray no. 5 
gray no. 8 
white 1300 
gray no. 50 
gray no. 24 
black 360° 15 
gray no. 13 
gray no. 29 
white 6o° 
used. The preexposure was in each case to gray of the same 
brightness as the stimulus at the illumination used. 
Results when white and black screens were used. When the 
stimulus color is gotten by reflection from a pigment surface, 
two factors operate to give a change of result when the illumina- 
tion is decreased. (1) There is a decrease in the amount of 
colored light coming to the eye. (2) There is an increase in the 
inductive action of the screen due to the change in the brightness 
relation of the stimulus to screen and to the increased sensitivity 
of the eye to brightness contrast at decreased illumination. 
In order to find out how much of our results with the white 
and black screens should be attributed to the decrease in the 
amount of colored light coming to the eye produced by the 
decreased illumination, and how much to the increased inductive 
actions of the screens, the limits of sensitivity were also deter- 
mined at both illuminations with the screens of the gray into 
which the color disappears in the peripheral retina. From the 
values obtained with the three screens at both illuminations, the 
amount of change due to decrease in the amount of colored 
light coming to the eye and the amount due to induction by the 
white and black screens were calculated as follows, (a) From 
"The gray No. 50 was in reality rendered blacker by the inductive action 
of gray No. 24 than the Hering black we used on the measuring-disc. A 
match could not be thus attained with black 360° as the table indicates. 146 
GERTRUDE RAND 
the number of degrees expressing the limits for a given color at 
standard illumination with a screen of the brightness of the color 
at that illumination was subtracted the number expressing its 
limit at decreased illumination, with a screen of the brightness 
of the color at the decreased illumination. That this gave the 
number of degrees the zone of sensitivity was narrowed by the 
decrease in the energy of the stimuli, may be said with the follow- 
ing qualification. If there is any influence upon color sensitivity 
of the local brightness-adaptation of the retina produced by the 
change in the general illumination, it is, of course, included in this 
effect. But, since this influence would have to be brought about 
by previous exposure to the illumination in question, it can be 
reduced to a minimum by guarding against an exposure to it 
for any considerable length of time. The effect of whatever 
adaptation there may be, however, can not be isolated or sepa- 
rated out from the above result, and the value expressing the 
amount the limit is narrowed by the actual decrease of the energy 
of colored light coming to the eye cannot, strictly speaking, 
be obtained. But it is probable that the adaptation effect is not 
sufficiently strong to influence the limits, since the sensitivity of 
the extreme peripheral retina falls off very abruptly from point 
to point. The difference, then, between the color limit obtained 
at standard illumination and the limit at decreased illumination, 
when in both cases there is no brightness induction from the 
screen, may be said to approximate the effect upon the limits 
produced by the decrease in the amount of colored light coming 
to the eye. (b) Figures can be obtained, however, from our 
results, which express the amount by which the zones are nar- 
rowed by the change in the inductive action of the white and 
black screens produced by decreasing the illumination, that are 
not open to theoretical questioning; for the influence of local 
brightness-adaptation, if there be any, is a constant for all screens 
at the same illumination. If then, the number of degrees which 
expresses the limits of sensitivity for either the white or the 
black screen at decreased illumination is subtracted from the 
number expressing the limit with a screen of the gray of the 
brightness of the color at this illumination, the result will rep- EXPERIMENTAL 
147 
resent the extent to which the limit was narrowed by the action 
of induction alone. 
The results show in general the following facts: 
i. At standard illumination, induction from the white screen 
narrows the limits of yellow and red; induction from the black 
screen narrows the limits of blue and green. The difference is 
in no case more than 40. 
2. At decreased illumination, the induction from the white 
screen narrows the limits of all the colors much more considerably 
than does the induction from the black screen.46 
3. The values expressing the narrowing of the limits caused 
by decrease of illumination without induction, are greatest in 
case of those colors which undergo maximum change of bright- 
ness in passing into the periphery, namely, for blue and red. 
We have shown by the results of the preceding section, that 
the increased induction produced by decrease of the general 
illumination is greater for the white screen than for the black, 
and, by the results of this section, that this increase is effective 
to the extent of narrowing the limits of sensitivity to all colors 
from 50 to 130 with this screen. With the black screen, the 
limits were narrowed from o° to 6°. At standard illumination, 
the limits were narrowed only from 10 to 40 with either the white 
or the black screen. 
Results in detail are given in Tables XV and XVI 
taken from the temporal meridians of the observers whose 
observations are recorded in Tables XII and XIII. In column 1, 
Tables XV and XVI, is given the stimulus. Column 2 shows 
the limit of sensitivity to the stimulus at standard illumination 
with a screen of a gray of the brightness of the color at standard 
illumination; column 3 shows the limit with a white screen; and 
column 4 with a black screen. Column 5 shows the limit at 
decreased illumination with a screen of the brightness of the 
4' For Observer A the results for green present an exception. At the 
decreased illumination used the green stimulus appeared bluish in the central 
retina. The induction of the black screen caused it to appear as a pale 
blue at a comparatively slight degree of excentricity. According to our 
definition of color limit, this point is the limit of green. It is, however, 
obvious that the exception is due rather to the qualitative than to the quanti- 
tative effect of brightness upon color. 148 
GERTRUDE RAND 
color at decreased illumination; column 6 shows the limit with 
a white screen; and column 7 with a black screen. 
Table XV. 
A. Showing the color limits at standard and decreased illum- 
ination (a) with gray screens of the brightnesses of the colors 
at the illumination used; and (b) with white and black screens. 
Standard Illumination 
Decreased Illumination 
Stimulus 
Limit with 
gray screen 
of brightness 
of color at 
standard 
illumination 
Limit with 
white screen 
Limit with 
black screen 
Limit with 
gray screen 
of brightness 
of color at 
decreased 
illumination 
Limit with 
white screen 
Lmit with 
black screen 
Yellow 
Green 
Red 
Blue 
44° 
37° 
43° 
53° 
42° 
36° 
42° 
50° 
45° 
34° 
44° 
49° 
43° 
36° 
40° 
49° 
35° 
3i° 
31° 
36° 
43° 
.27° 
40° 
43° 
Table XVI. 
Observer C. 
Yellow 
49° 
46° 
50° 
46° 
36° 
44° 
Green 
44° 
42° 
40° 
41° 
28° 
33° 
Red 
45° 
41° 
45° 
4i° 
34° 
41° 
Blue 
56° 
55° 
53° 
50° 
38° 
44° 
Tables XVII and XVIII to show the following facts: 
(a) How much the decrease of illumination narrowed the 
limits of color sensitivity by causing a decrease in the energy 
of the light-waves coming to the eye. This was determined by 
subtracting the value of the limit at decreased illumination with 
the screen of a gray of the brightness of the color at decreased 
illumination from its value at full illumination with the gray 
screen of the brightness of the color at full illumination. (&) 
How much the limits were narrowed by the action of the white 
and black screens at decreased illumination. This was ascer- 
tained by subtracting the values of the limit with the white 
and the black screen at decreased illumination from the value 
of the limit at decreased illumination with the gray screen of 
the brightness of the color at this illumination, (c) How much EXPERIMENTAL 
149 
more the limits were narrowed by the white and the black 
screens at decreased than at full illumination. This was com- 
puted for the white screen, for example, as follows: The 
quantity, limit at decreased illumination for gray screen of 
brightness of color at decreased illumination, minus limit for 
white screen at decreased illumination, is subtracted from the 
quantity, limit at full illumination for gray screen of brightness 
of color at full illumination minus limit for white screen at full 
illumination. A similar computation was made for the black 
screen. 
Table XVII. 
A. Showing (u) how much the limits were narrowed by 
decrease in the amount of colored light coming to the eye; (b) 
how much they were narrowed by increased induction of white 
and black screens at decreased illumination; and (c) how much 
more they were narrowed by induction of white and black screens 
at decreased than at full illumination. 
Stimulus 
How much 
limits were 
narrowed 
by decrease 
in amount of 
colored light 
coming to 
the eye 
How much 
limits were 
narrowed by 
induction of 
white screen 
How much 
limits were 
narrowed by 
induction of 
black screen 
How much 
more limits 
were nar- 
rowed by 
white screen 
at decreased 
than at full 
illumination 
How much 
more limits 
were nar- 
rowed by 
black screen 
at decreased 
than at full 
illumination 
Yellow 
i° 
8° 
o° 
6° 
x° 
Green 
i° 
5° 
9° 
4° 
6° 
Red 
3° 
9° 
o° 
8° 
i° 
Blue 
4° 
13° 
6° 
10° 
2° 
Table XVIII. 
Observer C. 
Yellow 
3° 
10° 
2° 
7° 
3° 
Green 
3° 
13° 
8° 
ix° 
4° 
Red 
4° 
7° 
o° 
3° 
o° 
Blue 
6° 
12° 
6° 
n° 
3° 
Results when a gray screen matching the color in brightness 
at standard illumination is used. In these experiments a deter- 
mination was made of the amount the limits of sensitivity are 
changed by the brightness induction caused by the alteration of 150 
GERTRUDE RAND 
the brightness relation between stimulus and screen with decrease 
of illumination, when a screen is used which matches the color 
in brightness at standard illumination. This determination was 
made as follows. 
An estimate was made of the amount the limits were narrowed 
by decrease of illumination when a screen of the brightness of 
the color at standard illumination is used for both standard and 
decreased illuminations. From this result was subtracted the 
amount the limits were narrowed by decrease of illumination 
when the screen is made in turn of the brightness of the color 
at standard and at decreased illumination. The difference 
obtained represents the value sought. It is given in Table XIX. 
Table XIX. 
A. Showing how much the color limits were narrowed at 
decreased illumination by the induction of the screen which 
matched the color in brightness at standard illumination. 
Stimulus 
Screen of 
brightness of 
color at de- 
creased il- 
lumination 
Limit 
Screen of 
brightness of 
color at 
standard 
illumination 
Limit 
Amount 
limit was 
narrowed by 
change in 
brightness 
relation be- 
tween stimu- 
lus and 
screen 
caused by 
decreased 
illumination 
Yellow 
gray no. 3 
43° 
gray no. 2 
4i° 
2° 
Green 
gray no. 5 
36° 
gray no. 8 
29° 
7° 
Red 
gray no. 50 
40° 
gray no. 24 
33° 
7° 
Blue 
gray no. 13 
49° 
gray no. 28 
46° 
3° 
(C.) The Effect of These Amounts of Induction upon the Limens 
of Color at Different Degrees of Excentricity. 
We have shown the effect of decreasing the general illumina- 
tion upon the color sensitivity of the peripheral retina with gray,, 
white, and black screens by the effect on the limits of sensitivity. 
This is only an indirect means of estimating its influence, for 
the results obtained cannot be translated into terms of direct 
measurement, owing to the irregular decrease in sensitivity of EXPERIMENTAL 
151 
the peripheral retina from the fovea outwards. In this section, 
we shall measure the influence of changes of illumination directly 
by the changes produced in the limen of sensation at various 
angles of excentricity. As in the previous section, measurement 
will be made of the effect upon sensitivity (a) of the decrease 
in the amount of colored light coming to the eye, produced by 
the decrease of illumination, (d) of the difference in the inducing 
power of the white and black screens, and (c) of the change in 
the brightness relation of stimulus to background. 
To determine the first of these three points, a campimeter screen 
had to be selected that gave no brightness contrast with the stimu- 
lus. To provide for differences in the brightness of the colors 
at the different points observed for the two illuminations at which 
we worked, a preliminary determination of the brightness of the 
sensation at these points was made at both illuminations by the 
flicker method. The brightness of the screen was chosen in each 
case of the brightness of the color according to these determina- 
tions. To eliminate the effect of preexposure, the stimulus previ- 
ous to exposure was in every case covered by a gray of the bright- 
ness of the color for the illumination used at the point of the 
retina at which we were working. Thus no brightness after-image 
was carried over to exert an inhibitive action upon the color sensa- 
tion. The stimulus was a disc compounded of the sectors of the 
color, and of the gray of the brightness of the color for the illum- 
ination used at the point of the retina under investigation. The 
proportions of the sectors were altered until the observer gave the 
judgment of just noticeable color. The average of judgments 
made in ascending and descending series was chosen as the final 
value of the limen. The difference between the limens at standard 
and decreased illumination was taken as the measure of the loss 
in intensity which the stimulus had sustained by the decrease of 
illumination. 
The effect upon the color limen of the increased induction from 
the white and black screens was shown by the same method, with 
the exception that the white and black screens were substituted 
for the gray of the brightness of the color. The stimulus was 
a disc composed of sectors of color and gray of the brightness of 152 
GERTRUDE RAND 
the color at the angle of excentricity at which the determination 
was made. 
The effect of the change in the brightness relation between 
the stimulus color and the screen produced by decrease of il- 
lumination was shown as follows. An estimate was made of 
the amount the limens are raised by the decrease of illumination 
when a screen was used for both standard and decreased illumina- 
tion that had a brightness-value equal to the color at standard 
illumination. From these results was subtracted the amount the 
Table XX. 
A. Showing how much the limens of sensitivity were raised 
at the fovea, and at points 150, 250, 300 from the fovea in the 
horizontal meridian on the temporal side by the decrease in the 
amount of colored light coming to the eye produced by the de- 
crease in the general illumination. 
Stimulus 
Point on 
horizontal 
temporal 
meridian at 
which limen 
was taken 
Limen at 
standard il- 
luminstiani 
with screen 
of brightness 
of color at 
standard 
illumination 
Limen at de- 
creased il- 
lumination 
with screen 
of brightness 
of color at 
decreased 
illumination 
How much 
limen was 
raised at 
decreased il- 
luminiation 
Yellow 
0° 
i8° 
20° 
2° 
15° 
22° 
32° 
10° 
25° 
35° 
40° 
5° 
30° 
50° 
65° 
15° 
Green 
o° 
20° 
20° 
0° 
15° 
27° 
28° 
i° 
25° 
40° 
50° 
10° 
Red 
0° 
9° 
11° 
2° 
15° 
9° 
13° 
4° 
25° 
17° 
25° 
8° 
30° 
25° 
45 
20° 
Blue 
0° 
9° 
10° 
1° 
15° 
IO° 
13° 
3° 
25° 
12° 
15° 
3° 
30° 
20° 
40° 
20° EXPERIMENTAL 
153 
limens were raised by decreasing the illumination when the 
screens were made in turn of the brightness of the color at stan- 
dard and at decreased illumination. The difference obtained 
represents the value sought. These results are of particular 
importance because they show that the influence of the brightness 
of the surrounding field can not be eliminated even when a 
screen of the brightness of the color is used unless some means 
be had of maintaining the general illumination of the room 
constant. 
Table XX shows how much the limens of sensitivity were 
raised at the fovea and at points 150, 250, and 300 from the 
Table XXL 
A. Showing the color limens at standard and decreased illum- 
inations with white and with black screens. 
Stimulus 
Point on 
horizontal 
temporal 
meridian at 
which limen 
was taken 
White screen 
Black screen 
Limen at 
standard 
illumination 
Limen at 
decreased 
illumination 
Limen at 
standard 
illumination 
Limen at 
decreased 
illumination 
Yellow 
0° 
22° 
25° 
28° 
30° 
15° 
25° 
50° 
35° 
45° 
25° 
50° 
8o° 
65° 
85° 
30° 
8o° 
125° 
95° 
U3° 
Green 
0° 
22° 
25° 
28° 
30° 
15° 
30° 
36° 
35° 
43° 
25° 
50° 
75° 
75° 
220° 
Red 
0° 
13° 
20° 
10° 
14° 
15° 
19° 
35° 
15° 
21° 
25° 
30° 
55° 
23° 
35° 
30° 
50° 
330° 
29° 
58° 
Blue 
0° 
17° 
22° 
IO° 
12° 
15° 
25° 
40° 
12° 
17° 
25° 
35° 
6o° 
i8° 
25° 
30° 
40° 
90° 
30° 
6o° 
fovea in the horizontal meridian on the temporal side by the 
decrease in the amount of colored light coming to the eye pro- 154 
GERTRUDE RAND 
duced by the decrease in the general illumination. The results 
of this table may be generalized as follows: 
I. The limen of color is higher in the periphery than in the 
center of the retina at both illuminations. 
2. The limen of color is higher at decreased illumination 
than at standard illumination. 
3. The direct effect upon the intensity of the sensation pro- 
duced by decreasing the illumination is shown by the limen 
determinations to be inconsiderable. In the central retina, the 
difference is but i° or 20. In the peripheral retina at the points 
considered there is a difference of from io° to 200. 
Table XXI shows the color limens at both standard and de- 
creased illuminations when white and black screens are used, at 
the fovea, and at points 150, 25°, and 300 in the horizontal 
meridian on the temporal side. 
Table XXII has been compiled from Tables XX and XXI 
to show how much greater the limens were for white and black 
screens at decreased than at full illumination; how much of the 
effect may be ascribed to the reduction of the amount of colored 
light coming to the eye; and how much to the increased induc- 
tion of the screens. It will be seen from the results of this table 
that the loss of the sensation in intensity due to the increased 
brightness induction is much greater than that caused by the 
reduction in the amount of colored light coming to the eye. 
It was shown in Table XIV that quite a great deal of bright- 
ness induction is caused by the change in brightness relation be- 
tween color and screen produced by decreasing the illumination. 
Table XIX shows how much this induction narrows the limits 
of sensitivity to the four colors used. Table XXIII, shows how 
much the limens are raised when the illumination is decreased by 
the inductive action caused by the change in the brightness re- 
lation between stimulus color and gray screen of the brightness 
of the color at standard illumination. Table XXII. 
A. Showing how much greater the limens were with white 
and black screens at decreased than at standard illumination and 
how much of this effect may be ascribed to the reduction in the 
amount of colored light coming to the eye and how much to the 
increased inductive action of the screens. 
White screen 
Black screen 
Stimulus 
Point on 
horizontal) 
temporal 
meridian at 
which limen 
was taken 
Total 
amount 
greater 
Amount due 
to decrease 
in amount of 
colored light 
coming to 
eye 
Amount due 
to increased 
induction of 
screen 
Total 
amount 
greater 
Amount due 
to decrease 
in amount 
of colored 
light coming 
to eye 
Amount due 
to increased 
induction of 
screen 
Yellow 
o° 
15° 
25° 
30° 
7° 
28° 
45° 
75° 
2° 
IO° 
5° 
15° 
8^ OOCn 
o o o o 
12° 
23° 
50° 
63° 
2° 
IO° 
5° 
15° 
10° 
13° 
45° 
48° 
Green 
o° 
15° 
25° 
5° 
9° 
35° 
f. 
10° 
5° 
8° 
25° 
IO° 
i6° 
i8o° 
o° 
i° 
IO° 
IO° 
15° 
1700 
Red 
0° 
15° 
25° 
30° 
n° 
26° 
38° 
305° 
2° 
4° 
8° 
20° 
9° 
22° 
30° 
285° 
5° 
12° 
i8° 
33° 
r 
O 004^ tO 
o o o o 
3° 
8° 
IO° 
13° 
Blue 
0° 
15° 
25° 
30° 
13° 
30° 
48° 
70° 
1° 
3° 
3° 
20° 
12° 
27° 
45° 
50° 
3° 
7° 
13° 
40° 
1° 
3° 
3° 
20° 
2° 
4° 
IO° 
20° 156 
GERTRUDE RAND 
Table XXIII. 
A. Showing how much the color limens were raised at de- 
creased illumination by the induction of the screens which 
matched the color at standard illumination. 
Stimulus 
Point on 
horizontal 
temporal 
meridian at 
which limen 
was taken 
Limen with 
screen of 
brightness of 
color at de- 
creased il- 
lumination 
Limen with 
screen of 
brightness of 
color at 
standard il- 
lumination 
Amount 
limen was 
raised by 
change in- 
brightness 
relation be- 
tween stimu- 
lus and 
screen 
caused by 
decrease of 
illumination 
Yellow 
0° 
20° 
20° 
0° 
15° 
32° 
32° 
0° 
25° 
40° 
40° 
o° 
30° 
65° 
n6° 
5i° 
Green 
o° 
20° 
20° 
0° 
15° 
28° 
40° 
12° 
25° 
50° 
igo° 
1400 
Red 
0° 
n° 
u° 
0° 
15° 
13° 
24° 
11° 
25° 
25° 
48° 
230 
30° 
45° 
150° 
105° 
Blue 
0° 
10° 
12° 
2° 
15° 
13° 
i6° 
3° 
25° 
15° 
23° 
8° 
30° 
40° 
55° 
15° 
(D.) The Influence of Change of Illumination upon the Action of 
the Preexposure on the Limens and Limits of Color. 
The brightness of the preexposure exerts an influence upon 
the color observation because the eye carries over an after-image 
from the preexposure into the color observation. If, for ex- 
ample, the preexposure is to black, a white after-image is 
aroused which fuses with the succeeding color sensation and 
strongly reduces its saturation. The effect of preexposure is 
especially strong in the peripheral retina because a very strong 
brightness after-image is aroused in the peripheral retina by a 
very short period of stimulation. It is very difficult for the 
writer to predict from the data she has at hand with regard to 
the effect of change of illumination upon the sensitivity of the EXPERIMENTAL 
157 
peripheral retina to the brightness after-image just what will 
be the effect of change of illumination upon the action of pre- 
exposure on the color sensitivity of the peripheral retina. But 
even though there be no change in the sensitivity of the peri- 
pheral retina to the brightness after-image with change of illu- 
mination, it is obvious that there will be some effect of the 
change of illumination because of the change in the brightness 
relation of the preexposure card to the colored stimulus. In 
case the stimulus light is gotten by reflection from pigment sur- 
faces, this change of brightness relation is due to the shift in 
the brightness of the colors produced by the change in the illumi- 
nation. In case transmitted light is used as stimulus, the bright- 
ness of the stimulus color is independent of changes in illumi- 
nation and will remain constant; but a change in the brightness 
relation of stimulus to preexposure will occur because the pre- 
exposure will lighten or darken with change of illumination. 
The writer hopes to make the quantitative investigation of 
this point the subject of a future study. At present she can only 
point out that if a guarantee is wanted that the effect of the 
brightness of the preexposure is eliminated from the results 
of the observation, the preexposure must be to a gray of the 
brightness of the color and the illumination of the room must 
be kept constant. 
The foregoing results show how strongly the changes in the 
illumination of the visual field influence the color sensitivity of 
the peripheral retina, particularly when the stimulus is sur- 
rounded by a white field. They also show that the influence 
of stirrounding field can not be eliminated even by means of a 
campimeter screen of the brightness of the color unless some 
means be had of keeping the general illumination of the room 
constant. It is obvious without further comment how import- 
ant it is that a method be devised to standardize this factor. 
The preceding experiments indicate that without this standardi- 
zation, no experiment can be repeated from time to time under 
the same conditions relative to any one of the brightness factors 
that influence color sensitivity. Results thus obtained are far 
from comparable. 158 
GERTRUDE RAND 
E. Methods of Standardising These Factors. 
We have shown in the preceding analysis of the color obser- 
vation that the factors which influence the limits of color sensi- 
tivity, and which, therefore, require standardization, are the 
brightness of the surrounding field, the brightness of the pre- 
exposure, and the general illumination of the retina. Standard 
conditions require either that the influence of a factor be elimi- 
nated, or that it be reduced to a constant. We have been able 
to treat all three of these variable factors in one or the other of 
these two ways. The effect of preexposure and campimeter 
screen has been eliminated and methods of measuring the gen- 
eral illumination and of keeping it constant have been provided. 
As we have seen, the surrounding field influences the color 
sensation by adding brightness to the stimulus by induction. 
When, for example, the surrounding field is black, white is in- 
duced by contrast across the stimulus color. Since the colors 
all differ in brightness, the induction takes place in different 
amounts for the different colors. This white in proportion to 
its amount reduces the action of the colors on the retina. Fur- 
ther, a given amount of white affects to different degrees the 
action of the different colors on the retina. The influence of pre- 
exposure is even more important than of surrounding field. If 
the preexposure is to black, white is added as after-image to the 
stimulus color. The effect of a black preexposure upon the 
stimulus color is greater than the effect of a black surrounding 
field because more white is added as after-image of preexposure 
than is induced by contrast from the surrounding field. Now, 
since brightness induction is greatest when there is maximal 
opposition between the inducing and induced fields, and since 
the brightness after-image also is most intensive when there is 
maximal opposition between the stimulus and the projection 
field, it is evident that no one screen nor preexposure can be 
found that will influence each color by an equal amount. The 
black preexposure and surrounding field concomitant upon work 
in the dark-room can be considered no exception to this state- 
ment. The influence of preexposure and surrounding field can 
not be successfully eliminated in work in the dark-room. By EXPERIMENTAL 
159 
using one screen and preexposure standard conditions of con- 
trast induction and brightness after-image can be maintained 
only if the colors are made of equal brightness. The objections 
to this procedure were pointed out in an earlier section. There 
remains the alternative of choosing in each case gray papers of 
the brightness of the colored stimulus for the screen and the 
preexposure. This necessitates changes of screen and of pre- 
exposure for each stimulus, but insures the complete elimination 
from the color excitation of all brightness influence due either 
to preexposure or to stimulation of the surrounding retina. In 
this way alone, then, may a proper regulation of these factors 
be obtained for any investigation whatsoever of the sensitivity 
of the peripheral retina. Further, the method gives a proper 
basis with regard to these two important factors from which 
to start all investigations of the effect of achromatic conditions 
upon color sensitivity. 
Standardization for either one of these factors, however, can 
be accomplished for one degree of illumination only. As the 
general illumination changes, the relation of the brightness of 
the preexposure and of the surrounding field to the brightness of 
the colored stimulus changes.47 It is obvious, then, that if stand- 
ardization is to be accomplished with regard to the influence of 
either of these factors, some means must be devised of main- 
taining the general illumination of the retina constant. 
In order to obtain a standard illumination, two things are neces- 
sary: (a) A means of controlling the illumination must be pro- 
vided, which is sufficiently sensitive to cause small changes. 
(b) A method of measuring the illumination produced has to be 
devised; at least, a means must be secured for determining when 
an illumination has been obtained that is equal to a given 
preceding illumination. We shall first discuss the method of 
measurement we adopted. As stated earlier in our paper, no 
47 When the colored light used to stimulate the retina is independent of the 
general illumination, e.g., when it is obtained from the spectrum, from mono- 
chromatic sources, or from standard filters, these two factors alone will 
modify the result of the color observation. If, however, light reflected 
from a pigment surface be used as stimulus, a change in the illumination 
will in addition change the amount of colored light coming to the eye. 160 
GERTRUDE RAND 
satisfactory means of determining the amount of daylight illum- 
ination in a room has been provided by the physicist, so there is 
little hope at this time of solving the problem from that side. The 
brightness induction of the peripheral retina, however, has been 
found by us to be extremely sensitive to changes in the general 
illumination. This phenomenon seems to provide us with a 
sensitive measure of these changes, while, at the same time, it 
represents the combined effects for sensation of the principal 
subjective factors that might vary from day to day. To apply the 
method in its most sensitive form, the inductive power of white 
was chosen because it is the most strongly affected by illumina- 
tion changes. For example, when No. 14 Hering gray was used 
as stimulus and white as campimeter screen, a noticeable change 
was produced in the induction when the white curtain of the 
optics-room was pulled forward 1 cm. from a position in which 
its edge was directly above the long axis of the campimeter. This 
caused a change in the illumination of the room so small that it 
could not be directly sensed. Further, at 11 o'clock in the morn- 
ing of a bright day in September, when a point at 25 0 on the nasal 
meridian was stimulated, Observer A reported that the white 
screen induced black across the stimulus No. 14 gray to an 
amount that caused it to equal in brightness 1070 of black and 
2530 of No. 14 gray; at 2 o'clock of the same day the induction 
was increased until the No. 14 gray matched 1500 of black and 
2100 of the gray; at 4 o'clock of the same day the No. 14 gray 
matched 1800 of black and 1800 of the gray.48 Working at 250 
in the temporal meridian, this observer reported at different times 
during one day and on different days, the wide variations shown 
by the following figures: 283° of black, 2250, 1450, 1900, 238°, 
etc. Observer C reported less induction, but her variations from 
time to time were equally great. At 250 in the temporal meridian, 
she found at different times 8o° of black, 1030, 1600, 1750, etc. 
After a careful study of the phenomenon with different screens 
and with different stimuli, the inductive action of the white 
screen upon a stimulus of No. 14 Hering gray, at 250 in the 
"This increase in the inductive action of the screen caused by the decrease 
in illumination, was accompanied by a shrinkage of the zones sensitive to 
color covering an area of 4 to 6°. EXPERIMENTAL 
161 
temporal meridian, was found to provide the best means of 
detecting changes in the illumination of the optics-room. At 
this point on the retina, the induction was by no means minimal, 
nor was it sufficiently great to cause the medium gray chosen for 
our stimulus to appear too dark to give a small j. n. d. of 
sensation. 
The sensitivity of this method of detecting changes in the general il- 
lumination was compared with the sensitivity of the Sharpe-Millar portable 
photometer. In this photometer one of the comparison fields is illuminated 
by the light of the room and the other by a standard tungsten lam'p enclosed 
in the photometer box. When the room is illuminated by daylight, the field 
receiving the light of the room is seen as white, while the field lighted by 
the tungsten lamp appears as a saturated orange. The difference in color 
between the two fields renders the photometric judgment difficult and 
makes the instrument very insensitive for daylight tests. For example, 
our tests showed that by the method for indentifying an illumination 
described in the text, a change in illumination could be detected which was 
produced by drawing the white curtain i cm. from a position in which its 
edge was directly above the long axis of the campimeter. But with the 
receiving surface of the portable photometer in precisely the same position 
as the stimulus screen of the campimeter, the edge of the curtain had to be 
moved 11.3 cm. in order that- the change of illumination might be detected. 
Moreover, this amount of change could be detected only in case the photo- 
metric field was continuously observed while the curtain was being drawn, 
in which case the comparison field was observed to become slightly darkened. 
The judgment was made, then, in terms of a just noticeably different bright- 
ness of the field which was illuminated by the daylight, rather than in terms 
of a disturbance in the brightness-equality of the two fields. When, on 
the other hand, the judgment was made in terms of a just noticeable 
disturbance in the equality of the two fields, as the judgment would 
have to be made if the photometer were to be employed for the reproduction 
of any former illumination taken as standard, the curtain had to be drawn 
44.2 cm. before the change could be detected. This j. n. d. represents an 
amount of illumination equal to 2.5 foot-candles. 
The next step was to procure a means of changing the illumin- 
ation of the room by very small amounts. This was accomplished 
by drawing the white curtain (described p. 86) across the 
skylight above the apparatus. The drawing of this curtain several 
inches made little difference in the illumination directly observable 
by the eye, although, as we have said, a change of 1 cm. when 
the edge of the curtain was directly above the apparatus, produced 
a noticeable change in the inductive action of the white screen. 
Having thus provided ourselves with a means of producing 162 
GERTRUDE RAND 
small changes of illumination and with a method of detecting 
them, we had in order to complete our work but to choose an 
illumination for each observer, which could be taken as standard. 
Since we wished to work on both light days and days of medium 
darkness, an average had to be chosen as our standard from the 
measurements obtained on a number of days ranging from light 
to dark, so that on bright days the room could be darkened, and 
on dark days it could be lightened until this value was obtained. 
For Observer A an illumination was selected which caused an 
induction of black across No. 14 gray stimulus viewed at 250 in 
the temporal meridian to an amount which caused the gray 
stimulus to equal in brightness 2100 of black and 1500 of No. 14 
gray; for Observer B 1800 of black and 1800 of No. 14 gray; 
and for Observer C 1450 of black and 2150 of No. 14 gray. The 
amount of black induction was identified in each case by means 
of a measuring-disc made up of sectors of black paper and No. 
14 gray of the Hering series. 
Previous to each series of observations the illumination of the 
room was changed until the amount of brightness induction was 
brought to the value chosen as standard. It was tested at 
intervals during the sitting and was readjusted when necessary. 
Details of the method of doing this are as follows: When the 
white screen and the No. 14 gray stimulus had been put in place, 
the observer took his position and adjusted the fixation-knot in 
front of the motor for the 250 point on the temporal meridian. 
The measuring-disc set at the standard value was mounted on 
the motor. The observer reported whether the stimulus appeared 
lighter or darker than the measuring-disc, or of a brightness equal 
to it. If the judgment lighter or darker was given, the curtain 
was drawn one way or the other until the stimulus accurately 
matched the measuring-disc in brightness. 
This method not only gives a sensitive measure of the changes 
of illumination of the visual field and a successful means of stand- 
ardizing the illumination of a room by daylight, but it has in 
addition advantages for work in psychological optics not possessed 
by an objective standardization, could that be successfully 
obtained. The problem of standardization, includes more for the EXPERIMENTAL 
163 
psychologist than it does for the physicist, for the former has 
variables to take into account in addition to the changes that may 
take place in the energy of the stimulus. Even though the illumin- 
ation of the room be made objectively constant, we should expect 
variations in the response of the retina to this illumination because 
of its own changes from time to time. Brightness contrast, for 
example, might be expected to vary from sitting to sitting even 
when the stimulus conditions are kept absolutely constant. Two 
factors would be concerned in these variations: changes in the 
inducing power of the surrounding parts of the retina, and 
changes in the sensitivity of the local area. These changes would 
take place even when the usual precautions known to the experi- 
menter in this field have been observed. Such precautions are 
commonly limited to fatigue, adaptation, etc. These precautions 
do not provide for the changes that occur in the retina from day 
to day. Moreover, they do not adequately guard against a change 
in a factor, unless some measure of that factor be had. So far as 
the writer knows, in these general precautions intended to keep 
the state of the retina constant, no measure of the variable factor 
has been provided to test the adequacy of the method. The 
method proposed by us, however, is planned with this in view. It 
takes into account not only the objective, but the subjective vari- 
ables, and reduces both to a constant. For example, when No. 14 
gray surrounded by the white field is made equal to the 
measuring-disc composed of 2100 of black and 1500 of the No. 14 
gray for Observer A, it means that the observation may be begun 
with the assurance that the total result of all the factors-the 
illumination of the room, the local sensitivity of the retina, and 
the inductive action of the surrounding parts of the retina-is the 
same as in the preceding observation. 
What has just been said should not be considered as more than 
a general statement of the application of the principles of the 
method. In actual practice a greater refinement of working may 
be attained. If, for example, one wishes to use a preexposure 
differing in brightness from that of the colored stimulus, and 
doubts whether a test which covers only the local sensitivity of 
the retina and the inductive action of the surrounding parts is a 164 
GERTRUDE RAND 
sufficient check upon the after-image sensitivity, he may make his 
standard include the effect of the preexposure he wishes to use. 
In short, if he does not consider adequate the more general test 
we have described, he may duplicate, in establishing his standard, 
any combination of brightness factors, due to preexposure, 
brightness of screen, or what not, that he may wish to use in 
his experiment proper. 
The test of a method is how well it works. The test of this 
method is that we shall be able closely to duplicate our results 
from sitting to sitting regardless of the changes in the outside 
illumination from day to day or from morning until afternoon. 
The method stands the test. Long series of observations in the 
peripheral retina show a very small M. V.-much less even than 
is shown in the ordinary color observations in the central retina 
where, as compared with the peripheral retina, the factors extran- 
eous to the stimulus exert little influence. 
The following table has been compiled from a number of 
observations to show the variations in the results of color limens 
and color limits (a) when the general illumination was controlled 
according to the method described above, and (&) when no further 
precautions were observed than were used by previous investiga- 
tors. In previous investigations of the color sensitivity of the 
peripheral retina, care has been taken-to work only at the same 
hours of days that appeared equally bright, or, if on days of 
different brightness, to make a rough approximation of preceding 
illuminations by means of curtains without using either a definite 
standard or means of measuring. For our work with the illumin- 
ation controlled, the gray of the brightness of the color at the 
illumination selected as standard was used for the preexposure 
and the campimeter screen. For the work without any especial 
control of the illumination, the gray of the brightness of the color 
on one of the days selected as typical was used throughout for 
preexposure and screen. This gave in the first case complete 
elimination of the effect of preexposure and surrounding field, 
and in the second case elimination as complete as could be gotten 
without accurate control of the general illumination. Results are 
given in the table for blue and green only because the sensitivity 
to these colors is affected most by changes of illumination. EXPERIMENTAL 
165 
Stimulus 
Illumination 
Screen and 
Preexposure 
Variation of 
limits on 
different 
days 
Variation of 
limens on 
different 
days 
Green 
Controlled 
Gray no. 8 
0° 
o049 
Uncontrolled 
Gray no. 9 
4°-6° 
6o°-82oc0 
Blue 
Controlled 
Gray no. 28 
0° 
2°-3° 
Uncontrolled 
Gray no. 30 
4°- 5° 
i8°-3O° 
At the conclusion of a piece of work the object of which has 
been the elimination of sources of error in one of the oldest and 
best developed fields of psychological investigation, the following 
comments having a more general application to other fields in 
which sensory determinations are required, may be justified. In 
all sensory determinations, investigators have been very much 
annoyed by the magnitude of the mean variation that has occurred 
in their results. This may be due to two sets of factors: errors 
in the control of the factors that influence the response of the 
sense-organ, and errors in judgment. To eliminate the latter 
source of errors, the psycho-physical methods have been devised. 
Before beginning her attempts to get a better control of the 
factors that influence the color sensitivity of the retina, the 
writer had used all the psycho-physical precautions known to her 
to eliminate errors in judgment, still her inability to reproduce her 
results rendered in her judgment any accurate investigation of 
the sensitivity of the peripheral retina hopeless. On the other 
hand, however, with the control she has been able to get of the 
factors that influence the sensitivity of the retina to color, and 
with only a casual observance of psycho-physical precautions, a 
very close reproduction of results has been rendered possible. 
With regard to work in the optics of color at least, then, she is 
forced to conclude that the major source of error is not in the 
factors that influence the judgment but in those that influence the 
response of the sense-organ. Moreover, she would suggest that 
if in other sensory fields more attention were paid to the factors 
49 The limen for green was taken in both cases at 250 on the temporal 
retina. 
" The limen for blue was taken in both cases at 400 on the temporal 
retina. 166 
GERTRUDE RAND 
that influence the response of the sense-organ and relatively less 
to the factors that influence the judgment, a higher degree of 
precision may be attained in our methods of working.51 
61 For a further discussion of this point, see Ferree, C. E., Transactions 
of the Illuminating Engineering Society, 1913, VIII. TPIE FLUCTUATION OF LIMINAL VISUAL STIMULI 
OF POINT AREA 
By C. E. Ferree 
Bryn Mawr College 
TABLE OF CONTENTS ' 
I. Introduction 378 
II. The accommodation theory 381 
III. The fluctuation of stimuli of point area. (The work of 
Heinrich and Chwistek.) 388 
IV. Experimental /... 396 
V. Conclusion 408 
I. Introduction 
In a series of articles published 1906-08/ the writer 
reported the results of an experimental study of the phe- 
nomena usually attributed to' fluctuation of attention. These 
phenomena, it was claimed, belong to three sense fields: visual 
sensation, auditory sensation, and cutaneous sensation. The 
problem was raised, it will be remembered, in 1888 by Nikolai 
Lange,2 who gathered together the instances of intermittence 
of minimal sensations and found for them a common explana- 
tion in the conception of an instable or fluctuating attention. 
The recurrent changes in the limen of sensation producing 
the intermittence are, he contended, due to involuntary changes 
in the degree of attention given to the stimulus. Previous to 
the series of articles mentioned above, two other explanations 
had also been given: (a) Involuntary changes in the adjust- 
ment of the sense organ in case of vision and audition, 
primarily accommodation in case of vision (Miinsterberg,3 
1C. E. Ferree: An Experimental Examination of the Phenomena 
Usually Attributed to Fluctuation of Attention, Amer. Jour. Psychol., 
XVII., 1906, 81-120; The Intermittence of Minimal Visual Sensations, 
Amer. Jour. Psychol., XIX., 1908, 59-129; The Streaming Phenomenon, 
Amer. Jour. Psychol., XIX., 1908, 484-503. 
2 N. Lange: Beitrage zur Theorie der sinnlichen Aufmerksamkeit 
und der activen Apperception, Philosophische Studien, IV, 1888, 389- 
422. 
3 Hugo Miinsterberg: Schwankungen der Aufmerksamkeit, Beitrage 
zur experimentellen Psychologie, Freiburg, 1889, 69-124. 379 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
Heinrich,4 and Heinrich and Chwistek) ;5 and (b) an over- 
flow of excitation from the circulatory and respiratory centers 
in the brain (Miinsterberg,6 Lehmann,7 Slaughter,8 Taylor,® 
etc.).10 
In the series of articles mentioned above it was shown on 
the negative side that in case of vision at least, intermittence 
cannot be ascribed to any of the previously mentioned causes; 
and on the positive side that it is a phenomenon of the adapta- 
tion and recovery of the sense organ. Intermittence was 
denied in case of minimal cutaneous sensation,11 and the 
4 W. Heinrich: Die Aufmerksamkeit und die Funktion der Sinnes- 
organe, Zeitsch, f. Psychol., XL, 1896, 59-76; and Ueber die Intensi- 
tatsanderungen schwacher Gerausche, ibid, XLL, abt. 2, 1907, 57-59; 
Zur Erklarung der Intensitatsschwankungen eben merklicher optischer 
und akustischer Eindriicke, Bulletin International de I'Academic des 
Sciences de Cracovie, Nov., 1898, 363-382. 
"W. Heinrich und L. Chwistek: Ueber das periodische Ver- 
schwinden kleincr Punkte, Zeitsch. f. Psychol., XLL, Abt. 2, 1907, 
59'74- 
8 Loc. cit. 
7 Alfred Lehmann. Ueber die Beziehung zwischen Athmung und 
Aufmerksamkeit, Philosophische Studien, IX., 1894, 66-95. 
8 J. W. Slaughter: The Fluctuations of the Attention in Some of 
their Psychological Relations, Amer. Jour. Psychol., XII., 1901, 
313'334- 
9 R. W. Taylor; The Effect of Certain Stimuli upon the Attention 
Wave, Amer. Jour. Psychol., XII., 1901, 335-345. 
10 Miinsterberg ascribed to this overflow, in case of respiration, an 
effect on the muscular control of the eye. During inspiration there 
was more accurate control of fixation and accommodation and during 
expiration a less accurate control of these adjustments. Lehmann 
leaves us in some doubt as to just how he believes the effect is 
produced. He says (op. cit., p. 84) : " Wir sahen dass die Reactionen 
am haufigsten sind in der Nahe des Inspirationsmaximums. Hier ist 
eben der Blutdruck am grossten, und von diesem Zustand muss 
angenommen werden, dass er fur die psychologische Arbeit des 
Gehirns giinstig sei. Wir wissen ja, dass das Blut, wahrend der 
Arbeit irgend eines Organes, demselben reichlicher zufliesst. Deshalb 
ist es hochst wahrscheinlich, dass auch die Arbeit eines Organes 
erleichtert werde wenn durch irgend eine Ursache eine Vergrosserung 
des Blutzuflusses herbeigefiihrt wird." Slaughter and Taylor are 
inclined to believe that the overflow affects the sensory cells directly. 
In their experiments a plethysmographic record of the peripheral 
blood pressure was taken while the fluctuations of the visual stimulus 
were being observed. They conclude that their results show a co- 
incidence between the maxima of the plethysmographic curve and the 
phase of visibility of the fluctuation record. Two kinds of maxima are 
found in the plethysmographic tracing, one due to inspiration and 
the other forming the crest of a long vaso-motor wave of unknown 
cause, commonly called the Traube-Hering wave. 
11 In 1907 the experiments in cutaneous sensation were repeated by 
Geissler (L. R. Geissler: Fluctuation of Attention to Cutaneous 
Stimuli, Amer. Jour. Psychol., XVIII., 1907, 318-321). A mistake FERREE 
380 
phenomenon was left open for further consideration in case 
of auditory sensation. A part of the work done at this time 
still remains unpublished. Some of it covers points still in 
dispute. For that reason two articles will be added to the 
former series. The first is in answer to an article by Heinrich 
and Chwistek entitled: " Ueber das periodische Verschwinden 
kleiner Punkte,"12 and is intended to clear up, if possible, at 
least so far as the writer's work is concerned, the last point 
in dispute between the adaptation and accommodation theories. 
Heinrich and Chwistek maintain that the fluctuation of mini- 
mal visual stimuli of point area is caused by periodic changes 
in the curvature of the crystalline lens and offer their 
results for stimuli of point area as evidence that the fluctua- 
tions of stimuli of all areas are caused by changes in accom- 
modation. In one of the former studies13 the present writer 
had worked with stimuli ranging from 2 mm. x 2 mm.- 
42 cm. x 38 cm. in area. He found that stimuli of these areas 
fluctuate just as readily for aphakial as for normal subjects, 
and that changes in accommodation, therefore, can not be 
considered an essential factor in the production of the 
phenomenon. It had never occurred to him to work with 
stimuli of point area. In the present study, however, stimuli 
was made by him in interpreting the writer's method of stimulating 
the tongue electro-cutaneously that has not yet been corrected. He 
says, " In repeating Ferree's experiment with electro-cutaneous stimula- 
tion of the tongue, we found some difficulty in eliminating the touch, 
pressure, and taste sensations set up by the electrodes. The best 
results were obtained by applying a i% solution of cocaine to the 
fore part of the tongue, upon which two strips of tin foil (Christmas 
tree foil), hammered as thin as possible, were laid. The strips were 
connected with the interrupter of a Du Bois-Reymond induction coil." 
Christmas tree foil was not used in the original experiments. This 
material was rejected at once by the present writer as unsuitable. It 
is much too stiff and gives rise to pressure sensations. Narrow 
strips of very thin and pliable wrapping foil were used instead. When 
these were placed on the fore part of the tongue moistened with 
spittle, the observer was utterly unable to tell whether or not they 
were in contact with the tongue when the coil was not working. 
Neither did they under the action of the current give rise to taste 
sensations. Of the two procedures the writer would prefer the one 
used in the original experiments. It seems to him obviously better 
to make the electrodes of wrapping foil than to use the stiffer material 
and cocainize the tongue into insensibility to contact, more especially 
since Geissler's observers report that the cocaine itself sets up dis- 
tracting sensations in the tongue. 
12 W. Heinrich and L. Chwistek: Zeitschr. f. Psychol., XLI., 1907, 
39-74. 
13 See An Experimental Examination of the Phenomena Usually 
Attributed to Fluctuation, 98-108. 381 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
of point area have been used. From the results of this 
study it will be shown that the fluctuations of these stimuli 
present no especial case; for (a) they occur just as readily 
for aphakial subjects as for subjects with normal eyes; and 
(b) identified by the tests used by the writer in his earlier 
experiments they correspond just as closely to adaptation 
phenomena as do the fluctuations of stimuli of larger area. 
In the second paper, work on the fluctuation of auditory 
stimuli will be reported. In this work the writer has suc- 
ceeded in getting conditions under which no fluctuations occur, 
whether the stimulus be tone or noise. His results also enable 
him to explain without recourse to central factors or the tensor 
mechanism of the middle ear the fluctuations which do occur 
under experimental conditions different from those he has 
used. The completion of these two pieces of work rounds 
up, so far as the writer knows, all of the outstanding points 
in his case against fluctuation of attention in its original 
meaning. 
II. The Accommodation Theory 
That involuntary changes in accommodation are a factor 
in causing the fluctuation of minimal visual stimuli was pro- 
posed first by Miinsterberg in 1889.14 Miinsterberg held that 
the fluctuation of these stimuli is due to two causes: unsteadi- 
ness of fixation and involuntary changes in accommodation. 
Although different views may be held with regard to the 
essential physiological and psychological factors in attention, 
all must agree, he says, that when a visual stimulus is 
attentively observed the eye is fixated and accommodated so 
as best to receive the impression on the retina. But this 
adjustment cannot be uniformly maintained for any length 
of time. Involuntary changes both in fixation and accom- 
modation occur. These changes weaken and confuse the 
light impression received on the retina, hence an object just 
noticeably different from its background will alternately dis- 
appear into this background and become distinct from it. 
The effect of lapses in accommodation is too obvious, he 
thinks, to need special explanation. The rays of light are no 
longer sharply focused on the retina and the image of the 
object blurs and becomes indistinguishable from its back- 
ground. For unsteadiness of fixation, however, the case is 
not quite so clear. The explanation is as follows. Fick, 
Kirschmann, and others have shown that the sensitivity of 
the retina to colorless light attains its maximum at a certain 
14 Hugo Miinsterberg: Schwankungen der Aufmerksamkeit, Bei- 
trdge zur experimentellen Psychologic, Freiburg, 1889, pp. 69-124. FERREE 
382 
distance from the fovea. Thus when the eye loses its fixation, 
the image of the object fixated travels towards a more sensitive 
part of the retina. This will cause the image of the rings 
on the Masson disc, for instance, which in the traditional 
fluctuation experiment are made just noticeably darker than 
their background, to lighten and become equal in brightness 
to the background. This gives the phase of invisibility. 
When fixation is regained, however, the ring again becomes 
noticeable. This gives the phase of visibility. These two 
factors, then, the lightening of the image of the ring due to 
involuntary changes of fixation and the blurring of its outlines 
due to involuntary changes in accommodation, should, accord- 
ing to Miinsterberg, be considered as the cause of the alternate 
appearance and disappearance of the rings on the Masson 
disc which were attributed by Lange to fluctuation of attention. 
Since the writer has already shown in his first article15 that 
involuntary changes in accommodation cannot be considered 
as an essential factor in these fluctuations,18 space will be taken 
here only to point out that changes in fixation should also 
not be considered essential factors in the sense in which 
Miinsterberg considers them factors. In the first place they 
could, in any event, have an effect only in case the stimulus 
was darker than the background. If the stimulus were lighter 
than the background, the brightening of the image would make 
it stand out more distinctly from the background than before, 
instead of causing it to disappear into the background as it is 
observed to do in fluctuating. Moreover, the explanation 
can have little or no application to the fluctuation of colored 
stimuli. Since both of the latter classes of stimuli fluctuate just 
as readily as the former, the principle can be regarded as 
having little value for purposes of explanation. And in 
the second place, this factor could not in all probability even 
cause the fluctuation of stimuli darker than the background, 
for the increased sensitivity of the extra-foveal retina would 
not only cause the ring to brighten, but also the background 
immediately surrounding it. The effect of the factor would 
16 See An Experimental Examination of the Phenomena Usually 
Attributed to Fluctuation of Attention, pp, 84 and 94-96. 
16 In the original article stimuli ranging in area from 2 mm. x 2 mm.- 
38 cm. x 42 cm. were used. For the writer's observers fluctuations 
never occurred when a stimulus 38 cm. X' 42 cm. or larger was 
observed at a distance of 1 meter. In the experimental portion of 
the present paper it will be shown that changes in accommodation 
are not an essential factor in the fluctuation of stimuli of smaller 
area than 2 mm. x 2 mm., namely stimuli of point area. Thus with the 
present paper the demonstration will have been finished for the whole 
range of areas for which fluctuation occurs. 383 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
thus be merely to raise both the gray of the stimulus and of 
the background in the brightness scale, not to make them 
equal, unless indeed one were affected more than the other 
by an amount that would be noticeable in sensation, which 
can hardly be possible since the difference between them is, 
to begin with, only just noticeable.17 
Miinsterberg supported his explanation by the following experimental 
evidence. The norm of the period of fluctuation was established for 
each of his subjects and the following variations of conditions were 
made, (i) A "prismatische Lorgnette" which moved the field of 
vision slightly to one side was placed before the eyes. When this 
was held steadily in position, the period of fluctuation was affected 
very little, but when it was removed and interposed every 2 seconds, 
causing the eye to move quickly to the side to follow the shift in the 
object fixated, the period was very noticeably lengthened. (2) 
Involuntary blinking was caused every 2 seconds by means of a sharp 
sound. Fluctuation was prevented. When the eyes were closed volun- 
tarily every 2 seconds, the same results were obtained. This, Munster- 
berg thinks, was because of the relief of muscular strain produced by 
the blinking. That is, as the lids are closed, the eyes move downwards 
and inwards; as they are opened, upwards and outwards (Bell's 
phenomenon). This frequent relief from the strain of fixating and 
accommodating so freshens the muscles, he thinks, and improves their 
action that disappearance never ensues. (3) The whole apparatus 
bearing the Masson disc was slowly moved back and forth, up and 
down, and sidewise. Each movement was executed in 2 seconds. 
Thus, in order to fixate the moving stimulus, the eye was kept continu- 
ously moving. The accommodation was also kept continuously chang- 
ing. In a companion series of experiments the head was moved slowly 
from side to side. In this case also, in order continuously to fixate 
the stimulus the eye was compelled to move.18 Fluctuation did not 
occur in either series of experiments. Again Miinsterburg thinks 
fluctuation was prevented because the muscles were kept in such a 
fresh condition that accurate fixation and accommodation could be 
maintained throughout the observations. A moment's reflection will 
show (1) that these assumptions cannot be wholly true. The attempt 
of the eye to follow the moving stimulus, whether the movement was 
apparent, produced by the interruption of the " prismatische Lorgnette " 
or actual, produced by the moving of the apparatus bearing the stimu- 
lus, must have resulted in the image falling now to this side, now 
to that side of the fovea. If so, the fixation maintained was far from 
accurate.18. Likewise when fixation was lost in blinking, it was doubt- 
If, tor example, one were very much lighter or darker than the 
other, the greater sensitivity of the extra-foveal retina might affect 
one more than the other enough to cause a noticeable change in the 
difference between them, but this can scarcely be assumed to be the 
case when one is only just noticeably different from the other. 
18 The movement was equal in amount to the movement of the head 
and in the opposite direction. 
18 This frequent shifting of the image from the position previously 
occupied by it on the retina would give abundant chance for the 
adapting retina to recover, and thus in terms of the adaptation theory 
to explain the absence of fluctuation. FERREE 
384 
less regained through a series of small oscillatory movements as 
commonly happens before the eye comes to rest in taking a new 
position. And (2) even were the assumptions true, the argument is 
not at all differential for Miinsterberg's theory. The same effect 
on fluctuation would be expected in terms of the adaptation theory. 
An abundant reason was given for the stimulus never disappearing 
in the effect of the eye-movement on restoring the adapting retina. 
Eye-movement, it will be remembered, exerts its effect on adaptation 
in two ways. There is (a) an indirect effect. As the result of the 
movement the image falls on a fresh area of the retina and the area 
previously stimulated is given a chance to recover, (b) There is a 
direct effect which is much greater than the indirect effect, namely, 
the influence of eye-movement upon the amount and direction of the 
lymph streams that are continually moving hither and thither in the 
retina. A detailed discussion of this effect was given in the writer's 
earlier work.10 Eye-movement is thus an essential factor in both 
theories. The relation to fluctuation ascribed to it in the two theories, 
however, is very different. In Miinsterberg's theory, eye-movement 
helps to cause the disappearance of the stimulus, while in the adaptation 
theory it is the most important cause of the reappearance of the 
stimulus. With regard to the relative merits of these two views, the 
writer will say that a few minutes' observation of a liminal stimulus 
should be enough to convince anyone that a voluntary eye-movement, 
for example, instead of causing the stimulus to disappear, will on 
the contrary serve to keep it distinct; and, if it has disappeared, will 
cause it to reappear. For a detailed demonstration that involuntary eye- 
movement acts in the same way and that it is the chief factor in render- 
ing adaptation intermittent, see the writer's earlier articles " An Ex- 
perimental Examination of the Phenomena Usually Attributed to 
Fluctuation of Attention," and " The Intermittence of Minimal Visual 
Sensations." 
Miinsterberg also conducted a series of experiments in which a 
comparison was made of the rate of respiration and fluctuation. The 
results showed that when the respiration was in short quick gasps, 
the rate of fluctuation was increased; and when it was slow, the rate 
of fluctuation was decreased. In explaining this result he attributes 
to breathing an influence on the muscular control of the eye. With 
the inspiration there is an increase of the muscular control; and with 
the expiration, a decrease. 
The accommodation factor was next taken up by Pace.21 
Pace compared the fluctuations obtained by his subjects before 
and after the paralysis of their ciliary muscles by a solution 
10 See The Intermittence of Minimal Visual Sensations, Amer. Jour. 
Psychol., XIX., 1908, 112-129; and The Streaming Phenomenon, Amer. 
Jour. Psychol.,. XIX., 1908, 484-503. 
In the blinking experiment ,in addition to the effect of the accom- 
panying eye-movement, the blinking would have itself produced an 
effect on fluctuation. That is, the closing of the lid shut off the light 
coming from the stimulus and gave the adapting retina a chance to 
recover. 
21 Edward Pace: Zur Frage der Schwankungen der Aufmerksamkeit 
nach Versuchen mit der Masson's Scheibe, Philosophische Studien, 
VIII., 1893, 388-403. 385 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
of sulphate of atropine, and found no significant difference 
in his results. He concluded, therefore, that changes in accom- 
modation could not be considered as essential to the 
phenomenon. 
The theory, however, would not down. It was revived by 
Heinrich, and Heinrich and Chwistek in a series of articles 
extending from 1896-1907.22 In the article entitled: "Die 
Aufmerksamkeit und die Funktion der Sinnesorgane," Hein- 
rich establishes the following principles which he considers 
of importance in explaining fluctuation. (1) When the atten- 
tion is directed away from optical impressions (a) the lens 
takes the curvature characteristic of far seeing, and (b) the 
lines of sight tend towards the parallel position. The 
demonstration of these principles, however, cannot be con- 
sidered as having any very direct bearing on the explanation 
of the phenomenon of fluctuation, for it may very well be 
conceived that the voluntary direction of attention away from 
visual impressions would cause changes in the accommodation 
and fixation of the eye of a magnitude that would be signifi- 
cant, while the involuntary lapses of attention occurring dur- 
ing the prolonged observation of a stimulus would not cause 
these changes at all. At least for the purpose of explanation 
of the phenomenon of fluctuation, the demonstration of the 
former cannot be considered the equivalent of the demonstra- 
tion of the latter. And (2) when the attention is directed 
away from all optical impressions, involuntary changes take 
place in the curvature of the lens. This conclusion is based 
upon the recurrent changes that take place in the breadth of 
the pupil and upon the behaviour of images reflected from 
the anterior surface of the lens. In drawing this conclusion 
from the first point of evidence, Heinrich obviously assumes 
a closer connection between the changes in the breadth of the 
pupil and changes in accommodation than can safely be done. 
Any one who has studied the reactions of the pupil under a 
very wide range of conditions can not help but know that this 
1 :i correlation does not exist. Moreover, the connection has 
not been found in a large enough percentage of cases to make 
it safe, even plausible, to assume that it exists in any situation 
in which it has not yet been demonstrated. In his second 
point of evidence, Heinrich does not describe the behaviour 
22 W. Heinrich: Die Aufmerksamkeit und die Funktion der Sinnes- 
organe, Zeitschr. f. Psychol., XL, 1896, 410-431; Ueber das periodische 
Verschwinden kleiner Punkte, ibid, XLL, 1907, 59-74, und Zur 
Erklarung der Intensitatsschwankungen eben merklicher optischer 
und akustischer Eindriicke, Bulletin International de l'Academie des 
Sciences de Cracovie, Nov., 1898, 363-382. FERREE 
386 
of the images observed. Apparently, however, the description 
is supplied in a later paper published in cooperation with 
Chwistek entitled " Ueber das periodische Verschwinden 
kleiner Punkte." At least a method of demonstrating changes 
in the curvature of the lens based on the behavior of the image 
reflected from its anterior surface is described here. But the 
validity of this demonstration is strongly open to question. 
In the experimental section of the present paper it will be 
shown that it is much more plausible to ascribe this behaviour 
to involuntary eye-movement than to involuntary changes in 
accommodation,-that in fact the same kind of behaviour has 
been described by de Schweinitz23 and others, in case of the 
images reflected from the cornea, as one of the common 
phenomena of ophthalmometry due to eye-movement. 
In the article entitled " Zur Erklarung der Intensitats- 
schwankungen eben merklicher optischer und akustischer 
Eindriicke,"24 Heinrich discusses the effect of variation of 
intensity, or differences in intensity between the stimulus and 
its background on the fluctuation of visual stimuli. Marbe25 
had found that an increase of intensity increases the phase of 
visibility, and conversely a decrease of intensity decreases the 
phase of visibility.26 
This, Heinrich thinks, is just what should be expected were the 
disappearances caused by recurring lapses in the adjustment of 
the lens. As will be shown in the next section of the present 
paper, however, these results offer no differential evidence 
in favor of the accommodation theory. They are just what 
would be expected in terms of any theory that has yet been 
advanced to explain the fluctuation of minimal visual stimuli. 
Heinrich also notes that for one of Marbe's observers the 
28 G. E. de Schweinitz: Diseases of the Eye, Philadelphia and 
London, 1902, p. 739. 
21 Op. cit., 366-369. 
25 Karl Marbe: Die Schwankungen der Gesichtsempfindungen, 
Philo so phis che Studien, VIII., 1893, 615-637. 
28 Marbe apparently was the first to make any separation of the 
phase of visibility from the phase of invisibility in drawing his con- 
clusions, and even he did not take any account of the phase of invisi- 
bility in making his comparisons. The total times of visibility of his 
stimuli under the different conditions alone were compared. This 
tendency to break up the total period of fluctuation into its phases 
for purposes of comparison was a step in the right direction, but it 
was little heeded by his successors. 
Marbe concludes that neither the fluctuation of the " Schrddersche 
Treppenfigur " nor the visual sensation is periodic. The phase of 
visibility of the visual stimulus increases with the increase in the 
difference in intensity between the stimulus and its background. The 
length of the period of fluctuation is a function of this increase. 387 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
phase of visibility was decreased when the image of the stimu- 
lus fell on the paraxial portions of the retina. This also, he 
says, is just what should be expected were the disappearance 
caused by changes in the curvature of the lens. So is it also 
what should be expected were adaptation the cause of the 
fluctuation, for one of the most conspicuous differences 
between the phenomena in the central and peripheral retina 
is the greater rapidity of adaptation in the peripheral retina.27 
Again, then, the evidence cannot be considered as differential. 
Heinrich also takes into consideration in this paper the results 
of Pace with the atropinized eye. He claims that changes 
in the curvature of the lens may still be observed when 
atropine has been used to paralyze the muscles of accommo- 
dation. While the present writer by no means contends 
that the muscles of accommodation are completely paralyzed 
by the use of atropine, still he would maintain that 
Heinrich's claim is strongly open to question if it is 
based on the kind of observation described by him 
and Chwistek in the article " Ueber das periodische Ver- 
schwinden kleiner Punkte." In any event the point is no 
longer of importance to the explanation of fluctuation, for 
it has been shown since that time by the present writer28 that 
eyes from which the lenses have been removed and which 
under careful test shown no residual accommodation, get the 
fluctuation apparently just as readily as the normal eye. 
In the article " Ueber das periodische Verschwinden kleiner 
Punkte," Heinrich and Chwistek, using stimuli of point area, 
attempt as their experimentum crucis to demonstrate the 
27 In fact in the writer's own experiments, designed to show the 
correspondence between adaptation and fluctuation in the peripheral 
retina, the farther the stimulus was moved towards the periphery of 
the retina the shorter became the phases of visibility and the longer 
the phases of invisibility. These experiments were made differential 
for the adaptation theory (a) by the method of variation of areas, 
and (b) by showing a rough correspondence between the phase of 
visibility in the fluctuation experiments with the adaptation time for 
different visual stimuli from center to periphery of retina. No. 27 
gray and the red, green, blue, and yellow of the Hering series of 
papers were used as stimuli. Only a partial list of the results obtained 
was published, however, because the writer did not at that time con- 
sider the phenomenon in indirect vision worthy of more space. The 
main arguments were established for direct vision and no more data 
were included for indirect vision that were needed to show in a 
general way that the phenomena here present no exception. (See " An 
Experimental Examination of the Phenomena Usually Attributed to 
Fluctuation of Attention," 116-119.) 
28 See An Experimental Examination of the Phenomena Usually 
Attributed to Fluctuation of Attention, 84 and 94-96. FERREE 
388 
coincidence of fluctuation and changes in accommodation. 
Heinrich had mentioned the desirability of making this 
demonstration in the preceding paper, but had discarded the 
idea as infeasible because of the conditions under which the 
experiment would have to be conducted. In choosing to work 
with point stimuli in this study, Heinrich and Chwistek have 
admittedly selected the conditions most favorable to the 
accommodation theory, for by Heinrich's own statement in the 
explanation of his theory in an earlier paper29 changes in 
accommodation would be more apt to cause the disappearance 
of point stimuli than of stimuli of larger area. Notwith- 
standing this admission, however, the results they get with 
point stimuli are advanced in the later paper as evidence that 
the fluctuation of stimuli of all areas is due to involuntary 
changes in accommodation. Their work with stimuli of point 
area will be taken up in detail in the next section of this paper. 
III. The Fluctuation of Stimuli of Point Area (the Work of 
Heinrich and Chwistek)30 
Heinrich and Chwistek maintain that the fluctuation of 
visual stimuli of point area is caused by periodic changes in 
the curvature of the crystalline lens. They also offer their 
results for stimuli of point area as evidence that the fluctua- 
tions of stimuli of all areas are caused by changes in accom- 
modation. Four arguments are advanced by them in support 
of this conclusion, (i) Periodic changes in the curvature of 
the lens are directly observable. Moreover, these changes are 
found roughly to coincide with the fluctuations of the point 
stimulus when both observations are conducted at the same 
time. They describe two methods of demonstrating this change 
of curvature. One may be considered popular, the other 
technical. The popular demonstration may be conducted as 
follows. Prick two holes in a cardboard nearer together than 
the breadth of the pupil of the eye. Hold the card close to 
the eye and look through the holes at a bright light. The holes 
will be seen as two dispersion circles with a bright overlapping 
area. When the curvature of the lens changes, the over- 
lapping area alternately contracts and expands. That is, as 
the lens becomes more convex, the dispersion circles become 
smaller and the overlapping area becomes narrower; and 
conversely, as the lens becomes less convex the circles become 
28 See Zur Erklarung der Intensitatsschwankungen eben merklicher 
optischer und akustischer Eindriicke, 366-67. 
80 W. Heinrich and F. Chwistek: Zeitschr. f. Psychol., XLL, Abt. 
2, 1907, 59-73- 389 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
larger and the overlapping area becomes broader. This change 
in the overlapping area, they say, can readily be observed. 
Their technical demonstration was accomplished by means of 
an ophthalmometer. Their method was as follows. Two 
spots of light were thrown on the eye of the observer by means 
of two mirrors reflecting the light from a lamp properly placed 
with reference to these mirrors and to the eye of the observer. 
These images were observed by means of an ophthalmometer. 
Their description of method is extremely meager. They say: 
" Lasst man die beobachtete Person den Punkt fixieren, dessen 
periodisches Verschwinden untersucht wird, und dreht die 
Glasplatten des Ophthalmometers, bis man in dem Instrument 
die beiden von der vorderen Linsenflache reflektierten 
Bildchen als drei Punkte sieht, so offenbart sich jede Kriim- 
mungsanderung der Linse dadurch, dass der mittlere Punkt 
bei grosseren Anderungen sich spaltet, bei kleiner breiter wird. 
Man beobachtet dann ohne weiteres, dass die Linsenein- 
stellung nicht stabil ist, sondern dass sie kleinen periodischen 
Aenderungen unterliegt. Diese Aenderung konnte mit unserem 
Instrument durch die Drehung der Flatten um hochstens 0.50 
kompensiert werden. Es war uns unmoglich die Aenderungs- 
richtung aus den Bewegungen des Punktes zu erkennen."31 
While these changes in the image reflected from the observer's 
eye were being recorded by a second person, the observer him- 
self recorded the fluctuations of a point stimulus. Simultan- 
eous records were thus obtained which could be compared in 
order to determine whether the phases of fluctuations coincided 
with the phases of changes in the image reflected from the eye. 
The point stimulus consisted of a small black point on a white 
ground or a small white point on a black ground, o. 1-0.3 mm. 
in diameter, observed at a distance of 70-150 cm. Two 
observers, Herr Sk. and Herr Zacz, were used. The eyes of 
both were normal, or emmetropic. Chwistek recorded the 
changes in the images reflected from the eye. Their results 
are stated as follows. For observer Sk., 776 phases were 
recorded. "Einseitige Notierung vom Herrn Sk., d. h. notiertes 
Verschwinden des Punktes ohne entsprechend notierte Ak- 
kommodationsschwankung ergab sich in 38 Fallen. Einseitige 
Notierung vom Herrn Chwistek, d. h. notierte Akkommoda- 
tionsanderung ohne entsprechende Aufzeichnung des Ver- 
schwindens des Punktes fand man in 40 Fallen." For 
observer Zacz, 296 phases were recorded. " Einseitige 
Notierung vom Herrn Zacz in 31 Fallen. Einseitige Notierung 
vom Herrn Chwistek in 32 Fallen." 
" Op cit., 60-61. FERREE 
390 
(2) Within the range of areas used by them an increase 
in the area of the stimulus was found to give longer phases 
of visibility and shorter phases of invisibility. And, con- 
versely, a decrease in the area of the stimulus was found to 
give shorter phases of visibility and longer phases of invisi- 
bility. Points were observed ranging for one observer 
(emmetropic) from .2 mm.-.5 mm. in diameter, at distances 
ranging from 100 011.-126.5 cm.; for another observer (2.5 
myopic), .2 mm.-.5 mm. in diameter at distances ranging 
from 35 cm.-39 cm.; and for a third observer (4 D. myopic), 
.2-1.5 mm. at distances ranging from 15 cm.-28-5 cm. 
(3) The phase of visibility was also found to vary with the 
intensity of the stimulus or with the brightness difference 
between the point and its surrounding field. The greater was 
this brightness difference, the longer the phase of visibility 
was found to be as compared with the phase of invisibility, 
and the less was this brightness difference, the shorter was 
the phase of invisibility. 
(4) When the stimulus was placed just beyond the far point 
for an observer with myopic eye, it was found to become 
periodically more and less distinct. Also two points placed 
at this distance were found alternately to blur into one and to 
separate into two. Two observers were used in these experi- 
ments. 
Before passing to his own experimental evidence that 
involuntary changes of accommodation are not an essential 
factor in the fluctuation of stimuli of point area, the writer 
has the following comments to make on the work of Heinrich 
and Chwistek. (1) In this work they have created for them- 
selves a special problem, that is, they employed stimuli of 
point area and strongly supraliminal intensity. The fluctua- 
tion of such stimuli has never been ascribed to the fluctuation 
of attention. Historically considered, then, they arei not 
working with the phenomenon to which they primarily make 
their conclusions apply; and, moreover, they have not in any 
way shown in a satisfactory manner the propriety of applying 
their conclusions to the phenomenon explained by Lange as due 
to the instability of attention. (2) Their popular demonstra- 
tion of involuntary changes in the adjustment of the lens is 
strongly open to question. Employing 124 subjects, the writer 
has not been able to make it work in a single case in which 
care was taken to rule out extraneous factors which would 
themselves cause the phenomenon. For example, extreme care 
must be taken to hold the card steady. Any variation in the 
distance of the holes from the pupil of the eye will cause a 391 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
variation in the breadth of the overlapping area. Especially 
must care be taken that the card does not touch the lid of the 
eye, for movements of the ball of the eye and more 
particularly of the lid change the distance of the card from the 
eye. These movements are often unnoticed unless the observer 
is especially looking for them, and are frequently of sufficient 
range to cause a change in the size of the dispersion circles. 
Without a doubt the phenomenon, when it has occurred, has 
been, so far as the writer's experience is concerned, an artifact 
due to the conditions under which the observations were made. 
(3) Their technical demonstration by means of the ophthal- 
mometer is, in the writer's opinion, just as strongly open to 
question. The writer criticizes this demonstration, however, 
with reluctance because of the meagerness with which they 
have described their method of working and observations. 
The following points, however, may be noted, (a) Working 
as they did, two images should have been observed, one 
reflected from the cornea, the other from the anterior surface 
of the lens.32 Both images should have been very much alike, 
with the exception that the one reflected from the cornea 
should have been larger and more distinct. Nothing is said 
in the article, however, that would give evidence to the reader 
that more than one image was observed, or that the image 
described was actually reflected from, the lens. But even if 
it were granted that the image observed was reflected from the 
lens, it would signify little, for the phenomenon described by 
them could have been caused just as well by involuntary eye- 
movements as by changes in the curvature of the lens. That 
is when the eye is accommodated, the anterior surface of the 
lens is hyperbolic in shape and varies in curvature considerably 
from point to point. A movement of the eye would, therefore, 
cause the rays of light forming the image to be reflected suc- 
cessively from points at which the surface had a different 
curvature. Each difference in curvature would give a differ- 
ence in the size of the image reflected. Eye-movement would, 
therefore, produce the same effect in the size of the image as 
changes in the convexity of the lens. That is, movements of 
greater range would correspond in effect to the changes in 
convexity of greater magnitude, and, conversely, movements 
of lesser range to the changes in convexity of lesser magnitude. 
In fact, the phenomenon they describe is one of common 
82 An image reflected from the posterior surface of the lens might 
also have been observed. But since this image is inverted and is 
besides very indistinct, it may be considered as having no bearing on 
the discussion. FERREE 
392 
observation in case of the corneal image, and in this case no 
attempt has been made to ascribe it to recurrent changes in the 
curvature. For example, de Schweinitz, in his treatise on the 
diseases of the eye, says :33 " Nothing is more common than 
to see the images of the mires [the mires correspond to the 
lights used by Heinrich and Chwistek] separate and overlap 
so that the apparent curvature of the cornea seems to change 
while under observation. The changes are due to slight move- 
ments of the eye which bring different portions of the cornea 
into view." We know that there are many involuntary eye- 
movements per minute even with the best control of fixation 
that can be obtained.34 It seems more plausible, therefore, to 
attribute the phenomenon observed by Heinrich and Chwistek 
to the involuntary eye-movements which we know occur in 
abundance, than to use it as a proof of a new phenomenon, 
namely, the involuntary changes in the curvature of the lens, 
even if it be granted that the image from the lens was 
observed. At least, it may be said that Heinrich and Chwistek 
were not warranted in concluding as they did, without having 
secured any differential evidence to bear out their conclusion 
or without even having considered eye-movement as a causal 
factor, (d) Since the corneal image is known also to double 
and overlap, a rare opportunity was given to Heinrich and 
Chwistek, in making these observations, to compare the be- 
haviour of the corneal image with that of the image reflected 
from the lens, if that really were the image they observed, 
and to determine by the presence or absence of coincidence in 
the two sets of changes, whether the doubling and overlapping 
of the images reflected from the lens has the same or a different 
cause from the doubling and overlapping of the images re- 
flected from the cornea. Had both images really been observed 
or had the characteristic doubling and overlapping of the cor- 
neal image even been known to Heinrich and Chwistek, one can 
hardly conceive that their conclusions would have been drawn 
without recourse to this means of determining whether or not 
both sets of changes should be ascribed to a common cause. 
In short, judging from their report as it stands; from the fact 
that the ophthalmometer as it is ordinarily constructed and 
used is intended only for the observation of the corneal 
images, and that such a phenomenon as they describe would 
83 G. E. de Schweinitz: Diseases of the Eye, Philadelphia and 
London, 1902, 739. 
34 See C. E. Ferree: An Experimental Examination of the Phe- 
nomena Usually Attributed to Fluctuation of Attention, Amer. Jour. 
Psychol., XVII., 1906, 113-115; also The Intermittence of Minimal 
Visual Sensations, ibid., XIX., 1908, 83-112. 393 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
have been extremely difficult to observe in case of an image 
reflected from the lens; and from the fact that descriptions of 
similar behaviour on the part of the corneal image are given 
by other observers, the writer cannot help but think, without 
any wish to be hypercritical, that considerable grounds are 
given for suspecting that Heinrich and Chwistek have 
observed the doubling and overlapping of the corneal image 
which is commonly attributed by de Schweinitz and others to 
involuntary eye-movement.35 Moreover, the crux of their 
argument is that they have actually observed a coincidence 
between the fluctuation of tiie visual stimulus and the changes 
in the adjustment of the lens. This, they contend, gives a 
certainty to their argument not yet attained in previous work 
on the problem. But even if the question whether or not it 
was a lens image that was observed be disregarded, it will be 
seen from the above discussion that is strongly probable 
that the coincidence they actually observed was between eye- 
movement and the fluctuation of the visual stimulus and not 
between changes in the curvature of the lens and the fluctua- 
tion of the visual stimulus. 
(4) Their explanation of the effect of variation of area 
on the fluctuation of a visual stimulus could apply only to 
stimuli of very small area. Moreover, even in the case of very 
small areas the effect they got is just what might be expected 
as the result of increase of area either in terms of Loria's 
explanation of the fluctuation of stimuli of point area36 or in 
terms of the writer's explanation: adaptation interfered with 
by eye-movement. They make two cases of their explanation 
of how changes of accommodation cause the fluctuation of 
stimuli of point area: (a) when the stimulus is a black point 
in a white ground and (b) when it is a white point in a black 
ground. In the former case the rays of light coming from 
the margin of the black point are not sharply imaged on the 
retina when the lens changes focus, hence they spread over 
the dark space on the retina corresponding to the black point. 
It is obvious that this spreading of the marginal light could 
blot out the dark space only in case the black stimulus were of 
very small area. Hence the explanation could not apply at 
all to stimuli of the size ordinarily used in the work on fluctua- 
tion. In the latter case the rays of light coming from the 
white point are not sharply imaged when the accommodation 
35 The writer leaves himself willingly open to correction on this 
point, however. 
36 See Heinrich and Chwistek: op. cit., p. 60; also Stanislaw Loria: 
Untersuchung uber das periphere Sehen, Zeitschr. f. Psychol., XL., 
1905, 160-186. FERREE 
394 
changes, and are spread over the surrounding dark space. 
Since strongly supraliminal stimuli were used, it is extremely 
doubtful whether even very small stimuli could be carried 
below the limen of sensation from this cause. Moreover, 
because strongly supraliminal stimuli were used and no 
attempt was made to control the intensity of the stimulus, 
an increase in the area of the stimulus would function for 
sensation as an increase of intensity.37 Therefore, from this 
cause alone, according to the theories advanced either by Loria 
or by the writer, an increase of area would produce an increase 
in the phase of visibility. Even in case of stimuli of point 
area, then, the effect of increase of area described by Heinrich 
and Chwistek offers no differential argument in favor of the 
explanation advanced by them. Furthermore, the theory of 
fluctuation of attention was meant to apply only to stimuli of 
liminal or approximately liminal intensity. When such stimuli 
are used, an increase of area produces just the opposite effect. 
For example, working in 1906 with liminal stimuli ranging in 
area from .5 x .5 cm. to 15 x 15 cm., the writer found that 
an increase of area caused a decrease in the phase of visibility 
and a corresponding increase in the phase in invisibility. And 
in the experimental section of this paper it will be shown that 
the same effect is produced in case of liminal stimuli of very 
small area. In both of these cases care was taken to keep the 
stimuli liminal in order that an increase in the area of the 
stimulus would not produce an increase in the intensity of the 
sensation. (4) The fourth argument advanced by Heinrich 
and Chwistek has no differential value whatever. It was first 
used by Heinrich in 1898, as applied to stimuli of larger area.38 
A more intensive stimulus, he thinks, is not so liable to be 
blotted out by involuntary changes in accommodation. There- 
fore, he concludes, the more intensive is the stimulus the 
longer should be the phases of visibility and the shorter the 
phases of invisibility. It is obvious, however, that this result 
is just what should be expected from adaptation as a causal 
factor. It should be expected even were it held that fluctua- 
tion is due to instability of attention. In fact an increase in 
the phase of visibility and a decrease in the phase of invisibility 
would be the natural consequence of an increase in the 
" We seem to have here a violation of one of the most fundamental 
principles in experimental procedure, namely, when it is wanted to 
determine the effect of a given factor, the effect of all other factors 
should, if possible, be eliminated from the results of the experiment. 
" W. Heinrich: Zur Erklarung der Intensitatsschwankungen eben 
merklicher optischer und akustischer Eindriicke, Bulletin International 
de l'Academie des Sciences de Cracovie, Nov., 1898, 363-382. 395 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
intensity of the stimulus in terms of any theory that has yet 
been advanced to explain fluctuation. 
(5) The writer is in some doubt as to what is meant by the 
fifth argument. " Befindet sich der Punkt, dessen Ver- 
schwinden man beobachtet, innerhalb des Akkommodations- 
bereiches der Linse, so beobachtet man nur das periodische 
Verschwinden desselben. Die Verhaltnisse sind komplizierter, 
wenn man den Punkt ausserhalb des Fernpunktes aufstellt 
was beim myopischen Auge leicht ausfiihrbar ist. In diesem 
Faile zeigt sich, dass der beobachtete Punkt, der jetzt nicht 
scharf gesehen wird, periodisch verschwindet, aber auch 
periodisch scharfer gesehen wird."39 In the first place he 
cannot understand why the above result should be expected, 
were changes in accommodation present, for when the far 
point is actually reached the ciliary system should be com- 
pletely relaxed. It is difficult then to see how the lens can 
be allowed to become any flatter, unless indeed it be held that 
the theory of accommodation commonly accepted for the 
human eye is incorrect. And in the second place, working 
under the conditions described by Heinrich and Chwistek, the 
writer has been unable to get anything that might be called 
three distinct and separate stages of clearness of his stimulus. 
Moreover, any stimulus of supraliminal intensity, fluctuating 
from any cause whatsoever and especially from causes purely 
retinal, would be apt to have, although not sharply defined, 
maximum, minimum, and intermediate degrees of distinctness. 
This the writer's observers were able to get at whatever dis- 
tance the stimulus was put from the eye, but they were utterly 
unable to detect the three distinct and separate stages that 
are reported by Heinrich and Chwistek. Nor were they ever 
able to see the stimulus as clearly beyond the far point as they 
were at the far point or nearer than the far point. In short, 
there was never at this point what could be considered a norm 
of clearness which was succeeded either periodically or even 
at irregular intervals by a degree of clearness in excess of 
this norm. Continuing, Heinrich and Chwistek say: " Das lasst 
sich am besten durch folgendes Experiment illustrieren: Stellt 
man nicht weit ausserhalb des Fernpunktes des myopischen 
Auges als Objekt zwei Punkte, die so nahe liegen dass sie als 
ein Fleck gesehen werden, so beobachtet man, dass die Punkte 
periodisch auf kurze Zeiten getrennt erscheinen." The writer 
has not succeeded in getting this phenomenon when working 
beyond the far point with the myopic eye. It is, however, of 
common occurrence for any eye when the points are placed 
Op. cit., 66. FERREE 
396 
at or slightly nearer than the limit of clear vision for these 
points and are regarded for any length of time. The points 
alternately blur into one and separate into two. In all prob- 
ability both retinal and accommodation factors are involved 
in this result, but no definite estimate can be made of how 
much importance should be assigned to either until comparative 
records be made for subjects without lenses and for normal 
subjects. In the writer's opinion, however, the above experi- 
ment comes the nearest of any yet described by Heinrich and 
Chwistek to giving tangible evidence that involuntary changes 
in accommodation occur. But even to demonstrate clearly that 
these changes occur, would not prove that they are essential 
or even important factors in the fluctuation of minimal visual 
stimuli even of point area.40 That they are not essential fac- 
tors will be shown by the writer in the next section of this 
paper. 
IV. Experimental 
In this section of our paper we propose to show (i) that 
involuntary changes in accommodation are not essential or 
even important factors in the fluctuation of minimal visual 
stimuli of point area, and (2) that, identified by tests used 
by the writer in his earlier experiments, these fluctuations 
correspond just as closely to adaptation phenomena as they 
do for stimuli of larger area. 
Probably the most convincing proof that one can offer that 
involuntary changes of accommodation are not essential to 
the fluctuation of stimuli of point area is the results obtained 
from aphakial subjects. Observations were made by the 
writer upon four aphakial subjects. They were all above 
sixty years of age, and three were above seventy. All of 
them had had the lenses removed from their eyes from 15-20 
years before. Both the advanced age of the subjects and the 
long period that had elapsed since their lenses were removed 
favored the absence of any residual accommodation. To 
make sure of this point, however, they were each tested as 
follows. The subject's head was clamped in a head-rest and 
a card bearing letters of very fine print (3% point type) was 
slid along a meter rod supported at the level of his eyes in the 
40 Lest it be thought that this experiment shows some coincidence 
between changes in accommodation and fluctuation, it may be pointed 
out that the cycle of changes experienced by the two points does not 
even include disappearance. The points merely blur into one and 
separate into two. That is, the only phenomenon cited by Heinrich 
and Chwistek that really gives any tangible evidence of involuntary 
changes in accommodation does not even occur in a series in which 
fluctuations are found. 397 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
median plane. The card was placed at his point of clearest 
vision as determined by the focus of his glasses and was 
moved both nearer and farther until just noticeable dimming 
took place. Every precaution was taken to secure accuracy. 
For one of the subjects the card could not be moved more 
than 2 mm. from the point of clearest vision without becoming 
less distinct. Very little more movement was required for any 
of the subjects. It may be safely said that all were practically 
without accommodation. Two of these observers were the 
same as were used by the writer in the earlier investigations 
made with stimuli of larger area. Opportunity was thus had 
to determine whether or not changes in accommodation play 
a more important role in the fluctuation of stimuli of 
point area than of stimuli of large area. So far as could 
be told from the records in both cases, they do not play 
a more important role. In cases of stimuli both of large and 
of very small area, the fluctuations occur for the aphakial 
subject with apparently no greater variation from the normal 
subject than is found from individual to individual with 
normal eyes. 
Of the methods used in the former work to demonstrate 
that fluctuation is a phenomenon of the adaptation of the 
sense organ, only three were available for stimuli of point 
area. In the first of these the stimuli were made of different 
colors. Speaking of this method in the first paper41 of the 
former series, the writer says, " Colors and grays were found 
to have an order of fluctuation times corresponding to their 
adaptation times. Four colors, red, green, blue, and yellow, 
gave very different fluctuation periods as compared with each 
other and with No. 27 Hering gray. The visibility times 
obtained were in the following order: red, green, blue, and 
yellow, the yellow being nearly four times as long as the red. 
" The complete adaptation times for sheets of the same 
colors were found to have the same order of length and a rough 
correspondence as to ratio of length. Further, a striking fact 
came out with regard to the phases of invisibility. Since red, 
for example, has a shorter phase of visibility than green, one 
might naturally expect that its phase of invisibility would also 
be shorter than the phase of invisibility of green. The reverse, 
however, is true. Red has a longer invisibility than green, 
and this peculiarity is especially marked if one considers the 
proportionality between the phases, i. e., the ratio invisibility: 
visibility. The same thing is true of the complementaries blue 
41 An Experimental Examination of the Phenomena Usually At- 
tributed to Fluctuation of Attention, p. 86. FERREE 
398 
and yellow. Clearly, we cannot look for a central explanation 
of this peculiarity; but it seems just what we might expect 
of adaptation from the standpoint of the compensation theory. 
The recovery process for the red is the green process. The 
green process is longer and seemingly more tenacious than 
the red, as is shown by the adaptation experiments proper, 
and is further borne out by the longer duration of the green 
after-image. A similar relation obtains in the blue-yellow 
process." In the earlier work, the stimuli were gotten as fol- 
lows. Squares of the color of the size that was wanted were 
pasted on a gray of the brightness of the color. The stimulus 
was rendered liminal by letting the light pass from the colored 
paper through a sheet of milk glass, matt on one side, placed 
at such a distance from the color as to render its intensity 
liminal. The intensity was easily regulated by slight changes 
in the distance of this glass from the colored paper. The light 
reflected from the colored papers could not be used, however, 
for stimuli of point area, because the milk glass mentioned 
above had to be used to reduce the intensity of the stimulus 
and it was impossible to get this glass thin enough to give 
noticeable color with stimuli of point area. Light was trans- 
mitted through color-filters instead. The stimulus was gotten 
as follows. A hole was pricked through a gray cardboard with 
a fine needle and covered with one or more layers of colored 
gelatin. In front of the card, in contact with it, was placed 
the sheet of milk glass, matt on one side. The hole was 
illuminated by a row of lights placed behind the cardboard, 
normal to its surface, at a distance sufficient to render the 
stimulus liminal. By this arrangement a just noticeable point 
of color was presented to the observer seated in front. 
A stimulus given by reflected light has always yielded more 
differential results in former experiments with the method 
of colors than a stimulus by transmitted light. This is prob- 
ably due to the fact that we were able to get from the former 
type of stimulus more color in proportion to the white light 
present, thus better bringing out the color differences in the 
liminal stimuli. The poorer method had to be used, however, 
because as stated above milk glass with one surface matt could 
not be obtained thin enough so that a point of colored paper 
pasted upon a background of equal brightness could be seen 
through it.42 
42 In case of the colored papers the liminal stimulus and surrounding 
field were of the same brightness, because the paper giving the stimulus 
was pasted on a gray of the brightness of the color. The only effect 
of the milk glass in front was to change the general scale of brightness 
of color and surrounding field. No brightness inequality was produced. 399 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
The registration of results was secured by means of a 
Ludwig-Baltzar kymograph, a telegraph key and an electro- 
magnetic recorder, a Jaquet chronograph (set to seconds), 
and a lamp rheostat to cut down the current from the lighting 
circuit. All of this apparatus was screened from the observer 
by means of a sliding curtain. The work was done in a long 
room with the windows all at one end. Thus cross lights, 
unequal illumination of the background, etc., could be avoided. 
The illumination of Ihe room was kept fairly constant by 
means of thin curtains covering the windows.43 The observer 
sat with his back to a high window and his head in a head-rest 
fastened to the edge of a long table, along which the frame 
bearing the stimulation apparatus was moved as required. 
The time used throughout was 1 sec. The following results 
were obtained. As in the earlier work with stimuli of a larger 
area, red showed a shorter phase of visibility and a longer 
phase of invisibility than green; and blue, a shorter phase of 
visibility and a longer phase of invisibility than yellow. In 
spite of the poorer method we were required to use, the results 
obtained were almost as strongly marked as they were when 
the same method was used with stimuli of a larger area. 
These results have been verified at the time this work was 
done and since by a large number of observers practiced and 
unpracticed. The results of three observers chosen as typical 
will be reported here. Tables I-III have been compiled from 
these results. 
TABLE I 
Obs. C.-Fluctuation with stimuli of the four principal colors of 
point area showing that the phases of visibility and invisibility have 
the characteristic adaptation and recovery peculiarities of these col- 
ors just as they have with stimuli of larger area. 
Stimulus 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red 
4-36 
•79 
1.68 
•65 
2-585 
•385 
6 .04 
Green 
5-3° 
•93 
1-13 
.42 
4.690 
■213 
6-43 
Blue 
7-75 
I . 12 
I .41 
. 71 
5 • 496 
.181 
0.16 
Yellow 
13.10 
I . 76 
1.32 
•59 
9-925 
. 100 
14.42 
In the case of the stimulus by transmitted light, this result was not 
so effectively secured because of the greater difficulty of equating the 
point of light and the surrounding field. 
"To keep the illumination constant presupposes a means of measure- 
ment. At the time the writer had at his command no means of 
measuring the illumination of a room by daylight. For a method of FERREE 
Obs. G. 
TABLE II 
Stimulus 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red 
5-i 
1.12 
1.26 
•54 
4 -047 
■ 247 
6 .36 
Green 
7 -43 
i-54 
1.14 
•37 
6.5^ 
•i53 
8-57 
Blue 
10.9 
1 -59 
1 .46 
•52 
7 .466 
•i33 
12 .36 
Yellow 
Did n 
ot flue 
tuate 
at all 
during 
period 
of ob- 
ser 
vation. 
400 
Obs. Ca. 
TABLE III 
Stimulus 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red 
i .82 
•59 
1.62 
•32 
1 .123 
.890 
3 -44 
Green 
2 .82 
•53 
1.27 
•36 
2 .243 
•45° 
4-°9 
Blue 
4.15 
•55 
1 .70 
•35 
2 .441 
•409 
5-85 
Yellow 
5 -22 
.86 
1-52 
•5i 
3 -434 
.291 
6-74 
In the second test strips of colored paper of the breadth of 
a point and 5 cm. in length were used.44 They were pasted 
on a gray background of the brightness of the color in each 
case and were observed as liminal color on the matt surface 
of the milk glass placed in front. They were arranged first 
with their longer dimension in the vertical plane, then in the 
horizontal plane. The former arrangement favored a maxi- 
mal disturbance of adaptation for observers having the 
greater range and frequency of eye-movement in the horizontal 
plane, and gave with these observers in the fluctuation experi- 
ments a corresponding increase in the phase of visibility and 
decrease in the phase of invisibility. Conversely, the latter 
arrangement favored a minimal disturbance of adaptation for 
these observers and gave a corresponding decrease in the 
phase of visibility and increase in the phase of invisibility. 
doing this, see C. E. Ferree and Gertrude Rand: An Optics-Room and 
a Method of Standardizing Its Illumination, Psychol., Rev., XIX., 
1912, 364-373- 
44 In this test we were able to use colored paper because strips, 
although only of the breadth of a point, could be seen when 5 cm. long 
through the sheet of milk glass we used. FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
401 
For each observer careful records were made of the frequency 
and range of movement and the total time the eyes were 
moving according to the methods described in the former 
papers/5 
Speaking of this test in the first paper of the former series, 
the writer says, pp. 84-90, " A more direct experimental con- 
firmation than was afforded by the method of variation of 
areas of this view that eye-movement interferes with the 
course of adaptation and is also the conditioning factor for 
the wide range of variability found in the phases of visibility 
and invisibility in the fluctuation experiments, is given by the 
following results. An examination of average frequency of 
eye-movement in the horizontal and vertical planes during 
fixation showed that three of our observers had a marked 
excess in both frequency and range in the horizontal, while 
the fourth had an excess of frequency in the vertical, but of 
range in the horizontal plane. This appeared to mean that 
for three observers, there was greater change of stimulation, 
and consequently greater relief for the adapted elements, in the 
horizontal than in the vertical direction; while the reverse 
was true, though probably to a lesser degree, for the fourth. 
To test this interpretation, stimuli longer than broad were 
used, e. g.} slips of paper 5 mm. x 40 mm. When these were 
placed with the longer dimension vertical, the shorter dimen- 
sion would fall in the direction of greater unsteadiness of 
fixation for the three observers who had the excess of eye- 
movement in the horizontal plane. Consequently, a maximal 
interference with adaptation for these stimuli would be ob- 
tained, and one might expect an increase in the phase of 
visibility and a decrease in the phase of invisibility. On the 
other hand, if the longer dimension were placed in the 
horizontal and the shorter in the vertical plane, the minimal 
interference possible for these stimuli would be secured, and 
a decrease in the phase of visibility and an increase in the 
phase of invisibility should ensue. For the fourth observer 
with the stimulus arranged as described above, the reverse 
should be true, but probably not in so marked a degree, since 
his range was greater in the horizontal, which fact to a certain 
extent would counteract the effect of frequency That 
these methods of arrangements of stimulus caused a marked 
change in the phases of visibility and invisibility for each 
" See An Experimental Examination of the Phenomena Usually 
Attributed to Fluctuation of Attention, 113-115; and The Intermittence 
of Minimal Visual Sensation, 84-87. FERREE 
402 
observer will be seen by inspecting the Tables. Indeed the 
Visibility 4- invisibility 
correspondence between the quantities : , 
visibility1 4- invisibility1 
frequency 
and , is much closer than was anticipated."46 
frequency1 
The results for the strips of point breadth are given in 
Tables IV-VI. For all the observers whose results are given 
in these tables, both the range and frequency of eye-movement 
were greater in the horizontal than in the vertical plane. 
The third test was based upon the fact that the time required 
for a colored stimulus to adapt depends to some extent upon 
the surrounding field. 
The question of what is meant by adaptation is logically raised here; 
among the followers of the Hering theory, it has come to mean, 
apparently, simultaneous induction, and Aall, reviewing the writer's 
first article,47 assumes that that is what is meant by adaptation in that 
TABLE IV 
Obs. H.-Fluctuation with horizontal and vertical arrangement of 
the stimulus. Showing how arrangements that favor maximal and 
minimal interference with adaptation affect the phases of visibility 
and invisibility. Stimulus 3 mm. x 50 mm. 
Stim- 
ulus 
Arrange- 
ment 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red. . . . 
Vertical.... 
2 -95 
.64 
1 .68 
•31 
1 -756 
•569 
4-63 
Cl 
Horizontal. . 
i .08 
•23 
2 .29 
• 4° 
.471 
2 .120 
3 -37 
Green.. . 
Vertical.... 
4 -°4 
•72 
1 -45 
.26 
2 .786 
•358 
5-49 
Il 
Horizontal. . 
1 .69 
•38 
1 .76 
-37 
.960 
1 .041 
3 -45 
Blue.... 
Vertical.... 
5 -40 
.86 
2 -°3 
•51 
2 .660 
•376 
7-43 
Il 
Horizontal. . 
2 -S1 
• 49 
3.10 
.46 
.806 
1 -235 
5.61 
Yellow.. 
Vertical. . . 
6-99 
•97 
1 -45 
.29 
4.82 
.207 
8-44 
a 
Horizontal. . 
3-55 
•72 
2 .IO 
•56 
1.928 
1 .690 
5-65 
48 For a more complete understanding why arranging the shorter 
dimension of the stimulus in the direction of the greatest eye-move- 
ment causes relatively long phases of visibility and short phases of 
invisibility; and conversely arranging the longer dimension of the 
stimulus in the direction of greatest eye-movement causes relatively 
short phases of visibility and relatively long phases of invisibility, see 
The Intermittence of Minimal Visual Sensations, 112-129; and The 
Streaming Phenomenon, 484-494. 
" Zeitschr. f. Psychol., XLIIL, Abt. 2, 1906, 456-457. 403 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
TABLE V 
Obs. R. 
Stim- 
ulus 
Arrange- 
ment 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red. . .. 
Vertical.... 
3 .60 
•75 
1 .46 
.24 
2.548 
.405 
5.06 
u 
Horizontal. . 
1.14 
.28 
3 -40 
•75 
•335 
2 .982 
4-54 
Green... 
Vertical.... 
4-3° 
.62 
1 -39 
.19 
3 -°93 
•323 
5-69 
Cl 
Horizontal. . 
2 .42 
•47 
2 .36 
•37 
1 .025 
•975 
4.78 
Blue.... 
Vertical.... 
4-93 
1.08 
1 .68 
•34 
2 -933 
• 340 
6.61 
(1 
Horizontal. . 
2 -95 
•54 
4-53 
.89 
.651 
1-535 
7 -48 
Yellow.. 
Vertical.... 
6.61 
1.23 
1 .46 
• 31 
4-527 
.220 
8 .07 
a 
Horizontal.. 
3 -26 
•79 
2 -75 
• 49 
1 .189 
• 843 
6 .01 
TABLE VI 
Obs. G. 
Stim- 
ulus 
Arrange- 
ment 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Hering 
gray 
No. 27 
Vertical.... 
8.48 
1 -32 
2 .03 
•39 
4-177 
•239 
10.51 
a 
Horizontal. . 
3-M 
•7i 
3 -21 
•56 
•978 
1.022 
6-35 
Red. ... 
Vertical. . . . 
5 -°9 
1 .12 
1 .84 
.29 
2 .766 
• 361 
6-93 
a 
Horizontal. . 
1 .65 
•38 
3 -6i 
.78 
•457 
2.187 
5 -26 
Green... 
Vertical.... 
7 .42 
1 .29 
1 -94 
• 4i 
3-824 
.261 
9-36 
a 
Horizontal. . 
2 .52 
.48 
2 .98 
.48 
.845 
1.182 
5-5° 
Blue.... 
Vertical.... 
8.99 
1 -47 
2 .84 
•39 
3-i65 
•3*5 
11.83 
a 
Horizontal. . 
3.88 
.84 
5-99 
•97 
.647 
1 -542 
9-87 
Yellow.. 
Vertical.... 
11 -49 
I -97 
1 .90 
•32 
6.047 
.165 
13 -39 
u 
Horizontal. . 
4 -84 
•97 
3-86 
.89 
1 .254 
•797 
8.70 
article. The writer, however, by no means believes that the tendency 
of a color to lose its saturation on prolonged exposure to the eye of 
of all grays to become mid-gray is due entirely or even to any 
considerable extent to simultaneous induction. He grants an influence 
to the surrounding field when there is a surrounding field, and is at 
present making a quantitative study of that influence, but it is obvious 
that the influence of the surrounding field can have no part in the 
phenomenon called general adaptation, for in that case the whole retina 
is stimulated by the same kind of light. It can apply to local adapta- FERREE 
404 
tion alone and even in local adaptation it cannot be considered a factor 
of primary importance. In 1838-1840 the loss of saturation experienced 
by a color on prolonged exposure to the eye was explained by 
Fechner48 as due to the exhaustion or fatigue of the retinal elements. 
This explanation was adopted by Helmholtz, and became a feature 
of the Young-Helmholtz theory. Hering,40 however, following a 
suggestion made by Godart, 1776,50 and elaborated by Plateau, 
1833-1835," chose rather to consider the retina compensating in func- 
tion. A compensating retina, it is obvious, should not exhaust. 
Hering bore himself out in this general position by claiming that the 
eye is ordinarily exposed to stimulation by white light from 15-18 
hours during the course of a day, and yet at the end of that time it 
has not noticeably lost in its sensitivity to white light. 
Hering himself apparently has not based his claim on experimental 
evidence. At least he neither offers results of his own nor quotes 
from the work of others. His conclusion seems to be drawn wholly 
from general observation. He says (Ueber Ermiidung und Erholung 
des Sehorgans, Arch. f. Ophthal., XXXVII., 1891, (3), p. 2) : 
" Anderseits ist es eine bekannte Thatsache, dass wir des Abends 
nicht merklich schlechter sehen als des Morgens und dass dies auch 
dann noch der Fall ist, wenn dem Tage eine in hellen Raumen durch- 
wachte Nacht und ein neuer schlafloser Morgen folgt. Also einerseits 
fortwahrende Ermiidung und zwar eine so schnell-vor sich gehende, 
dass schon nach einer wenige Secunde wahrenden-Fixierung eines 
weissen Objects auf dunklem Grunde sich die Folgen der " Ermiid- 
ung" durch ein deutliches negatives Nachbild verrathen, und ander- 
seits trotz solcher fortwahrenden raschen Ermiidung keine merkliche 
Beeintrachtigung des Lebens selbst bei tagelanger Belichtung der 
Netzhaut." He contends (p. 1) that according to the theory of fatigue, 
advocated by Helmholtz and Fick, this should not be. During an 
exposure of several hours to white light, the eye never has a chance 
completely to recover, hence should become from beginning to end of 
the period progressively more fatigued. 
This conclusion is not at all in agreement with experimental results 
obtained by C. F. Muller (Versuche fiber den Verlauf der Netzhauter- 
miidung. Diss, inaug., Zurich, 1866), for example, who from the 
results of his tests of the loss of sensitivity of the eye to white light 
from morning to night, concludes: " Am Abende erscheint der Retina 
irgend ein Object nur in 0.49 derjenigen Helligkeit, in welcher es ihr 
am Morgen erschienen ware." Moreover, he found that the shape of 
the curve of fatigue undergoes a very decided change during the 
course of the day. Aubert also disagrees with Hering. He says 
48 G. T. Fechner: Pogg, Ann., XLIV., 1838, 221, 513; XLV., 1838, 
227; L., 1840, 193, 427. The theory was conceived earlier by Scherffer 
(Abhandlung von den zufdlligen Farben, Wein, 1765; also Journal de 
Physique de Rozier, XXVI., 175, 273), who explained the negative 
after-image by the conception that the retina is diminished in sensi- 
tivity by fatigue produced by previous stimulation. 
49 Ewald Hering: Zur Theorie vom Lichtsinne, 1874; von Graefe's 
Archiv, XXXVII., 1891, (3), I, and 1892, XXXVIII., (2), 252. 
00 de Godart: Journal de Physique de Rozier, VIII., 1776, (1), 269. 
"Plateau: Ann. de Chimie et de Physique, LIIL, 1833, 386; LVIL, 
1835, 337; P°gg- Ann., XXXII., 1834, 543. More fully in Essai d'une 
theorie generale, etc. Mem. de l'Acad. de Belgique, VIII., 1834. 405 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
(Moleschott's Untersuchungen, VIII., 1862, 251; see also Beitrage zur 
Physiologic der Netzhaut. Abhandlungen der Schlesischen Gesell- 
schaft, Breslau, 1861, 39) : " Es erscheint mir also aus obiger 
Bemerkung hervorzugehen, dass im Laufe des Tages dur ch die 
Einwirkung des Lichtes die Empfindlichkeit unserer Retina fortwahrend 
abnimmt, so dass wir am Abende weniger empfindlich gegen Licht 
sind, als des Morgens." Moreover, without supporting evidence either 
from general observation or from experiments on color, in fact in 
complete disregard of this evidence, Hering, as he has done in many 
other cases in his work on the optics of color, has generalized with 
regard to the retina's response both to white light and to colored light 
from the results of observations with white light alone. For example, 
it is scarcely necessary to point out that the eye cannot be 
exposed from 15-18 hours to colored light without loss of sensi- 
tivity to color. Without dwelling further, however, on the evidence 
for and against a compensation theory, it will be sufficient for our 
purpose here to point out that if one were to hold to a compensation 
theory in the Hering sense, it would be necessary for him to seek 
some other explanation than exhaustion for the loss of sensitivity 
of the eye, apparent or real, to color or brightness. Hering apparently 
conceives that this happens only in case two surfaces of different 
quality are juxtaposed, and then all that takes place is that each is 
induced over the other and the qualitative difference between the two 
tends to disappear. There is, then, no real loss of sensitivity of the 
eye to either. Both become alike because by induction they are mixed 
to equality. The following objections may be offered to the explana- 
tion. (1) As stated before, it cannot apply to general adaptation. 
Yet it is well known that the eye loses its sensitivity to color when 
the whole retina is stimulated by that color, in fact more rapidly 
than when only a part is stimulated, except perhaps in case of certain 
combinations of color and surrounding field. (2) It can apply to local 
adaptation only in case the two fields juxtaposed both belong to the 
brightness series. For example, when the eye is exposed for some 
length of time to a white surface contiguous to a black or a light gray 
to a dark gray, the lighter surface is observed to darken and the 
darker to lighten. This might be explained by the mixture of the 
two qualities by induction. The evidence afforded by the observation, 
however, is not at all differential, for the phenomenon may be ex- 
plained just as well by exhaustion. A different situation entirely is 
presented, however, when the two contiguous surfaces are colored. 
In this case there is very little in the phenomenon that could by the 
most favourable interpretation be construed as a mixture to an inter- 
mediate color quality. For example, when red and blue are stared at 
in juxtaposition, we should expect, in terms of Hering's explanation, 
both surfaces to become purple with no more loss of saturation than 
would be attendant upon distributing each color uniformly upon both 
surfaces. This, however, is not at all what takes place. The promi- 
nent effect is loss of saturation. The two surfaces tend to become 
alike for the most part only because both tend towards gray. The 
blue, it is true, does acquire a tinge of violet, but it does this as the 
result of adaptation even when red is not juxtaposed. It probably 
does become slightly more reddish by being alongside the red, but 
the evidence of induction is not great. The red, likewise, may be 
modified a little by being alongside the blue, but the effect is even less 
noticeable than it is for the blue. Similar results are gotten with FERREE 
406 
green and yellow. In case the colors juxtaposed are complementary 
colors, the results of induction should be towards a cancellation to 
gray. But again the tendency towards gray which is actually observed 
affords no differential evidence for this theory of induction, because 
the shift towards gray can be explained just as easily in terms of 
the exhaustion theory. And that induction can have little to do with 
the phenomenon may be shown by the facts (i) that the tendency 
would have been towards gray had the whole retina been stimulated 
by one of the colors alone, and (2) that so far as can be told, the 
process is hastened little, if any, by the juxtaposition of the two 
colors. In the Lichtsinne, 1878, pp. 36-37 Hering describes the experi- 
ment upon which he bases his explanation of adaptation in terms of 
simultaneous induction. His device for stimulating the eye consists 
of a white and black surface juxtaposed. No attempt is made to 
extend the experiment to color. Moreover, in drawing his conclusions, 
no heed whatever is given to what would happen were the whole retina 
stimulated by light of one quality. This is a truly remarkable instance 
of a broad generalization made from a slender basis of fact. 
The writer, then, does not wish it to be understood that he explains 
the fluctuation of minimal visual stimuli in terms of simultaneous 
induction.. He has called this fluctuation a phenomenon of the adapta- 
tion and recovery of the sense organ, meaning by adaptation here, as in 
the original article, the progressive loss of sensitivity of the eye to 
colored and to colorless light caused by prolonged exposure. Just 
what the factors are in adaptation, will be made the subject of a 
further paper. They vary under different circumstances. In case of 
local adaptation, simultaneous induction is one of the factors, and in 
certain especial cases it may exert considerable influence, as is recog- 
nized in the test described above; but to make it the sole cause of the 
adaptation of the eye to its stimulus seems to the writer, in the face 
of the experimental evidence, to be little short of absurd. 
In the earlier experiments it was found that by keeping the 
surrounding field constant and varying the stimulus, or con- 
versely, by keeping the stimulus constant and varying the 
surrounding field, a difference in the period of fluctuation was 
obtained, showing itself chiefly in the phase of visibility. The 
same thing held in the recognized adaptation experiments. 
The variations in the phases of visibility and invisibility that 
were produced in the one, were produced in the other; the 
only departure from precise correspondence being that the 
differences were more marked in case of the recognized adapta- 
tion experiments, as would be expected from the longer dura- 
tion of the process. The old series of Hering papers was 
used both in these experiments and in the experiments with 
stimuli of point area because combinations more favorable to 
rapid adaptation could be found in this series. Some of the 
combinations most favorable were the vermilion of the series 
on the blue-green, and the vermilion upon Hering gray No. 27; 
and some of the most unfavorable combinations were dark 407 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
red on yellow, and dark blue on yellow. The combinations 
favoring rapid adaptation gave in the fluctuation experiments 
a short phase of visibility and a long phase of invisibility, 
and conversely, the combinations unfavorable to rapid adapta- 
tion gave long phases of visibility and short phases of invisi- 
bility. Although the writer had carefully determined in an 
earlier experiment with large areas which were the favorable 
and which the unfavorable combinations, still in order to make 
the correspondence between fluctuation and adaptation still 
more complete, in cases of stimuli of very small area both 
adaptation and fluctuation experiments were conducted in the 
present study. As was the case in the earlier experiments, 
the advantages of a stationary stimulus and surrounding field 
had to be sacrificed in these experiments, because the use of 
the milk glass to reduce the saturation of the stimulus, as was 
done when a stationary system was used, would also have 
reduced the saturation of the color in the surrounding field. 
This would not have been desirable for the purpose of the 
experiment. Accordingly, the Masson disc with the broken 
radius of point breadth was substituted for the stationary 
system. In case of the adaptation experiments, a point of color 
of full intensity was pasted upon the various backgrounds 
and observed at the proper distance. In conducting this 
adaptation series with stimuli of point area, we were not only 
getting the results needed for comparison in our fluctuation 
series, but by using stimuli of full intensity, we were applying 
our test under precisely the same conditions used by Heinrich 
in his fluctuation experiments. In both cases the effect of the 
favorable and unfavorable combinations was plainly marked 
in the results. For the results of these experiments see 
Tables VII-X. 
TABLE VII 
Obs. R.-Showing that combinations that influence adaptation 
time correspondingly influence fluctuation for stimuli of point area 
just as they do for stimuli of larger area. Fluctuation series, stim- 
ulus-ring 0.3 mm. broad and of liminal intensity. 
Stim- 
ulus 
Background 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red. . . . 
Blue 
4-258 
.81 
1 -93° 
.83 
2 .220 
• 45° 
6.215 
Red.. . . 
Orange 
6 .041 
.98 
1 -327 
•59 
4 - 552 
.219 
7-368 
Red. . . . 
Yellow-green 
6.10 
.91 
•933 
•32 
6.538 
.152 
7 -°33 
Yellow.. 
Red 
9.166 
1 . IO 
1 .125 
.26 
8.147 
.122 
10.291 FERREE 
408 
TABLE VIII 
Obs. B. 
Stim- 
ulus 
Background 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red.... 
Blue 
3-3° 
.82 
2.871 
.88 
1.149 
.870 
6.171 
Red. . . . 
Orange 
5 -083 
-94 
2 .416 
•75 
2.103 
•475 
7 -499 
Red. . . . 
Yellow-green 
4-125 
•59 
1.540 
.64 
32 .678 
•373 
5 -665 
Yellow.. 
Red 
6.230 
• 94 
1.050 
•32 
5 -933 
.168 
7.28 
TABLE IX 
Obs. R.-Showing that combinations that influence adaptation 
time correspondingly influence fluctuation for stimuli of point area 
just as they do for stimuli of larger area. Adaptation series, stimuli 
of point area and of full intensity. 
Stim- 
ulus 
Background 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
Vis. 
Period 
Red... 
Blue 
13 -o 
1.09 
2 -577 
.76 
5 -°44 
.198 
15-577 
Red... 
Orange 
11 -323 
.90 
1.112 
•58 
10.182 
.098 
12 .435 
Red.. . 
Yellow-green 
11 .032 
.87 
.781 
•32 
14.122 
.070 
11-813 
Yellow. 
Red 
200 . 
No 
fluctu 
ation. 
TABLE X 
Obs. B. 
Stim- 
ulus 
Background 
Vis. 
M.V. 
Invis. 
M.V. 
Vis. 
Invis. 
Invis. 
vis. 
Period 
Red... 
Blue 
5-76i 
•59 
5 -833 
• 654 
•987 
I .012 
u-594 
Red... 
Orange 
6.636 
.98 
4-723 
•76 
1 .405 
.711 
11 -359 
Red... 
Yellow-green 
10.751 
1.05 
4-854 
•97 
2 .215 
• 451 
15-605 
Yellow. 
Red 
200 . 
No 
fluctu 
ation. 
V. Conclusion 
In conclusion the following points may be reviewed, (i) 
The work offered by Heinrich and Chwistek in support of the 
accommodation theory for the fluctuation of stimuli of point 
area was done with stimuli of full intensity. In using stimuli 
of this intensity Heinrich and Chwistek have created for them- 
selves a special problem. The doctrine of fluctuation of atten- 
tion has never been applied to the fluctuation of stimuli 409 
FLUCTUATIONS OF LIMINAL VISUAL STIMULI 
strongly supraliminal in intensity. (2) Their strongest and 
most direct argument for the accommodation theory is their 
claim of having directly demonstrated a coincidence between 
involuntary changes of accommodation and fluctuation. The 
validity of this claim, however, rests primarily upon whether 
or not they have given a valid demonstration of the in- 
voluntary changes in accommodation. Their demonstration of 
involuntary changes in accommodation is strongly open to 
question. Employing 124 observers, the writer has been un- 
able in a single case to make their popular demonstration work 
when care was taken to rule out extraneous factors which 
would themselves cause the phenomenon. And their technical 
demonstration with the ophthalmometer is in terms of a 
phenomenon which is described by de Schweinitz and others 
as one of the common phenomena of ophthalmometry due to 
eye-movement. The coincidence, then, which they claim to 
have observed between the fluctuation of stimuli of point area 
and changes in the curvature of the lens, is in all probability 
a coincidence between eye-movement and fluctuation. (3) 
Moreover, none of the evidence they have offered as indirectly 
proving the accommodation theory can be considered in any 
sense differential. All of it can be explained just as easily 
either in terms of the writer's adaptation theory or in terms 
of Loria's theory for the fluctuation of stimuli of point area. 
Some of it can even be explained in terms of any theory that 
has yet been advanced to account for the fluctuation of minimal 
visual stimuli. (4) The fluctuation of stimuli of point area 
presents no especial case. For (a) involuntary changes in 
accommodation are not an essential factor in these fluctua- 
tions. They take place for aphakial subjects apparently just 
as readily as for subjects with normal eyes. And (b) identi- 
fied by the tests used by the writer in his earlier work these 
fluctuations correspond just as closely to adaptation phe- 
nomena as do the fluctuations of stimuli of larger area. (5)The 
fluctuation of minimal visual stimuli whether of large or small 
area is a phenomenon of the adaptation and recovery of the 
sense organ. And by adaptation is meant the progressive loss 
of sensitivity to colored and colorless light caused by prolonged 
exposure of the eye to these lights. It is not simultaneous 
induction. Simultaneous induction can be considered only as 
a minor factor in the adaptation of the eye to its stimulus. [Reprinted from Science, N. S., Vol. XL., No. 1020, 
Pages 84-91, July 17, 19 If] 
THE PEOBLEM OF LIGHTING IN ITS EE- 
LATION TO THE EFFICIENCY OF 
THE EYE1 
Up to the present time the work on the prob- 
lem of lighting has been confined almost en- 
tirely to the source of light. The goal of the 
lighting engineer has been to get the maxi- 
mum output of light for a given expenditure 
of energy. Until recent years little attention 
has been given to the problem in its relation 
to the eye. It is the purpose of this paper to 
outline in a general way some of the more im- 
portant features of this phase of the subject, 
and to give some of the results of work that is 
now being done on the problems that these 
features present. 
Confronting the problem of the effect of 
lighting systems on the eye, it is obvious that 
the first step towards systematic work is to 
obtain some means of making a definite esti- 
mate of this effect. The prominent effects of 
bad lighting systems are loss of efficiency, 
temporary and progressive, and eye discom- 
fort. Three classes of effect may, however, 
be investigated: (1) the effect on the general 
level or scale of efficiency for the fresh eye; 
(2) loss of efficiency as the result of a period 
of work; and (3) the tendency to produce dis- 
comfort. Of these three classes of effect the 
last two are obviously the more important, 
for the best lighting system is not the one 
that gives us the maximum acuity of vision 
for the momentary judgment or the highest 
1 This paper, with some changes, was read be- 
fore the American Philosophical Society of Phila- 
delphia, April 4, 1913. 2 
level of efficiency for the fresh eye. It is 
rather the one that gives us the least loss of 
efficiency for a period of work, and the maxi- 
mum of comfort. 
In 1911 the American Medical Association 
appointed a committee to study the effect of 
different lighting systems on the eye. The 
writer was asked to share in the work of this 
committee. The problem presented to him 
was to furnish tests that would show the ef- 
fect of different lighting systems on the eye, 
and more especially to devise, if possible, a 
test that would show loss of efficiency as a re- 
sult of three or four hours of work under 
an unfavorable lighting system. In his work 
directed along these lines he has succeeded in 
getting methods of estimating effect which 
after eighteen months of trial seem sufficiently 
sensitive to differentiate between good and 
bad lighting systems with regard to these 
points. Tie has undertaken, therefore, to 
determine (1) the lighting conditions that 
give in general the highest level or scale of 
visual efficiency; (2) the conditions that give 
the least loss of efficiency for continued work; 
and (3) the conditions that cause the least 
discomfort. This plan of work, it is scarcely 
needful to remark, will involve a wide range 
of experimentation. The crux of the problem 
is, however, to secure reliable methods of esti- 
mating effect. Having these methods, the 
factors, whatever they may be, distribution, 
intensity, quality, position of the light rela- 
tive to the eye, etc., can be varied one at a 
time and the effects be determined. From these 
effects it should not be difficult to ascertain 
what lighting conditions are best for the eye, 
and what is the relative importance of the 
factors that go to make up these conditions. 
Further, it should be possible on the practical 
side to test out and perfect a lighting system 
before it is put on the market; also to deter- 3 
mine the best conditions of installation for a 
given lighting system; to investigate the effect 
of different kinds of type and paper on the 
eye; to study the effect of different kinds of 
desk lighting, etc. In short, it is obvious that 
the usefulness of such tests is limited along 
these lines only by their sensitivity. 
A detailed description of the tests we are 
using has already appeared in print.2 Time 
can not be given to them here. A brief report 
only of some of the results of the work in 
which they have been employed is possible in 
the time placed at my disposal. 
In the study of the problems presented to 
us in this field it has been thought best to 
conduct the investigation at first along broad 
lines in order to determine in a general way 
the conditions that affect the efficiency and 
comfort of the eye. Later a more detailed 
examination will be made of the ways in 
which these conditions have been worked out 
in the various types of lighting systems in use 
at the present time. The following aspects of 
lighting sustain an important relation to the 
eye: the evenness of the illumination, the 
diffuseness of light, the angle at which the 
light falls on the object viewed, the evenness 
of surface brightness, intensity and quality. 
The first four of these aspects are very closely 
interrelated, and are apt to vary together in 
a concrete lighting situation, although not 
in a 1:1 ratio. For the purposes of this paper 
these aspects will be grouped together and 
referred to as the distribution of light and 
surface brightness in the field of vision, or 
still more generally as distribution. The ideal 
condition with regard to distribution is to have 
the field of vision uniformly illuminated with 
2 ' ' Tests for the Efficiency of the Eye Under Dif- 
ferent Systems of Illumination and a Preliminary 
Study of the Causes of Discomfort," Transactions 
of the Illuminating Engineering Society, 1913, 
VIII., pp. 40-60. 4 
light well diffused and no extremes of sur- 
face brightness. When this condition is 
attained, the illumination of the retina will 
shade off more or less gradually from center to 
periphery, which gradation is necessary for 
accurate and comfortable fixation and accom- 
modation. 
The factors we have grouped under the 
heading distribution can be most conveni- 
ently discussed perhaps with reference to four 
types of lighting systems in common use 
to-day: illumination by daylight, direct light- 
ing systems, indirect lighting systems and 
semi-direct systems. In the proper illumina- 
tion of a room by daylight, we have been able 
thus far to get the best conditions of distribu- 
tion. Before it reaches our windows or sky- 
lights daylight has been rendered widely dif- 
fuse by innumerable reflections; and the 
windows and skylights themselves, acting as 
sources, have a broad area and low intrinsic 
brilliancy, all of which features contribute 
towards giving the ideal condition of distri- 
bution stated above, namely, that the field of 
vision shall be uniformly illuminated with 
light well diffused and that there shall be no 
extremes of surface brightness. Of the sys- 
tems of artificial lighting the best distribu- 
tion effects, speaking in general terms, are 
given by the indirect systems. In this type of 
system the source is concealed from the eye 
and the light is thrown against the ceiling or 
some other diffusely reflecting surface, in such 
a way that it suffers one or more reflections 
before it reaches the eye. In some of the 
respects most important to the eye, this system 
gives the best approximation of the distri- 
bution effects characteristic of daylight of 
any that has yet been devised. The direct 
lighting systems are designed to send the 
light directly to the plane of work. There is 
in general in the use of these systems a 
tendency to concentrate the light on the work- 5 
ing plane or object viewed rather than to dif- 
fuse it, and, therefore, a tendency to emphasize 
brightness extremes rather than to level them 
down. Too often, too, the eye is not properly 
shielded from the light source and frequently 
no attempt at all is made to do this. The semi- 
indirect systems are intended to represent a 
compromise between the direct and indirect 
systems. A part of the light is transmitted 
directly to the eye through the translucent 
reflector placed beneath the source of light, and 
a part is reflected to the ceiling. Thus, de- 
pending upon the density of the reflector, this 
type of system may vary between the totally 
direct and the totally indirect as extremes and 
share in the relative merits and demerits of 
each in proportion to its place in the scale. 
By giving better distribution this type of 
system is supposed also to be a concession to 
the welfare of the eye, but our tests show that 
the concession, at least for the type of 
reflector we have used,3 is not so great as it is 
supposed to be. In fact, installed at the inten- 
sity of illumination ordinarily used or at an 
intensity great enough for all kinds of work, 
little advantage is gained for the eye in this 
type of lighting with reflectors of low or 
medium densities; for with these intensities 
of light and densities of reflector, the bright- 
ness of the source has not been sufficiently 
reduced to give much relief to the suffering 
eye.4 Until this is done in home, office and 
s The reflectors we used were supplied to us by a 
prominent lighting corporation, interested neither 
in the manufacture nor the sale of lighting fixtures, 
in response to a request for a representative semi- 
indirect lighting system. Obviously, however, final 
conclusions should be reserved until the tests are 
extended to other types of reflectors. 
4 The semi-indirect system used by us was but 
little better for the eye than the direct sys- 
tem. The direct system we employed was the one in 
general use throughout the building in which our 
tests were made. It was installed about six years 6 
public lighting we can not hope to get rid of 
eye-strain with its complex train of physical 
and mental disturbances. 
It is not our purpose, however, at this time 
to attempt a final rating of the merits of 
lighting systems. For that our work is still 
too young. Moreover, there are relatively 
good and bad systems of each type, and good 
and bad installations may be made of any 
system. What we hope to do is by making an 
appropriate selection and variation of condi- 
ago and is, therefore, not of the most modern 
type. It seems to the writer safe to say, however, 
that it gives effects fully as good as most direct 
lighting in actual use in the country to-day. Fur- 
thermore, it is difficult to believe that any great 
injustice has been done to direct lighting, so far as 
this principle of lighting has been commercialized 
up to this time, by the selection of this system, be- 
cause of the fact that very little less loss of effi- 
ciency was obtained from the semi-indirect lighting 
system, which on account of its similarity to indi- 
rect lighting represents, we have good reason to be- 
lieve from our results, a greater modification of di- 
rect lighting for the welfare of the eye than any 
that is found within the class of direct systems. 
However, a final conclusion will be reserved until a 
more extensive investigation of the direct systems 
has been made. The writer further does not wish 
to be understood as contending that direct lighting 
can not be accomplished in a way that is not ex- 
cessively damaging to the eye. Doubtless great im- 
provement can be made in this type of lighting if 
proper attention is given to the fundamental prin- 
ciples governing the effect of light on the eye. It 
does not seem to the writer, however, that the prin- 
ciple of direct lighting offers as great possibilities 
in this direction as the indirect; still he permits 
this also to remain an open question in his mind. 
It is obvious that much can be accomplished for the 
welfare of the eye in cases both of the direct and 
semi-indirect systems by using sources of large 
area and of low intrinsic brilliancy, by removing 
them as much as possible from the field of vision, 
by employing better means of diffusing the light, 
etc. 7 
•bions to find out what the factors are that 
are of importance to the eye, and from this 
knowledge as a starting point to work towards 
reconstruction. 
With regard to the effect of the distribu- 
tion of light and surface lightness on 
the eye a brief statement will be given here 
only of its effect on efficiency; and in the 
consideration of efficiency loss of efficiency 
will receive the major part of our attention. 
No attempt will be made, for example, to pre- 
sent the results of the study of the factors 
producing discomfort. The study of these 
factors has constituted for us an entirely 
separate and independent piece of work inves- 
tigated by separate and independent methods. 
Our tests for loss of efficiency5 show that 
when the intensity and quality of the light 
5 The tests were made in a room 30.5 feet long, 
22.3 feet wide, and 9 feet high. The artificial 
lighting was accomplished by means of two rows 
of fixtures of four fixtures each. Each row was 6 
feet from the side wall and the fixtures were 6 
feet apart. The reflectors were in the different 
cases 19-26 inches from the ceiling. Clear tungsten 
lamps were used as source. The voltage was kept 
constant by means of a voltmeter and a finely grad- 
uated wall rheostat placed in series with the light- 
ing circuit. In case of the direct system two bulbs 
making an angle of 180° were used for each fixture 
and the distribution was obtained by means of white 
slightly concaved porcelain reflectors 16 inches in 
diameter fastened directly above. In case of the in- 
direct system corrugated mirror reflectors, enclosed 
in brass bowls, were used. For the semi-indirect 
system the distribution was obtained by means of 
inverted alba reflectors 11 inches in diameter 
which threw a part of the light against the 
ceiling and transmitted the rest directly to 
the room, minus a rather large absorption quan- 
tity. The daylight illumination came from 
three windows all on one side of the room and 
situated in a line parallel with the line of 
sight used when making the tests. These windows 
were so sheltered that it was never possible for 
them to receive light directly from the sun or 
from a brightly illuminated sky. Moreover, the 8 
are equalized at the point of work, the eye 
loses practically nothing in efficiency as the 
result of three to four hours of work under 
daylight. It loses enormously for the same 
period of work under the system of direct 
lighting selected for our work and almost as 
much under the system of semi-indirect 
lighting. Under the system of indirect 
lighting, however, the eye loses but little 
more than it loses in daylight. The results 
of these tests show also that acuity of vision 
as determined by the momentary judgment is 
light from one of them, the one nearest the ob- 
server, was further diffused by passing through 
a diffusion sash made of double thick glass ground 
on one side. The intensity in foot-candles was 
made equal at the point of work for all the sys- 
tems employed. In making this equalization the 
light was photometered in five directions at the 
point of work: with the receiving surface of the 
photometer in the horizontal plane, at angles of 
45° and 90° pointing towards the observer, and at 
angles of 45° and 90° pointing in the opposite di- 
rection. In installing the lights in the different 
systems it was impossible to make the intensity 
equal in all of these directions. Care was taken to 
make it equal in the plane of the test card, i. e., 
the vertical plane, and as nearly as possible equal 
in the other planes. The Sharpe-Millar portable 
photometer was used to make these measurements, 
also another method mentioned in a former paper 
(op. cit., p. 49) which is more sensitive to day- 
light illumination than is the Sharpe-Millar 
method. The effect of varying distribution of 
light was thus tested under conditions in which 
quality and intensity were reduced as nearly to a 
constant as was possible with the systems em- 
ployed. The intensity in the vertical plane was 
made in each case 1.4 foot-candles or approximately 
so. Space can not be taken here for an engineer- 
ing specification of the installations used and the 
lighting effects produced. A full report of the 
work including detailed brightness and illumina- 
tion measurements, photographs showing the il- 
lumination effects obtained, descriptions of installa- 
tions, etc., will be published in the Transactions of 
the Illuminating Engineering Society. 9 
higher for the same foot-candles of illumina- 
tion for the daylight system than for the 
systems of artificial lighting, and that for the 
latter systems, it is highest for the indirect 
system, next highest for the semi-indirect 
system, and lowest for the direct. It will thus 
be seen that for all purposes of clear seeing, 
whether the criterion be maximum acuity or 
the ability of the eye to hold its efficiency for 
a period of work, the best results are given in 
order by the systems that give the best dis- 
tribution of light and surface brightness. The 
effect of distribution is not so great, however, 
on the ability of the fresh eye to see clearly as 
it is on its power to hold its efficiency. 
The loss of efficiency found in the above 
work seems to be predominantly, if not en- 
tirely muscular, for the tests for the sensitiv- 
ity of the retina show practically no loss of 
sensitivity as the result of work under any of 
the lighting systems employed. The following 
reasons are suggested why the muscles of the 
eye giving both fixation and accommodation 
should have been subjected to a greater strain 
by the systems of direct or semi-direct light- 
ing, than by the system of indirect light- 
ing or daylight. (1) The bright images 
of the sources falling on the peripheral 
retina which is in a perpetual state of dark- 
ness-adaptation, as compared with the cen- 
tral retina, and is, therefore, extremely sensi- 
tive in its reaction to such intensive stimuli, 
set up a reflex tendency for the eye to fixate 
them instead of, for example, the letters which 
the observer is required to read. (2) Like- 
wise, a strong reflex tendency to accommodate 
for these brilliant sources of light, all at 
different distances from each other and the 
lettered page, is set up. (3) These brilliant 
images falling on a part of the retina that 
is not adapted to them, causing as they do 
acute discomfort in a very short period of time, 
doubtless induce spasmodic contractions of the 
muscles which both disturb the clearness of 10 
vision and greatly accentuate the fatiguing 
of the muscles. The net result of all these 
causes is excessive strain, which shows itself 
in a loss of power to do work. In the illu- 
mination of a room by daylight, however, 
with a proper distribution of windows, the 
situation is quite different. The field of 
vision contains no bright sources of light to 
disturb fixation and accommodation and to 
cause spasmodic muscular disturbances due 
to the action of the intensive light sources on 
the dark-adapted and sensitive peripheral 
retina. As has already been pointed out, the 
light waves have suffered innumerable reflec- 
tions and the light has become diffuse. The 
field of vision is comparatively speaking uni- 
formly illuminated and there are no extremes 
of surface brightness. The illumination of 
the retina, therefore, falls off more or less 
gradually from center to periphery, as it should 
to permit of fixation and accommodation for a 
given object with a minimum amount of strain. 
It is not our purpose, however, to contend 
that distribution is the only factor of impor- 
tance in the illumination of a room. We have 
chosen to begin our work with types based on 
distribution, only because it has seemed to us, 
both from our own work and from a survey of 
the work done by others, that this is the most 
important factor with which we have yet to 
deal in our search for the conditions that give 
minimum loss of efficiency and maximum 
comfort in seeing. The quality of light and 
its intensity at the source are already pretty 
well taken care of, apparently better taken care 
of, at least in general practise relative to their 
importance to the eye, than is distribution. 
A systematic study of factors, however, can 
not stop with an investigation of the effect of 
distribution alone. The intensity and quality 
of light must also be taken into account. For 
example, one of the most persistent questions 
asked by the illuminating engineer is, " How 
much light should be used with a given 11 
lighting system to give the best results for 
seeing?" We have undertaken, therefore, to 
determine the most favorable range of inten- 
sity for the four types of distribution men- 
tioned above. Curves have been obtained 
showing the effect on the efficiency of the eye 
of three or four hours of work under different 
intensities of light, for the direct and semi- 
indirect systems; and rough comparisons have 
been made for the indirect system and for day- 
light. Detailed tests will be made for these 
latter two systems early next year. Our tests 
show, in general, the following results. A very 
wide range of intensity is permissible for day- 
light and the indirect system. For the semi- 
indirect system the eye falls off heavily in effi- 
ciency for all intensities with the exception of 
a narrow range on either side of 2.2 foot- 
candles, measured at the level of the eye at 
the point of work with the receiving surface 
of the photometer in the horizontal plane. 
For the direct system no intensity can be 
found for which the eye does not lose a very 
great deal in efficiency as the result of work. 
Thus it seems that distribution is funda- 
mental. That is, if the light is well distri- 
buted and there are no extremes of surface 
brightness as is the case for daylight and the 
indirect systems of artificial lighting, the 
ability of the eye to hold its efficiency is, 
within limits, independent of intensity. In 
short, the retina is itself highly accommoda- 
tive or adaptive to intensity, and if the proper 
distribution effects are obtained, the condi- 
tions are not present which cause strain and 
consequent loss of efficiency in the adjustment 
of the eye. 
Details of the conditions of installation and 
of the methods of working can not be given 
here. It will be sufficient to state that the 
work was done in the same room, with the 
same fixtures, and in general with the same 
conditions of installation and methods of 
working as were used in the tests for distri- 12 
bution. Nor can a full statement of results 
be made. Time will be taken, however, for a 
more detailed examination of the results ob- 
tained for the direct and semi-indirect sys- 
tems. For the semi-indirect systems, our test 
showed that the intensity most favorable to 
the eye was secured when the photometric 
reading with the receiving surface in the 
horizontal plane showed 2.2 foot-candles of 
light at the point of work, 1.52 foot-candles 
in the 45° position, and .58 foot-candle in 
the vertical position. At this intensity of 
illumination, the semi-indirect system,. so far 
as its effect on the eye's loss of efficiency is 
concerned, compares fairly well with the in- 
direct system at such ranges of intensity as 
we have employed. At intensities appreciably 
higher than this most favorable value, or lower, 
the loss of efficiency is very great. At the 
intensity commonly recommended in lighting 
practise, the semi-indirect system is almost, 
if not quite, as damaging to the eye as the 
direct system. The intensity recommended 
by the Illuminating Engineering Society, for 
example, in its primer issued in 1912, ranges 
from 2-3 to 7-10 foot-candles, depending upon 
the kind of work. Five foot-candles is taken 
as a medium value. This medium value, it 
will be noted, is more than double the amount 
we have found to give the least loss of effi- 
ciency for the type and installation of semi- 
indirect system we have used. The intensity 
we have found to give the least loss of effi- 
ciency for this type of lighting, does not, 
however, give a maximum acuity of vision 
as determined by the momentary judgment. 
At an intensity that does give maximal acuity 
for the momentary judgment the eye runs 
down rapidly in efficiency. That is, in this 
type of lighting, one or the other of these 
features must be sacrificed. High acuity and 
little loss of efficiency can not be had at the 
same intensity. They could both be had only 
under the indirect system and daylight. How- 13 
ever, the amount of light we find to give the 
least loss of efficiency seems to be sufficient for 
much of the work ordinarily done in the home 
or office. It is not enough, though, for draft- 
ing or work requiring great clearness of 
detail. 
In case of the direct system, we were able to 
improve the conditions, so far as loss of 
efficiency is concerned, by reducing the inten- 
sity; but the system never proved so favorable 
in this regard as even the semi-indirect system. 
In the tests made under the direct system care 
was taken to have the fixtures in the 
same position in the room in every case as 
they were for the semi-indirect system. The 
most favorable intensity is secured by an in- 
stallation that gave 1.16 foot-candles in the 
horizontal, .85 in the 45° position and .45 
in the vertical. At this intensity, however, 
the loss in the efficiency of the eye for three 
hours of work was almost four and one half 
times as great as for a wide range of inten- 
sities for either the indirect system or day- 
light. 
Two facts, then, may be emphasized at this 
point. (1) Of the lighting factors that influ- 
ence the welfare of the eye, those we have 
grouped under the heading distribution appar- 
ently are fundamental. They seem to be the 
most important we have yet to deal with in our 
search for the conditions that give us the mini- 
mum loss of efficiency and the maximum com- 
fort in seeing. If, for example, the light is 
well distributed in the field of vision and there 
are no extremes of surface brightness, our 
tests seem to indicate that the eye, so far 
as the problem of lighting is concerned, is 
when the proper distribution is present, inten- 
sities high enough to give the maximum dis- 
crimination of detail may be employed with- 
out causing appreciable damage or discomfort 
to the eye. (2) For the kind of distribution 
effects given by the majority of lighting 
systems in use at the present time, our results 14 
show that too much light is being employed 
for the welfare and comfort of the eye. 
The effect of quality of light on the eye has 
been the subject of much discussion and much 
misunderstanding. There seems to be a feel- 
ing even among lighting engineers and oph- 
thalmologists that colored light gives better 
results for seeing than white light. Some, for 
example, hold that the kerosene flame furnishes 
the ideal source of light and that its virtues 
are due largely to the yellow quality of the 
light it gives off. While the writer has not 
as yet begun a systematic study of the effect 
of quality of light, and while he is, therefore, 
not as yet willing to commit himself on this 
point, he will say that when intensity and dis- 
tribution are equalized, an installation of clear 
carbon lamps, which gives a . light compara- 
tively rich in yellow and red, causes the eye 
to fall off more in efficiency as the result of 
3-4 hours of work than an installation of 
clear tungsten lamps, the light from which is 
more nearly white. In short, the question 
whether or not white or colored light is better 
for the eye can not be answered until definite 
tests are made of this point alone under con- 
ditions in which all other factors are rendered 
constant. The effects of the kerosene flame, 
for example, as compared with other sources 
of illumination, must be tested under a system 
of installation that gives the same intensity 
at the source, and, as nearly as possible, the 
same distribution in the field of vision as is 
given by other illuminants. This has not been 
done at all. Our judgment of the compara- 
tive merits of the color quality of the light 
given by it have been based on the roughest 
kinds of impression, obtained under condi- 
tions of installation in which there has been 
no attempt at control of the other factors that 
influence the effect of light on the eye. The 
work that has been done up to this time on the 
relation of quality of light to seeing has been 
confined to visual acuity as determined by the 15 
momentary judgment, and even this work 
which alone can give no safe grounds at all 
for drawing general conclusions as to the 
effect of light on the welfare of the eye, shows, 
whenever the comparison has been made, that 
white light gives a greater acuity of seeing 
than light with a dominant color tone. If, as 
has been maintained by some on the grounds 
of their working experience, the kerosene flame 
is easier on the eye than the more modern 
sources of illumination, the writer would be 
inclined, more especially in view of his results 
on the effect of differences in intensity on the 
efficiency of the eye, to ascribe the benefit, 
whatever there may be, to the low intrinsic 
brilliancy of the kerosene flame. For, as has 
already been stated, it may be safely said that 
for the kind of distribution effects we are 
getting from the large majority of our light- 
ing systems, too much light is being used for 
the welfare and comfort of the eye. Added 
to this is the effect of the position of 
the light in the field of vision. The kero- 
sene lamp may be placed at the back or side 
of the person using it, and, if in the field of 
vision, it is usually at or near the level of 
the eye. In the two former cases the effect of 
concealed lighting is given, and in the latter 
case the lamp occupies the most favorable posi- 
tion possible for an exposed source. That is, if 
the source of light is to be in the field of 
vision at all, it should be as nearly as possible 
at the level of the eye. This is because of the 
greater tendency of a light source to produce 
discomfort and loss of efficiency when its 
image falls on the upper and lower halves of 
the retina than when it falls in the horizontal 
meridian. These facts have been clearly 
brought out in our work on the effect of posi- 
tion of the light in the field of vision. 
In addition to studying the conditions that 
give us maximum efficiency, it is important to 
determine the lighting conditions and eye 
factors that cause discomfort. In fact, it 16 
might well be said that our problem in light- 
ing at present is not so much how to see better 
as it is how to see with more comfort and with 
less damage to the general health on account 
of eye-strain. Any comparative study of the 
conditions producing discomfort necessitates 
a method of estimating discomfort. As stated 
earlier in the paper, our method of estimating 
discomfort is entirely distinct and separate 
from our method of studying efficiency. Time 
can not be taken here to go into details of 
either the method or of the results of this 
study. It will be sufficient to say that the 
effect of distribution of light and surface 
brightness, intensity, and quality are also being 
studied in their relation in the comfort as 
well as to the efficiency of the eye. 
In conclusion, the writer wishes to point 
out that no one of the factors he has men- 
tioned can be safely omitted in the search 
for the most favorable conditions of lighting. 
Nor can one be investigated and a correla- 
tion between it and the others be taken for 
granted. We have been content, heretofore, 
to base our conclusions with regard to the 
relation of a lighting system to seeing on the 
conventional visual acuity test. While this 
test may tell us something about the general 
level or scale of efficiency of the fresh eye, it 
can tell us nothing of loss of efficiency, because 
the muscles of the eye, although they may have 
fallen off enormously in efficiency, can under 
the spur of the will be whipped up to their 
normal power long enough to make the judg- 
ment required by the test. Moreover, it tells 
us nothing of the conditions that produce dis- 
comfort. In short, the general level or scale 
of efficiency of the fresh eye, loss of efficiency 
as the result of work, and the tendency to pro- 
duce discomfort constitute three separably 
determinable moments, no one of which should 
be neglected in installing a lighting system. 
C. E. Ferree 
Bryn Mawr College A PRELIMINARY STUDY OF THE DEFICIENCIES OF 
THE METHOD OF FLICKER FOR THE 
PHOTOMETRY OF LIGHTS OF 
DIFFERENT COLOR [Reprinted from The Psychological Review, Vol. XXII, No. 2, March, 1915.] 
A PRELIMINARY STUDY OF THE DEFICIENCIES OF 
THE METHOD OF FLICKER FOR THE 
PHOTOMETRY OF LIGHTS OF DIF- 
FERENT COLOR1 
BY C. E. FERREE AND GERTRUDE RAND 
Bryn Mawr College 
Synopsis 
A satisfactory method of photometry should combine the following features. 
(1) It should enable one to detect small differences in luminosity and to reproduce 
results for a given observer with a small mean variation and for a number of observers 
with a comparatively small mean variation. That is, the method should possess an 
adequate degree of sensitivity. (2) It should be known either to possess of itself logical 
sureness of principle or its results must agree in the average with those of some method 
which can be shown to have this sureness of principle. The method of flicker probably 
satisfies the first of these requirements better than the equality of brightness method. 
It does not, however, possess of itself the needed sureness of principle, nor have its 
results been shown to agree in the average with any method which is accorded sureness 
of principle. Points are enumerated in the paper appended which raise doubt with 
regard to the correctness of the photometric balance obtained by the method of flicker. 
Only one of these is discussed, namely, the influence of the time element in the exposure 
of the eye to the lights to be compared. With regard to this point, it is shown from 
experimental data (1) that the sensations aroused by lights differing in color value rise 
to their maximum brightness at different rates; and (2) that the single exposures used 
in the method of flicker are much shorter than is required for these sensations to rise 
to their full value. The eye, therefore, is very much underexposed to its stimulus by 
the method of flicker. That is, the rate of succession used in the method of flicker is 
too fast for the single impressions to arouse their maximum effect in sensation and too 
slow for the successive impressions to add or summate as much as they would need 
to do to rise to their full value or perhaps even to a higher value than would be given 
by the individual exposures. Only one other possibility for a correct balance remains,- 
equality is attained at some value lower than the full value. This can not be assumed, 
however, without violating well-known laws relating to the factors which influence 
persistence of vision. 
The principal point of discussion, then, is to what degree it should be held that the 
difference in lag between the sensations aroused by the single exposures used in the 
method of flicker is obliterated in a succession of exposures. Broadly considered, three 
positions are possible with regard to the point for the rates of succession that are 
employed in the method of flicker. (1) The difference is not obliterated at all. In 
this case the photometric balance should deviate from the true balance in direct pro- 
1 Paper read by C. E. Ferree at the Philadelphia section of the Illuminating 
Engineering Society, January 16, 1914. 
110 FLICKER PHOTOMETRY 
111 
portion to the difference in lag for the single exposures. (2) The difference is in part 
obliterated, but it is still present to a degree which renders the method untenable for 
precise work. And (3) the difference is entirely obliterated or so nearly so as to be of 
no practical consequence to the validity of the method. The second is approximately 
the position taken in this paper. The following evidence is offered in support of this 
position, (a) At high intensities of light the writers get by the method of flicker a 
deviation from the true photometric balance, as determined by the equality of bright- 
ness method, in a direction which corresponds to the difference in lag between these 
colors at high intensities as determined both in their own laboratory1 and by Broca 
and Sulzer, (i) At low intensities they get a difference in lag for the colors which is in 
the same direction as the deviation obtained by Ives and Luckiesh at low intensities 
(the reverse Purkinje effect), (c) A change in the relative lengths of exposure to the 
two lights in the method of flicker produces a deviation from the equality of brightness 
balance which again corresponds in direction to the changes that are produced in the 
sensations aroused by the single exposures when similar changes are made in the 
relative lengths of exposure. And (d) determinations made at several intensities of 
light by the method of flicker show a deviation from the equality of brightness balance 
which is many times the smallest difference in brightness that can be detected by the 
method. Moreover, in their own results the writers find that these deviations in every 
case correspond to the difference in lag given by lights of the same order of magnitude 
of intensity, so far as can be judged from the determinations of lag that have been made 
up to this time. When, however, determinations have been made on a larger number 
of observers, individual differences may be found in the amount and distribution of lag 
just as they have been found in the amount and direction of the deviation of the flicker 
from the equality of brightness balance. Later in the interests of a fairer comparison 
the writers hope to make in every case compared the determination of lag and the photo- 
metric determinations on the same observer. 
The writers have preferred to call the work of which this 
paper is a brief report a preliminary study for the following 
reasons. (1) Only one of the points directly pertaining to 
the method of flicker that should be investigated has been 
taken account of in the work. And (2) to complete the chain 
of evidence needed for this point, a more especially directed 
and perhaps more careful determination should be made 
than has yet been done of the time required for visual 
sensations colored and colorless to rise to their maximum of 
intensity. Such a study with especial reference to the needs 
of photometry is now in progress in our laboratory, but is as 
yet unfinished.2 
1 See this paper, footnote I, pp. 125-130. 
2 In the work now in progress in our laboratory, attention will be paid to the 
following points. In case of colors, care will be taken to use lights of a small range of 
wave-length. The intensities of the lights used will be specified photometrically and 
radiometrically. The white light will in addition be specified either spectro-photo- 
metrically or spectro-radiometrically. For the sake of the comparisons needed in 112 
C. E. FERREE AND GERTRUDE RAND 
A satisfactory method of photometry should combine the 
following features, (i) It should enable us to detect small 
differences in luminosity and to reproduce our results for a 
given observer with a small mean variation and for a number 
of observers with a comparatively small mean variation. 
That is, the method should possess an adequate degree of 
sensitivity. (2) It should be known either to possess of itself 
logical sureness of principle, or its results must in the average 
agree with those of some method which can be shown to 
possess this sureness of principle. Methods having these 
features have been developed for the photometry of colorless 
light. The problem of the photometry of colored light, 
however, has presented great difficulty. 
Methods of Photometering Colored Light. 
The methods for photometering colored light may be 
grouped under two headings: the direct methods and the in- 
direct methods. In the former class we have the method of 
direct comparison or, as it is sometimes called, the equality of 
brightness method. Of the latter class the method of flicker 
has received the greatest amount of attention and has been the 
most favored. It will be the purpose of this paper (1) briefly 
to compare the relative advantages and disadvantages of 
the method of flicker and the equality of brightness method 
with regard to sensitivity; (2) to show that the method of 
flicker, so far as it has been developed up to the present time, 
does not seem to possess of itself the sureness of principle 
needed to meet the requirements of a satisfactory method; 
and (3) to show that as yet its results have not been found 
satisfactorily to agree in the average with those of any method 
which can be shown to have this sureness of principle. In a 
the photometric work, all determinations for lights differing in composition will be 
made at the different intensities employed with stimuli equalized photometrically. 
Comparative results will be obtained for the same observers for the best of the methods 
already in use, and three new methods will be introduced. In part, results will be 
obtained for observers who have also been employed in the work on the method of 
flicker. The work will be done for different intensities of light, and both under 
dark and light room conditions. In a survey of the work done up to the present 
time, one can not help but note that too little care has been taken to observe even 
some of the most essential of the above conditions. FLICKER PHOTOMETRY 
113 
later paper a new method of photometry will be described which 
possesses approximately as high a degree of sensitivity for color 
work as the method of flicker and gives results which agree 
much more closely in the average with those obtained by the 
equality of brightness method. The second of the above points 
will be shown as follows. It will be pointed out that at the rate 
of speed at which the impressions are given in the method of 
flicker, the eye is very much underexposed to its stimulus. It 
can reasonably be assumed that this underexposure causes a 
reduction of the intensity of the sensation, and should lead, 
therefore, to a false estimation of the brightness of the colors. 
In fact, at the rate of rotation of the exposure apparatus 
required for lights of the order of intensity employed in 
practical work, this reduction produces for the observers we 
have used an effect similar to the Purkinje phenomenon.1 
At least a deviation from the equality of brightness values 
is found in our results for such intensities which accords well 
with the Purkinje phenomenon. That is, reds and yellows 
are underestimated in brightness, and blues and greens are 
overestimated. And (Z?) it will be shown that flicker itself, 
the phenomenon on which the equalization at the photo- 
metric screen is based, is subject to variations depending upon 
a number of factors the effect of which has not in all cases been 
adequately studied and in some cases not even recognized. 
An investigation of one of these alone, the effect of varying 
1 We do not mean to draw too close an analogy here between the effect on the 
brightness of sensation produced by keeping the intensity of light constant and reduc- 
ing the time of exposure of the eye to the light, and the effect produced by keeping the 
time of exposure of the eye constant and reducing the intensity of the light employed 
(the Purkinje phenomenon). In attempting to interpret the effect produced by the 
short exposures used in the method of flicker, our data should be taken primarily 
from the results showing the relative rise of sensation to its maximum for white light 
and lights of the different colors. (See discussion of the development time of sensa- 
tion, pp. 118-130). It is quite possible and in fact quite probable from Broca and 
Sulzer's results, for example, that for a part of the upward course blue and green 
rise faster than red, and conversely for a part of the course red rises faster than blue 
and green. (Yellow was not used by Broca and Sulzer.) The results of Broca and 
Sulzer are cited on this point, not by any means because their method of making 
the determination is the freest from criticism of any that have yet been used, but 
because they alone have attempted to plot the comparative curves for the different 
colors and white light at different points from the threshold to the maximum. 114 
C. E. FERREE AND GERTRUDE RAND 
the ratio of the time of exposure of the eye to the lights to be 
compared, is enough to lead one seriously to question whether 
the method of flicker can be safely used in the work of hetero- 
chromatic photometry, at least not without calibration, and 
perhaps not without an amount of calibration which is in 
itself prohibitive of the use of the method in practical work.. 
The third point will be covered in the following way. (i) It 
will be pointed out that the only method that has thus far 
been used as a standard with which to compare the method 
of flicker has been the equality of brightness method. The 
selection of this method as a standard has been recommended 
among others by Whitman, Wilde, and Schenck, and a com- 
parison of the results of the two methods, more or less com- 
plete, has been made by a number of experimenters. And (2) 
it will be shown both from our own work and from a very great 
preponderance of the work done by others who have made the 
comparison, that the results by the method of flicker do not 
agree in the average with those obtained by the equality of 
brightness method; and, therefore, that justification for the 
adoption of the method of flicker can not yet, at least, be 
fairly claimed through its agreement in result with the 
equality of brightness method. 
The Equality of Brightness Method.-With regard to sensi- 
tivity in the photometry of lights of different color, the 
equality of brightness method has the following disadvan- 
tages. (1) Small differences in luminosity can not be de- 
tected because the actual difference present is '.masked by 
the difference in color quality. (2) Results for a given 
observer can not be reproduced within a small limit of 
variation, because the ability to do this in turn presupposes 
the ability to detect small differences which, as has just been 
stated, can not be done. (3) Results can not be reproduced 
from observer to observer within a small limit of variation 
because (a) the sensitivity to color varies more among ob- 
servers than does, for example, the sensitivity to brightness, 
hence there is a variable amount of the disturbing factor of 
color present for different observers; and (IT) because the 
standard or pattern for the judgment of equality differs FLICKER PHOTOMETRY 
115 
more from individual to individual when the factor of color is 
present than when it is not. That is, in any photometric 
judgment the observer must decide for himself what he will 
call equality and make all his judgments conform to this 
pattern or standard. When color is present to interfere with 
the judgment of equality, the selection of this standard varies 
more for different observers than it does when no color is 
present. With regard to all the points on which sensitivity 
depends, therefore, the equality of brightness method may 
be said to possess a low degree of sensitivity. 
The Method of Flicker.-The method of flicker possesses 
greater sensitivity than the equality of brightness method. 
That is, smaller differences in the luminosity of the photo- 
metric surfaces can be detected, and the judgment of equality 
is surer and more reproducible.1 This is because the disturb- 
ing factor of color difference in the impressions to be compared 
is eliminated from the judgment. That is, instead of being 
given simultaneously, the stimuli are given in succession and 
at such a rate that all color differences between them disap- 
pear, and the brightness impressions are permitted to develop 
in sensation unobscured by differences in color quality. 
The use of the phenomenon of flicker to detect a difference 
in brightness between two illuminated surfaces can best be 
understood possibly by considering the phenomena that take 
place when successive impressions of colored and colorless 
light are made upon the retina at different rates of speed. 
When the retina is exposed successively to colorless lights 
differing in brightness, the following phenomena take place. 
When the rate of succession is low, the impressions remain 
1 This higher degree of reproducibility can be claimed perhaps only for the judg- 
ments given by a single observer. It does not seem to obtain to any considerable 
extent, so far as results are available for comparison, when results are compared from 
observer to observer. For example, in a group of eighteen observers Ives gets differences 
as great as 159 per cent, for .481 /*, 114 per cent, for .498 26 per cent, for .518 
18 per cent, for .537 13 per cent, for .556^; 10 per cent, for .576 n; 28 per cent, for 
•595 65 per cent, for .615 g; 86 per cent, for .635 and 122 per cent, for .655 
The percentage of average variation from the mean for these observers is 17 per cent 
for .481 g; 13.4 per cent, for .498 m; 6 per cent, for .518 y; 3 per cent, for .537 n; 2.75 
per cent, for .556 n; 2.2 per cent, for .576^; 5.4 per cent, for .595 /z; 9.5 per cent, for 
.615 13.2 per cent, for .635 and 19.3 per cent, for .655^. {Philos. Mag., 1912, 
24, Ser. 6, pp. 853-863.) 116 
C. E. FERREE AND GERTRUDE RAND 
more or less separate and distinct. At rates higher than this,, 
we have in order Fechner's colors,1 flicker, and the fusion of 
the two impressions into a uniform gray. When the eye is 
exposed successively to colored and colorless light, the follow- 
ing phenomena take place. At low rates, we have again the 
more or less separate successions of the two impressions. At 
rates slightly higher than this, we have first a phenomenon 
that may be called by analogy color flicker, and then an inter- 
mingling of color and brightness flicker. At still higher rates 
we have color fusion, brightness flicker, and complete color 
and brightness fusion. Thus, both in case of colored and 
colorless light, brightness flicker seems to be a phenomenon 
due solely to the succession at certain rates of speed of im- 
pressions differing in luminosity or brightness. Moreover, 
the phenomenon is very sensitive to changes in the luminosity 
of the successive impressions. That is, a very slight change 
in one of the impressions will produce flicker when there is no 
flicker, or will cause a noticeable change in its amount when 
there is flicker. Flicker thus becomes a very sensitive means 
of detecting brightness difference. This sensitivity, however, 
is not so great in case of colored as it is in case of colorless 
light. It would in fact in all probability be very low were it 
not for the fortunate fact that color fusion takes place at a 
very much lower rate of succession than brightness fusion. 
Concerning the ease and sureness of making the judgment, 
then, the case with regard to the method of flicker may be 
summed up as follows: By giving the impressions to be 
compared to the retina successively at a certain rate of speed, 
the disturbing element of color difference, which so interferes 
with the detection of brightness difference when the im- 
pressions are given simultaneously, is eliminated, and the 
phenomenon of brightness flicker stands out clearly in a field 
uniform as to color quality. That is, by using a method of 
successive impressions we have succeeded in eliminating the 
1 Fechner's colors are best observed when the successions are made by rotating 
discs made up of white and black sectors, or by discs specially constructed for the 
purpose. This phenomenon occurs at a rate of speed near the upper limit required 
to give separate impressions, and consists of impressions of color mingled with the 
more or less separate impressions given by the white and black sectors. FLICKER PHOTOMETRY 
117 
feature that renders the comparison of the brightness of the 
simultaneous impressions so difficult to make, namely, the 
difference in color quality between the impressions to be 
compared. The judgment, then, is easy, and the principle 
on which the equalization is based seems to be clear. The 
method has come to have many supporters, but other things 
besides the sureness of judgment must be taken into account. 
This brings us to a consideration of our second point, namely, 
the method of flicker when applied to the photometry of 
lights of different color does not seem to possess the sureness 
of principle needed to meet the requirements of a satisfactory 
method. We have two reasons for making this assertion. 
In the first place, as we have already stated, at the rate of 
speed at which the impressions are given in the method of 
flicker, the eye is very much underexposed to its stimulus. 
And in the second place, flicker, the phenomenon on which 
the equalization is based at the photometric surface, is subject 
to many variations depending upon a number of factors the 
bearing of which on the application of the phenomenon to 
photometry, has not in all cases been adequately studied, and 
in some cases not even recognized. A few of these may be 
suggested in passing, (i) The intensity of illumination and 
the influence it exerts on the speed of alternation that has to be 
used in order to give the method maximum sensitivity. 
(2) The different rates of speed required for the fusion of the 
different colors, and the varying lower limit this difference 
puts upon the rates of speed that can be used. (3) The effect 
of the saturation of the colors used on the fusion rate. (4) 
The effect of field size. (5) The effect of the ratio of the time 
of exposure of the eye to the lights to be compared; etc. A 
better knowledge than we now have of the effect of these 
factors is, the writers believe, of fundamental importance in 
the employment of the phenomenon of flicker in the photom- 
etry of lights of different colors. At a later date they hope 
to report the results of a systematic study of these factors. 
In the present paper, the effect of only one of them will be 
considered, namely, the ratio of the time of exposure of the 
eye to the lights to be compared. An investigation of this 118 
C. E. FERREE AND GERTRUDE RAND 
point alone is enough to lead one seriously to question whether 
the method of flicker can safely be used in the work of hetero- 
chromatic photometry, at least not without calibration, and 
perhaps not without an amount of calibration which is in 
itself prohibitive of the use of the method in practical work. 
The Action of Light on the Eye under the Conditions 
Imposed by the Method of Flicker. 
Both of the above points will probably be more easily 
understood if a brief consideration is given to the way the 
eye responds to colored and colorless lights when the im- 
pressions are given to it in the manner they are given in 
the method of flicker. The eye is not an ideal sense organ, 
that is, it does not respond at once with its full intensity of 
sensation at the beginning of stimulation, nor does the sensa- 
tion cease with the cessation of stimulation. It takes, for 
example, an interval of time for the sensation proper to a 
given stimulus to rise to its maximum; and also an interval 
to die away after the stimulation has ceased, depending for 
its length upon several factors.1 
The interval of time required for a sensation to rise to its 
maximum will be called in this paper the development time of 
sensation. Plateau in 18342 first expressed the belief that 
1 There are two phases to this after-effect, positive and negative. The positive 
alone concerns us here. In this phase, which is often called the persistence of vision, 
the original sensation tends to persist in its original color and brightness. More 
accurately described, however, it rapidly loses in color and rapidly darkens. In the 
negative phase there is a brightness reversal, that is, what is light in the original 
sensation becomes dark, and the color changes to the complementary color. The 
negative phase is much longer than the positive. The length of the positive depends 
upon many factors: the intensity of the stimulus, the time of exposure of the eye 
to the stimulus, the state of adaptation of the eye, the general illumination of 
the field of vision, the brightness of the local preexposure and post-exposure, etc. 
Unless the eye is put under very especial conditions of stimulation, the duration of the 
positive phase is very short indeed, in fact, momentary. For a further discussion of 
this point with reference to the method of flicker, see appendix. 
2 As early as the time of Bacon it was noted that there is a period of inertia in 
vision. ("At in visu (cujus actio est pernicissima) liquet etiam requiri ad eum 
actuandum momenta certa temporis: idque probatur ex iis, quae propter motus 
velocitatem non cernunter; ut ex latione pilae ex sclopeto. Velocior enim est praeter- 
volatio pilae, quam impressio speciei ejus quae deferri poterat ad visum."-'Novum 
Organum,' lib. IL, Aph. XLVI.). Later Beudant ('Essai d'un Cours Elementaire FLICKER PHOTOMETRY 
119 
color sensation does not come at once to its maximum. He, 
and later Fick1 in 1863, showed that when a sector of white 
paper passes very rapidly only once before the eye it looks 
to be a dark gray. With the experiments of Exner in 1868, 
the work of determining the development time of visual 
sensation was definitely begun. Different methods of making 
the determination have been used by different investigators, 
and different results have been obtained. There is, however, 
among the different results a certain amount of agreement. 
At least the order of magnitude of the development time can 
be fixed within certain limits. The chief points of interest 
in these investigations have been (1) to compare the develop- 
ment time of the different sensations of color with each other 
and with that of colorless sensation; and (2) to determine 
whether the intensity of the stimulus has any effect upon 
the development time. All who have made the comparison 
have found that each of the color sensations has a development 
time different from the colorless sensations; all with the excep- 
tion of Durr and Berliner, that each of the colors has a different 
development time; and all with the exception of Durr, that an 
increase of intensity shortens to some extent the development 
time of all sensations. 
A table (Table I.) ha,s been prepared showing the develop- 
ment time obtained by each of these men for the different 
et General des Sciences Physique: Partie Physique,' p. 489, 3me edition) also stated 
that an object which moves with extreme rapidity before the eye is not perceived 
because impressions are not made on the eye instantly. Plateau (' Nouveaux Memoirs 
de l'Academie Royale des Sciences et Belles Lettres de Bruxelles,' 1834, 8, p. 53) 
made the observation that when a bit of white paper passes very rapidly before the 
eye, it appears not white but gray. He was the first to express the belief that color 
sensation also does not come at once to its maximum of intensity. Swan {Trans. 
Roy. Soc. Edinb., 1849, 16, pp. 581-603) observed that the "light of the sky seen 
immediately over a ball in its descent through the air, seemed less bright than at those 
parts of the retina where the action of light had not been interrupted by the passage 
of the dark body"; and conducted some experiments to determine the intensity of 
light sensation with short exposures. Exposing the eye to lights of different intensities 
for intervals ranging from 1/100 to 1/16 of a second, colorless 'lights of different 
intensity produce like portions of their total effect on the eye in equal times.' While 
he does not directly determine the interval required for the light sensation to come to 
its maximum, he estimates it from the results of his experiments with short exposures 
to be about 1/10 of a second. 
iFick, A. Archiv fur Anatomie und Physiologie, 1863, p. 739. 120 
C. E. FERREE AND GERTRUDE RAND 
Table I 
Showing a Comparison of the Development Time of Visual Sensation with 
the Average Time of Exposure of the Eye to Its Stimulus Used in our 
Experiments with the Method of Flicker. 
Sensation 
Development Time 
Average Time of 
Exposure of Color by 
Method of Flicker 
Exner1 
1868 
White 
5 intensities: 
.118-2.87 sec. 
.0178 sec., when 
the value of the 
colored sector was 
1800. 
Kunkel2 
1874 
Different colors 
•057- 133 
Charpentier3.... 
1887 
White, 
5 intensities 
.014-.049 
Lough4 
1896 
White, 
5 intensities 
.090-. 148 
Durr5 
1902 
White, 
Different colors 
2 intensities: 
.266 
•54i 
Martins6 
1902 
White 
Different colors 
6 intensities: 
.013-093 
.020-.090 
.0213 sec., when 
the value of the 
colored sector ran- 
ged from 45°-3i5°. 
Broca and Sulzer7 
1903 
White 
Different colors 
8 intensities: 
.031-125 
•07 -125 
McDougall.8... . 
1904 
White 
Different colors 
12 intensities: 
.049-2 
.100-. 108 
Buchner9 
1906 
White 
3 intensities: 
.033-230 
Berliner10 
1907 
Different colors 
• 13° 
1 Exner, S., 'Ueber die zu einer Gesichtswahrnehmung nothige Zeit.,' Sitzungs- 
berichte der Kaiserlichen Akademie der Wissenschaften, Math.-Phys. Classe, 1868, 
58, pp. 601-632. 
2 Kunkel, A., 'Ueber die Abhangigkeit der Farbenempfindung von der Zeit,' 
Pfluger's Archiv, 1874, 9, p. 197. 
3 Charpentier, ' Sur la periode d'addition des impressions illuminismes,' Comptes 
Rendus Societe de Biologic, 1887, 4, pp. 192-194. 
4 Lough, 'The Relations of Intensity to Duration of Stimulation in Our Sensations 
of Light,' Psychological Review, 1896, 3, pp. 484-492. 
6 Durr, E., 'Ueber das Ansteigen,der Netzhauterregung,' Philo sophische Studien, 
1901-1903, 18, pp. 215-273. 
6 Martius, G., 'Ueber die Dauer der Lichtempfindungen,' Beitrdge zur Psychologic 
und Philosophic, Leipzig, 1902, I, Heft 3. 
7 Broca, A., and Sulzer, D., Comptes Rendus der Seances de 8 Academic des Sciences, 
1902, 134, pp. 831-834; 1903, 137, pp. 944-946; 977-979; and 1046-1049. 
8 McDougall, W., 'The Variation of the Intensity of Visual Sensation with the 
Duration of the Stimulus,' British Journal of Psychology, 1904-1905, 1, pp. 151-189. 
9 Buchner, M., 'Ueber das Ansteigen der Helligkeitserregung,' Psychologische 
Studien, 1906-1907, 2, pp. 1-29. FLICKER PHOTOMETRY 
121 
colored and colorless sensations; and, for comparative pur- 
poses, the average exposure time that was used in our experi- 
ments for all the colors in the determination of their bright- 
ness by the method of flicker. In choosing this time of ex- 
posure for the method of flicker, in order to secure for the 
method the greatest possible sensitivity, we used the slowest 
rate of succession of colored and colorless sectors that could 
be employed. 
An inspection of this table will show that while the results 
for the development time of sensation differ quite a little 
among themselves, they agree in one very important partic- 
ular, namely, they are all much greater than are the intervals 
that are used in the longest exposures that are permissible 
by the method of flicker. That is, by the method of flicker, 
the eye is very much underexposed to its stimulus. The 
effect of this under exposure is obviously to cause a reduction 
in the intensity of sensation. That is, the rate of succession 
of impressions used in the method of flicker is too fast for the 
single impressions to arouse their maximum effect in sensa- 
tion and too slow for the successive impressions to add or 
summate as much as they would need to do to cause the in- 
tensity of the sensation aroused by each light to rise to its 
full value, or perhaps even to rise to a higher value than 
would be given by the individual exposures. In fact as will 
be shown in an appendix to this paper the sensation can 
not be expected to rise to its full value through summa- 
tion if the Talbot-Plateau law be true, however rapid is 
the rate of succession of the individual impressions (see ap- 
pendix). Even when a rate is reached at which complete 
fusion takes place, both for the color and brightness com- 
ponents in sensation, there is according to the Talbot- 
10 Berliner, 'Der Ansteig der reinen Farbenerregung im Sehorgan,' Psychologische 
Siudien, 1907, 3, pp. 91-155. 
W. Swan in an article entitled 'On the Gradual Production of Luminous Impres- 
sions on the Eye; Part II., being a description of an instrument for producing isolated 
luminous impressions on the eye of extremely short duration, and for measuring their 
intensity,' Trans. Roy. Soc. Edinb., 1861, 2, pp. 33-40, has described a very ingenious 
but complicated apparatus for getting short periods of stimulation of the retina, but 
apparently neither he nor any one else has ever used the apparatus described. 122 
C. E. FERREE AND GERTRUDE RAND 
Plateau law, a reduction in the intensity of each sensation 
which is the same as would be gotten were the intensity of 
each light to be reduced in proportion to the time of ex- 
posure of that light to the total time of exposure of both 
lights, and no further increase in the rate of succession pro- 
duces any change in the effect.1 The possibility then of the 
sensations which, as is shown by the work on development 
times, are unequal for the single exposures used in the 
method of flicker, reaching equality by rising to their full 
value seems to be ruled out. In terms of the Talbot-Plateau 
law they could not reach their full intensity through an effect 
of summation, however fast the rate of succession be made, 
1 Ewald ("Versuche zur Analyse der Licht- und Farbenreaktionen eines Wirbel- 
losen " (Daphnia pulex), Ztschr.f. Psychol, u. Physiol, d. Sinnes., 1914, 48, pp. 285-325; 
and "The Applicability of the Photochemical Energy Law to Light Reactions in 
Animals," Science, 1913, 38, pp. 236-238) has made an interesting contribution with 
regard to the effect of the intermittent action of light on the eye which it may not be 
out of place to mention here. The facetted eye of the daphnia was used in his experi- 
ments. When exposed to light this eye responds by turning towards the light, and 
when lights of different intensities are used it turns towards the stronger light. After 
having determined the sensitivity of this response to difference in intensity of light by 
exposing the eye to a number of lights of different intensities acting continuously on 
the eye, he undertook to make a comparison of the effect of light acting continuously 
and intermittently. The intermittence was gotten by rotating a sectored disc in front 
of one of the lights. The lights were so chosen that the same amount of energy acted 
upon the eye in a given unit of time from both the continuous and intermittent sources. 
That is if a ratio of total open to closed sector of the value 1/10 was used, the light in 
front of which these sectors were rotated was made ten times as intense as the light 
acting continuously. The sectored disc was then rotated at different speeds. When 
a speed of 30 revolutions per second was attained the eye remained stationary. That 
is at this speed of rotation the two lights produced equal effects on the eye,-which is, 
of course, no more than a demonstration of the Talbot-Plateau law for the primitive 
eye. But when the speed was made slower than this, the eye invariably turned towards 
the light which was acting continuously. That is when the rate at which the im- 
pressions were given to the eye was made slower the result was to weaken the effect 
on the eye even though the same amount of light was received by the eye in a unit of 
time in both cases. Ewald's results show then that, so far as the primitive eye is 
concerned, when light impressions are given to the eye at certain high rates of succession 
(analogous to the fusion rates for the human eye) there is a reduction in the amount of 
response aroused which is the same as would be produced were the intensity of the 
light reduced by an amount proportional to the ratio of the time of exposure to the 
light to the total time of the observation; and when they are given to the eye at rates 
slower than these the effect on the eye is the same as if the light acting on it had been 
still further reduced in intensity. FLICKER PHOTOMETRY 
123 
let alone attain it at the rates which are employed in the 
method of flicker.1 
1 There seems to be only one other possibility that the method of flicker should 
give the true photometric balance between lights of different color values, namely, that 
the sensations aroused should reach equality at some value lower than the full value. 
That this is extremely improbable is shown by the following consideration. The 
weaker sensation or the sensation which has the slower rate of development for a single 
exposure would have to rise in value because of summation effect resulting from the 
succession of exposures until it became equal to the stronger sensation. To produce 
any effect of summation each individual impression would have to last over in sensation 
until the next impression of its kind is received which, since the impressions alternate, 
would be the next impression but one. And to produce the particular effect required 
here, not only would each excitation have to last over until the next one is aroused, but 
the weaker one would have to last over more strongly than the stronger one, else the 
effect of the summation would not be to produce the gain of the weaker on the stronger 
which is required to bring the two to the true photometric balance. That is, the 
advocate of this point of view would say that even though for the single exposure one 
color is weighted more than the other, the effect of this is obliterated in a succession of 
impressions and the two rise to equal value, because the weaker sensation would carry 
over more strongly hence would gain more relatively in the process of summating than 
would the stronger sensation. This is not at all in accord with the experimental evi- 
dence available at this time on the relation of the positive after-effect or persistence of 
sensation to the original sensation. Goldschmidt ("Quantitive Untersuchungen uber 
positive Nachbilder," Psych. Studien, 1910, 6, pp. 159-252) and others show, for ex- 
ample, that the stronger the original sensation, the more strongly does it tend to 
carry over after the light is cut off. Goldschmidt also concludes from his experiments, 
which is a very important point for this discussion, that the tendency of the sensation 
to carry over is, so far as its brightness is concerned, independent of the color. That 
is, suppose that a photometric balance was obtained for green and red lights of com- 
paratively high intensities by the method of flicker. Then according to Broca and 
Sulzer's curves, also the results obtained in this laboratory, green would attain to a 
higher brightness value for the single exposure than would be attained by red. Hence if 
green is not to be overestimated by the method of flicker, red must carry over more 
strongly as the impressions succeed each other than does green, and thus make up by 
a summation effect the deficiency shown in the single exposure. But according to 
Goldschmidt's results this greater tendency to carry over could not be assumed for 
red, either because of its color value or because of its weaker intensity, and there is 
no other aspect of the sensation which could have any bearing on the question in hand. 
Moreover, this hypothesis is rendered still more untenable by the experimental fact 
that the situation at low intensities is reversed. That is, at low intensities red, as 
shown by the curves for difference in lag (see Fig. 2, p. 127), attains to a higher 
value than green for the single exposure. Then if red is not to be overestimated 
by the method of flicker and in direct proportion to the values given to the two sensa- 
tions in the single exposure, green must be carried over more strongly in the succession 
of impressions than is red. The explanation of both of these points would require 
not only that the color value of the stimulus exerts an influence on the carrying over 
of the brightness aspect of the sensation, but that this influence reverses in passing 
from high to low intensities. For a discussion of how highly improbable it is for the 
rates of succession used in the method of flicker that one impression could last over 
until the next impression but one is received in any amount that could be of con- 
siderable consequence to the method, see appendix. 124 
C. E. FERREE AND GERTRUDE RAND 
It seems fair to conclude, then, that instead of getting by 
the method of flicker the sensations that should be aroused by 
the lights with which we are working, we get sensations of lower 
intensity. But it may be asked what if there is a reduction of 
the intensity of the impressions received? Equalization is all 
we are working for and the intensity of both impressions is 
reduced. Is it not possible, therefore, to find a ratio of time of 
exposure to each light such that the amount of reduction in 
the intensity of both impressions will be equal ? This would be 
comparatively simple if the rate of development for all the 
colored were the same as for all the colorless sensations. The 
intervals of exposure could be made equal as is ordinarily done 
when sectored discs are used and as apparently must be done 
when the exposures are given by means of a rotating prism. 
But the development time for color sensation is not the same 
as for colorless sensation, and, moreover, the consensus of evi- 
dence is that the rate of development is not the same for any 
two of the color sensations. Thus from the standpoint of the 
unequal reduction in intensities produced by the method of 
flicker, the task of selecting a proper ratio of exposure time of 
colored to colorless light, in case of the different colors, is one 
that requires a great deal of accurate knowledge if the method 
is to have the sureness of principle needed,-more, the writers 
think, than we now possess.1 
1 One scarcely needs point out in this regard that there is apparently no point 
in the intensity scale for which a given reduction in intensity for colored light gives 
the same change in luminosity in sensation that it does for white light. Beginning 
with the spectrum of fully saturated colors and comparing the effect of reduction by 
equal amounts of colored and white lights equal in photometric value, the blues and 
greens are found not to decrease in luminosity so fast as the white light, and the reds 
and yellows are found to decrease faster. Or as the phenomenon is ordinarily ex- 
pressed, there is a relative lightening of the blues and greens and a relative darkening 
of the reds and yellows. Nor is the phenomenon of unequal change confined to the 
lower intensities. It is more striking for these intensities, but it occurs also for the 
higher intensities. This conclusion is drawn from the statement made by several 
writers that beginning with the spectrum of fully saturated colors and increasing the 
intensity of light, all the colors are found to tend towards white, and in so doing to 
change their luminosities at different rates. (For example, see Helmholtz, H., 'Ueber 
Hrn. D. Brewster's neue Analyse des Sonnenlichts,' Pogg. Ann., 1852, 86, p. 520; also 
Handbuch der physiologischen Optik, zw. Aufl., 1896, pp. 465-466; Chodin, A., 'Ueber 
die Abhangigkeit der Farrbenempfindungen von der Lichtstarke,' Sammlung physio- 
logischer Abhandlungen von Preyer, 1877, 1, p. 33 ff.; Brucke, E., 'Ueber einige FLICKER PHOTOMETRY 
125 
We have discussed here, moreover, the effect of underex- 
posure at only one rate of rotation of the exposure apparatus. 
The situation becomes still more complicated when this rate 
is changed. If it were changed, as it must be to preserve the 
sensitivity of the method in passing from high to low illu- 
mination, the whole scale of magnitude of the underexposure 
would change, and a shift in the relative evaluation of the 
luminosities of the different colors might very well be ex- 
pected from the shape of the sensation curves as they rise to 
their maximum. In fact this shift is found in the work of 
previous investigators1 who have made the comparison at 
Empfindungen im Gebiete der Sehnerven,' Sitzungsber. der Wiener Akademie, Math.- 
Natur. Klasse, 1878, 77, Abth. 3, p. 63.) As we have already stated, however (p. 113), 
we do not mean to draw too close an analogy here between the effect on the bright- 
ness of sensation produced by keeping the intensity of light constant and reducing 
the time of exposure of the eye to the light, and the effect of keeping the time of 
exposure of the eye to the light constant and reducing the intensity. The degree to 
which the analogy holds can scarcely be considered as fixed until more work is done 
showing the way in which the luminosity curves for the different colors rise to their 
maximum as the time of exposure of the eye to the different colored lights is increased. 
1 See, for example, the phenomenon called by Ives the "reverse Purkinje" effect 
{Philos. Mag., 1912, 24, Ser. 6, pp. 170-173); later demonstrated and discussed by 
Luckiesch {Electrical World, March 22, 1913, p. 620). These writers have found that 
the red end of the spectrum shows a relatively higher luminosity value as compared 
with the green end by the method of flicker at low than at high illuminations. From 
the shape of Broca and Sulzer's curves for the rise of visual sensation, to its maximum, 
for example, this result might very well be due to the difference in the relative lumin- 
osity value of the colors caused by the difference in the length of exposure given to the 
eye in the method of flicker at the faster rates of speed required for the higher illumina- 
tions and at the slower rates for lower illuminations. That is, the longer exposures 
given by the slower speeds of rotation allow the colors to attain a higher intensity. 
For example, the speeds used by Ives for what he calls 10 Illumination Units range, 
for the different colors for five observers, from 7 to 10 cycles per second, and for 250 
Illumination Units from 10 to 22 cycles per second. 
Broca and Sulzer's curves (Fig. 1) are appended here for one order of intensity 
of stimulus {Comptes Rendus, 1903, 137, p. 978. For other intensities see p. 945). 
The curves given were selected because they alone show a comparison between the 
results for colored and white light. It will be seen from these curves that for exposures 
less than .07 sec. (approximate value), blue and green rise to a higher value than red; 
for exposures ranging from .07 to .11 sec., blue rises to a higher value than red, and 
red higher than green; and for exposures ranging from .11 sec. to about .25 sec., red 
rises to a higher value than blue or green. 
There is also a very strong probability that the relative lag in sensation for the 
different wave-lengths is not the same for lights of low intensity as for lights of higher 
intensities. In fact the results that have been obtained so far in this laboratory in 
determining the development time of the sensations aroused by red, yellow, green, and 126 
C. E. FERREE AND GERTRUDE RAND 
intensities low enough to necessitate a decided reduction in 
speed of rotation. And that the shift is different from the 
normal effect on the brightness of the colors produced by a 
decrease of illumination, is shown by the fact that according 
to the results of these investigators it is in a different direction 
from that given by the equality of brightness method. 
Moreover, in the later paper it will be shown that a change in 
this evaluation amounting to several times the smallest dif- 
ference in luminosity that can be detected by the method, is 
also produced by working the reverse variation, that is, by 
keeping the rate of rotation constant and changing the ratio 
of value of colored to colorless sector. It is difficult, there- 
fore, to avoid the conclusion that the type of exposure used in 
the method of flicker is an important factor in the cause of its 
disagreement with the results obtained by other methods. 
blue lights of spectrum purity show that at low intensities red and yellow rise more 
rapidly in photometric value than green and blue. This result is quite marked in those 
parts of the curves representing an exposure time of the same order of magnitude as 
is used in the method of flicker. In order to show this point we have appended here 
Fig. i 
three curves representing the relative rates of development of red, yellow, green, and 
blue at intensities which we will designate for the present as low and intermediate; and 
of red, yellow, and green for a higher intensity. These determinations were made by FLICKER PHOTOMETRY 
127 
Reexamining the case, then, with regard to the underex- 
posure of the eye by the method of flicker, we find that the 
short exposure times necessary to the method cause a re- 
M. A. Bills of this laboratory. Later, results will be given for red, green, blue, and 
yellow at a number of intensities, and specifications will be made of the intensities 
employed in both photometric and radiometric units. The colored lights used in 
determining the curves given below were obtained from a spectrum of good definition 
and were in each case equal in photometric value, as they should be if results are to be 
used in interpreting the action of light on the eye under the conditions imposed by the 
Fig. 2 
method of flicker when the photometric balance is attained. In constructing these 
curves time of exposure is plotted along the abscissa and brightness of color along the 
ordinate. The curves in Fig. 2 are for the low intensity; in Fig. 3 for the intermediate 
intensity; and in Fig. 4 for the higher intensity. 
It is very probable that there is considerable individual difference in the amount 
and distribution of lag. A rigid test of the correspondence of difference in lag to the 
direction of deviation of results gotten by the method of flicker from those obtained by 
the equality of brightness method would require that the photometric determinations 
and the determination of lag should be made for a given quality and intensity of light 
by the same observer. 
If it should be found that there is an individual difference in the amount and dis- 
tribution of lag, the result would supplement very nicely the explanation why a much 
higher degree of reproducibility is gotten by the method of flicker than by the method 126 
C. E. FERREE AND GERTRUDE RAND 
intensities low enough to necessitate a decided reduction in 
speed of rotation. And that the shift is different from the 
normal effect on the brightness of the colors produced by a 
decrease of illumination, is shown by the fact that according 
to the results of these investigators it is in a different direction 
from that given by the equality of brightness method. 
Moreover, in the later paper it will be shown that a change in 
this evaluation amounting to several times the smallest dif- 
ference in luminosity that can be detected by the method, is 
also produced by working the reverse variation, that is, by 
keeping the rate of rotation constant and changing the ratio 
of value of colored to colorless sector. It is difficult, there- 
fore, to avoid the conclusion that the type of exposure used in 
the method of flicker is an important factor in the cause of its 
disagreement with the results obtained by other methods. 
blue lights of spectrum purity show that at low intensities red and yellow rise more 
rapidly in photometric value than green and blue. This result is quite marked in those 
parts of the curves representing an exposure time of the same order of magnitude as 
is used in the method of flicker. In order to show this point we have appended here 
Fig. i 
three curves representing the relative rates of development of red, yellow, green, and 
blue at intensities which we will designate for the present as low and intermediate; and 
of red, yellow, and green for a higher intensity. These determinations were made by FLICKER PHOTOMETRY 
127 
Reexamining the case, then, with regard to the underex- 
posure of the eye by the method of flicker, we find that the 
short exposure times necessary to the method cause a re- 
M. A. Bills of this laboratory. Later, results will be given for red, green, blue, and 
yellow at a number of intensities, and specifications will be made of the intensities 
employed in both photometric and radiometric units. The colored lights used in 
determining the curves given below were obtained from a spectrum of good definition 
and were in each case equal in photometric value, as they should be if results are to be 
used in interpreting the action of light on the eye under the conditions imposed by the 
Fig. 2 
method of flicker when the photometric balance is attained. In constructing these 
curves time of exposure is plotted along the abscissa and brightness of color along the 
ordinate. The curves in Fig. 2 are for the low intensity; in Fig. 3 for the intermediate 
intensity; and in Fig. 4 for the higher intensity. 
It is very probable that there is considerable individual difference in the amount 
and distribution of lag. A rigid test of the correspondence of difference in lag to the 
direction of deviation of results gotten by the method of flicker from those obtained by 
the equality of brightness method would require that the photometric determinations 
and the determination of lag should be made for a given quality and intensity of light 
by the same observer. 
If it should be found that there is an individual difference in the amount and dis- 
tribution of lag, the result would supplement very nicely the explanation why a much 
higher degree of reproducibility is gotten by the method of flicker than by the method 128 
C. E. FERREE AND GERTRUDE RAND 
duction in the action of the standard and comparison lights 
on the eye. If this reduction were equal in amount, quite 
enough difficulty would be encountered. But it is not 
of equality of brightness only when the results of a single observer are considered (see 
footnote, p. 115). That is since the factor of color difference which so disturbs the 
judgment in the equality of brightness method is eliminated in the method of flicker, 
we should get correspondingly a higher degree of reproducibility for a single observer 
and for different observers, were there not some factor present in the method of flicker 
and not in the equality of brightness method, which varies from individual to individual. 
So supplemented the explanation would be as follows. In the method of flicker the 
judgment for a single observer shows a higher degree of reproducibility than in the 
equality of brightness method because of the elimination of the disturbing factor o 
color difference; but a false balance is established by the method, the deviation from 
the true balance depending in direction and amount for different observers upon the 
difference in the amount and distribution of lag in the rise of the sensations towards the 
maximum. The difference in the amount and direction of this deviation from the true 
balance from observer to observer is the cause of the relatively low degree of reproduci- 
bility of results when the work of different observers is compared. 
Fig. 3 Fig. 4 130 
C. E. FERREE AND GERTRUDE RAND 
equal in amount, and we have no adequate information as to 
the amount of the inequality. Until we have more informa- 
tion with regard to the amount of the reduction and the 
effect it produces, any successful attempt to regulate the 
relative duration of the exposure to colored and colorless light, 
even for a single color at a single intensity, can scarcely be 
more than the result of chance. Surely to expect to accom- 
plish this for all intensities of all colors by a single ratio of 
exposure, more especially by means of a 1:1 ratio, as has been 
the practice in the past, is, it would seem to the writers, to 
ignore the sensation principles which underlie the method. 
Obviously this ratio requires calibration, and to give the 
method the sureness of principle required, it would seem that 
the calibration might have to be made for each color at the 
particular intensity at which the work is to be done. This 
calibration might possibly be accomplished by means of an 
accurate knowledge, if that knowledge could be secured in 
sufficient detail, of the temporal course of visual sensation as 
it rises to its maximum for the given intensity of light used; or 
it might be done by comparing the results obtained by the 
method of flicker with the results obtained by some other 
method adopted as a standard. So far the tendency seems 
to have been to look for this standard within the subject of 
photometry itself. As has already been stated, several 
writers have signified a desire to make the equality of bright- 
ness method a standard, and comparisons have been made of 
the results obtained by the method of flicker and the equality 
of brightness method., 
A consideration of the results of these comparisons together 
with the data collected by one of the writers in ten years use of 
the flicker and equality of brightness methods in making the 
brightness matches needed for the work in color sensitivity, has 
influenced us to open the question anew in the interests of 
the work on color sensitivity. In the work with these two 
methods all done at comparatively high illuminations and 
with a large number of observers, agreement has been rare. 
For this reason the chief incentive to make the present study 
has not been to establish disagreement, but to investigate FLICKER PHOTOMETRY 
131 
further the causes of disagreement. The results of this study 
seem to indicate that the type of exposure of the eye to its 
stimulus by the method of flicker is an important cause of 
disagreement. 
So much for the theoretical considerations relating to the 
method of flicker. To test the accuracy of some of the more 
important points that have come up, a plan of experimentation 
has been formulated and in part carried out. So far as the 
results of that experimentation will be reported upon in this 
paper, the following things will be shown, (i) By comparing 
the time required for the sensations aroused by colored and 
colorless light to reach their maximum of intensity, we have 
already shown that as a general case the eye is very much 
underexposed to its stimulus by the method of flicker, and 
we have concluded that the effect of this underexposure 
on the brightness component of sensation will be unequal 
in amount for colored and colorless light, and should lead, 
therefore, to a false estimation of the brightness of the 
colors. That this conclusion is justified so far as our work is 
concerned, will be demonstrated in part by comparing the 
results of the method of flicker with those obtained by the 
equality of brightness method, in which case the eye is fully 
exposed to its stimulus, and showing that for the method of 
flicker there is for our observers for the intensities of light 
used and for the rate of rotation of the photometer head 
required for these intensities, a characteristic underestimation 
of the brightness of red and yellow and overestimation of the 
brightness of blue and green. That this characteristic devia- 
tion is due to the type of exposure used in the method of 
flicker and not to some other factor will be further shown in 
the consideration of our second point. (2) We have said 
that the ratio of the time of exposure should be considered 
as a factor influencing the results obtained by the method of 
flicker. In order to confirm this judgment of the case, we 
have varied this ratio, keeping the other conditions constant, 
and have found that a corresponding variation is produced 
in the results. That is, by changing the value of the colored 
and colorless sectors in the rotating disc we have used to 132 
C. E. FERREE AND GERTRUDE RAND 
regulate the time of exposure in the method of flicker, corre- 
sponding variations are obtained in the characteristic under- 
estimations of red and yellow and the overestimations of blue 
and green. These variations, it will be shown, moreover,, 
are very much greater than the changes in luminosity that 
are required to be detected by the method of flicker, and are, 
therefore, worthy of being taken into account in an evaluation 
of the usefulness of the method, whatever method be adopted 
as a standard for comparison. And (3) we have contended 
that if the equality of brightness method be adopted as the 
standard for work in color photometry, the method of flicker 
does not satisfy the requirements, for it does not give results 
which agree in the average with those obtained by the equal- 
ity of brightness method. This will be shown both from 
results of our own work and from a preponderance of the 
work done by others who have made the comparison. In 
our own work the comparison has been made for a series 
of intensities which may be considered as at least fairly repre- 
sentative of the higher intensities, they being considered more 
favorable to agreement by Dr. Ives.1 Especial care has been 
taken in this series to duplicate at one point the intensity 
which Dr. Ives finds the most favorable to agreement. 
The remainder of the paper will be taken up with the 
demonstration of these three points. 
I. The Underestimation of the Luminosities of Red 
and Yellow and the Overestimation of the 
Luminosities of Blue and Green 
Special tables have not been prepared for this point 
because the results can readily be seen in the tables for points 
II. and HI. In these tables taken collectively the comparison 
will be shown for a representative series of variations, both 
of the ratio of the time of exposure to colored and colorless 
light and of the intensity of the lights employed. In every 
case underestimation is found to be characteristic for red and 
yellow, and overestimation for blue and green. 
1 Ives, H. E., ' Studies in the Photometry of Lights of Different Colors, Philos. 
Mag., 1912, 24, Ser. 6, pp. 149-188. FLICKER 'PHOTOMETRY 
133 
IL The Variation of the Ratio of the Time of Exposure 
to the Colored and Colorless Light Causes a 
Corresponding Variation in the Characteristic 
Underestimation of Red and Yellow and 
Overestimation of Blue and Green 
As has already been stated, the work under this heading 
has been undertaken in part to show the preceding point, 
and in part to show that the amount of this underestimation 
and overestimation is a variable function of the ratio of the 
time of exposure to the colored and the colorless light. The 
effect of the variation was determined both when the com- 
parison was made between colored and colorless pigment 
surfaces, and between colored and colorless lights. For the 
pigment surfaces the standard red, green, blue, yellow, white, 
and black of the Hering series of papers were used. For the 
colored lights, two sources have been used: the Wratton and 
Wainwright color filters, and the light of the spectrum. 
Since the work with the spectrum as source has not yet been 
finished, results will be given at this point from the work with 
the filters. Of these filters, only the Alpha and Eta were 
used. The former transmits a band of red from the end of 
the spectrum to .65 p., the latter, a band in the blue-green 
from .52 /z to .465 These two alone were used for the 
following reasons: (1) They are fairly representative of the 
colors that show a relative change in luminosity with change 
of intensity. And (2) the yellow, green, and blue filters each 
transmits components that undergo opposite luminosity 
changes with a change of intensity of the source. That is, 
the best yellow of the series transmits also a green com- 
ponent; the best green, a yellow component; and the best 
blue transmits some of the violet. 
The photometric apparatus employed was for the sake of 
comparison made to conform very closely in its essential 
features to that described by Dr. Ives.1 The general plan 
of our apparatus is indicated in Fig. 5. It consists of a photo- 
meter bar carrying the standard white light {A\ a second bar 
carrying the colored light (R), a sectored disc (C), and a 
1 Ives, H. E., op. cit., p. 161. 134 
C. E. FERREE AND GERTRUDE RAND 
screen (D) provided with a small aperture (0) through which 
the light comes to the eye. The standard white light was 
enclosed in a black light-proof box (£), which was provided 
Fig. 5 
in front with a circular opening 4 cm. in diameter for the trans- 
mission of the light. In passing to the sectored disc, the 
light was screened both from the observer's eye and from the 
colored source by black screens properly placed. The light 
which was passed through the colored filters was placed in a 
similar light-proof box (T) provided with an opening 4 cm. 
square for the transmission of the light. Above and below 
this opening were grooves into which the color filters were 
slid. The sectored discs were made of aluminum. The edges 
of these discs were carefully bevelled and the surface was kept 
freshly covered with magnesium oxide deposited from the 
burning metal. The aperture in the screen through which 
the light passed to the observer's eye was 3 mm. square. 
The visual angle subtended by this aperture at the observer's 
eye at 20 cm. distance was very small. A small angle was 
needed to guard against the unequal sensitivity of the central 
and paracentral portions of the retina to flicker, and against 
the difference in their brightness sensitivity to colored and 
colorless light. A 13-candle-power Mazda lamp was used as 
source for the colorless light, and 13-cp., 52-cp., and 130-cp. 
lamps for the colored light. These lamps were operated on a 
no D.C. circuit in series with an ammeter and finely gradu- 
ated rheostat to guard against fluctuations in the current and FLICKER PHOTOMETRY 
135 
loss of efficiency in the lamps. Also fresh lamps were sub- 
stituted at the beginning of each series of observations. As 
a check on the results obtained from these lamps, several 
series of observations were made using a standardized tungsten 
lamp, street series, 16.6-cp. operated at 11.43 volts by a 
storage battery for the colorless light, and a similar lamp of 67- 
cp. operated by a storage battery at 10.35 volts for the 
source of the colored light. 
The method of making the flicker judgment was as follows: 
A preliminary determination was made of the approximate 
setting of the light which was being moved, to give equaliza- 
tion. The speed of rotation of the sectored disc was then 
reduced until flicker was obtained. The position of the light 
was again adjusted until no flicker was obtained, and so on. 
This variation in the speed of rotation of the disc and the 
position of the light was continued until the position was 
ascertained that gave no flicker for the lowest speed of rota- 
tion. The final determination of this point was made by 
moving the light in both directions until noticeable flicker was 
obtained, and taking the average of these two readings. 
The movement required to give flicker on either side of this 
average position ranged usually from 2 to 9 mm. depending to 
some extent upon the observer and the intensity of illumina- 
tion used. Employing the above apparatus and method, 
results were obtained for the highest intensity of colored light 
used for a total open sector of 3150, 2700, 2250, 1800, and 450; 
and for the other intensities, for a total open sector of 315°, 
1800, and 450. In making the comparisons by the equality 
of brightness method, the disc was rotated until one of its 
edges bisected horizontally the photometric field. The 
results are shown in Tables II.-V. They will be sum- 
marized briefly as follows: (1) For all values of open sector 
and for all intensities of light, there was an underestimation 
of the luminosity of the red light and an overestimation of the 
luminosity of the blue-green. (2) As the size of the open 
sector was decreased, there was a corresponding increase in 
the amount of the underestimation of the luminosity of the 
red for all the intensities employed, and of the overestimation 136 
C. E. FERREE AND GERTRUDE RAND 
Table II 
Observer A 
Showing that the Underestimation of Red and the Overestimation of Blue-green is a Characteristic of the Method of Flicker, for Lights of the In- 
tensity Used in this Work, and that the Amount of this Underestimation and Overestimation is a Variable Function 
of the Ratio of the Time of Exposure of the Eye to the Colored and the Colorless Light 
Source of 
White 
Light 
Source of Col- 
ored Light 
Color 
Equality of 
Brightness 
Method 
Flicker Method 
Difference by 
Equality of 
Brightness 
Method and by 
Flicker Method 
with 1800 Open 
Sector 
Change Pro- 
duced by Vary- 
ing Sectors 
Amouut of Change that 
Can be Detected by 
No. of 
Revolutions 
per Second, 
Flicker 
Method 
Distance of 
White Light 
Giving Equality 
of Illumination 
Value of Col- 
ored Sector 
Distance of 
White Light 
Giving no 
Flicker 
Equality of 
Brightness 
Method 
Flicker 
Method 
13 cp. 
13 cp. 151 cm. 
Red 
I25.5 
315° 
133.9 cm. 
- 10.4 cm. 
- 3-4 cm. 
3-5 cm. 
.4 cm. 
3-6 
distant from 
1800 
135-9 
-3 
4-7 
photometric 
Blue- 
45° 
137-3 
•3 
4-i 
screen 
191 
• 315° 
I83-45 
+ IO.55 
+5-4 
4 
•35 
4.2 
green 
1800 
180.45 
•35 
6 
Red 
45° 
178.05 
•25 
5-2 
52 cp. 151 cm. 
93-4 
315° 
102.05 
- II.I 
-4 
3 
•25 
3-7 
distant from 
1800 
104.5 
.2 
5-9 
photometric 
45° 
106.05 
•25 
5-4 
screen 
Blue- 
103 
315° 
94-5 
+ l3 
+5-65 
4 
•3 
4-i 
green 
1800 
90 
•3 
6.1 
130 cp. 89 cm. 
Red 
45° 
88.85 
•35 
5-9 
75-3 
315° 
82.9 
- 9-7 
-4 
3 
•4 
6-3 
distant from 
270° 
84 
•45 
7-4 
photometric 
2250 
84.6 
•4 
9-2 
screen 
1800 
85 
•5 
9.2 
45° 
86.9 
•5 
8.9 
Blue- 
82 
315° 
76 
+ 7 
+3-25 
3-5 
•5 
6.6 
green 
2700 
77 
•3 
7-4 
225° 
76.3 
•3 
8.1 
1800 
75 
•3 
7-7 
45° 
73-75 
•35 
7 FLICKER PHOTOMETRY 
137 
Source of 
White 
Light 
Source of Col- 
ored Light 
Color 
Equality of 
Brightness 
Method 
Flicker Method. 
Difference by 
Equality of 
Brightness 
Method and by 
Flicker Method 
with 1800 Open 
Sector 
Change Pro- 
duced by Vary- 
ing Sectors 
Amount of Change that 
Can Be Detected by 
No. of 
Revolutions 
per Second, 
Flicker 
Method 
Distance of 
White Light 
Giving Equality 
of Illumination 
Value of Col- 
ored Sector 
Distance of 
White Light 
Giving no 
Flicker 
Equality of 
Brightness 
Method 
Flicker 
Method 
16.6 Cp. 
67 cp. stan- 
Red 
86.5 
315° 
96.I 
-13 
-4.2 
2-5 
•5 
6.1 
standard 
dard lamp, 
1800 
99-5 
•3 
8-7 
lamp 
89 cm. dis- 
45° 
IOO.3 
•4 
8.2 
tant from 
Blue- 
8S 
315° 
77-4 
+II-55 
4-5-4 
3 
•3 
5-6 
photometric 
green 
1800 
73-45 
•45 
7-2 
screen 
45° 
72 
•5 
6-9 
13 cp. 
c;2 cp. iqi cm. 
Red 
88 cm. 
315° 
99.18 cm. 
- 17.5 cm. 
- 8.82 cm. 
3-5 cm- 
.8 cm. 
7- 
distant from 
1800 
105.5 
•85 
7-8 
photometric 
45° 
108 
•9 
10.9 
screen 
Blue- 
98 
315° 
96.2 
+ 7-4 
+7-7 
4 
.6 
7a 
green 
1800 
90.6 
.8 
7.2 
45° 
88.5 
-6-5 
•9 
6.15 
16.6 cp. 
67 cp. stan- 
Red 
86 
3iS° 
96.5 
"IS 
3 
.8 
9 
standard 
dard lamp, 
1800 
IOI 
•85 
11.8 
lamp 
89 cm. dis- 
45° 
103 
+ 10.8 
.8 
11.4 
tant from 
Blue- 
84-3 
315° 
78.5 
+8 
4-3 
•9 
5-3 
photometric 
green 
1800 
73-5 
•85 
7-4 
screen 
45° 
7°-5 
.88 
7 
Table IL-Continued 
Observer A 
Table III 
Observer B 138 
C. E. FERREE AND GERTRUDE RAND 
of the blue-green. (3) The amount of change in the photo- 
metric value of the color produced by varying the ratio of 
exposure to colored and colorless light was many times the 
smallest amount of change that can be detected by the method 
of flicker, and, therefore, must be considered of consequence 
in relation to the application of the method to practical work. 
Column 1 of these tables represents the source of white 
light; Column 2, the source of colored light; Column 3, the 
color used; Column 4, the distance of the white light from the 
disc when judgment of equality is given by equality of bright- 
ness method; Column 5, the value of the colored sector for 
the method of flicker; Column 6, the distance from the disc 
at which the white light has to be placed to give the judgment 
of no flicker; Column 7, the difference in the distance the 
white light was placed for the equality of brightness method 
and the flicker method with 1800 of colored sector; Column 8, 
the change in the distance of the white light produced by 
varying the value of the colored sectors in the method of 
flicker; Column 9, the distance the white light has to be moved 
in the equality of brightness method to change the judgment 
from equality to just noticeably lighter or darker; Column 10, 
the distance the white light has to be moved in the flicker 
method to change the judgment from no flicker to just 
noticeable flicker; and Column 11, the number of revolutions 
per second of the sectored disc for the method of flicker. 
Tables IV. and V. represent the results of Tables II. and 
III. expressed in percentage of luminosity at the photometric 
screen. 
Pigment papers are still used in a great many laboratories 
for the investigation of color sensitivity, and because of their 
convenience and ease of manipulation, they probably will be 
used for many years to come for preliminary work and for a 
certain class of investigations in which only comparative 
results are wanted. In estimating the brightness or luminos- 
ity of these pigment colors, the method of flicker is now 
much more extensively used perhaps than any of the other 
methods of making brightness comparisons. For this reason 
we have considered it worth while to extend our work to the FLICKER PHOTOMETRY 
139 
Table IV 
Observer A 
Showing the Results in Table II. Expressed in Percentage of Luminosity 
Source of 
White Light 
Source of 
Colored Light 
Color 
Disagreement 
Between 
Equality of 
Brightness 
Method and 
Flicker Method 
with 1800 
Open Sector 
Change 
Produced by 
Varying 
Sectors 
Amount of Change that 
Can Be Detected by 
Equality of 
Brightness 
Method 
Flicker 
Method 
13 cp. 
13 cp. 151 cm. 
distant from 
Red 
Blue- 
-15 % 
- 4-8% 
5 % 
•4% 
photometric 
screen 
green 
+ 12.3 
+ 6.3 
37 
.6 
52 cp. 151 cm. 
distant from 
Red 
Blue- 
- 20 
- 7-5 
7 
•4 
photometric 
screen 
green 
+ 30 
+13 
7-3 
.6 
130 cp. 89 cm. 
distant from 
Red 
Blue- 
- 21 
- 9 
7 
•9 
photometric 
screen 
green 
+ I9-7 
+ 7-3 
6.8 
7 
16.6 cp. 
standard 
67 cp. standard 
lamp, 89 cm. 
Red 
Blue- 
-23.6 
- 8.3 
87 
1 
lamp 
distant from 
photometric 
screen 
green 
+ 33 
+ 8.3 
8-5 
.8 
Table V 
Observer B 
13 cp. 
«J2 cp. iqi cm. 
Red 
-30% 
-16% 
7-5% 
i-7% 
distant from 
Blue- 
photometric 
screen 
green 
+ 17- 
-J- 18 
7-7 
1.2 
16.6 cp. 
standard 
67 cp. standard 
lamp, 89 cm. 
Red 
Blue- 
- 28. 
- 12.4 
6-5 
1.6 
lamp 
distant from 
photometric 
screen 
green 
+3I-7 
+24. 
9 
2 
investigation of the effect of varying the value of the colored 
and the colorless sectors on the brightness of the pigment colors 
as determined by the method of flicker. Of the devices 
available for applying the method to these colors, the Schenck 
apparatus was selected as best suited to our purpose. As 
colors to be investigated, the red, green, blue, and yellow of 
the Hering series of standard papers were chosen. Sectors 
of the value of 1800, 2700, and 3000 were used. Values lower 140 
C. E. FERREE AND GERTRUDE RAND 
than 1800 were not used because they could not be accurately 
obtained with the type of photometer employed. Two in- 
tensities of illumination were used, one of 390 foot-candles 
(vertical component) received directly under a skylight and 
diffusion sash of ground glass; the other, 5 foot-candles, the 
illumination of a room lighted by windows. Space will be 
given here only for the results for the higher illumination. 
This illumination was carefully chosen far above the range of 
intensities at which the Purkinje phenomemon occurs when 
the eye is fully exposed to its stimulus, in order to subject 
our demonstration to a rigid test. We were seeking, for 
example, to ascertain whether an intensity might not be 
found so high that the underexposure of the eye to its stimulus 
by the method of flicker would not cause an underestimation 
of the brightness of red and yellow and an overestimation of 
the brightness of blue and green.1 That these underestima- 
tions and overestimations occur at this high illumination and 
by amounts many times the smallest brightness difference that 
can be detected by the method, will be shown in Table VI. 
Column 1 of this table shows the color used; Column 2, the 
black-white value of the color estimated by the equality of 
brightness method; Column 3 gives the value of the colored 
sector; Column 4, the white-black value of the color estimated 
by the flicker method; Column 5 gives the difference in the 
result by the equality of brightness and flicker method with 
i8o° colored sector; Column 6 gives the change produced in 
the result by the method of flicker by varying the size of the 
colored sector; Column 7 gives the amount of change that 
can be detected by the equality of brightness method; and 
Column 8, by the flicker method. 
III. The Method of Flicker Does not Give Results 
Which Agree in the Average with those Obtained 
by the Equality of Brightness Method 
Nothing will be added in this section except to make our 
comparisons at the intensity of illumination found to be most 
1 We have been careful to choose high intensities because Dr. Ives has contended 
that at high intensities the disagreement between the methods of flicker and equality 
of brightness tends to disappear. FLICKER PHOTOMETRY 
141 
Table VI 
Observer A 
Showing that the Underestimation of Red and Yellow and the Overestimation of Blue and 
Green is a Characteristic of the Method of Flicker for Light of the Intensity Used 
in this Work, and that the Amount of this Underestimation and Over- 
estimation is a Variable Function of the Ratio of the Time of 
Exposure of the Eye to the Colored and the Colorless 
Light. 
Equality of 
Brightness 
Method 
Flicker Method 
Difference by 
Equality of 
Brightness 
Change Pro- 
Amount of 
Change that 
can be Detected 
by 
Color 
Value of 
Colored 
Sector 
Method and by 
duced by 
Varying 
Equal- 
White-black 
Value 
White-black 
Value 
with 1800 
Colored Seetor 
Sectors 
ity of 
Bright- 
ness 
Method 
Flicker 
Method 
Red 
White 64° 
Black 296° 
3 00° 
270° 
180° 
White 58.9° 
Black 301.i° 
White 56.2° 
Black 303.8° 
White 46.6° 
Black 313.4° 
-17-4° 
-12.3° 
8° 
1.8° 
1.2° 
1.8° 
Yellow 
White 332° 
Black 28° 
300° 
270° 
180° 
White 328.3° 
Black 31.7° 
White 321.7° 
Black 38.3° 
White 313.4° 
Black 46.6° 
-18.60 
-14-9° 
9-5° 
2° 
1.8° 
1.8° 
Green 
White 88.5° 
Black 251.50 
300° 
270° 
180° 
White 99° 
Black 261° 
White 105.5° 
Black 254.50 
White 114.9° 
Black 245.1° 
+26.40 
+ l5-9° 
9° 
•45° 
•9° 
•45° 
Blue 
White 12.50 
Black 347-5° 
3 oo° 
270° 
180° 
White 14.50 
Black 345.5° 
White 19.1° 
Black 340.9° 
White 23.2° 
Black 336.8° 
4-10.7° 
4-8.7° 
5-3° 
2.2° 
1.4° 
•9° 
favorable for agreement by Dr. Ives.1 The plan of the ap- 
paratus used in this work is indicated in Fig. 6. A spectro- 
scope was used to give the colored light; a 32-cp. carbon 
lamp (F) was used as the source of the colorless light. This 
lamp gave a light of the same quality as that used by Dr. Ives, 
namely, the quality of the carbon standard of 4.85 watts per 
mean spherical candle. When placed at 32.6 cm. from the 
sectored disc (D), 270 meter candles of light were reflected 
1 Ives, H. E., op. cit., p. 173. 142 
C. E. FERREE AND GERTRUDE RAND 
from the disc. The eye piece was removed from the spectro- 
scope and a lens system was used in its place consisting of two 
lenses {A} and (5), one to render the light emerging from the 
objective slit (C) parallel, and the other to focus it on the eye 
30 cm. distant. Between the eye and the focusing lens (5) 
was interposed the sectored disc (D). Thus the light reflected 
from the sectored disc suffered no absorption in passing to the 
eye. A stimulus-opening (E) 16 mm. in diameter was placed 
in front of the disc 20 cm. from the eye. This subtended the 
Fig. 6 
same visual angle as the field size that Dr. Ives found to be 
the most favorable. A pupillary aperture I mm. square 
placed in front of the eye reduced the light reflected from the 
white disc to the intensity called by Dr. Ives 270 illumination 
units.1 The colors used were a very narrow band of the 
spectrum in the region of .68 /z, .57 n, .52 n, and .47 /j,, giving 
the four pure colors red, yellow, green, and blue. The method 
of making the comparison was as follows: The sectored 
disc was turned so that its edge bisected horizontally the 
photometric field, and the luminosity of the colored field was 
altered by changing the width of the collimator-slit until it 
equalled by the equality of brightness method the 270 illumi- 
nation units. Using this slit width, then, the disc was rotated 
and the position of the white light was adjusted until no 
1 By using a pupillary aperture I mm. square, Dr. Ives has reduced the light 
entering the eye by an amount which, so far as we can see, can not be determined. 
He has established an arbitrary unit which he calls an illumination unit. We can 
not, therefore, compare the intensities of light used by us in the preceding experi- 
ments (pp. 134 ff.) with the 270 illumination units used by Dr. Ives. If one were to 
judge, however, by the apparent brightness of the disc in the two cases, he would have 
to say that the amount of light entering the eye was considerably greater for our FLICKER PHOTOMETRY 
143 
flicker was obtained. The flicker determinations were made 
with 3150, 1800, and 450 total open sector as before. The 
results are shown in Tables VII and VIII. 
Table VII 
Observer A 
Showing that the Underestimation of Red and Yellow and the Overestimation of Blue and 
Green is a Characteristic of the Method of Flicker for Lights of the Intensity Used 
in this Work, and that the Amount of this Underestimation and Over- 
estimation is a Variable Function of the Ratio of the Time of 
Exposure of the Eye to the Colored and the Colorless 
Light. Intensity same as was used by 
Ives 
Equality of 
Brightness 
Method 
Flicker Method 
Difference by 
Equality of 
Brightness 
Change 
Produced 
by Varying 
Sectors 
Amount of Change 
that Can Be 
Detected by 
No. of 
Wave- 
length 
Distance of 
White Light 
Giving 
Equality of 
Value of 
Colored 
Sector 
Distance of 
White Light 
Giving no 
Method and 
by Flicker 
Method 
with 1800 
Colored 
Equality 
of Bright- 
ness 
Flicker 
Method 
tionsper 
Second, 
Flicker 
Method 
Illumination 
Flicker 
Sector 
Method 
.68 fl 
32.6 cm. 
315° 
41.5 cm. 
- 11.6 cm. 
-3.6 cm. 
2.4 cm. 
.5 cm. 
9.2 
1800 
44-2 
•45 
12 
32.6 
45° 
45-i 
- 6.9 
•5 
II 
•57 m 
315° 
37-4 
-3-4 
2.8 
•4 
9.8 
1800 
39-5 
•4 
14 
45° 
40.8 
•4 
12 
•52 M 
32.6 
315° 
23 
+ 13-6 
+4-9 
2 
•5 
9-7 
1800 
19 
•4 
13 
32.6 
45° 
18.1 
•4 
11.7 
•47 m 
315° 
23 
+ 13-8 
+5 
2-3 
•45 
9-9 
1800 
18.8 
•4 
14.1 
45° 
18 
•45 
12 
higher intensities than the 270 illumination units used by Dr. Ives. Thus it seems prob- 
able that most of our preceding tests were made with an intensity of light equal to or 
greater than that used by him. His claim, it will be remembered, was that one of the two 
causes of disagreement between the results obtained by the methods of flicker and 
equality of brightness in preceding experiments is the low intensity of the lights used. 
(The other was the lack of proper regulation of the size of the photometric field.) 
We do not believe that either one of these factors is the fundamental cause of disagree- 
ment, as is attested in our experiments by the fact that strong disagreement remains 
when both of them have been eliminated, at least, as completely as they were eliminated 
by Dr. Ives. A consideration of the functioning of the eye under very short exposures 
to light, shows, we believe, a much more fundamental cause of disagreement, namely, 
the difference in the way in which the eye responds to light stimuli when presented 
for the lengths of time used in the two methods. 144 
C. E. FERREE AND GERTRUDE RAND 
Table VIII 
Observer B 
.68 n 
32.6 
315° 
41-3 
-11.6 
-4 
3 
.8 
12 
1800 
44-2 
•7 
14 
45° 
45-3 
.8 
13-4 
•57 V 
32.6 
3*5° 
38 
- 8.9 
-4-3 
2.9 
•9 
12.5 
1800 
41-5 
.6 
14.2 
45° 
42.3 
.8 
i3-i 
■52M 
32.6 
315° 
23-4 
+ 12.1 
+3-9 
3-3 
.8 
12 
1800 
20.5 
•7 
14 
45° 
19-5 
•7 
12.2 
•47 m 
32.6 
3i5° 
22.8 
+ 12.6 
+4 
3-5 
•9 
11.1 
1800 
20 
•7 
14-4 
45° 
18.8 
.8 
12.8 
It was stated in the beginning of the paper that disagree- 
ment between the results of the method of flicker and equality 
of brightness would be shown from a preponderance of the 
work done by others who have made the comparison. As a 
general case the fact scarcely needs more than the pointing 
out. Before the work of Ives, disagreement was pretty 
generally admitted. Bell1 says: "That the flicker and equal- 
ity of brightness methods do not give coincident results when 
we consider the general case of flicker photometers, as com- 
pared with equality of brightness photometers, is a fact that 
has been too long familiar to photometrists to admit of a dis- 
cussion." Comparisons of the two methods have been made 
by Whitman, Wilde, Dow, Bell, Stuhr, Luckiesh and Ives. 
Whitman2 compared the luminosities of a red and green light 
placed 6 ft. apart on a photometer bar. He found that the 
setting of the photometer for equality of illumination differed 
for the equality of brightness and flicker methods by 1.2 ft. 
for one observer, and .8 ft. for another. Wilde3 photometered 
a tungsten lamp against a carbon by the methods of flicker 
and equality of brightness, and found a difference of 6 per 
cent, in the result. Bell4 compared the ratio of lumin- 
1 Bell, L., 'Acuity in Monochromatic Light,' Electrical World, Sept. 9, 1911, 58, 
P- $37- 
2 Whitman, F. P., 'On the Photometry of Differently Colored Lights and the 
Flicker Photometer,' Physical Review, 1896, 3, pp. 241-249. 
3 Wilde, L. W., 'The Photometry of Differently Colored Lights,' The Electrician, 
July 16, 1909, 63, pp. 540-541. 
4 Bell, L., 'Chromatic Aberration and Visual Acuity,' Electrical World, May 11 
1911, 57, pp. 1163-1166. FLICKER PHOTOMETRY 
145 
osities of a mercury vapor lamp with that of a tungsten 
lamp by means of the flicker method and found it to be 5.42. 
These same lights by the equality of brightness method gave 
a ratio ranging from 6.86 to 10.93 for different observers. 
Stuhr1 compared red and green lights by several methods 
including the method of flicker and equality of brightness. 
He found that the mean deviation of the values obtained by 
the method of flicker from those obtained by the equality of 
brightness method amounted to 14.14 per cent. Luckiesh2 
photometered a red against a blue-green light by the methods 
of flicker and equality of brightness, and found a difference of 
62 per cent, in the ratios of the luminosities of the two lights 
by the two methods. 
Two factors have in the main been assigned to the cause 
of the disagreement: the effect of intensity and of size of the 
photometric field. Lauriol, Dow, Millar, Ives, and Luckiesh 
have investigated the former factor, and Schenck, Dow, and 
Ives the latter. These are both factors which affect the 
results of both methods. In comparison little attempt has 
been made to find the factors that affect the results of each 
method alone. As a general case these, it would seem, might 
be more apt to prove a source of disagreement than those 
which affect both methods. 
With regard to the intensity of the light as a factor, 
Lauriol3 and Dow4 claim that the relative shift in the bright- 
ness of the different colors at low illuminations is shown by 
both methods. The shift for Dow, however, is more pro- 
nounced in the equality of brightness than the flicker deter- 
minations. For Lauriol the shift for the different colors 
varies in magnitude by the two methods and in some cases 
in direction. Millar,5 on the other hand, claims that the 
1 Stuhr, J., 'Ueber die Bestimmung des Aequivalenzwertes verschiedenfarbiger 
Lichtquellen,' Kiel, Philos. Diss., Vol. 19, Okt., 1908, p. 50. 
2 Luckiesh, M., 'Purkinje Effect and Comparison of Flicker and Equality of 
Brightness Photometers,' Electrical World, March 22, 1913, p. 620. 
3 Lauriol, 'Le photometre a papillotement et la photometric heterochrome,' 
Bull. Soc. Intern, des Electriciens, 1904, pp. 647-652. 
4 Dow, J. S., 'Color Phenomena in Photometry,' Philos. Mag., 1906, 12, Ser. 6, 
p. 131. 
B Millar, P. S., 'The Problem of Heterochromatic Photometry,' Trans. Illuminating 
Engineering Society, 1909, 4, p. 769. 146 
C. E. FERREE AND GERTRUDE RAND 
Purkinje phenomenon is not shown at all by the flicker 
method at low illuminations, while Ives1 and Luckiesh2 go 
to the other extreme and declare that a reverse Purkinje 
effect is obtained by the flicker method. With regard to size 
of field as a factor, Schenck3 found that a decrease in size 
lowered the mean variation for the flicker method and de- 
creased the luminosity value obtained for all the colors. 
Dow4 found that as the size of the field was decreased, red 
and yellow lightened relatively to green and blue. This 
effect was more pronounced for the equality of brightness 
than for the flicker method. Ives5 found this effect for the 
equality of brightness method, but the reverse effect for the 
flicker method. 
Ives, admitting the disagreement between the two- 
methods and accepting size of field and intensity of the 
stimulus as the cause of the disagreement, sought to determine 
whether a field size and intensity could not be found for 
which the two methods agree. He photometered different 
portions of the spectrum against carbon lamps at a number of 
intensities and with a number of field sizes. He found in 
general for five observers that the luminosity curves obtained 
by each method differed. This difference, however, was less 
for high intensities than for low. 
A table is appended (Table IX) in which is shown in per- 
centage the difference in results gotten by the five observers 
used by Dr. Ives at the intensity of light which he calls most 
favorable to agreement for the two methods (250 Illumination 
Units). It will be seen that the disagreement for these 
observers is in the average as great, if not greater than was 
gotten by our own observers. Percentage of overestimation 
by the method of flicker is designated by +, and under- 
estimation by -. 
1 Ives, H. E., op. cit., p. 171. 
2 Luckiesh, M., op. cit., p. 620. 
3 Schenck, F., 'Ueber die Bestimmung der Helligkeit grauer und farbiger Pigment- 
papiere mittels intermittirende Netzhautreizung,' Pfinger's Archiv, 1896, 64, pp. 607- 
628. 
4 Dow, J. S., op. cit., pp. 130-134; 'Physiological Principles Underlying the 
Flicker Photometer,' Philos. Mag., 1910, 19, Ser. 6, pp. 58-77. 
6 Ives, H. E., op. cit., p. 172. FLICKER PHOTOMETRY 
147 
Table IX 
Showing in ■percentage the difference in results between the methods of flicker and 
equality of brightness for the five observers used by Dr. Ives at the intensity 
of light which he calls most favorable. 
•653m 
•643m 
.632g 
.622g 
.6l2n 
•594m 
•574m 
•555m 
•545m 
•536m 
.526M 
•5I7m 
> 
-12. % 
- 3-6 
- 4-3 
- 7-3 
- 10. 
- 1. 
+ 0.5 
- 0.4 
~ 3-i 
- i-9 
0 
+ 0.6 
H. E. I. 
1111111+++++ 
HH JO Ln tO 
O 00 4* 00 00 tO O 00 }0 p Ox'p 
00M O 4^ O 4» Cn Co Cn ' 
| M. L. 
-18. % 
- 7-o 
- 15-5 
- 4.2 
- 6.5 
- 0.5 
- 2-5 
-11.9 
-13-8 
-12.6 
- 21.4 
-13-8 
P. w. c. 
111+1++11111 
LO M to 4k 4k lo Ln 
LO LO N 4^ M M M to Ln hH H-t 
GO ' LnCn ' Cn ' 
I C.F.L 
+++++1++1111 
LO LO w H H to in 
0 LO 4^ 00 00 Ln Ln 0 Ln to LO 0 
b Cn Co ' Cd Cd bo Co Cd Cd *Cj 
F. E. C. 
It has been our purpose in general in this part of the 
paper to indicate a field of investigation in the department 
■of physiological optics about which little is known as yet with 
certainty, rather than to report a finished piece of work or to 
attempt to draw positive conclusions. When functioning 
under the conditions imposed by the method of flicker, too 
little is known of the characteristics of the eye, we believe, 
to render safe its use as a measuring instrument. Our 
purpose in particular has been to point out and show the effect 
of a factor which we believe to be an important source of 
disagreement between the equality of brightness and the 
flicker methods, and to suggest that a more careful study be 
made of the factors that influence the method of flicker before 
it is adopted in its present form as the method for the stan- 
dardizing laboratories. Just as one factor has been over- 
looked, so there may be others the influence of which should 
not be ignored. 
Appendix 
Three other points which may be of interest in connection 
with the above work are appended here. The first two were 
discussed by Dr. Ives in a series of articles on the method of 
flicker in the Philos. Mag., 1912, 24, Ser. 6, pp. 149-188, 
352-370, 744-751, 845-853, 853-863. (1) In the third of his 
■series of articles, he apparently wishes to show that the cause 148 
C. E. FERREE AND GERTRUDE RAND 
of the disagreement between the results of the methods of 
flicker and equality of brightness lies on the side of the latter 
method. That is, the difficulty of making the judgment is so 
great that not an equalization, only an 'appraisement' is 
accomplished. To demonstrate this, he attempts to get rid 
of the disturbing factor of color difference in the equality of 
brightness method by making his comparisons always between 
lights differing only slightly in composition. That is, a 
green is compared with a green slightly shifted toward the 
yellow or blue, etc. (See his work with the 'cascade' 
method, p. 748.) A curve of luminosity for the spectrum 
obtained in this way is found to agree more closely with the 
flicker curve than one obtained in the ordinary way. The 
following things may be said of this demonstration, however. 
In the first place, he states that the cumulative errors are so 
great in the method that he could not begin at one point in 
the spectrum having a given luminosity and work in a given 
direction, then reverse this direction of working and obtain 
at all a close approximation to the luminosity value for the 
point at which he started. For this reason he drops the 
point by point procedure of the 'cascade' method, and plots 
his curve by taking his observations at twelve points in the 
spectrum. From the observations of these points the whole 
curve is constructed. In the second place, his method does 
not entirely accomplish his purpose of getting rid of all differ- 
ence in color quality between the lights compared. In order 
to add some further data bearing upon the question whether 
the lack of agreement hitherto found between the results 
obtained by the equality of brightness and flicker methods 
could have been due to the difficulty of making the equality 
of brightness judgments of fields differing in color quality, 
we have thought it worth while to make the comparison 
using an equality of brightness method which for the pur- 
poses of this investigation presents, we believe, some points 
of advantage over the method used by Dr. Ives.1 That is, 
1 We do not, however, mean to propose this as an entirely satisfactory method 
of heterochromatic photometry for the reason given in the discussion of the relation 
of the method to the Talbot-Plateau law (see footnote p. 149). We are using the 
method here merely to show that when the disturbing factor of color difference in the FLICKER PHOTOMETRY 
149 
the method we have used offers even less chances for errors 
in judgment, is simpler, and entirely eliminates the presence 
of a second color in the fields to be compared. The method 
is as follows: The sectored disc was adjusted so that its 
outer edge bisected vertically the photometric field. A 
standard colorless light was moved to the position on the 
photometer bar that gave the judgment of equality by the 
method of flicker, and the disc was rotated at the fusion rate. 
Half of the field was thus of color of the original saturation 
and luminosity, and the other half was a fusion of the 
colored sector of the original saturation and luminosity and a 
gray sector of the luminosity of the color as determined by 
the method of flicker. Now, if the luminosity of the color by 
the method of flicker were the same as by the equality of 
brightness method, the two halves of the photometric field 
should match in luminosity (within the limits imposed by the 
Talbot-Plateau law).1 That is, the addition of the colorless 
fields to be compared is eliminated from the equality of brightness method, there is 
still a large, in fact an apparently undiminished characteristic difference between 
the results of the equality of brightness and flicker methods, which, so far as one 
can see, can in no way be ascribed to the equality of brightness method employed. 
The degree to which the influence of color difference on the judgment of the bright- 
ness equality of the fields compared is removed by this method is shown by the greatly 
increased reproducibility of the judgment. For our observers, the reproducibility 
is almost as great as it was for the method of flicker. There was thus but little more 
of the element of appraisement in this method than there was in the method of flicker, 
while the characteristic difference in the results obtained by the two methods was not, 
so far as could be determined, appreciably lessened. 
1 A few words are needed to explain what is meant above by "within the limits 
imposed by the Talbot-Plateau law." It could scarcely be expected from a considera- 
tion of this law that the two fields would match especially under the dark-room condi- 
tions under which photometry is done, even when the gray sector was chosen equal in 
brightness to the color by the equality of brightness method. That is, when the 
colored is mixed with the gray sector by the method of successive impressions, there 
is a reduction of the intensity of each impression which is the same as would be 
gotten were the intensity of each light to be reduced in proportion to the time of ex- 
posure of the eye to each light to the total time of exposure of the eye to both lights. 
(See the discussion of the Talbot-Plateau law, p. 121.) That is, if the value of each 
sector is 1800, the impression made upon the eye by each light is the same, according 
to the Talbot-Plateau law, as if both lights were reduced one-half in intensity. But 
in suffering the reduction, the luminosity of the colored sector is not changed the same 
in amount as is that of the gray sector. If it is blue or green, for example, its bright- 
ness is not reduced so much as is that of the gray sector, and its fusion with the gray 
sector tends to lighten that sector and to make the second half of the field lighter than 150 
C. E. FERREE AND GERTRUDE RAND 
to the colored sector would produce no change in its luminos- 
ity, and the two halves of the field would present a fully 
saturated color of a given luminosity and a less saturated 
color of the same luminosity (within the limits imposed above). 
But if there were an underestimation or an overestimation of 
the luminosity of the color by the method of flicker, the 
brightness of the second half of the field would be modified 
in this direction in proportion to the value of the colored and 
colorless sector; and if the underestimation or overestimation 
were great enough the two halves would not match. In 
proportion as the colorless sector is made larger in the second 
half of the field, the color of the mixture loses saturation, and 
the comparison with the fully saturated half of the fields 
becomes more difficult to make. On the other hand, in 
proportion as the colored sector is made larger, the effect on 
the brightness of the mixture of the difference between the 
flicker value and the true sensation value, if such a difference 
exists, is lost. After considerable preliminary investigation 
it was decided to use in turn colored sectors of the value of 
3000, 2700, and 1800. The comparison was made for lights 
of the intensities specified in the preceding sections of the 
paper. In all cases when the color was red or yellow, the 
the first. If, however, the colored sector is red or yellow, it is reduced more in bright- 
ness than is the gray sector, and its fusion with that sector tends to darken it and so 
to render the second half of the photometric field darker than the first. We have 
conducted experiments to determine whether the above effect, which is a direct corol- 
lary of the Talbot-Plateau law, actually takes place in observable amounts. When 
the light of the spectrum or light of the purity given by the Wratten and Wainwright 
filters was used, we found that it did. That is, when the second half of the field was 
green or blue and was fused with a gray of the luminosity of the color employed, 
determined by the equality of brightness method, this half of the field was observably 
lighter than the first half. Conversely, when red or yellow was used, the second half 
of the field was darker than the first. The effect, however, was not nearly so great as 
it was when the gray sector was made of the brightness of the color as determined 
by the method of flicker. That is, if two experiments are conducted, one in which 
the second half of the field is made by fusing the colored sector with a gray sector 
of the brightness of the color as determined by the equality of brightness method, and 
the other in which this half of the field is made by fusing the colored sector with a gray 
sector of the brightness of the color as determined by the method of flicker, the differ- 
ence in brightness between the two halves of the field is quite appreciably greater 
in the second case than in the first. For example, when the colors are green and blue, 
the second half of the field is more too light in the second case than in the first; and 
when red and yellow, it is more too dark. FLICKER PHOTOMETRY 
151 
second half of the field was darker than the first; and when 
either blue or green, was lighter than the first half of the field. 
Determinations were made also of how much the colorless 
light had to be moved to make the two halves of the field 
match. These distances were not much different from those 
contained in the tables in the preceding sections of the paper 
expressing the difference in the estimation of the luminosity of 
the colors by the methods of flicker and equality of brightness 
(see pp. 136, 137),-certainly not any more than should be 
expected when it is remembered that a part of the effect 
of the difference is lost by mixing the colorless light repre- 
senting the flicker determination with a sector of the colored 
light in its true luminosity value. The work was done also 
with pigment papers with a similar result. Thus it seems 
reasonable to conclude that the cause of the disagreement 
between the two methods can not be attributed entirely at 
least to the difficulty of making the equality of brightness 
judgment due to the difference in color quality between the 
fields compared, for in the above cases the color quality of 
the lights compared was the same. In the third place, dis- 
regarding the results of the above experiments, the writers 
scarcely need point out that it would be extremely difficult 
to explain such a systematic drift of luminosity in one direc- 
tion in one part of the spectrum, and in the opposite direction 
in the other part, as we obtained, in terms of errors due to a 
false judgment of the sensations actually aroused. More- 
over, it would be just as difficult to explain Dr. Ives's own 
reverse Purkinje effect in terms of a false judgment of the 
actual brightness values presented in sensation; or the closer 
agreement he obtains between the results by the methods of 
flicker and equality of brightness at high illuminations, in 
which case there is the maximum amount of color present 
and, therefore, the maximum color difference to disturb the 
equality of brightness judgment between colored and color- 
less light. Moreover, the kind of errors that one finds as 
due to uncertainty of judgment is a deviation on either side of 
a mean. This occurs when all other factors are eliminated if 
several judgments of the same sensation are made. Such 152 
C. E. FERREE AND GERTRUDE RAND 
errors are compensated for by taking the average or mean of 
the determinations. If it is not conceded that they are com- 
pensated for, how, for example, can the average of the results 
by the equality of brightness method be taken as a standard 
in terms of which to evaluate the results obtained by other 
methods? (See Whitman, Schenck, Wilde, etc.).1 Surely 
this should not be allowed if there were a consistent deviation 
in any one direction from the true brightness value for a given 
color due to errors in judgment. Moreover, such a character- 
istic drift due to errors in judgment is unknown in all previous 
work in psychophysics, and not only unknown, but unsus- 
pected. 
(2) In the fourth paper of the series,2 Dr. Ives applies as 
a test to the method of flicker what he calls two axioms of 
measurement. These are (a) things which are equal to the 
same things shall be equal to each other; and (b) the whole 
shall be equal to the sum of its parts. He finds that the 
method of flicker satisfies these axioms better than the equali- 
ty of brightness method. We would point out that these 
tests would not be expected to reveal to any considerable de- 
gree the influence of the factor we are discussing. They are 
tests which would apply as a check on the power to make the 
judgment of the brightness of the sensation properly, or to 
any tendency of this brightness equality to drift in one direc- 
tion in any part of the spectrum without a compensating drift 
in the opposite direction in some other part of the spectrum; 
but they are not tests that could be expected to show whether 
or not there is underestimation in one half of the spectrum 
and overestimation in the other half. For example, the area 
of the curve of the spectrum plotted by the method of flicker 
might very well sum up to the value of the reassembled white 
light because of the compensating effect of the underesti- 
mation of one half of the spectrum and the overestimation of 
the other half. 
(3) Since the foregoing paper was presented, the writers 
1 While Dr. Ives does not explicitly state that he takes the equality of brightness 
method as a standard in terms of which to evaluate the correctness of the results by 
other methods, the point of view is strongly implied in his first paper (loc. cit.). 
2 Philos. Mag., 1912, 24, Ser. 6, pp. 845-853. FLICKER PHOTOMETRY 
153 
have met with the contention from a prominent advocate of 
the method of flicker that the effect of a reduction of intensity 
is not given by the method of flicker because each individual 
impression is carried over until ;|the next is given, with suffi- 
cient intensity to preclude the effect of reduction. Whether 
or not each individual impression can be considered as carry- 
ing over with sufficient intensity to preclude the effect of re- 
duction is an important point and should, lest the issue be in 
doubt, be included in a discussion of the principles underlying 
the method of flicker. It may not be out of place, therefore, 
for us to consider the question here briefly, even though it 
has not as yet, so far as we know, been discussed in print. 
As evidence that each individual impression should be 
considered as carrying over with sufficient intensity to pre- 
clude the effect of reduction, it was contended, as the case 
was presented to us, that the rate used in the method of flicker 
is the fusion rate of the two impressions. Two reasons were 
given for considering this rate as the fusion rate, (i) If the 
two impressions be red and green, for example, yellow is pro- 
duced at the rate of succession used in the flicker method. 
Yellow, it was pointed out, is a fusion of red and green, and, 
therefore, the rate used must be considered as the fusion rate 
for these colors. In answer to this point we would again call 
attention to the phenomena (see p. 116) which are produced in 
sensation when two impressions differing in color and bright- 
ness are given to the eye successively at different rates of 
speed.1 When the rate is very slow, the effect of separate 
and distinct impressions is given, each in its proper color and 
brightness. When a little faster rate is used, the impressions 
become confused and a flickering effect is produced both in 
the color and brightness components of the sensation. When 
the rate is made still faster, the flickering of color dies out, 
leaving only brightness flicker; that is, the color components 
of the two sensations have been fused. That the brightness 
1 We wish at this point to state very emphatically that our account of the fusion 
of the color and brightness components of sensation at different rates of speed is not 
based on any theoretical conception of a separate brightness and color sense, but upon 
actual observation of the phenomena that take place when light impressions differing 
in color and luminosity are combined at different rates of succession. These phe- 154 
C. E. FERREE AND GERTRUDE RAND 
components have not been fused, however, is attested by the 
presence of brightness flicker, which is now left outstanding 
in a field uniform as to color quality. As the rate of succes- 
sion is made still faster, brightness flicker becomes less and 
less pronounced and finally disappears.1 The rate at which 
this disappearance takes place is the fusion rate for the bright- 
ness components for the two sensations, and is much higher 
for all the colors than is the rate at which the fusion of the 
color components takes place.2 (Interpreted in terms of the 
nomena may be readily demonstrated by any kind of flicker photometer head if a 
sufficiently sensitive control of speed of rotation is had. (We have used for the control 
of speed of rotation a rheostat and motor especially constructed to give fine changes.) 
It can be very plainly and perhaps most conveniently demonstrated by rotating 
sectors of pigment papers at the proper gradations of speed in a good daylight illu- 
mination. 
1 We find that Kriiss {Physical. Zeitschr., 1904, 5, p. 67) gives a description of 
the phenomena that take place in sensation when two impressions differing in color 
and brightness are given to the eye successively at different rates of speed, very 
similar to that we have given here. He says: "If we slowly alternate the illumination 
from two differently colored light sources, for example, from a Hefner lamp and a gas 
burner, we clearly distinguish a succession of reddish and bluish bands with weak 
washed-out limits between them. As the rate of succession is increased it becomes 
progressively more difficult to distinguish the two colors from each other. At a com- 
paratively low rate they begin to lose themselves in each other. At a slightly higher 
rate the difference in color disappears altogether and we have a color mixture. In this 
mixture, however, a brightness succession, a flicker, is observable which disappears 
only by a further increase in the rate of succession. Physiologically, it is of great 
interest that the distinguishing of separate colors ceases at a much slower rate of 
succession than the rate at which completely continuous sensation begins." 
2 The following values will serve to give a rough comparative showing of the 
rates at which the phenomena described above take place. The colors used were red 
and green. They were obtained from pigment papers of the Hering series of standard 
papers and from gelatine filters. Two intensities of color were employed in each case. 
The brightness of the Hering green for the lower intensity of illumination was .000814 
cp. per sq. in.; of the red, .000594 cp. per sq. in. The phenomenon of separate 
impressions occurred from the lowest speed up to 6.9 revolutions per second. The 
impression of an intermingled color and brightness flicker was given from this rate 
up to 9.6 revolutions per second, at which rate the color components of the sensation 
fused, giving a field uniform as to color quality but with a strong outstanding brightness 
flicker. Brightness flicker was present until a speed of 22 revolutions per second was 
obtained. At this speed the brightness components in sensation were completely 
fused and the rotating disc presented a surface uniform both as to color and brightness. 
In making these determinations, the same sized field was used as was employed in our 
work with the method of flicker, i. e., the disc was viewed through an aperture 3 mm. 
X 3 mm. in a gray screen (Hering No. 24) 20 cm. from the eye. For the higher inten- 
sity the green surface was illuminated to a brightness of .00242 cp. per sq. in.; the 155 
FLICKER PHOTOMETRY 
duration of the impression after the light has been cut off, 
this means, of course, that the brightness component in the 
sensation does not carry over under these conditions with as 
little loss of intensity as does the color component.) It is 
evident, then, that the rate of succession which is used in the 
method of flicker is at or near the fusion rate for the color 
components of the two sensations, not for the brightness com- 
ponents; nor is it anywhere near the fusion rate for the bright- 
ness components. But it is the brightness components in 
which we are interested in photometry. That is, it is in 
terms of the brightness component that all photometric judg- 
ments are made. The color components, when they differ 
in tone, only serve to confuse the judgment. It is, therefore, 
our object in all methods of photometry as much as possible 
to get rid of difference in the color components. This can be 
accomplished in the method of flicker only because of the 
fact we have just pointed out, namely, that the fusion of the 
color component in sensation comes at a much lower rate of 
succession than the fusion of the brightness component. That 
is, all color differences, whether sensed as distinct or as flicker- 
ing sensations, disappear at a rate of succession that has little 
or no effect on eliminating the brightness factor, or in this 
case the equivalent of this elimination, the fusion of the bright- 
ness components of the two sensations. In fact, if there 
were no difference in the fusion rate of the color and bright- 
ness components, the flickering color impressions would so 
mask the presence of brightness flicker at any rate of succes- 
sion that could be used, that the method would doubtless 
red, .00167* CP- Per sq. in- The phenomenon of separate impressions occurred from 
the lowest speed to 6 revolutions per second, at which rate color flicker began. Color 
fusion took place at 12.4 revolutions per second, and brightness fusion at 29.3 revolu- 
tions per second. At the lower intensity for the filters, the brightness of the green 
was .154 cp. per sq. in.; for the red, .099 cp. per sq. in. As compared with their 
brightness these colors were much more poorly saturated than were the Hering pig- 
ments. The phenomenon of separate impressions ceased and color flicker began at 
6.5 per second. Color fusion took place at 11.5 revolutions per second, and brightness 
fusion took place at 35.4 revolutions per second. At the higher intensity for the 
filters the brightness of the green was .22 cp. per sq. in.; for the red, .143 cp. per 
sq. in. Color flicker began at 7 revolutions per second; color fusion took place at 
12.9 revolutions per second; and brightness fusion was complete at 38.3 revolutions 
per second. 156 
C. E. FERREE AND GERTRUDE RAND 
have little if any greater sensitivity than the equality of 
brightness method. (2) The second point that was cited 
in support of the contention that the rate used in the method 
of flicker is the fusion rate for the two sensations aroused, is 
that no brightness flicker is present when in terms of the 
method the two impressions are adjudged of the same bright- 
ness. This to the present writers seems indeed a strange 
confusion of meanings. Fusion is a term used to represent 
what takes place when two impressions or sensations differing 
in quality are combined into one, the same or homogeneous 
as to quality. This combination may be obtained in case of 
light stimuli, for example, by mixing two lights evenly and 
allowing them to act simultaneously on the eye; or it may be 
obtained by giving two lights to the eye in succession at such 
a rate that the sensation aroused by the one lasts over until 
the next one is set up with a sufficient degree of intensity to 
give the effect of continuity or homogeneity of quality. It 
may add to the clearness of our discussion, then, to consider 
what takes place in this regard when two impressions differing 
in brightness are given to the eye at the different rates of 
succession mentioned in the preceding paragraph. At the 
rate at which distinct and separate impressions are given, 
each sensation obviously dies away completely before the 
next one is aroused. If a rate slightly faster than this is 
selected, the sensation does not die away completely before 
the next one is set up. 
There is a slight lasting-over from one impression to the 
next. This when the two impressions differ in brightness 
gives the effect of a wavering or flickering sensation. At the 
lowest speed at which flicker is produced, the effect of this 
lasting-over has its minimum value. As the speed is further 
increased it becomes greater and attains its maximum value 
at the rate of complete brightness fusion.1 (See discussion of 
Talbot-Plateau law, pp. 121.) It is obvious, then, that the 
rate of speed employed in the method of flicker, which is, 
roughly speaking, the lowest rate at which brightness flicker 
1 At the fusion rate neither sensation rises to its maximum value, for example, nor 
has a chance to die away until the next one develops. The effect is that of a con- 
tinuous sensation homogeneous as to color and brightness. FLICKER PHOTOMETRY 
157 
■can be obtained unmixed with color flicker, is not the fusion 
rate for the brightness components in sensation nor is it 
anywhere near this rate.1 It is equally obvious also that the 
absence of flicker when the final adjustment of the lights has 
been made for a photometric balance, can not be adduced as 
any evidence that this rate is the fusion rate for the brightness 
component of the two sensations, or, what is more significant 
in relation to the above mentioned claim, that it is a rate to 
which more than a minimum of lasting-over effect from 
impression to impression can be ascribed. Flicker is absent 
merely because, in accord with the purpose of the method, 
such an adjustment of the distance of the lights from the 
photometer head is made that the sensations aroused by the 
two lights are of equal brightness. Such sensations do not 
flicker whatever may be their rate of succession. It can, 
therefore, be considered as little more than absurd to adduce 
the absence of flicker when the photometric balance has been 
attained as evidence that the rate used is the fusion rate for 
the brightness components of sensation, and to pass from this 
to the conclusion that the same amount or anywhere near the 
same amount of carrying-over effect is present for this rate 
as obtains when the fusion rate is used. In fact, if this 
carrying-over effect were present to any considerable degree, 
the whole point of the flicker method would be lost. That 
is, it is the purpose in the method of flicker to select a rate of 
succession that will give the eye the maximum of sensitivity 
to brightness difference (or flicker), namely, the lowest 
rate at which flicker can be produced, rather than a rate that 
will fuse out this difference in sensation. 
But supposing it could be established, as was contended, 
that we have in the rate used in the method of flicker a com- 
plete color and brightness fusion of the sensations aroused 
by the two lights, little would be gained for the claim that 
there is no reduction in the effect on sensation of the two 
lights employed, if it be granted, for example, that the 
Talbot-Plateau law is true. In substance this law is as 
1 Flicker and fusion are in fact antithetical terms, and the rates of succession 
which are favorable for each are widely separated in the scale of frequencies. 158 
C. E. FERREE AND GERTRUDE RAND 
follows. When once the rate of rotation is sufficient to give 
a uniform sensation, the color and brightness of the disc are 
the same as they would be if all the light reflected from the 
sectors were evenly distributed over the surface of the disc; 
and no further increase in rapidity produces any effect on its 
appearance.1 In terms of this law it is seen that the effect on 
1 See H. F. Talbot, 'Experiments on Light,' Philos. Mag., 1834, Ser. 3, 5, pp. 321- 
334- 
Talbot phrases this law as follows (pp. 328-329): "Since then these two things- 
the intensity of light and the time of the body's remaining in any given part of the 
circle-are each inversely proportional to the circumference of the circle it describes, 
it follows that they must be directly proportional to each other; that is to say, an 
irregular intermittent luminary whose observations are too frequent and too transitory 
for the eye to perceive, loses so much of its apparent brightness from this cause as is 
indicated by the proportion between the whole time of observation and the time 
during which it disappears." "The rapidity of the rotation does not affect the argu- 
ment." To verify this reasoning, Talbot conducted experiments with reflected light 
using pigment surfaces and mirrors to send the light to the eye; and with transmitted 
light using sectored discs to cut down the time of exposure of the eye to various luminous 
sources. 
In 1835 Plateau repeats and verifies Talbot's experiments. ('Betrachtungen 
fiber ein von Hrn. Talbot vergeschlagenes photometrisches Princip,' Poggen. Annal., 
1835, 35, pp- 457-468). He concludes from his experiments as follows (pp. 462-463) 
"Nun muss zufolge des am Anfange dieses Aufsatzes dargelegten Princips die schein- 
bare Helligkeit der Scheibe sich zu der des Papiers verhalten wie die Voriibergangsdauer 
eines weissen und eines schwarzen Sectors; odor was dasselbe ist, wie die Winkelbreiten 
eines weissen Sectors zur Summe der Winkelbreiten eines weissen und schwarzen 
Sectors, oder endlich, was auch noch dasselbe ist, wie die Breite sammtlicher weisser 
Sectoren zum ganzen Kreisumfang." 
Swan, apparently working in ignorance of the writings of Talbot and Plateau, 
in substance formulates the law anew in 1849 (see W. Swan, 'On the Gradual Produc- 
tion of Luminous Impressions on the Eye and Other Phenomena of Vision,' Trans. 
Roy. Soc. Edinb., 1849, 16, pp. 581-603. See also F. Boas, 'Ein Beweis des Talbot'- 
schen Satzes und Bemerkungen zu einigen aus demselben gezegonen Folgerungen,' 
Wiedem. Ann., 1882, 16, 359-362; A. M. Bloch,'Experiences sur la vision,' 'Comfit. 
Rend, de la Soc. de Biol., 1885, 2, p. 495; A. Charpentier, 'Loi de Bloch relative aux 
lumieres de courte duree,' ibid., 1887 4, p. 5; etc. 
For a more modern statement of this law and one also more consistent with the 
relation of changes in light energy to changes in sensation, see Helmholtz, 'Handbuch 
der physiol. Optik,' zw. Aufl., 1896, p. 483, "Wenn eine Stelle der Netzhaut von 
periodisch veranderlichem und regelmassig in derselben Weise wiederkehrendem Lichte 
getroffen wird, und die Dauer der Periode hinreichend kurz ist, so entseht ein continuir- 
licher Eindruck, der dem gleich ist, welcher entstehen wiirde, wenn das wahrend einer 
jeden Periode eintreffende Licht gleichmassig uber die ganze Dauer der Periode 
vertheilt wiirde"; or E. C. Sanford, 'Experimental Psychology,' 1898, p. 146, "When 
once the rate of rotation is sufficient to give a uniform sensation, the color and bright- 
ness of any concentric ring are the same that they would be if all the light reflected FLICKER PHOTOMETRY 
159 
sensation is the same as is gotten by reducing the intensity of 
each light by an amount proportional to the ratio of the 
exposure time of that light to the total time of exposure to 
both lights; or in case the photometer head is a sectored disc, 
in proportion to the value of the given sector or set of sectors 
to 360°. That is, with a total value of each sector or set 
of sectors of 1800, the effect on sensation is the same as if each 
light were reduced one-half in intensity; if the total value of 
one sector or set of sectors is 900, the effect on sensation is the 
the same as if the light illuminating that sector were reduced 
to one-fourth of its intensity; if the total value were 450, the 
same effect is produced as if the light were reduced to one- 
eighth of its intensity; etc. Thus, even if the rate of suc- 
cession that is used in the method of flicker could be con- 
sidered as the fusion rate for the brightness component of the 
sensation aroused, little advantage could be gained for the 
position in question. For the conclusion most certainly could 
not be avoided that the effect on sensation would be the same 
as if the lights were reduced in intensity, and by an amount 
proportional to the ratio of exposure time of each light to 
the total time of exposure to both lights. 
The position under discussion seems also to involve to 
some extent a confusion of principle of the method of flicker 
with the method of critical frequency. For example, in the 
method of critical frequency, the impressions are given to the 
eye at the fusion rate. We need scarcely call to mind the 
procedure. One sector or set of sectors of the disc is illu- 
minated by one of the lights to be compared and the other is 
black or of a very low luminosity. The disc is rotated at a 
rate which completely fuses the sectors in sensation. This 
light is then removed and the other light to be compared is 
substituted for it. The distance of this light from the disc 
is then adjusted until the rate of rotation required to produce 
fusion is the same as it was in the previous case. When this 
adjustment is obtained the intensity of illumination of the 
disc by the two lights is said to have been the same, and the 
from it were evenly distributed over its surface, and no further increase in rapidity 
produced any effect on its appearance." 160 
C. E. FERREE AND GERTRUDE RAND 
relative brightnesses of the lights themselves are calculated 
by the law of inverse squares. The situation is, however, 
quite different for the method of flicker. Both sectors or 
sets of sectors of the disc are illuminated by the lights to be 
compared, and the rate of rotation is to be made such that 
if there were any brightness difference between the sectors, 
the maximum of flicker, not fusion, would be produced. If a 
rate were used that would produce fusion, for example, for 
any given amount of brightness difference, it is obvious that 
no difference in brightness equal to or less than this amount 
could be detected by the method. That is, the whole point 
of the method is to use a rate of speed that could not possibly 
be the fusion rate for any appreciable amount of brightness 
difference between the impressions to be compared; and in so 
far as this purpose can be realized in the different cases in 
which the method is employed, sensitivity for the method is 
obtained. 
What our critic really needs to establish in order to support 
his position is that summation instead of fusion takes place. 
That is, if the total effect of each light on sensation is to rise 
to a higher level than is given by each individual impression, 
the individual impressions must in proportion to the rise 
summate or add their individual intensities. To produce this 
effect of summation, each individual impression would have 
to last over in sensation until the next impression of its kind 
is received, which, since the impressions alternate, could be 
the next impression but one. For example, when red and 
green lights are being compared, if the value of the red sensa- 
tion is to rise to a higher level than that given by a single 
impression, the sensation aroused by one exposure to red 
would have to last over until sensation is aroused by the next 
exposure to red; that is, would have to last through the inter- 
val of exposure to green and into and wholly or partly through 
the succeeding interval of exposure to red. How highly 
improbable it is that this could happen to any degree that 
would be of saving consequence to the method, is shown by the 
following two considerations, (a) The wavering character of 
the sensation which we call flicker is due to the fact that a FLICKER PHOTOMETRY 
161 
given sensation does not carry over without a great loss of 
intensity through the next succeeding interval, let alone through 
the next interval but one. And (b') even at the rate at which 
complete color and brightness fusion takes place, there is 
according to the Talbot-Plateau law no effect of summation 
great enough to cause each individual sensation to attain to a 
higher intensity than that fixed by the ratio of the time of 
exposure of its stimulus light to the total time of exposure of 
both lights, nor to produce a noticeable change in this inten- 
sity, however great is the speed of the succession. That is, 
we have a reduction of the intensity of the sensation aroused 
by each light which is the same as would be gotten were the 
intensity of each light to be reduced by an amount propor- 
tional to the ratio of the time of exposure of that light to the 
total time of exposure of both lights, and no further increase 
in the rapidity of the succession produces any change in this 
effect.1 
With regard to the method of flicker, then, the case 
apparently stands as follows. The individual impressions 
are so short that the eye is very much underexposed to its 
stimulus, and the rate of succession is so slow that there is 
1 If one were permitted to interpret the Talbot-Plateau law with regard to what 
takes place when a rate of succession is employed greater than the fusion rate for both 
the colored and brightness components of sensation, two possibilities would be opened 
for explaining why no change in sensation is produced as the rate of succession is 
increased, and the length of each individual exposure is correspondingly decreased, 
(i) Either the increase in the reduction of the exposure-time causes no further reduction 
in the sensation aroused by the individual exposures; or (2) there is, owing to the 
increased rate of succession, a summation effect which just compensates for the reduc- 
tion of the individual impressions. Now even if we were to accept as true the one of 
these alternatives which is the more favorable for the case of flicker, namely, that a 
compensating summation action takes place, and assume that this compensating 
summation obtains clear down to the rate of succession that is used in the method of 
flicker, we would have to expect as much reduction in the sensation aroused by each 
of the lights as is expressed by the Talbot-Plateau law. That is, the reduction for 
each would be the same as would be gotten were the intensity of each light to be re- 
duced in proportion to the exposure-time of each to the exposure-time of both. As we 
have already pointed out, however, it is extremely improbable that there could be a 
compensating summation action at the flicker rate great enough to be of any consid- 
erable consequence to the method, because the wavering character of the sensation 
which we call flicker is due to the fact that a given sensation does not carry over 
without great loss until the next one develops, let alone until the next but one 
develops, which it would have to do to produce any summation effect. 162 
C. E. FERREE AND GERTRUDE RAND 
not enough carrying-over from impression to impression to 
produce fusion, let alone the summation effect which is 
needed to cause the intensity of the sensation to rise to its full 
value or perhaps even to a higher level than would be given 
by a single exposure. Moreover, according to the Talbot- 
Plateau law a summation effect great enough to cause the sen- 
sation to rise to its full value is never produced, however fast 
is the rate of succession; for once the fusion rate is obtained, 
there is a reduction of the intensity of the sensation aroused by 
each light which is the same as would be gotten were the 
intensity of each light to be reduced in proportion to the 
time of exposure of that light to the total time of exposure 
to both lights, and there is no change in this effect however 
much the rate of succession is increased. 
1 Since the above discussion was presented to the Illuminating Engineering 
Society, Ives in collaboration with Kingsbury, has published a sixth article on the 
method of flicker {Philos. Mag., Nov., 1914, 28 (167), pp. 708-728) in which a theory 
of flicker photometry is developed based on an analogy drawn between the response 
of the eye under successive stimulation to the action of incandescent lamp filaments 
under a fluctuating current. The gist of the article is that if the eye behaves under 
the conditions obtaining in flicker photometry as do lamp filaments (subject to cer- 
tain modifications which are not in accordance with what is known of the function- 
ing of the eye) under a fluctuating current, the method of flicker should give with 
high intensities of light at the photometer screen the same results on the average as 
the equality of brightness method. It is our purpose here merely to note the article, 
not to give a detailed discussion. The theory will be discussed in a later paper in con- 
nection with further experimental data. It may not be out of place to state at this 
time, however, that the analogy of the eye and the incandescent lamp filament is not 
based on experimental examination of the eye's manner of response, but is assumed. 
Moreover, considerable evidence is offered in the present paper, we think that the 
eye does not react to its stimulus given to it in succession at the flicker rate according 
to the laws which govern the temperature response of lamp filaments, more especially 
when the impressions differ widely as to wave-length. It has not been claimed, for 
example, that the flicker method does not give the same results as the equality of 
brightness method when the lights compared do not differ as to wave-length. [Reprinted from the Transactions of the Ittuminating 
Engineering Society, No. 6, 1915.] 
THE EFFICIENCY OF THE EYE UNDER DIFFERENT 
CONDITIONS OF FIGHTING: THE EFFECT 
OF VARYING THE DISTRIBUTION 
FACTORS AND INTENSITY* 
BY C. E. FERREE AND GERTRUDE RAND, 
BRYN MAWR COLLEGE. 
Synopsis: In a previous paper** a plan of work was outlined by one 
of the writers for the study of the effect of different kinds of lighting 
conditions on the eye. The problem was divided into three parts: (1) the 
determination of the conditions that give in general the highest level or 
scale of visual efficiency; (2) the conditions that give the least loss of 
efficiency for continued work; and (3) the determination of the conditions 
that cause the least discomfort. Tests were described especially designed 
to meet the requirements of each of these divisions of the work and 
results were given to show in a general way the sensitivity of the tests 
employed. The work of the present paper is confined to the second divi- 
sion of the problem and should be considered as an explorative investiga- 
tion for the determination of factors. Six aspects of lighting are con- 
sidered provisionally as sustaining an important relation to the eye: the 
evenness of the illumination, the diffuseness of light, the angle at which 
the light falls on the object viewed, the evenness of surface brightness, 
intensity and quality. Only the first five of these are dealt with in this 
paper. The first four are called, for convenience of reference, distribution 
factors. In order to produce the variation in the distribution factors 
needed for the purposes of the test, three types of reflectors in common use 
were employed-a direct, a semi-indirect, and an indirect. These reflec- 
tors were selected with reference to the object of the investigation rather 
than as representative in every case of any particular principle of lighting. 
The illumination effects produced in each case were specified in the fol- 
lowing ways: (1) A determination was made of the average illumination 
of the room under each of the three installations. (2) The brightness of 
prominent objects in the room, such as the test card, the reflectors for the 
semi-indirect installation, the reading page, specular reflection from sur- 
faces, etc., was given. (3) Photographs were made of the room from 
three positions under each kind of installation. These effects were then 
correlated with the results obtained with the eye test. 
In order to determine the effect of varying intensity with a certain 
grouping of distribution factors, lamps of different wattage were used 
with each type of reflector employed in the distribution series. The 
*A brief report of the work described in this paper was read by one of the 
writers (Ferree) at the seventh annual convention of the Illuminating Engineering 
Society held at Pittsburgh, September 22-25, 1913. 
The Illuminating Engineering Society is not responsible for the statements or 
opinions advanced by contributors. 
** Trans. I. E. S., p. 40, vol. VIII (1913). 410 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
dency to emphasize brightness extremes in the field of vision 
rather than to level them down. Too often, too, the eye is not 
properly shielded from the primary source of light, and fre- 
quently no attempt at all is made to do this. The semi-indirect 
systems are intended to represent a compromise between the 
direct and the indirect systems. A part of the light is transmitted 
directly to the plane of work through the translucent reflectors 
placed beneath, and a part is reflected to the ceiling. Thus, de- 
pending upon the density of the reflector, this type of system may 
vary between the totally direct and totally indirect as extremes and 
share in the relative merits and demerits of each in proportion to 
its place in the scale. It is not our purpose, however, at this time 
to attempt a final rating of the comparative merits of types of 
lighting systems. For that our work is still too young. More- 
over, there are relatively good and bad fixtures of each type, and 
good and bad installations may be made of any system. What 
we hope to do is by the appropriate selection and variation of 
conditions to find out what the factors are that are of importance 
to the eye in lighting, and from this knowledge as a starting point 
to work towards reconstruction. 
It was stated also in our former paper that the problem dealing 
with loss of efficiency presents two phases. We may investigate 
(a) whether the eye shows a loss of efficiency after three or four 
hours of work under a given lighting system, and (b) whether 
there is progressive loss of efficiency in working several months 
or years under a given system. We have confined and purpose to 
confine our work for the present to the former aspect of the 
problem, because it alone falls within the scope of laboratory 
studies and because we believe that the problem should be worked 
out first in miniature with all the conveniences of manipulation 
and possibilities of precision obtaining under laboratory con- 
ditions. 
II. THE EFFECT OF VARIATION IN THE DISTRIBUTION OF 
LIGHT AND SURFACE BRIGHTNESS ON THE EFFI- 
CIENCY OF THE EYE FOR A PERIOD OF WORK. 
In order better to understand the data given in the tables of 
results, the nature of the tests used in this part of the work will 
again be briefly called to mind. It will be remembered that the 
conventional tests for the eye's responsiveness to its stimulus, FERREk AND rand: EFFICIENCY OF THE EYE 
411 
namely, tests for brightness sensitivity, color sensitivity, and 
visual acuity, were found to be practically useless for this work. 
Modified and rendered sensitive in the ways described in the 
previous paper, they were found to serve as a measure of the 
general level of efficiency of the unfatigued eye under different 
conditions of lighting; but they failed to show loss of efficiency 
as the result of a period of work. This is due to the following 
reasons. (a) There is doubtless very little, if any, loss of 
sensitivity to brightness and color during this length of time.2 
It is commonly believed, in fact, that the brightness and color 
processes are compensating in nature. And (&) the visual acuity 
test, in spite of the fact that its results may be ascribed prac- 
tically entirely to changes in the muscular control of the eye, 
is not adapted to show loss in muscular efficiency, because 
the muscles of the eye, while they may have fallen off enormously 
in efficiency, can under the spur of the will be whipped up to 
their normal power long enough to make the judgment required 
by the test. But they can not long sustain this extra effort. This 
consideration, it will be remembered, led us to continue the test 
through an interval of time. After considerable experimenta- 
tion an interval of three minutes was chosen as best suited for 
our purpose. When the observer is required to look at the test 
card for three minutes, the test objects, even when the eyes are 
fresh, are not seen clearly for the whole time. They are seen 
alternately as clear and blurred. The time they are seen clear and 
blurred is recorded on a rotating drum upon which a line regis- 
tering seconds is also run. From this record the ratio of time 
seen clear to time seen blurred is determined. This ratio may 
be fairly taken as a measure of the efficiency of the eye for three 
minutes of clear seeing at the time the test is taken. In applying 
the test to our problem, a record is taken at the beginning and at 
the close of work, and the ratios of the time clear and the time 
2 That there is practically no loss of sensitivity to brightness and color for this 
period of time was shown in our former paper by the results of our tests for bright- 
ness and color sensitivity with and without the time element as an aid to the test. 
(See also in connection with tests for brightness and color sensitivity, Ferree 
and Rand: A Note on the Determination of the Retina's Sensitivity to Co'ored 
Light in Terms of Radiometric Units, Amer. Jorir. of Psychol., 1912, Vol. XXIII, 
pp. 328-332; An Optics-Room and a Method of Standardizing its Illumination, 
Psychol. Rev., 1912, Vol. XIX, pp. 364-373; Colored After-Image and Contrast 
Sensations from Stimuli in which no Color is Sensed, ibid, pp. 195-239; Rand: 
The Factors that Influence the Sensitivity of the Retina to Color: A Quantitative 
Study and Methods of Standardizing, Psychol. Rev. Monog., 1913, 166 pp.; The 
Effect of Changes in the General Illumination of the Retina upon its Sensitivity to 
Color, Psychol. Rev., 1912, Vol. XIX, pp. 463-490. 412 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
blurred are compared for the two cases to determine how much 
the eye has lost in efficiency as the result of work. Two values 
were used for the distance at which the test card was placed 
from the eye: (a) the maximal distance at which the test objects 
could be seen clearly in the momentary judgment, and (&) a 
distance less than this. The latter distance was finally chosen be- 
cause for the maximal distance, towards the close of the test even 
when the eyes were fresh, the value of the time blurred became 
too high, it was found, to make the most effective comparison 
of the ratios obtaining at the beginning and at the close of work. 
In order to eliminate the memory and fatigue factors which 
make it impossible to reproduce results in a series of tests with 
the same observer when the conventional Snellen test of visual 
acuity is employed, it will be remembered that the test card was 
made to consist of one or more simple objects, and the type of 
judgment was changed so that results were rendered in terms 
of clearness of vision instead of in terms of the ability to recog- 
nize a series of letters or characters.3 That is, in this type of 
test the observer knows what the objects are, and he records the 
time during which he sees them clear and the time he sees them 
blurred. A number of test objects were used in the work of last 
year: two vertical parallel lines stamped I mm. apart on a white 
■ I I 
card ; the letters li printed in small type, the figures • , • , 
I I 
- • -, etc. To these was added this year the figure = 
I . 
This form of test object was suggested by the one used in Dr. 
Ives' visual acuity apparatus.4 While apparently it gives ex- 
3 For a further explanation of this point see Tests for the Efficiency of the 
Eye under Different Systems of Lighting and a Preliminary Study of the Causes 
of Discomfort, Trans. I. E- S., Vol. VIII, 1913, pp. 43-45. 
4 The writers wish to state that the test object used by them was similar to 
that employed by Dr. Ives only with regard to form. One of the prominent features 
of the apparatus used by Dr. Ives, for example, is a device for the control of the 
width of the parallel lines and the interspaces, while the figure used by us was 
printed on a white card with a fixed width of line and interspace. All that the 
writers wish to point out here is that a figure made up of parallel lines and inter- 
spaces is not, they believe, the most suitable for work of the kind we are doing 
because of the comparatively large mean variation it gives in the ratio, time clear 
to time blurred. 
The figure was at first made 7 mm. in diameter; but this figure was found to 
be too large. It would blur, irregularly over its surface, i. e., the edges would 
become indistinct when the center was clear and vice versa. The figure finally 
adopted was 3.5 mm. in diameter. This size was found to be more satisfactory for 
our work. FERREE and rand: efficiency of THE Eye 
413 
cellent results for the purpose for which it was adopted by 
Dr. Ives, it gives too large a mean variation of ratio, time 
clear to time blurred, when the element of time is introduced 
into the visual acuity test to be of maximal service in our 
work. This is probably because a figure of this form is more 
influenced by adaptation, the streaming phenomenon,5 and other 
variable physiological conditions of the retina than are, for 
example, the letters li. This latter object was found to be far 
the most satisfactory for our purpose. When used as test object 
the mean variation of the ratio, time clear to time blurred, for 
the same observer working under conditions as nearly constant as 
possible, is very small indeed.6 Results will be given, therefore, 
in this report only for the work in which the letters li were used 
as test object. 
In our work on distribution the tests were made in a room 
30.5 ft. (9.29 m.) long, 22.3 ft. (6.797 m-) wide, and 9.5 ft. 
(2.895 m.) high. The artificial lighting was accomplished by 
means of two rows of fixtures of four fixtures each. Each row 
was 6 ft. (1.828 m.) from the side wall, and the fixtures were 
6 ft. apart. The reflectors were 29 in. from the ceiling for the 
direct system, and 16 in. for the indirect and semi-indirect. Clear 
tungsten lamps were used as source. The voltage was kept 
constant by means of a voltmeter and a finely graduated wall 
rheostat placed in series with the lighting circuit. 
In order to get the desired variation in the distribution of light 
and surface brightness in the field of vision required for the 
purposes of the test, four types of lighting were selected. One 
may be called a direct system; one an indirect system; one a 
semi-indirect system; and one was the illumination of a room by 
daylight. In case of the direct system, two bulbs making an 
angle of 180 deg. were used for each fixture. Directly above the 
5 Ferree, C. F-, The Streaming Phenomenon, Amer. Jour, of Psychol., 1908, 
19, pp. 484-503; also The Intermittence of Minimal Visual Sensation, Amer. Jour, of 
Psychol., 1908, Vol. XIX, pp. 112-130. 
c The order of magnitude of the mean variation of the test for the fresh eye 
was obtained as follows. Beginning at 9 a. m., five three-minute records were run 
with a rest period of 20 minutes between each test. This was done with all ob- 
servers on several days under each system of lighting employed. The rest period 
was taken in each case in a room lighted by daylight facing a wall with an evenly 
lighted mat surface. For a single series of five tests, the variations of the time 
seen clear in the three-minute period have always fallen within 1 per cent, for all 
of the observers we have used and all systems of lighting. 414 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
lights was fastened a slightly concaved porcelain reflector 16 in. 
in diameter. This type of fixture was not chosen with an especial 
reference to its representative character in any system of com- 
mercial classification. It was chosen rather with reference to 
the purpose of the test. It may be said, however, that it was the 
one in use throughout the building in which the tests were made 
and gives effects very similar to much of the lighting in actual 
use at the present time. In case of the indirect system, corrugated 
mirror reflectors were used enclosed in a brass bowl. For the 
semi-indirect system inverted alba reflectors n in. in diameter 
were employed. The daylight illumination came from three win- 
dows all on one side of the room. These windows were so shel- 
tered that it was never possible for them to receive light directly 
from the sun or from a brilliantly illuminated sky. Moreover, 
the light from one of them, the one nearest the observer, was 
further diffused by passing through a diffusion sash made of 
double thick glass ground on one side. 
In order to get the effect of the distribution factors on the eye's 
loss of efficiency as the result of a period of work, the tests should 
be conducted with the quality and intensity of light made as nearly 
equal as possible. The quality of light was made approximately 
the same for the three installations of artificial light by using 
clear tungsten lamps in each case. It was decided to make the 
intensity of light as nearly equal as possible at the point'of test, 
and to give a supplementary specification of the lighting effects 
in the remainder of the room for the three installations of arti- 
ficial light.7 At the point of test the light was photometered in 
several directions. It was made approximately equal in the plane 
of the test card and as nearly as possible equal in the other direc- 
tions. 
The specification of the lighting effects in the remainder of the 
room has been accomplished as follows. (i) A determination 
7 We have not as yet made the fuller photometric specification of the room 
lighted by daylight with our present arrangement of windows, curtains, etc. We hope 
to make the effect of varying the distribution factors in daylight illumination 
(employing windows, skylights, etc.) the study of a future study. In this study a 
photometric analysis of the illumination effects produced will be made an especial 
feature. FERREE AND RAND: EFFICIENCY OF THE EYE 
415 
has been made of the average illumination of the room under 
each of the three installations. The room was laid out in 3-ft. 
squares, and illumination measurements were made at 66 of the 
intersections of the sides of these squares. Readings were made 
in a plane 122 cm. above the floor with the receiving test-plate of 
the illuminometer in the horizontal, 45 deg. and 90 deg. positions, 
measuring respectively the vertical, 45 deg., and horizontal com- 
ponents. The 122 cm. plane was chosen because that was the 
height of the test object. (2) A determination was made of the 
brightness of prominent objects in the room, such as the test card, 
the reflectors for the semi-indirect installation, book of the ob- 
server, specular reflection from surfaces, etc. The brightness 
measurements were made by means of a Sharp-Millar illumino- 
meter with the receiving test plate removed. The instrument was 
calibrated against a magnesium oxide surface obtained by de- 
positing the oxide from the burning metal on a white card. By 
this method the reflecting surfaces were used as detached test 
plates. The readings were converted into candlepower per sq. in. 
by the following formula: Brightness = Foot-candles/vr X 144. 
(3) Photographs were made of the room from three positions 
under each system of illumination. 
In Fig. 1 (see "Further Experiments on the Efficiency of the 
Eye under Different Conditions of Fighting," Trans, of the Ill. 
Eng. Soc., 1915, X, p. 452a)8 the test room is drawn to scale: Plan 
of room, north, south, east, and west elevations.9 In the drawing 
plan of room, are shown the 66 stations at which the illumination 
measurements were made and the position of the outlets for the 
lighting fixtures A, B, C, D, E, F, G, H. In the drawing, east 
elevation, the position of the observer at one of the points at 
s The present paper is the second one in a series of three on the efficiency of 
the eye under different conditions of lighting. Before it was printed the third 
paper had been read at the eighth annual convention of the Illuminating Engineering 
Society and printed in the papers for that convention. In this paper it had been 
found necessary to repeat some of the data of the second paper for reference. Since 
both the second and third papers are now appearing simultaneously, the data that 
was repeated in the third paper has been omitted from the second. Wherever this 
has been done a cross reference is given to the third paper. 
0 For the scale drawing of the test-room, for the measurements for the direct 
and semi-indirect systems given in Table II, and for the photographs of the test- 
room, we are indebted to Mr. C. W. Jordan of the United Gas Improvement Co. 416 
TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
which the tests were taken is represented.10 The other three 
positions are indicated by X. 
Table I (see Table I, op. cit., p. 454) shows the number and 
wattage of the lamps used at outlets A, B, C, D, E, F, G, H; and 
Table II (see Table II, op. cit., 454-455) gives the illumination 
measurements for each of the 66 stations represented in Fig. 1, 
made with the receiving test plate of the photometer in the hori- 
zontal, vertical, and 45 deg. planes. 
Table III has been compiled as a supplement to Table II for the 
purpose of making a comparative showing of the evenness of il- 
lumination at the 122 cm. level given by the three systems of light- 
ing. Two cases may be made of this : (1) A comparison may be 
made of a given component from station to station; or (2) the 
difference between the components may be compared. To facil- 
itate the comparisons, (a) the mean variation from the average 
of each of the components has been computed; and (&) the dif- 
ference in the averages of the three components has been deter- 
mined. Results for the first of these points are shown in Division 
A of the table; for the second in Division B.11 
10 The track along which the test card was moved was parallel to the east and west 
walls of the room. During the three hours of reading which intervened between 
the two tests the observer moved just far enough back from the upright supporting 
the mouthboard to give room for the book to be held and to permit of a comfortable 
reading position. The book was elevated and held approximately at an angle of 45 deg. 
When taking the test, the observer _ faced the north wall of the room, in such a 
position that with the eyes in the primary position, the lines of regard were parallel 
with the east and west walls of the room. Care was taken-to have print of uniform 
size and distinctness for use with the three systems, and to have a page which 
gave a comparatively small amount of specular reflection. The brightness values of 
the page in the horizontal and 45 deg. positions for the three systems, are given in 
the legends for Figs. 8, 9, and 10. 
11 It would be interesting to make this comparison for- other levels in the room 
and for a greater number of components. But unfortunately we have not been able 
to make the number of measurements needed for this comparison. The evenness of 
the illumination, it will be remembered, is not only of importance to the efficiency 
of the eye with reference to the object directly viewed, but also in its influence on 
the distribution of surface brightness. The evenness of surface brightness depends 
in general upon two sets of factors: (a) the nature and position of the reflecting 
surfaces in the room; and (b) the type of delivery of light to these surfaces. 
We realize that the evenness of the illumination on the 122 cm. plane given by 
the indirect and semi-indirect units was somewhat interfered with by the reflectors 
of the direct system which were beneath and a little to the right of these units 
when in position for the test. Also the evenness of surface brightness on the 
ceiling for the direct system was interfered with by the indirect and semi-indirect 
reflectors, which were above and a little to the side of the direct units. The in- 
fluence of this "dead apparatus" will be eliminated in the next series of installa- 
tions. Moreover, the installation in each case was not such as to give the best 
effects obtainable from the type of reflector used. For example, the indirect re- 
flectors were too close to the ceiling to give the maximum evenness of illumination 
and of surface brightness for the type of reflector used. The above analysis of 
effects is, therefore, not made for the purpose of drawing general conclusions with 
regard to the type of reflector employed. It is made solely for the sake of the 
comparison of the illuminating effects obtained with the corresponding results for 
loss of efficiency. TABLE III.12-(Distribution Series). 
Compiled from Table II to show a comparison of the evenness of the illu- 
mination at the 122 cm. level given by the direct, semi-indirect, and indirect 
systems. Division A shows the mean variation from the average for each of 
the three components of illumination ; Division B, the difference in the 
average value of the three components. 
Division A. 
System 
Mean variation of the components 
Percentage of mean variation of 
components 
Vertical 
Horizontal 
45° 
Vertical 
Horizontal 
45° 
Direct 
1.88 
I.09 
i-53 
38% 
47% 
32% 
Semi-indirect ■. 
1.68 
0.66 
1.32 
39% 
42% 
36% 
Indirect 
1.1 
0.4 
0.61 
30% 
37% 
19% 
Division B. 
System 
Difference between components 
Percentage of difference between 
components 
Vertical 
and 
Horizontal 
Vertical 
and 
45° 
45° 
and 
Horizontal 
Vertical 
and 
Horizontal 
Vertical 
and 
45° 
45° 
and 
Horizontal 
Direct 
2.68 
O.23 
2-45 
54% 
5% 
51% 
Semi-indirect .. 
2.68 
0.64 
2.04 
63% 
15% 
56% 
Indirect 
2.13 
O.3I 
I.82 
59% 
9% 
55% 
12 For Tables I and II, see Tables I and II, Further Experiments on the 
Efficiency of the Eye, etc., Trans. I. E. S., 1915, Vol. X, pp. 454-455. Fig. 2.-Showing the test room illuminated by the direct system. The photograph was- 
taken from the south end of the room at a point 4 ft. from the west wall. 
Fig. 3.-Showing the test room illuminated by the semi-indirect system. The photograph, 
was taken from the south end of the room at a point 4 ft. from the west wall. Fig. 4.-Showing the test room illuminated by the indirect system. The photograph 
was taken from the south end of the room at a point 4 ft. from the west wall. 
Fig- 5--Showing the illumination effects for the west wall of the room, direct system. Fig. 6.-Showing the illumination for the west wall of the room, semi-indirect system. 
Fig. 7.*-Showing the illumination effects for the west wall of the room, indirect system. 
* For Figs. 8, 9, and 10, see Figs. 2, 3, and 4, " Further Experiments on the Efficiency of 
the Eye, etc." Trans, of the I. E. S., 1915, Vol. X, pp. 452a-452b. FERREE and rand: FFFICIFNCY OF THE FYF 
417 
In Figs. 2 to io are given photographs showing the illumination 
of the room and the distribution of surface brightness for the 
three systems. Figs. 2, 3 and 4 are taken from the south end of' 
the room at a point 4 ft. from the west wall. These photographs 
were taken so as to comprehend as much of the room as was pos- 
sible in one view. They include the greater part of the ceiling, 
floor, and north wall; six of the fixtures; and about one-half of 
the east wall. The difference in surface brightness for the various 
points of the room (including the lighting units) is, it will be 
noted, greatest for the direct system, next greatest for the semi- 
indirect system, and least for the indirect system. The indirect 
and semi-indirect reflectors were attached to arms of approxi- 
mately equal length which could be revolved about the fixture 
stem as an axis. When the tests were taken, these reflectors were 
turned in each case to the inside position indicated in the photo- 
graph, the object being to have the two types of reflectors as 
nearly as possible in the same position in the field of vision for the 
comparative tests. The direct fixtures, it will be noted, were below 
and slightly outside this position. In our next series of experi- 
ments, arrangements have been made such that the reflectors can 
be placed in exactly the same position for each type of installation 
when it suits the needs of the experiment to have it so. The 
slight deviation from exact coincidence found in these experi- 
ments is, however, perhaps of no great consequence for the pur- 
pose of the present work especially in the case of the indirect and 
semi-indirect reflectors. In Figs. 5, 6 and 7, are represented the 
illumination effects for the west half of the room. These photo- 
graphs show the distribution of light and shade on the greater part 
of the west wall, and the adjacent ceiling, and include two of the 
fixtures. In Figs. 8, 9 and 10 (see Figs. 2-4, op. cit., pp. 452a-4£2b) 
are shown the brightness measurements of all surfaces having 
very high or very low brilliancy. The spot measured is indicated 
by a cross, and the numerical value of the brightness measurement 
in candlepower per square inch is printed nearby. These spots 
are also lettered for convenience of reference in the intensity 
series. That is, since several installations were used in the in- 
tensity series it was found convenient to express these values in 
tabular form and to identify them with the surfaces measured 420 
transactions of ieeuminating engineering society 
very small indeed. This seems to indicate (a) that for the scale 
of magnitudes present in this series of experiments, the gradation 
of surface brightness for the indirect system is very close to what 
the eye is prepared to stand without loss of efficiency; and (&) 
that an increase in differences in brightness above this point is 
followed at first by a rapid increase in loss of efficiency and 
later by a much slower increase. In the intensity series the 
following points also come out. (i) The effect of size of ratio 
on loss of efficiency is different for different orders of magnitude 
of brightness. That is, for the range of scale of magnitudes we 
have used, the lower is the order of magnitude, the greater is the 
ratio that is permissible. And (2) the size of the brilliant object 
as well as its brilliancy is of importance. That is, within certain 
limits, as yet undefined, an increase in the area of the brilliant 
surface causes an increase in loss of efficiency. 
Supplementary to Tables IV-VII we have computed for the 
three systems the mean variation of the several brightness values 
from their average values. While important from the standpoint 
of showing the variation from the mean for the different systems, 
such a comparison is, however, probably not so important from 
the standpoint of the eye as are the comparisons given in Tables 
IV-VII. That is, from the standpoint of effect on the eye it is 
probably more important to give a representation of the bright- 
ness of individual surfaces, more especially of surfaces showing 
extremes of brightness, than it is the mean variation from the 
average brightness of all the surfaces. In order to make possible 
the comparison with and without the source and the spot above 
the source, the table is made to show separately the mean varia- 
tion for the following measurements: (0) for all; (&) for all 
but the source; and (c) for all but the source and the spot above 
the source. Results are given in Table VIII. 
Obviously the effect of these installations on the eye's ability 
to maintain its efficiency for a period of work will vary with 
the position of the observer in the room. The tests have been 
made, therefore, at four positions: one in which six fixtures were 
in the field of view, one in which four were in the field of view, 
one in which two were in the field of view, and one in which 
none were in the field of view. This variation of position at FERREE AND rand: EFFICIENCY OF THE EYE 
421 
which the observation was made accomplishes two purposes, 
(i) It gives us a more representative idea of the difference in 
the effect on the eye of the four types of lighting. And (2) it 
shows the effect of varying the number of surfaces showing 
brightness differences, particularly the number of primary 
sources in the field of view. 
TABLE VIII.14-(Distribution Series). 
Compifed from Table IV to show the mean variations in surface bright- 
ness for the direct, semi-indirect, and indirect systems.15 
Measurements 
considered 
Mean variation for the three 
systems 
Percentage of mean variation for 
the three systems 
Direct 
Semi- 
indirect 
Indirect 
Direct 
Semi- 
indirect 
Indirect 
All 
All but the 
94-977 
0.075 
O.O235 
189% 
145% 
135% 
source 
All but the 
source andthe 
spot abovethe 
0.0018 
0.01817 
0.0235 
33% 
120% 
135% 
source 
0.0016 
O.OOI3 
O.OOI2 
32% 
30% 
35% 
14 For Tables IV-VII, see Tables III-VI, Further Experiments on the Ef- 
ficiency of the Eye, etc., Trans. I. E. S., 1915, Vol. X, pp. 457, 463-464. 
15 It is scarcely necessary to point out that the above results seem to indicate 
that the great advantage of the indirect over the other systems of lighting we have 
used with regard to the factor: evenness of surface brightness, comes primarily at 
least from its provision for shielding the eye from the light source rather than 
from any conspicuously greater evenness of illumination given by it to the objects 
in the field of view. In fact all of the systems give a fairly even distribution of 
surface brightness outside of the source and the surfaces immediately surrounding 
it. 
The need of keeping the surface brightness within certain limits and the 
primary importance of properly shielding the eye from the source to the accom- 
plishment of this desideratum are both obvious Doubtless many ways will be de- 
vised in course of time for cutting down useless and harmful brightness differences in 
lighting effects. For example, the possibility is here suggested of producing a still 
smaller brightness difference than is given by the indirect reflectors of the type we 
have employed, by using semi-indirect reflectors of such a density as to give a 
surface brilliancy equal to that of the spot of light cast upon the ceiling. The 
.value of this brilliancy, because of the larger area of luminous surface presented, 
could then be made smaller than that of the ceiling spot cast by the indirect 
reflector and still give the same amount of light to the room. A similar effect 
may be obtained with the indirect reflector by using lamps of lower wattage and 
adding the light needed to make up the deficiency by installing directly beneath the 
reflector lamps of low wattage in translucent enclosures of a density that will 
give a surface brilliancy equal to that of the ceiling spots. The effect of both 
of these devices would be to lower the surface brilliancy for a given light flux 
by increasing the area of the luminous surface. Whether either would be advisable 
from other standpoints we are not at present prepared to say. 422 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
Results will be given in this paper only for the position with 
six fixtures in the field of view. The results for the other posi- 
tions will be given in a later paper. When working at the posi- 
tion with six fixtures in the field of view, our tests show that the 
eye loses practically nothing in efficiency as the result of three to 
four hours of work under daylight, it loses enormously for the 
same period of work under the direct installation, and almost as 
much under the semi-indirect installation. Under the 'indirect 
installation the eye loses a little more than under daylight, but 
not nearly so much as under the other installations. 
The results of the work on distribution are given in Tables 
IX and X. Early in the work it was found that nearly as 
much difference in result was gotten for two as for three hours 
of work. In Table IX is shown the loss in efficiency for Ob- 
server R for three hours of work under the four systems; and 
in Table X, the loss in efficiency for Observer G for two hours 
of work. These tables are typical of the results obtained from 
all of our observers for these periods of work.10 Column I of 
these tables gives the type of lighting system. Column 2 gives 
the total wattage of the lamps used, and Column 3 the voltage at 
which these lamps were operated. Columns 4, 5, and 6 give the 
foot-candles of illumination at the point of work measured re- 
spectively in the horizontal, vertical, and 45 deg. planes. Column 
7 gives the maximal distance at which the test object could be 
seen clearly, and Column 8 the distanc'e chosen at which to con- 
duct the test for loss of efficiency. Care was taken in every case 
to choose this working distance of such a value that the ratio it 
sustained to the maximum distance was always approximately 
the same. Column 9 gives the total time the test object was seen 
clear in the three minutes of observation and Column 10 the total 
time it was seen blurred. Column 11 gives the ratio of the total 
time seen clear to the total time seen blurred, and Column 12 
gives the comparative values of these ratios in terms of a com- 
mon standard. These ratios were reduced to a common scale or 
standard in order to make the comparison of the amounts of 
16 Obviously in the consideration of the effect of a given lighting system on 
the ability of the eye to hold its efficiency for a period of work, the age of the 
observer and the condition of his eyes should be taken into account. For a full 
clinic report of the eyes of the observers employed, see op. cit., foot-note 14, p. 460. FpRRpp AND RAND: PFFICIPNCY OF THp pYp 
423 
change in their ratios easier. They express the comparative 
ability of the eye to sustain its power of clear seeing for three 
minutes before and after work for the four conditions of light- 
ing used. 
It will also be noted from Column 8 of the above tables that 
the visual acuity tests show that acuity of vision as determined by 
the momentary judgment is higher for the same foot-candles of 
illumination under daylight than under artificial light, and of the 
artificial lights it is very slightly highest for the indirect system, 
next highest for the semi-indirect system, and slightly lowest for 
the direct. It will thus be seen that for all the purposes of clear 
seeing, whether the criterion be maximal acuity or the ability of 
the eye to hold its efficiency for a period of work, the best re- 
sults are given in order by the systems that give the best distri- 
bution. The effect of distribution, however, on the ability of the 
fresh eye to see clearly, is not nearly so great as it is on its power 
to hold its efficiency for a period of work. 
In order to give a typical representation in graphic form of 
the effect on the efficiency of the eye of a period of work under 
these four conditions of lighting, the results of the above tables 
will also be given in the form of a chart made up of straight 
lines showing in each case the loss of efficiency from beginning 
to close of work. In constructing these charts, the length of time 
of work is plotted along the abscissa, and the ratio of the time 
the test object is seen clear to the time it is seen blurred is 
plotted along the ordinate. Each one of the large squares along 
the abscissa represents one hour of work and along the ordinate 
an integer of the ratio. Chart A shows the results for Table IX, 
and Chart B for Table X. An inspection of these charts will 
show how widely different in amount is the loss in efficiency 
under the specified conditions for the direct and semi-indirect 
systems as compared with the indirect system and daylight, and 
how close is the correspondence between the results for the 
direct and semi-indirect system, and between the results for the 
indirect system and daylight. 
The loss in efficiency found in the above work seems to be 
predominantly, if not entirely, muscular, for the tests for the 
sensitivity of the retina show practically no loss in sensitivity 424 
transactions of irruminating ENGINEERING society 
TABLE IX.-(Distribution Series). 
Observer R, showing the eye's loss in efficiency as the result of 3 hours of work under the systems of indirect, semi-indirect, 
and direct lighting employed as compared with daylight. 8 lamps, indirect and semi-indirect systems; 16 lamps, direct sys- 
tem. Intensities equalized on test card. Clear tungsten lamps. 
fighting system 
Watts 
Volts 
Foot-candles 
Time 
Maximal 
distance 
at which 
test object 
can be 
seen 
clear 
Working 
distance 
Total 
time 
clear 
Total 
time 
blurred 
Total 
time 
clear 
total 
time 
blurred 
Ratios 
reduced 
to com- 
mon 
standard 
Hori- 
zon- 
tal 
Verti- 
cal 
45° 
Daylight 
- 
- 
5-5 
1-32 
4-2 
9 
12 
A.M. 
M. 
90.0 
90.0 
74-o 
74-o 
137 
135 
43 
45 
3.18 
3-0 
3 5 
3-3 
Indirect 
800 
107 
5-2 
1.36 
3-5 
9 
12 
A.M. 
M. 
84-5 
84-5 
67-5 
67-5 
135 
132 
45 
48 
3-o 
2-75 
3-5 
3-2 
Semi-indirect 
760 
107 
5-8 
i-45 
4-0 
9 
12 
A.M. 
M. 
80.5 
79-5 
68.5 
68.5 
142 
92 
38 
88 
3-73 
1.04 
3-5 
0-97 
Direct 
880 
107 
4.2 
1.41 
2.6 
9 
12 
A.M. 
M. 
81.0 
78.0 
68.0 
68.0 
139 
71 
4i 
109 
3-39 
0.65 
3-5 
0.67 
TABLE X.-(Distribution Series). 
Observer G, showing the eye's loss in efficiency as the resultof 2 hours of work under the systems of indirect, semi-indirect, 
and direct lighting employed as compared with daylight. 8 lamps, indirect, and semi-indirect systems; 16 lamps, direct sys- 
tem. Intensities equalized on the test card. Clear tungsten lamps. 
Daylight 
Indirect 
800 
107 
5-5 
5-2 
1-32 
1.36 
4-2 
3-5 
9 A.M. 
II A.M. 
9 A.M. 
II A.M. 
91.0 
91.O 
85.O 
85.O 
77-o 
77-o 
72.0 
72.0 
161 
160 
159 
i57 
19 
20 
21 
23 
8-45 
8.00 
7-57 
6.83 
3-50 
3-3i 
3-50 
3-15 
Semi-indirect 
760 
107 
5-8 
i-45 
4.0 
9 A.M. 
II A.M. 
83.O 
8l.O 
72.0 
72.0 
i59 
150 
21 
30 
7-55 
5-oo 
3-50 
2.30 
Direct .. 
880 
107 
4.2 
1.41 
2.6 
9 A.M. 
II A.M. 
8l.O 
78.O 
69-5 
69-5 
i45 
I31 
35 
49 
4-14 
2.67 
3-5° 
2.20 FERREE AND RAND: EFFICIENCY OF THE EYE 
425 
Chart A (Observer R).-Showing the eye's loss in efficiency as the 
result of three hours of work under the systems of indirect, semi-indirect 
and direct lighting employed as compared with daylight. 
Foot-ca ndles 
lighting system 
Watts 
Volts 
Horizontal 
Vertical 
45° 
A-Daylight 
. - 
5-5 
1-32 
4.2 
B-Indirect 
■. 8ob 
107 
5-2 
I.36 
3-5 
C-Semi-indirect . 
• • 760 
107 
5-8 
1-45 
4.0 
D-Direct 
.. 880 
107 
4.2 
I.41 
2.6 
Chart A. 
Chart B (Observer G).-Shows the eye's loss in efficiency as the result 
of two hours of -work under the systems of indirect, semi-indirect and direct 
lighting employed as compared with daylight. 
Foot-candles 
lighting system 
Watts 
Volts 
Horizontal 
Vertical 
45° 
A-Daylight 
.. - 
- 
5-5 
I.32 
4.2 
B-Indirect 
• 800 
107 
5-2 
I.36 
3-5 
C-Semi-indirect . 
• • 760 
107 
5-8 
1-45 
4.0 
D-Direct .... 
.. 880 
107 
4.2 
I.41 
2.6 
Chart B. 426 
TRANSACTIONS OF IREUMINATING ENGINEERING SOCIETY 
as the result of work under any of the installations employed.17 
The following reasons are suggested why the muscles of the eye 
giving both fixation and accommodation should be subjected to 
a greater strain by the direct and semi-indirect installations than 
by the indirect installation or daylight, (i) The bright images 
of the sources falling on the peripheral retina, which is in a 
perpetual state of darkness-adaptation as compared with the 
central retina, and is, therefore, extremely sensitive in its re- 
action to such intensive stimuli, set up a reflex tendency for the 
eye to fixate them instead of, for example, the letters which the 
observer is required to read. (2) Likewise a strong reflex tend- 
ency to accommodate for these brilliant sources of light, all at 
different distances from each other and the lettered page, is set 
up. (3) These brilliant images, falling on a part of the retina 
that is not adapted to them, causing as they do acute discomfort 
in a very short period of time, doubtless induce spasmodic con- 
tractions of the muscles which both disturb the clearness of 
vision and greatly accentuate the fatiguing of the muscles. The 
net result of all these causes is excessive strain which shows 
itself in a loss of power to do work. In the illumination of a 
room by daylight, however, with a proper distribution of win- 
dows, the situation is quite different. The field of vision con- 
tains no bright sources of light to disturb fixation and accommo- 
dation and to cause spasmodic muscular disturbances due to the 
action of the intensive light sources on the dark-adapted and 
sensitive peripheral retina. As we have already pointed out, 
the light waves have suffered innumerable reflections and the 
light has become diffuse. The field of vision is, comparatively 
speaking, uniformly illuminated, and there are no extremes of 
surface brightness. The illumination of the retina, therefore, 
falls off more or less gradually from center to periphery, as it 
should to permit of fixation and accommodation for a given 
object with a minimum amount of strain. 
17 In the next paper of the series it is shown (see op. cit. pp. 484-490) that the 
loss in muscular efficiency is confined largely to the accommodation muscles. The 
fixation muscles apparently suffer little loss for the period of work we have used. pprrel and rand : pfficilncy op ths pyE 
427 
III. THE EFFECT OF VARIATION IN THE INTENSITY OF 
LIGHT ON THE EFFICIENCY OF THE EYE FOR A 
PERIOD OF WORK. 
It is not our purpose, however, to contend that distribution 
of light and surface brightness in the field of vision is the only 
factor of importance in the illumination of a room. The in- 
tensity and quality of light must also be taken into account. For 
example, one of the most persistent questions asked by the il- 
luminating engineer is: "How much light should be used with a 
given lighting installation to give the best results for seeing?" 
We have undertaken, therefore, to determine the most favor- 
able range of intensity for the four types of lighting we have 
used. Our work shows in general the following results. A 
very wide range of intensity is permissible for daylight, and a 
comparatively wide range for the indirect installation. For 
the semi-indirect installation the eye fell off heavily in efficiency 
for all intensities with exception of a narrow range on either 
side of 2.2 foot-candles measured at the level of the eye at the 
point of work with the receiving surface of the photometer in 
the horizontal plane. For the direct installation no intensity 
could be found for which the eye did not lose a very great deal 
in efficiency as the result of work. Thus it seems that the factors 
we have grouped under the heading distribution are funda- 
mental. That is, if the light is well distributed and diffuse, as it 
was in case of the daylight and indirect installations we used, 
and there are no extremes of surface brightness, the ability of 
the eye to hold its efficiency is, within limits, independent of 
intensity. In short, the retina is itself highly accommodative 
or adaptive to intensity, and if there is the proper distribution 
and diffuseness of light and the proper gradation of surface 
brightness in the field of vision, the conditions are not present 
which cause strain and consequent loss in efficiency in the ad- 
justment of the eye. The results of this series of tests, then, 
accomplish two purposes. (I) They show that when the dis- 
tribution and diffuseness of light and the distribution of surface 
brightness in the field of view are properly taken care of, the 
eye, so far as the problem of lighting is concerned, is practically 
independent of intensity. And (2) they show the effect on the 
efficiency of the eye of the variations in surface brightness pro- 428 
transactions of illuminating engineering society 
duced by varying intensity in case of the direct and semi-indirect 
installations we have used. 
The tests were made in the same room, with the same fixtures, 
and in general with the same conditions of installation and meth- 
ods of working as were described in the work on distribution. 
To secure the various degrees of intensity needed, lamps of 
different wattage were used. These were selected from a series 
of tungsten lamps ranging from 15-100 watts. In order to keep 
the distribution factor as nearly constant as possible for a given 
type of system, the lamps used in making the test for that type 
of system were all of one wattage, i. e., all 15's, 25's, 40's, 6o's, 
or 100's. 
For the semi-indirect system the total range of intensity of 
illumination employed is shown by the following figures. The 
series was begun with 25-watt lamps18 and consisted of 25, 40, 
60, and 100-watt lamps.19 For the 25-watt lamps the photometer 
reading at the point of work with the receiving test plate of 
the photometer in the horizontal plane, showed 1.6 foot-candles; 
with the test plate in the vertical position 0.45 foot-candle; in 
the 45 deg. position 1.15 foot-candles. For the 100-watt lamps, 
6.8 foot-candles were obtained with the test plate horizontal; 
1.82 foot-candles with the test plate vertical; and 4.5 foot-candles 
with it in the 45 deg. position. The tests for loss in efficiency29 
showed that the intensity most favorable to the eye was secured 
13 Since the most favorable intensity was given by the 40-watt lamps and since 
the 15-watt lamps gave so little light as to be extremely trying to the eyes, it was 
thought best to begin the series with the 25-watt lamps instead of the 15 as was 
done in case of the direct system. 
19 Owing to their smaller size, socket extenders had to be used for the 25 and 
40-watt lamps. That is, without the extenders these lamps came so low in the re- 
flector as to change the distribution effects given by the reflector. 
20 In conducting these tests it was found necessary to allow a period of 
adaptation without work to the illumination of the room before the first test was taken. 
If this were not done, especially in case of the lower intensities of light used, the 
changing sensitivity of the eye to the intensity of light employed produced a notice- 
able change in the visual acuity between the times the tests before and after work were 
taken. Since the distance of the test card was kept the same for the two tests, this 
change in the visual acuity tended to influence the ratio: time clear to time blurred. 
To determine the length of time needed with a given intensity of light to insure a con; 
stant acuity so far as adaptation is concerned, preliminary tests were made as follows. 
The acuity of the observer was taken every three minutes until no noticeable 
change was found. This length of time was then always allowed for that observer 
as an adaptation period prior to the loss of efficiency test conducted for the intensity 
of illumination. FERREE AND RAND: EFFICIENCY OF THE EYE 
I 
429 
when the photometric reading with the test plate in the horizontal 
plane showed, 2.2 foot-candles; in the vertical plane, 0.58 foot- 
candle; and in the 45 deg. plane, 1.52 foot-candles. The total 
wattage in this case was only 320. r\t this intensity of illumin- 
ation the semi-indirect installation, so far as its effect on the eye 
is concerned, compares very favorably with the indirect installa- 
tion at such ranges of intensity as we have employed. At in- 
tensities appreciable higher than this most favorable value, 
however, or appreciably lower, the loss in efficiency is very great. 
At the intensity commonly recommended in lighting practise, this 
semi-indirect installation is almost, if not quite as damaging as 
the direct installation. The intensity recommended by the Il- 
luminating Engineering Society, for example, in its primer issued 
in 1912, ranges from 2-3 to 7-10 foot-candles, depending upon 
the kind of work; 5 foot-candles is taken as a medium value. 
This medium value is more than double the amount we have 
found to give the least loss in efficiency for the type and installa- 
tion of semi-indirect lighting we have used. The intensity we 
have found to give the least loss in efficiency for this type of 
lighting does not, however, give maximal acuity of vision as 
determined by the momentary judgment. At an intensity that 
does give maximal acuity of vision as determined by the momen- 
tary judgment, the eye runs down rapidly in efficiency. That 
is, in this type of lighting one or the other of these features must 
be sacrificed. High acuity and little loss in efficiency can not 
both be had at the same intensity. These features can both be 
had only under daylight and, in case of the installations, we used, 
with the indirect system. However, the amount of light we 
find to give the least loss in efficiency seems to be sufficient for 
much of the work ordinarily done in the office or home. It is 
not enough, though, for drafting or other work requiring great 
clearness of detail. By giving better distribution effects this sys- 
tem is supposed also to be a concession to the welfare of the eye, 
but our tests show that this concession is not so great as it is 
supposed to be. In fact, installed at the intensity of illumination 
ordinarily used, or at an intensity great enough for all kinds of 
work, little advantage is gained for the eye in this type of light- 
ing with reflectors of low or medium densities; for with these 430 
TRANSACTIONS OF ILLUMINATING LNGINPFRING SOCIETY 
intensities of light and densities of reflector, the brightness of the 
source has not been sufficiently reduced to give much relief to the 
suffering eye. Until this is done in home, office, and public 
lighting, we can not hope to get rid of eye strain with its complex 
train of mental and physical disturbances. If the semi-indirect 
principle of lighting is to be used with benefit to the eye, a density 
of reflector and type of installation must be employed that will 
give a gradation of brightness in the field of view in conformity 
with the limits of difference that the eye can stand without loss in 
efficiency or comfort. 
In case of the direct system of lighting, we were able to im- 
prove the conditions so far as loss of efficiency of the eye is 
concerned, by reducing the intensity; but this system never proved 
to be so favorable in this regard as even the semi-indirect system. 
In the tests made under the direct system care was taken to have 
the fixtures as nearly as possible in the same position as they 
were for the semi-indirect system. Our fixtures for the direct 
system were so installed that either one or two lamps could be 
used in each fixture, totalling respectively 8 and 16. In order 
to get a wider range of intensity both numbers of lamps were 
used, i. e., one series was made with 8 lamps and aonther with 
16. Four intensities of light were used in each case. These 
intensities were secured in the 8-lamp system by using lamps 
totalling 120, 200, 320, and 480 watts. The foot-candles at the 
point of work ranged from 0.64 with the receiving test plate of 
the photometer in the horizontal, 0.32 in the vertical, and 0.49 
in the 45 deg. position with the lamps totalling 120 watts, to 2.6 
with the test plate in the horizontal, 1.02 in the vertical, and 2.0 in 
the 45 deg. position with the lamps totalling 480 watts. The four 
intensities were secured in the 16-lamp system by using lamps 
totalling 240, 365, 400, and 880 watts. The foot-candles at the 
point of work with the 16-lamp system ranged from 1.23 with 
the test plate in the horizontal, 0.54 in the vertical, and 0.935 in 
the 45 deg. position with the lamps totalling 240 wattts, to 4.2 
with the test plate in the horizontal, 1.41 in the vertical, and 2.6 
in the 45 deg. position with the lamps totalling 880 watts. The 
most favorable intensity was secured by an installation that gave 
1.16 foot-candles with the test plate in the horizontal, 0.45 in the FERREE AND rand: efficiency of the Eye 
431 
vertical, and 0.85 in the 45 deg. position. This intensity was 
given by the 8-lamp system with a total wattage of 200. At this 
intensity, however, the loss in the efficiency of the eye for three 
hours of work was almost four and one-half times as great as 
for the most favorable intensity for the semi-indirect system; and 
more than four and one-half times as great as for a wide range 
of intensities for either the indirect system or daylight. 
The following specification was made of the illumination effects 
for the intensity series. (1) Illumination measurements were 
made for the highest intensity employed at the 66 stations in 
the test room. These measurements were made in the way de- 
scribed in the preceding section. For the other intensities em- 
ployed, measurements were made at 9 representative stations to 
show in a general way the order of magnitude of reduction 
produced by using the lamps of lower wattages. (2) Brightness 
measurements were made of prominent objects in the room, such 
as the test card, the book of the observer, and all surfaces show- 
ing very high or very low brilliancy, for all intensities for all 
systems. 
In Table XI are given the illumination measurements for the 
highest wattages used made with the receiving test plate of the 
photometer in the horizontal, vertical, and 45 deg. planes. Tables 
XII, XIII and XIV show the illumination measurements for 
the other wattages employed in the series at nine representative 
stations. These measurements are intended to show the order 
of magnitude of reduction of the illumination of the room pro- 
duced by using the lamps of lower wattage. They conform in 
each case pretty closely, it will be noted, to the simple ratio 
of the wattages employed. Tables XV, XVI and XVII give 
the brightness measurements for these installations for the dif- 
ferent intensities used. The points at which the measurements 
were taken are indicated by the letters A, B, C, D, E, F, etc., 
see Figs. 8 and 9. In Tables XVIII, XIX and XX are given the 
prominent brightness ratios for the different intensities used. 
It was stated in the preceding section that the order of magnitude 
of the brightness scale exerts an influence on the effect of bright- 
ness ratio' on the eye's loss of efficiency. This influence is readily 
seen on comparing the results of Table XVIII with those of 432 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
Table XXI. That is, while the various brightness ratios remain 
pretty much the same for the different intensities of light em- 
ployed, the least loss of efficiency was given by the 40-watt lamps. 
This loss was, for example, very much less than was given by the 
100-watt lamps, not quite so much less than was given by the 
60-watt lamps, and very little less than was given by the 25 watt 
lamps. The loss in efficiency for the 25-watt lamps can also 
doubtless be attributed in part to an insufficient amount of light. 
At least the testimony of the various observers was that not 
enough light was given by these lamps for ease and comfort 
in reading. The results of these experiments seem to show, then, 
that a given order of magnitude of brightness difference in the 
field of view has more effect on the efficiency of the eye when 
the general scale of brightness values is higher than when it is 
low. 
A comparison of Tables XIX and XX with Tables XXII and 
XXIII shows the influence of the area of the bright surface on 
the ability of the eye to hold its efficiency for a period of work. 
For example, although it is shown in Table XIX that the ratios: 
lightest surface to darkest surface, and lightest surface to test 
card and reading page, are greater in the 16-lamp system for the 
15 than for the 25-watt lamps, Table XXII shows that greater 
loss of efficiency is caused by the 25-watt lamps. Similarly, 
Table XX shows that in the 8-lamp system these ratios are 
greater for the 25 and 40 than for the 60-watt lamps, while 
Table XXIII shows that the 60-watt lamps cause the greater loss 
in efficiency. This may be explained as follows. The brightest 
surfaces in the field of vision for the direct system are the fila- 
ments of the lamps. The brightness measurements given in the 
table are in terms of candlepower per square inch. The candle- 
power per square inch is the same, for example, for the filaments 
of the 15 as for those of the 25-watt lamps. But since the darkest 
surfaces, the test card, and the reading page, are darker for the 
15-watt than for the 25-watt system, the ratios: lightest to 
darkest surface, lightest surface to test card, and lightest sur- 
face to reading page, are greater for the 15 than for the 25-watt 
system. While, however, the candlepower per square inch is the 
same for the 15 as for the 25-watt filaments, the actual candle- FERREE AND RAND ! EFFICIENCY OF THE EYE 
433 
power is less for the 15-watt filaments because of their smaller 
area of surface. That is, the area of the brilliant surface or 
in terms of luminous effects, its actual candlepower must be 
taken into account in estimating the effect on the eye as well as 
the candlepower per square inch. The effect of area on sensa- 
tion is well known in physiological optics (for example, see 
Abney, Philos. Trans., 1897, CXC, A, p. 169), and is expressed 
in the law that within limits an increase of area of the stimulus 
functions as an increase of intensity, although not in a simple 
ratio. Apparently, too, in its effect on the eye's power to main- 
tain its efficiency for a period of work, an increase of area of 
the brilliant surface also functions within limits as an increase 
in intensity.21 Ratios expressed in candlepower per square inch 
do not seem therefore, in all cases to be an adequate specification 
of surface brightness, so far as its effect on the efficiency of the 
eye is concerned, unless the areas compared be the same. 
21 The above explanation is, however, not complete. It shows only that the 
ratios: lightest to darkest surface, and lightest surface to test card and reading 
page, are greater for the 15 than for the 25-watt lamps because the candlepower 
per square inch not the actual candlepower was used in computing the ratios. 
We are not at present able to give the ratio of actual candlepower of lightest to 
darkest, lightest to reading page, etc., because we did not measure the actual candle- 
power of the darkest surface, the reading page, etc., only the candlepower per 
square inch. However, since the test card and the reading page were of the same area 
in case of the different intensities, and the darkest surface of approximately the same 
area, ratios based on the total candlepower of the lightest surface (the lamp 
filament) and the candlepower per square inch of the darkest surface, test card, and 
reading page have comparative values. These ratios are very little different for the 
15- and 25-watt lamps. That is, the ratio lightest to darkest for the 15-watt 
lamps = 28,698, for the 25-watt lamps = 28,933; lightest to test card for the 
15-watt lamps = 11,828, for the 25-watt lamps = 12,616; lightest to reading page 
for the 15-watt lamps = 7,352; for the 25-watt lamps = 7,483. 
A complete explanation of the result will doubtless involve two factors (1) the 
ratio of the actual candlepower of the lightest and darkest surfaces; and (2) the 
point brought out in connection with Tables XVIII and XXI, namely that a given 
order of magnitude of brightness difference in the field of view has more effect on 
the loss of efficiency of the eye when the general scale of brightness values is high 
than when it is low. From this we would expect, for example, that if the ratio 
lightest to darkest surface and lightest surface to test card and reading page were 
equal or approximately so for the 25- and 15-watt lamps, for example, the greater 
loss of efficiency should come with the lamps of higher wattage. Similarly for the 
8-lamp system, the 60- and 40-watt lamps should cause a greater loss of efficiency 
than the 25-watt lamps. The 15-watt lamps with this system gave too little light to 
read with ease and comfort hence are ruled out of count in the comparison. 
For investigating in detail the effect of area of the brilliant surface on the eye's 
loss of efficiency, the campimeter may prove of convenience and of service. This is 
one of the instances where the abstract may be used to advantage to supplement the 
concrete method of investigation. (See Memorandum on the Report of the Research 
Committee, Trans. I. E. S., 1914, Vol. IX, No. 4, p. 358.) 
The great difficulty with the abstract type of investigation, as the writers see the 
case at this time, is that a determination of what is permissible with regard to one 
factor in isolation may not be at all permissible in conjunction with other factors. 
A more feasible plan seems to us to be to vary the factor over a certain practical 
range in an actual concrete situation. By a proper selection of the concrete situations 
employed the ground of all that is practicable in lighting can be covered, and the 
results obtained can have a safe application. 434 
transactions or illuminating engineering society 
TABLE XI.-(Intensity Series.) 
Showing the illumination measurements in foot-candles at each of the 66 stations represented in Fig. 1 for the highest 
wattage used in the intensity series for the semi-indirect system, and the direct system (16 lamps and 8 lamps). 
Direct 
(16 lamps) 
400 watts 
I 1 1 1 1 1 1 1 1 I | <n mco cm 0 0 cm in m-co 0 0 
1 1 1 1 1 1 1 1 1 1 1 M-cn o> ^00 io h ncm cm no 
0 
in 
Direct 
(8 lamps) 
480 watts 
III 1 II 1 1 III 
Semi- 
indirect 
800 watts 
1.00 
2.85 
3-40 
3-40 
3.60 
i-74 
1.11 
i-94 
3.80 
4-30 
4.10 
4-35 
4.20 
Direct 
(16 lamps) 
400 watts 
I 1 1 1 I 1 1 1 1 1 1 £8 
OOOOOOOOOi-iOO>-i 
Vertical 
Direct 
(8 lamps) 
480 watts 
r 1 1 1 1 1 1 1 1 1 1 ino om 0 « w 1^ 0 co ch 
1 1 1! 1 1 1 1 1 1 1 V" •st 10 v- itio 4 - x c-c 
0 0 0000000^0-,^- 
Semi- 
indirect 
800 watts 
0.55 
0.67 
0.74 
0.74 
0.69 
0.67 
0.60 
1.14 
1.13 
1.60 
1.48 
1.50 
1.62 
Direct 
(16 lamps) 
400 watts 
0.75 
0.63 
0.76 
0.76 
1.40 
1-25 
1.20 
1.90 
1.25 
0.70 
0.90 
2.60 
3-40 
1.70 
i-73 
4.00 
2.40 
1.10 
1.40 
2.50 
3.00 
1.90 
2.00 
2.90 
Horizontal 
Direct 
( 8 lamps) 
480 watts 
1.17 
1.10 
i-33 
0.91 
1-50 
1.90 
2.10 
3-30 
2.20 
1.12 
1.42 
2.80 
5-oo 
2.60 
2.25 
4-40 
2.60 
1.10 
i-39 
2.40 
3.80 
2.40 
3-oo 
5.00 
Semi- 
indirect 
800 watts 
i-74 
1.86 
1.80 
1.66 
2.68 
4.00 
4-4° 
3.80 
2.85 
1.86 
1.88 
2.90 
5-50 
6.80 
6.80 
6.80 
3-50 
2.00 
2-95 
'4-50 
7.00 
6.70 
6.60 
6.40 
Station 
1 
2 
3 
4 
5 
6 
7 
8 
9 
10 
11 
12 
13 
U 
15 
16 
17 
18 
19 
20 
21 
22 
23 
24 FERREE AND RAND: EFFICIENCY OF THE EYE 
435 
Horizontal 
Vertical 
45° 
Semi- 
Direct 
Direct 
Semi- 
Direct 
Direct 
Semi- 
Direct 
Direct 
Station 
indirect 
(8 lamps) 
(16 lamps) 
indirect 
(8 lamps) 
(16 lamps) 
indirect 
(8 lamps) 
(16 lamps) 
800 watts 
480 watts 
400 watts 
800 watts 
480 watts 
400 watts 
800 watts 
480 watts 
400 watts 
25 
4.20 
2.80 
2.30 
I.40 
I-I5 
0.82 
2.50 
2.00 
i-45 
26 
2.23 
t.46 
I.IO 
- 
- 
- 
- 
- 
27 
2-45 
I.42 
1.20 
- 
- 
- 
- 
- 
- 
28 
3-95 
2.80 
2.20 
i-54 
1.22 
0.83 
-3.10 
2-30 
1.82 
29 ' 
6.80 
5.80 
4. IO 
1.82 
I.48 
1.00 
4-30 
4.20 
2.82 
30 
6.70 
3.20 
1.70 
1.78 
I.II 
0.75 
5-io 
2.50 
1.40 
31 
7-40 
2.85 
2.10 
2.20 
I.06 
0.81 
5-20 
2-35 
1.60 
32 
7.10 
5-15 
4.00 
2.10 
1.18 
1.20 
5-20 
3.80 
2.90 
33 
4.10 
3-5° 
2.40 
i-75 
1-05 
O.98 
3-20 
2.70 
2.05 
34 
4.60 
2.65 
2-35 
1.76 
1-35 
I.4I 
3-4° 
2 20 
2.10 
35 
6.80 
4.00 
3-10 
2.40 
2.00 
i-44 
5-20 
3-30 
2-5° 
36 
6.40 
3.60 
2.00 
2-45 
1.61 
1.00 
4-7o 
2.50 
1.68 
37 
6.20 
3-20 
1.70 
2.48 
1.64 
o.95 
4-75 
2.70 
1.42 
38 
7.00 
430 
3-20 
2.52 
2-15 
1.50 
5-4o 
3-50 
2.40 
39 
4-40 
3-oo 
2.70 
i-95 
1.50 
1.44 
3-68 
2.60 
2.20 
4° 
2-37 
1.26 
I.IO 
- 
- 
- 
- 
- 
- 
4i 
1.70 
i-43 
0.95 
- 
- 
- 
- 
- 
- 
42 
3-3° 
2.70 
2.40 
2.00 
1.42 
i-i5 
3-30 
2.40 
2.25 
43 
7.00 
4-9° 
3.60 
2.38 
1.87 
1.17 
5-62 
4.10 
3.00 
44 
7.10 
3-70 
2.00 
2.88 
1.48 
1.00 
5-78 
2.65 
2.00 
45 
6.80 
3.60 
2.05 
2.65 
i-54 
i-i5 
5.60 
2.35 
2.10 
46 
7.10 
4-7o 
4.00 
2.58 
i-45 
1.20 
5-4o 
3-70 
2-95 
47 
3-65 
2.50 
2.45 
1.82 
1.24 
1.17 
3-20 
2-3° 
2.20 
48 
3-7° 
2.60 
2.50 
1.92 
i-45 
1.50 
3-50 
2.15 
I.80 
49 
6.20 
3.60 
3-°o 
2.45 
2.10 
1.60 
5-oo 
3-io 
2.l8 
50 
6.00 
2.50 
2.00 
2.70 
1.80 
1.24 
5-20 
2-55 
1.85 
TABLE XI.--(Intensity Series.)- {Continued}. 436 
TRANSACTIONS OF ILLUMINATING LNGINLFRING SOCIETY 
Horizontal 
Vertical 
45° 
Semi- 
Direct 
Direct 
Semi- 
Direct 
Direct 
Semi- 
Direct 
Direct 
Station 
indirect 
800 watts 
(8 lamps) 
480 watts 
(16 lamps) 
400 watts 
indirect 
800 watts 
(8 lamps) 
480 watts 
(16 lamps) 
400 watts 
indirect 
800 watts 
(8 lamps) 
480 watts 
(16 lamps) 
400 watts 
51 
5-5° 
2.50 
I.90 
2.52 
1.78 
I.26 
5.10 
2.50 
I.90 
52 
4-85 
3.60 
3.IO 
2-55 
2.10 
I.60 
4-65 
3-30 
2-75 
53 
3-20 
2.30 
2.30 
1-73 
i-54 
1.26 
• 3'10 
2.25 
I.92 
54 
• 2.47 
3.00 
4.00 
i-77 
1.80 
2.10 
2.65 
3-10 
3-20 
55 
4-70 
4.00 
3-55 
2.28 
1.68 
1-3° 
4-35 
3.80 
2.80 
56 
5-4o 
2.10 
1.90 
2.98 
1.82 
1-25 
5.60 
2.50 
2.10 
57 
5-io 
2-3° 
1.67 
2-95 
1.74 
1-35 
5-30 
2.60 
2-15 
58 
5-20 
4-40 
4-30 
2 60 
1.69 
1.88 
5-28 
3.60 
3-35 
59 
2.90 
2-45 
2.50 
2-35 
1-38 
1.20 
3.60 
2.50 
2.22 
60 
2.52 
i-94 
1.70 
1.82 
i-49 
1-37 
2.72 
2-35 
2.51 
61 
3-92 
2.20 
2.08 
2.90 
2.25 
1.80 
4.80 
3-40 
2.70 
62 
3-52 
1.64 
i-47 
2.87 
1.86 
i-5o 
4-7o 
2-35 
1.96 
63 
3-42 
1.66 
1.40 
2.62 
1.86 
1.40 
4-30 
2.40 
1.96 
64 
3-32 
2.10 
2.20 
2-37 
2.20 
1.72 
4.10 
3-i° 
3.00 
65 
2.30 
i-43 
1.62 
1.66 
1-74 
1.38 
3-00 
2-35 
2.25 
66 
1.40 
0.90 
0-97 
- 
- 
- 
- 
- 
- 
Average 
4-44 
2.72 
2.14 
i-93 
1.48 
1.12 
4-03 
2.55 
2.06 
TABLE XI.-(Intensity Series.)-{Continued}. fbrree and rand: ebficibncy or the; rye 
437 
TABLE XII. -(Intensity Series.) 
Showing the illumination measurements in foot-candles at nine representa- 
tive stations for the different intensities used for the semi-indirect system. 
Station 
Horizontal 
Vertical 
45° 
8oo 
watts 
480 
watts 
320 
watts 
200 
watts 
800 
watts 
480 
watts 
320 
watts 
2G0 
watts 
800 
watts 
480 
watts 
320 
watts 
200 
watts 
Card • • • • 
6.8 
3-3 
2.2 
1.6 
1.82 
0-94 
0.58 
0.45 
4-5 
2.4 
1.52 
i-i5 
12 
2.9 
i-79 
1.0 
0.75 
0-55 
o-35 
0.22 
0.17 
1-49 
0.76 
0.51 
0.4 
16 
6.8 
3-2 
2.0 
1.32 
0.69 
0-44 
0.23 
0.15 
3-6 
i-75 
1-03 
0.68 
31 
7-4 
3-6 
2.2 
1.6 
2.2 
o-94 
O.58 
0.44 
5-2 
2.4 
i-53 
I-I5 
34 
4.6 
2.1 
i-45 
o-95 
I.76 
0-9 
O.58 
0.43 
3-4 
1.68 
1.12 
0-74 
39 
4-4 
i-79 
1.17 
0-93 
1-95 
0.86 
0.54 
0.41 
3.68 
1-52 
0-95 
0.73 
45 
6.8 
3-3 
2.2 
i-4 
2.65 
i-3 
0.82 
0.58 
5-6 
2-7 
i-7 
1.19 
54 
2-47 
1.31 
0.87 
o-7 
1.77 
1.15 
0.62 
0.48 
2.65 
i-47 
0.95 
0.76 
58 
5-2 
2.82 
1-9 
i-44 
2.6 
i-35 
O.87 
0.63 
5-28 
2.68 
1.8 
i-3 
Average • 
4-44 
2.66 
1.78 
1.11 
i-93 
1.16 
0-77 
0.48 
4-03 
2.42 
1.61 
1.01 
TABLE XIII.-(Intensity Series.) 
Showing the illumination measurements in foot-candles at nine representa- 
tive stations for the different intensities used for the direct system (16 lamps). 
Station 
Horizontal 
Vertical 
45° 
400 
watts 
365 
watts 
240 
watts 
400 
watts 
365 
watts 
240 
watts 
400 
watts 
365 
watts 
240 
watts 
Card 
1.86 
2.1 
1.23 
0.8 
0.6 
0.54 
1.46 
1-33 
0.935 
12 
2.6 
i-45 
i-43 
0.44 
0.275 
0.265 
1.42 
0.68 
0.85 
16 
4.0 
4-4 
2.6 
0.47 
0-34 
0.385 
2.4 
2.4 
i-55 
31 
2.1 
2.1 
i-45 
0.81 
0-735 
0.575 
1.6 
1.68 
1.1 
34 
2-35 
2-3 
1.84 
1.41 
1.49 
1.08 
2.1 
2-5 
1.57 
39 
2.7 
2.6 
i-5 
i-44 
1-55 
0.825 
2.2 
2.4 
1.19 
45 
2.2 
i-75 
i-45 
1-27 
i-3 
0.77 
2.1 
i-95 
1.23 
54 
4.0 
1.6 
1-34 
2.1 
i-i3 
0.78 
3-2 
1.64 
1.38 
58 
4-3 
2-3 
2-5 
1.88 
1.25 
1.02 
3-35 
2.1 
2.1 
Average 
2.14 
- 
1.26 
1.12 
- 
O.67 
2.06 
- 
1.21 438 
TRANSACTIONS OP UvRUMINATING BNGINppRING SOCIETY 
TABLE XIV.-(Intensity Series.) 
Showing the illumination measurements in foot-candles at nine representa- 
tive stations for the different intensities used for the direct system (8 lamps). 
Station 
Horizontal 
Vertical 
45° 
480 
watts 
320 
watts 
200 
watts 
120 
watts 
480 
watts 
320 
watts 
200 
watts 
120 
watts 
480 
watts 
320 
watts 
200 
watts 
120 
watts 
Card • • • • 
2.6 
i-97 
16 
0.64 
1.02 
0.65 
0.45 
0.32 
2.0 
i-39 
0.85 
0.45 
12 
2.8 
1.94 
.36 
0.69 
0-45 
O.41 
0.21 
0.15 
1-7 
1.12 
0.76 
0.41 
16 
4.4 
2.8 
2.4 
1.22 
O.4I 
0.36 
0.2 
O.II 
2-4 
1 88 
i-44 
0.66 
3i 
2.95 
2.1 
1-34 
0.71 
I.06 
O.71 
0.28 
0.32 
2.35 
i-45 
1-03 
0.6 
34 
2.65 
1.76 
i-3 
O.76 
i-35 
I.OI 
O.92 
0.4 
2.2 
i-45 
1.24 
0.56 
39 
3.0 
2.2 
1-25 
0-7 
i-5 
1.0 
0.66 
o-44 
2.6 
1.8 
1.05 
0.65 
45 
3-6 
2.1 
1.2 
O.69 
i-54 
1.07 
0.72 
0.46 
2-35 
1.84 
1.15 
0.65 
54 
3-o 
2.2 
1.22 
0.68 
i-54 
1.1 
0-59 
0.4 
3-1 
2.4 
1.18 
1-63 
58 
4-4 
3-i 
2.1 
1.3 
1.69 
1.22 
0-77 
0.52 
3-6 
2-5 
1.68 
0.96 
Average. 
2.72 
i.8r 
1.13 
0.68 
1.48 
0-99 
0.62 
0-37 
1.12 
c-75 
o-47 
0.28 
TABLE XV.-(Intensity Series.) 
Showing the brightness measurements in candlepower per square inch for 
the different intensities used for the semi-indirect system at points indicated 
by the letters A, B, C, D, etc., see Fig. 3, Further Experiments on the Effi- 
ciency of the Eye, etc., Trans. I. E. S. (1915), vol. X, p. 452b. 
Position 
8oq watts 
480 watts 
320 watts 
200 watts 
A 
0.687 
0.370 
0.180 
O.1428 
B 
0.0461 
0.0219 
O 01402 
O.OIOO8 
c 
0.0858 
0.0504 
0.0346 
0.02414 
p 
0.0461 
0.0219 
O.OI63 
O.OIOO8 
P 
0.00264 
0.00177 
0.0008 
O.OOO61 
p 
O.OO34 
0.00187 
O.OOIO34 
O.OOO792 
G 
0.0058 
0.00242 
0.00187 
O.OOI23 
H 
0.00662 
0.00259 
O.OOI62 
O.OOI44 
I 
0.00638 
0.00237 
O.OOI87 
O.OOI23 
T 
O.OOI49 
0.00076 
O.COO484 
O.OOO325 
J 
K 
0.00462 
0.00189 
O.OOI4 
0.000902 
0.00255 
0.00173 
O.OOIO85 
0.00063 
M 
O.OO572 
0.00224 
0.001408 
O.OOII 
N 
0.00286 
0.00173 
O.OOIO85 
0.00063 
o 
0.00704 
O 00462 
0.00264 
0.00176 
p 
0.00616 
0.003196 
O.OOI98 
0.00154 ' 
x 
O.OO3432 
0.00176 
O.OOIO5 
0.000814 
O.OIO7 
0.00462 
0.0029 
0.002024 
0.00654 
0.00316 
O.OOI93 
0.00176 TABLE XVI.-(Intensity Series.) 
Showing the brightness measurements in candlepower per square inch for 
the different intensities used for the direct system (16 lamps) at points indi- 
cated by the letters A, B, C, D, etc., see Fig. 2, Further Experiments on the 
Efficiency of the Eye, etc., Trans, of the I. E. S., 1915, X, p. 452a. 
Position 
400 watts 
365 watts 
240 watts 
A 
B 
0.1897 
0.1232 
0.1232 
c 
0.00253 
O.OOI5I 
0.00151 
D 
0.00277 
O.OOI45 
0.00119 
E 
0.00097 
O.OOO67 
0.000545 
F 
0.00277 
O.OO185 
0.00156 
G 
0.00303 
0.00246 
0.00172 
H 
0.00303 
O.OO229 
0.00174 
I 
J 
0.00075 
O.OOO4 
0.000453 
K 
0.00252 
O.OOI67 
0.00176 
L 
0.00191 
O.OOI49 
0.00154 
M 
0.00273 
O.OOI98 
0.00194 
N 
0.00176 
O.OOI45 
0.00136 
o 
0.0026 
0.00242 
0.00143 
P 
0.00215 
O.OOI67 
O.OOI19 
Q 
0.00184 
O.OOIO3 
0.00103 
x 
0.00172 
O.OOI32 
O.OOI 
Reading page horizontal 
0.00396 
0.00405 
O.OO2II 
Reading page 450 position 
0.0029 
O.OO273 
O.OOI76 
TABLE XVII.-(Intensity Series). 
Showing the brightness measurements in candlepower per square inch for 
the different intensities used for the direct system (8 lamps) at points indi- 
cated by the letters A, B, C, D, etc., see Fig. 2, Further Tests for the Effi- 
ciency of the Eye, etc., Trans, of the I. E. S., 1915, X, p. 452a. 
Position 
480 watts 
320 watts 
200 watts 
120 watts 
A 
IOOO.OOOCO 
1000.00000 
1000.00000 
1000.00000 
B 
0.2953 
0.2398 
0.1657 
0.08998 
c 
0.00317 
0.00299 
0.00154 
0.00097 
D 
0.00454 
0.0033 
0.00185 
0.000704 
F 
0.001848 
0.00118 
0.00059 
0.0003 
F 
0.00198 
0.00272 
0.00145 
0.00063 
G 
O.CO347 
0.00361 
0.00189 
0.00074 
H 
0.00391 
0.00334 
O.OOI22 
0.001 I 
I 
0.00405 
0.0029 
O.OOI67 
0.00092 
J 
0.00069 
0.00046 
O.OOO37 
0.00023 
K 
0.00308 
0.00167 
O.OOI22 
0.00073 
L 
0.00229 
0.00141 
O.OOIO3 
0.00056 
M 
0.00387 
0.00229 
0.00141 
0.00068 
N 
0.00192 
0.00128 
O.OOO96 
0.00054 
o 
0.00246 
0.00252 
O.OOIOI 
0.00065 
P 
0.00192 
0.00185 
O.OOO83 
0.00051 
Q 
0.00325 
0.00222 
O.OOI36 
0.000704 
x 
0.002376 
O.OOI4I 
O.OOO924 
0.00062 
Reading page hori- 
zontal 
0.00528 
O.OO334 
0.00229 
0.00123 
Reading page 45° 
position 
0.003696 
0.0022 
0.00149 
0.00077 Ratio 
800 
480 
320 
200 
Lightest to darkest 
Lightest to test card 
Lightest to reading page • • • 
2nd lightest to darkest 
2nd lightest to test card - • •. 
2nd lightest to reading page. 
3rd lightest to darkest 
3rd lightest to test card 
3rd lightest to reading page- 
0.687 /0.00149 = 455 
0.687 /o.00343 = 200 
0.687 /o.00654 = 105 
0.0858/0.00149 = 57 
0.0858/0.00343 = 25 
0.0858/0.00654 = 13 
0.0461/0.00149 = 31 
0.0461/0.00343 = 13 
0.0461/0.00654 = 7 
0.37 /o.00076 = 486 
0.37 /o.00176 = 210 
0-37 /O.OO316 = 113 
0.0504/0.00076 = 66 
0.0504/0.00176 = 28 
0.0504/0.00316 = 15 
0.0219/0.00076 = 29 
00219/0.00176= 12 
0.0219/0.00316 = 6.9 
0.18 /o.000484 = 372 
0.18 /o.00105 - 
0.18 /0.00193 = 93 
0.0346/0.000484 = 71 
0.0346/0.00105 = 32 
0.0346/0.00193 = 17 
0.0163/0.000484 = 34 
0.0163/0.00105 = 15 
0.0163/0.00193 = 8 
0.1428 /0.000325 = 439 
0'1428 /o.000814 = 175 
0.1428 /o.00176 = 81 
0.02414/0.000325 = 74 
0.02414/0.0008 r4 = 29 
0.02414/0.00176 = 13 
0.01008/0.000325 = 31 
0.01008/0.000814 = 12 
0.01008/0.00176 = 6 
Ratios 
400 
365 
240 
Lightest to darkest 
Lightest to test card 
Lightest to reading page 
2nd lightest to darkest 
2nd lightest to test card 
2nd lightest to reading page 
3rd lightest to darkest 
3rd lightest to test card 
3rd lightest to reading page 
1000.00000/0.00075 = L333.333 
1000.00000/0.00172 = 581,395 
1000.00000/0.0029 = 344.828 
0.1897 /o.00075 =• 253 
0.1897 /o.00172 = IIO 
0.1897 /0.0029 = 65 
0.00316/0.00075 = 4.2 
0.00316/0.00172 = 1.8 
0.00316/0.0029 = 1.1 
1000.00000/0.0004 = 2,500,000 
1000.00000/0.00132 - 757,575 
1000.00000/0.00273 = 366,300 
0.1232 /0.0004 == 308 
0.1232 /o.00132 = 93 
0.1232 /o.00273 = 46 
0.00246/0.0004 = 6.15 
0.00246/0.00132 = 1.8 
0.00246/0.00273 = 0.9 
1000.0000/0.000453 - 2,207,505 
1000.0000/0.0011 = 909,090 
1000.0000/0.00176 = 568,181 
0.1232/0.000453 = 272 
0.1232/0.0011 = 112 
0.1232/0.00176 = 70 
0.0018/0.000453 = 4 
0.0018/0.0011 - 1.6 
0.0018/0.00176 = 1.02 
TABLE XIX.-(Intensity Series). 
Showing some prominent ratios of surface brightness for the different intensities used for the direct system (16 lamps). 
TABLE XVIII.-(Intensity Series.) 
Showing some prominent ratios of surface brightness for the different intensities used for the semi-indirect system. Ratio 
480 
320 
200 
120 
Lightest to darkest 
1000.00000/0.00069 = 
L449,275 
1000.0000/0.00046 = 
2,173,9'3 
1000.00000/0.00037 = 
2,702.702 
1000.000000/0.00023 = 
4,347,826 
Lightest to test card 
1000.00000/0.002376 = 
1000.0000/0.00141 - 
1000.00000/0.000924 = 
1000.000000/0.00062 = 
Lightest to reading page • • • 
420,168 
709,220 
1,086,960 
1,612,903 
1000.00000/0.003696 = 
270,270 
1000.0000/0.0022 = 
454,545 
1000.00000/0.00149 = 
671,141 
1000.000000/0.00077 = 
1,298,701 
2nd lightest to darkest 
0.2953 /o.00069 - 
428 
0.2298/0.00046 - 
5'9 
0.1657 /o.00037 = 
449 
0.08995 /o.00023 - 
390 
2nd lightest to test card .... 
0.2953 /o.002376 = 
124 
0.2298/0.00141 = 
170 
0.1657 /o. 00092 4 = 
179 
0.08995 /0.00062 = 
128 
2nd lightest to reading page • 
0.2953 /o.003696 = 
80 
0.2298/0.0022 = 
109 
0.1657 /o.00149 = 
III 
0.08995 /0.00077 = 
117 
3rd lightest to darkest 
0.00454/0.00069 = 
6.6 
0.0033/0.00046 - 
7-1 
0.00185/0.00037 = 
5-o 
0.000704/0.00023 = 
3-1 
3rd lightest to test card 
0.00454/0.002376 = 
i-9 
O.OO33/O-OOI4I - 
2.3 
0.00185/0.000924 = 
2.0 
0.000704/0.00062 = 
1.1 
3rd lightest to reading page- 
0.00454/0.003696 = 
1.2 
0.0033/0.0022 = 
i-5 
0.00185/0.00149 = 
1.2 
0.000704/0.00077 = 
0.9 
TABLE XX. -(Intensity Series.) 
Showing some prominent ratios of surface brightness for the different intensities used for the direct system (8 lamps). 442 
transactions or illuminating engineering society 
Watts 
Volts 
Foot-candles 
Time 
Maximal 
distance 
at which 
test object 
can be 
seen 
clear 
Working 
distance 
Total 
time 
clear 
Total 
time 
blurred 
Total 
time 
clear 
total 
time 
blurred 
Ratios 
reduced 
to com- 
mon 
standard 
Hori- 
zon- 
tal 
Verti- 
cal 
45° 
320 
107 
2.2 
0.58 
1.52 
9 A.M. 
12 M. 
73-5 
7^-5 
58.5 
58.5 
117 
112 
63 
68 
1.86 
1.65 
3-5 
3-i 
3 20 
IIO 
2.31 
0.62 
1.6l 
9 A.M. 
12 M. 
74-0 
72.0 
57-0 
57-o 
140 
135 
40 
45 
3-5 
3-o 
3-5 
3-o 
200 
IIO 
I.72 
0 484 
I.29 
9 A.M. 
12 M. 
72.5 
72.5 
57-5 
57-5 
138 
13° 
42 
50 
3-27 
2.6 
3-5 
2.78 
200 
107 
1.6 
0-45 
I-15 
9 A.M. 
12 M. 
65-5 
65-5 
5i-5 
5i-5 
141 
I23 
39 
57 
3-6r 
2.14 
3-5 
2.07 
480 
107 
3-3 
0-94 
2-4 
9 A.M. 
12 M. 
79-° 
76.5 
63-5 
63-5 
124 
96 
56 
180 
2.21 
1.11 
3-5 
i-75 
800 
107 
6.8 
1.82 
4-5 
9 A.M. 
12 M. 
85-5 
82.5 
66.5 
66.5 
126 
62 
54 
118 
2-33 
0.525 
3-5 
0.78 
760 
107 
5-8 
i-45 
4.0 
9 A.M. 
12 M. 
80.5 
79-5 
68.5 
68.5 
142 
92 
38 
88 
3-73 
1.04 
3-5 
0.97 
TABLE XXI.-(Intensity Series.) 
Observer R, showing the effect on the efficiency of the eye of varying the intensity of light in the 
semi-indirect lighting system. FERREE AND RAND : EFFICIENCY OF THE EYE 
443 
TABLE XXIL-(Intensity Series.) 
Observer R, showing the effect on the efficiency of the eye of varying the intensity of light in the direct 
lighting system. (16 lamps.) 
Watts 
Volts 
Foot-candles 
Time 
Maximal 
distance 
at which 
test object 
can be 
seen 
clear 
Working 
distance 
Total 
time 
clear 
Total 
time 
blurred 
Total 
time 
clear 
Total 
time 
blurred 
Ratios 
reduced 
to com- 
mon 
standard 
Hori- 
zon- 
tal 
Verti- 
cal 
45° 
240 
107 
1-23 
0-54 
o-935 
9 A.M. 
74-0 
62.0 
132 
48 
2-75 
3-5 
12 M. 
73 0 
62.0 
97 
83 
1.17 
i-49 
365 
107 
1.6 
0.6 
i-33 
9 A.M. 
74-5 
64 0 
III 
69 
1.6l 
3-5 
12 M. 
74-o 
64.0 
68 
112 
0.607 
i-32 
400 
107 
1.86 
0.8 
1.46 
9 A.M. 
76.0 
65-0 
142 
38 
3-68 
3-5 
12 M. 
75-o 
65-0 
102 
78 
1-3 
1-23 
880 
107 
4.2 
1.41 
2.6 
9 A.M. 
8t.o 
68.0 
139 
41 
3-39 
3-5 
12 M. 
78.0 
68.0 
7i 
IO9 
0.65 
0.67 
TABLE XXIII. -(Intensity Series.) 
Observer R, showing the effect on the efficiency of the eye of varying the intensity of light in the direct 
lighting system. (8 lamps.) 
200 
107 
i.i6 
045 ' 
0.85 
9 A.M. 
12 M. 
72.0 
71.0 
56.0 
56.0 
122 
105 
58 
75 
2.1 
i-4 
3-5 
2-3 
120 
io7 
0.64 
0 32 
0.49 
9 A.M. 
12 M. 
70.5 
70.0 
56.5 
56.5 
141 
IIO 
39 
7o 
2.87 
i-57 
3-5 
1.91 
320 
io7 
i-97 
0.65 
i-39 
9 A.M. 
12 M. 
73-5 
73-o 
60.5 
60.5 
137 
107 
43 
73 
318 
1.46 
3-5 
1.6 
480 
107 
2.6 
1.02 
2.0 
9 A.M. 
12 M. 
76.0 
73-5 
63-5 
635 
159 
128 
21 
52 
7-57 
2-5 
3-5 
I-I5 444 
transactions or illuminating engineering society 
Showing the effect on the efficiency of the eye 
Showing the effect on the efficiency of the eye 
Showing the effect on the efficiency of the eye 
of varying the intensity of light in the direct 
of varying- the intensity of light in the 
semi- 
of varying ihe 
intensity of light in the direct 
indirect lighting system. 
lighting system 
. (16 lamps.) 
lighting system. (8 lamps.) 
Foot-candles 
Foot-candles 
Foot-candles 
Hori- Verti 
Hori- Verti- 
Hori- Verti- 
Watts Volts zontal cal 
45° 
Watts 
Volts zontal cal 450 
Watts Volts zontal cal 45° 
A . . 200 107 1.6 0.45 
1.15 
A . . 240 
107 1.23 0.54 0.935 
A . . 120 107 0.64 0.32 0.49 
B . . 200 no 1.72 0.484 
1.29 
B . . 365 
107 1.6 0.6 1.33 
B . . 200 107 1.16 0.45 085 
C . . 320 107 2.2 0.58 
1-52 
C . . 400 
107 1.86 0.8 1.46 
C . . 320 107 1.97 0.65 1.39 
D . . 320 no 2.31 0.62 
1.61 
X . . 880 
107 4.2 1.41 2.6 
D . . 480 107 2.6 1.02 2.0 
B . . 480 107 3.3 0.94 
2.4 
F . . 800 107 6.8 1.82 
4-5 
X . . 760 107 5.8 1.45 
4-o 
I 
Chart E. 
Chart D. 
Chart C. FERREE AND rand: EFFICIENCY OF THE EYE 
445 
Chart F.-Showing the effect on the efficiency of the eye of varying the 
intensity of light for the semi-indirect system of lighting. Foot-candles at 
the point of the test card are plotted along the abscissa; loss of efficiency 
along the ordinate. X - points where the change in intensity was pro- 
duced by changing the voltage (see Table XXI). 
Chart F. 
Chart G.-Showing the effect on the efficiency of the eye of varying the 
intensity of light for the direct system of lighting. Foot-candles at the 
point of the test card are plotted along the abscissa; and loss of efficiency 
along the ordinate. A - curve for 16 lamps; B, for 8 lamps. 
Chart G. 446 
TRANSACTIONS OR ILLUMINATING ENGINEERING SOCIETY 
The results of the tests for the intensity series are shown 
in Tables XXI-XXIII. Three hours was selected as the period 
of work in all of these experiments. Briefly stated the procedure 
was as follows. First the most favorable intensity was deter- 
mined and then variations were made on either side of this 
intensity until it was certain that the characteristic effect of 
increase and decrease of illumination was obtained. Table XXI 
gives the results for Observer R under the semi-indirect system. 
Seven variations of intensity were used. These results are typical 
of the effect of variations of intensities for this system. Tables 
XXII and XXIII show the results for the direct system for the 
same observer. • For the direct system the most favorable inten- 
sity, it will be noted, was secured with the 8-lamp system with 
a total wattage lower than could be gotten with the 16-lamp 
system, i. e., a system totalling 200 watts caused the least loss of 
efficiency to the eye, while 240 was the smallest total of wattage 
that could be secured with the 16-lamp system. 
Charts have been constructed also to give a graphic representa- 
tion of these tables. Chart C shows the results of Table XXI; 
Chart D, of Table XXII; and Chart E, of Table XXIII. 
In these charts loss of efficiency was plotted against time of 
work. In Charts F and G loss of efficiency is plotted against 
intensity of light in foot-candles at the point of the test card. 
Chart F shows the results for Table XXI; Chart G for Tables 
XXII and XXIII. 
IV. CONCLUSION. 
Two facts may be emphasized at this point. (1) Of the light- 
ing factors that influence the welfare of the eye, those we have 
grouped under the heading distribution are apparently funda- 
mental. They seem to be the most important we have yet to deal 
with in our search for the conditions that give us the minimum 
loss of efficiency and the maximum comfort in seeing. If, for 
example, the light is well distributed in the field of vision and 
there are no extremes of surface brightness, our tests seem to 
indicate that the eye, so far as the problem of lighting is con- 
cerned, is practically independent of intensity. That is, when the 
proper distribution effects are obtained, intensities high enough 
to give maximum discrimination of detail may be employed FERREE AND RAND: EFFICIENCY OF THE EYE 
447 
without causing appreciable damage or discomfort to the eye. 
(2) For the kind of distribution effects given by reflectors of 
the type employed in our direct and semi-indirect installations, 
our results show that unquestionably too much light is being used 
for the welfare of the eye. 
Before concluding our paper we wish again to state that the 
units we have employed were not selected as fully representative 
of the classes direct, semi-indirect, and indirect. Agreement in 
fact has not yet been reached with regard to what falls within 
each of these classes. The units employed were chosen rather 
to show the effect on the ability of the eye to maintain its ef- 
ficiency for a period of work of varying the factors we have 
grouped under the heading distribution. We hope ultimately to 
determine the limits between which each of these factors may vary 
without damage to the eye in a selected range of lighting situa- 
tions, especially the factor surface brightness. These most fav- 
orable conditions will then serve as a goal to be attained what- 
ever principle of lighting is employed. 
Our next step in this division of the work will be to determine 
the effect on loss in efficiency of using reflectors of different 
degrees of opacity when the light is distributed to the plane of 
work both by the direct and indirect principles of lighting. 
That is, reflectors of different densities: prismatic, alba, opalux, 
totally opaque, etc., will be used turned up and down. In each 
case the installation will be made with special reference to giving 
the best results obtainable for the particular type of unit em- 
ployed ; and the factors: evenness of illumination, diffuseness of 
light, the angle at which the light falls on the work, and the 
evenness of surface brightness will be varied separately in turn, 
and the effect on loss of efficiency will be determined. More- 
over, if it is found that the factors in question can not be studied 
in sufficient detail in the concrete lighting situation, the work 
will be supplemented by more abstract investigations. The re- 
sults of this series of tests should give us among other things, 
for example, a still better idea of what amount of brightness dif- 
ference the eye is adapted to stand, and the comparative effect of 
different ratios of surface brightness on loss of efficiency. [Reprinted from the Transactions of the Illuminating 
Engineering Society, No. 6, 1915.] 
FURTHER EXPERIMENTS ON THE EFFICIENCY OF 
THE EYE UNDER DIFFERENT CONDITIONS 
OF LIGHTING* 
c. X. Ferree; and g. rand. 
Synopsis: This paper is a continuation of the papers presented to 
the Society in 1912 and 1913. It describes the completion of the plan of 
work outlined in the preceding papers for one set of lighting conditions 
for three of the tests thus far devised by one of the writers (Ferree) - 
namely, a test of the ability of the eye to hold its efficiency for a period 
of work; a test for loss of efficiency of the fixation muscles; and a test 
for the comparative tendency of different conditions of lighting to produce 
discomfort. A report is also given of some miscellaneous experiments 
related to the hygienic employment of the eye in which the following 
points are taken up: the effect of varying the area and conversely the 
intrinsic brightness of the ceiling spots above the reflectors of an indirect 
system of lighting; the effect of varying the angle at which the light 
falls on the work in a given lighting situation; the effect of using an 
opaque eye shade with dark and light linings with each of the installations 
of artificial lighting employed in this and the previous work; the effect 
on the efficiency of the fixation muscles of three hours of work under 
these installations; the effect of motion pictures on the eye for different 
distances of the observer from the projection screen; a determination of 
the tendency of different conditions of lighting to produce discomfort, 
and a comparison of the tendency of these conditions to produce dis- 
comfort and to cause loss of efficiency. 
INTRODUCTION. 
The present paper is the third in a series of papers presented 
to this Society on the subject of lighting in its relation to the 
eye. In the first paper of this series1 it was pointed out that if 
we are to make a comparative study of the effect of different 
conditions of lighting on the eye, we must have a means of 
estimating effects. Work was described in this paper in which 
* A paper read at the eighth annual convention of the Illuminating Engineering 
Society, Cleveland, O., September 21-24, W- 
The Illuminating Engineering Society is not responsible for the statements or 
opinions advanced by contributors. 
1 Tests for the Efficiency of the Eye Under Different Systems of Illumination and 
a Preliminary Study of the Causes of Discomfort, Trans. I. E- S., vol. VIII, 1913, pp. 40-60. 449 
FERREE AND rand: EFFICIENCY OF THE EYE 
the tests already known to physiological optics had been applied 
to the problem with negative results. New tests were proposed 
and brief results were given to show their feasibility for the 
problem in hand and to some extent their sensitivity. The sug- 
gestion was made that a systematic investigation of the effect of 
different conditions of lighting on the eye should include a study 
of the following points: (1) the efficiency of the fresh eye, (2) 
the loss of efficiency as the result of a period of work, and (3) 
the tendency to produce discomfort. In the second paper of 
the series,2 presented to the Society last year, a plan of work was 
outlined and in part carried out in which the first two of the 
above points were covered for a given set of lighting conditions. 
The following factors of importance to the eye were enumerated: 
the evenness of illumination, the diffuseness of light, the angle 
at which the light falls on the object viewed, the evenness of 
surface brightness, intensity and quality. The first four of these 
factors are very closely interrelated and are apt to vary together 
in a concrete lighting situation, although not in a 1 : 1 ratio. It 
was convenient, therefore, for the purpose of this first investiga- 
tion, which was primarily explorative in character, to group them 
together under one heading and to refer to them as distribution 
factors. In order to investigate the effect of certain wide varia- 
tions in these factors, tests were conducted under four types of 
lighting in common use: one was the lighting of a room by day- 
light from windows; the others were the lighting of the same 
room by units commonly called direct, semi-indirect, and indirect, 
selected to serve the purposes of the test.3 
2 "The Efficiency of the Eye Under Different Conditions of Lighting-The Effect 
of Varying the Distribution Factors and Intensity." Trans, of the Ill. Eng. Soc., 
1915, vol. X, pp. 407-447. 
3 According to the plan as the investigation proceeds, the effect of varying each 
of these factors separately will be studied. No especial attempt was made to do 
this in the previous study. In making the experimental variations necessary to the 
investigation, it was stated as our purpose to keep as close as possible to actual 
lighting situations. More abstract investigations will be resorted to only when it 
becomes necessary to supplement the results by details that cannot be gotten from 
the concrete investigation. The objection to the abstract type of investigation, as 
the writers see the case at the present time, is that its results are very apt to be 
misleading. That is, what is permissible with regard to one factor in isolation, may 
not be at all permissible in conjunction with other factors. A more feasible plan 
seems to us to be to vary the factor over a certain practical range in actual concrete 
situations. By a proper selection of the proper situations employed, the ground of 
all that is practicable in lighting may be covered, and the results obtained can have 
a safe application. TRANSACTIONS 01? IEEUMINATING ENGINEERING SOCIETY 
450 
For the systems of artificial lighting the tests were made at 
four positions in the room; one at which six of the eight lighting 
units employed were in the field of view, one at which four were 
in the field of view, one at which two were in the field of view, 
and one at which none was in the field of view. This variation 
of position at which the observation was made accomplishes, it 
was pointed out, two purposes, (i) It gives a more representa- 
tive idea of the difference in the effect on the eye of the four 
types of lighting employed. And (2) it shows the effect of 
varying the number of surfaces in the field of view presenting 
brightness differences, more particularly the number of primary 
sources. The effect of varying intensity under each of the above 
conditions of distribution was also tested. The two sets of ex- 
periments were called respectively the distribution and intensity 
series. Results were given in the preceding paper for only one 
of the above positions in the distribution series, and for only the 
direct and semi-indirect systems for the intensity series. The re- 
sults of the remainder of these two series of experiments, to- 
gether with the report of some miscellaneous experiments will 
constitute the subject matter of the present paper. In these 
miscellaneous experiments, the following points have been taken 
up: the effect of varying the area and conversely the intrinsic 
brightness of the ceiling spots above the reflectors for the indirect 
system; the effect of varying the angle at which the light falls on 
the work; the effect of using an eye shade with dark and light lin- 
ings with each of the three installations of artificial lighting; the 
effect on the efficiency of the fixation muscles of the eye of three 
hours of work under each of the conditions of lighting described 
in the distribution and intensity series; the effect of motion 
pictures on the eye for different distances of the observer from 
the projection screen; a determination of the tendency of each 
of the conditions of lighting that have been used in these ex- 
periments to produce discomfort, and a comparison of the tend- 
ency to produce discomfort and to cause loss of efficiency. Be- 
sides including some additional matter, these experiments, in 
connection with those of the preceding paper, complete the 
plan of work we had outlined for one set of lighting conditions 
for three of the tests we have thus far devised, namely, a test 451 
jTrw: and rand: efficiency or the EyE 
for the ability of the eye to hold its efficiency for clear seeing 
for a period of work, a test for loss of efficiency of the fixation 
muscles, and a test for the comparative tendency of the different 
conditions of lighting to produce discomfort, with the exception 
that in a further. analysis of the loss of efficiency caused by 
these lighting conditions, which will be carried out in part by 
means of these tests, data will be added later to show still more 
clearly the relative amounts of loss that are sustained by the dif- 
ferent functions of the visual apparatus. 
DISTRIBUTION SERIES. 
As was pointed out in the former paper, in order to get the 
effect of variation in the distribution factors on the eye's loss of 
efficiency as the result of a period of work, the test should be 
conducted with the quality and intensity of light made as nearly 
equal as possible. The quality of light was made approximately 
the same for the three installations of artificial lighting employed 
by using clear tungsten lamps in each case. It was decided to 
make the intensity of light as nearly equal4 as possible at the 
point of test, and to give a supplementary specification of the 
lighting effects in the remainder of the room for the three in- 
stallations of artificial light. 
At the point of test the light was photometered5 in several 
directions. It was made approximately equal in the plane of the 
test card and as nearly as possible equal in the other directions. 
The specification of the lighting effects in the remainder of the 
room was accomplished as follows: (i) A determination was 
made of the average illumination of the room under each of the 
three installations. The room was laid out in 3-ft. (0.9 m.) 
4 This equalization was made at the point of test for the position of the observer with 
six of the fixtures in the field of view. For the other positions illumination measure- 
ments were made in several directions at the test card, and brightness measurements 
were made of the surface of the test card and of the observer's book held in the horizontal 
and 45 deg. positions. Equalization could not have been made at all of these points with- 
out having changed the relation and magnitude of the distribution factors, which would 
not have been in accord with the purpose of the test, namely, to determine the effect of a 
certain grouping or relation of these factors for the four positions in the room. 
6 We have not as yet made the fuller photometric specifications of the room lighted by 
daylight with our present arrangement of windows, curtains, etc. We hope to make the 
effect of distribution factors in daylight illumination (employing windows, skylights, 
etc.) the subject of a future study. In this study a photometric analysis of the illumina- 
tion effects produced will be made an especial feature. , TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
452 
squares and illumination measurements were made at 66 of the 
intersections of the sides of these squares. Readings were taken 
in a plane 122 cm. above the floor with the receiving test plate 
of the illuminometer in the horizontal, the 45 deg. and 90 deg. 
positions measuring respectively the vertical, the 45 deg., and 
horizontal components of illumination. The 122 cm. plane was 
chosen because that was the height of the test object. (2) A 
determination was made of the brightness of prominent objects 
in the room, such as the test card, the reflectors for the semi- 
indirect installation, the reading page, the specular reflection 
from surfaces, etc. The brightness measurements were made by 
means of a Sharp-Millar illuminometer with the receiving test 
plate removed. The instrument was calibrated against a mag- 
nesium oxide surface obtained by depositing the oxide from the 
burning metal on a white card. By this method the reflecting 
surfaces were used as detached test plates. The readings were 
converted into candle-power per sq. in. by the following formula: 
Brightness = Foot-candles/br X 144. 
(3) Photographs were made of the room from three positions 
under each system of illumination. 
A complete specification of the test room, the types of installa- 
tion used, and the illumination effects produced for the systems 
of lighting, is given in the previous paper which appears else- 
where in this number of the Transactions (pp. 413-422). Only 
such data will be repeated here as are necessary for reference.7 
In Fig. 1 the test room is shown drawn to scale: plan of 
room, north, south, east and west elevations. In the draw- 
ing, plan of room, are shown the 66 stations at which the 
illumination measurements were made, and the positions of 
the outlets for the lighting fixtures, A, B, C, D, E, F, G, and H. 
In the drawing, east elevation, the observer in position at one 
of the points (Position I) at which the tests were taken is repre- 
7 For a description of the test see the previous article referred to above 
(pp. 410-413); also Tests for the Efficiency of the Eye Under Different Systems of 
Illumination and a Preliminary Study of the Causes of Discomfort, Trans. I. E. S., 
vol. VIII (1913). PP- 4i-5i- Fig. 1.-Plan of test room. 
Fig. 2.-Showing brightness measurements of all surfaces having very high or very 
low brilliancy, direct system. The brightness of the printed page from which the 
observer read was, when held in the horizontal position, 0.0057 cp. per sq. in.; 'n 
the 45 deg. position, 0.004 cp. per sq. in.6 
6 The bright spots on the doors of the apparatus case rated at 100 cp. persq. in., shown 
in Fig. 2, were not in the field of view when the tests were taken. That is, when the tests 
were taken, the doors were thrown open, and all of the apparatus which might give 
specular reflection was removed. Fig. 3.-Showing brightness measurements of all surfaces having very high or very 
low brilliancy, semi-indirect system. The brightness of the printed page from 
which the observer read was, when held in the horizontal position, 0.0058 cp. per 
sq. in.; in the 45 deg. position, 0.0039 cp. per sq. in. 
Fig. 4-Showing brightness measurements of all surfaces having very high or very 
low brilliancy, indirect system. The brightness of the printed page from which 
the observer read was, when held in the horizontal position, 0.0088 cp. per sq. in.; 
in the 45 deg. position, 0.0043 cp. per sq. in. 453 
FERREE AND rand: EFFICIENCY OF THE EYE 
sented.8 The other three positions are indicated in the photo- 
graphs by (x). They will be referred to in the tables and charts, 
in order, by the numerals II, III, and IV. 
Table I shows the number and wattage of the lamps used at 
the outlets A, B, C, D, E, F, G, and H; and Table II shows the 
illumination measurements for each of the 66 stations repre- 
sented in Fig. I. These measurements were made with the re- 
ceiving test plate of the photometer in the horizontal, vertical 
and 45 deg. planes.9 
8 The track along which the test card was moved was parallel to the east and 
west walls of the room. During the three hours of reading which intervened be- 
tween the two tests, the observer moved just far enough back from the upright sup- 
porting the mouth-board to give room for the book to be held and to permit of a 
comfortable reading position. The book was elevated and held approximately at an 
angle of 45 deg. When taking the test, the observer faced the north wall of the 
room, in such a position that with the eyes in the primary position, the lines of 
regard were parallel with the east and west walls of the room. Care was taken to 
have print of uniform size and distinctness for use with the three systems and to 
have a page which gave a comparatively small amount of specular reflection. 
9 See also Table III, The Efficiency of the Eye Under Different Conditions of 
Lighting, etc., Trans. I. E. S., vol. X, 1915, p. 416a. This table was compiled as a 
supplement to Table II for the purpose of making a comparative showing of the 
evenness of illumination at the 122 cm. level given by the three systems of lighting. 
Two cases were made of this. (1) Comparisons were made of each component from 
station to station; (2) the difference between the components was compared. To 
facilitate these comparisons (a) the mean variation from the average of each of the 
components was computed; and (b) the difference between the averages of the three 
components was determined. The evenness of the illumination, it will be remembered, 
is not only of importance to the efficiency of the eye with reference to the object 
directly viewed; but also in its influence on the distribution of surface brightness. 
The evenness of surface brightness depends in general upon two sets of factors; (1) 
the nature and position of the reflecting surfaces in the room; and (2) the type of 
delivery of light to these surfaces. 
We realize that the evenness of the illumination on the 122 cm. plane given by 
the indirect and semi-indirect units was somewhat interfered with by the reflectors of 
the direct system which were beneath and a little to the right of these units when in 
position for the test. Also the evenness of surface brightness on the ceiling for the 
direct system was interfered with by the indirect and semi-indirect reflectors which 
were above and a little to the side of the direct units. The influence of this "dead 
apparatus" will be eliminated in the next series of installations. Moreover, the in- 
stallation in each case was not such as to give the best effects obtainable from the 
type of reflector used. For example, the indirect reflectors were too close to the 
ceiling to give the maximum evenness of illumination and surface brightness for 
the type of reflector employed. The analysis of the effects given in the former paper 
was not made, therefore, for the purpose of drawing general conclusions with regard 
to the type of reflector used. It was made solely for the sake of the comparison of 
the illumination effects obtained with the corresponding results for loss of efficiency. TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
454 
TABLE I. 
Showing the number and wattage of the lamps used at outlets 
A, B, C, D, E, F, G and H. 
Outlet 
Direct 
Semi-indirect 
Indirect 
Watts 
Watts 
Watts 
A 
2-6o 
I-IOO 
I-IOO 
B 
2-60 
I-IOO 
I-IOO 
c 
2-60 
I-IOO 
I-IOO 
D 
2-40 
I-IOO 
I-IOO 
E 
2-60 
I-IOO 
I-IOO 
F 
G 
2-60 
I-IOO 
I-IOO 
H 
■ 2-40 
I- 60 
I-IOO 
TABLE II.-Distribution Series.10 
Showing the illumination measurements in foot-candles for each of the 66 
stations represented in Fig. 1 for the direct, semi-indirect, and indirect 
systems used. 
Horizontal 
Vertical 
45° 
Semi- 
Semi- 
Semi- 
Station 
Direct 
indirect 
Indirect 
Direct 
indirect 
Indirect 
Direct 
indirect 
Indirect 
I 
I.4C 
I.44 
1.22 
2 
I.32 
i-47 
1.26 
3 
I.IO 
1.40 
1-32 
4 
i-37 
1.10 
1-47 
5 
2.03 
2.58 
2.20 
6 
2.50 
3-20 
2-95 
7 
2.51 
3-60 
2.9O 
8 
3-3° 
3-75 
3-00 
9 
2.78 
2-53 
2.20 
IO 
1.50 
i-59 
i-35 
ii 
2.12 
1.64 
1.66 
12 
4.20 
2.65 
2.70 
0-47 
0.48 
0-47 
2.40 
I.25 
1-43 
13 
6.10 
5-25 
4.10 
0-47 
0.48 
O.42 
3-3° 
2.25 
1.96 
14 
3-7o 
4-95 
4-40 
0.48 
0-47 
0-47 
2.00 
2.40 
2.30 
15 
3.00 
4-85 
4-5o 
0-44 
0.48 
o-47 
1-97 
1.88 
2.30 
16 
6.60 
4-25 
4.10 
0.70 
0-37 
0.48 
3-58 
1.60 
2.10 
17 
4-65 
2-35 
3-15 
0.48 
O.24 
0.46 
1.80 
0.69 
1.60 
18 
2.15 
1.69 
2.20 
0-49 
0.38 
o-47 
1.10 
0.66 
1.63 
19 
2-95 
2.10 
2.50 
20 
5-30 
3-20 
3-40 
I.62 
0.53 
0.86 
3-0° 
2-30 
2.20 
21 
6.60 
4.80 
4.60 
2.00 
O.71 
0-94 
3.60 
1.85 
3.00 
22 
2.25 
4-40 
4.80 
O.6l 
0.69 
1.07 
I-I5 
1.80 
2.90 
23 
4-50 
6.00 
5-io 
1.20 
1.14 
1.10 
2.18 
3-3° 
2.90 
24 
6-95 
5-40 
5-00 
I.76 
I.30 
1.04 
3.60 
3-10 
3-oo 
25 
4-85 
3-72 
3-50 
1-33 
0.78 
0-75 
2-75 
1.85 
2.10 
26 
2.50 
1.82 
2.20 
27 
2.81 
2.05 
2.40 
28 
6.50 
3-28 
3-70 
1.30 
I.II 
1.12 
4-40 
2.10 
2.50 
29 
9.00 
6.40 
5-20 
i-45 
1-50 
1.48 
6.30 
3.60 
3-40 
30 
4.95 
6-95 
5-40 
1-36 
I.46 
1.40 
3-i5 
4-15 
3.60 
31 
4.80 
6.20 
5-20 
0-77 
1.20 
1.24 
2.78 
3-85 
3.60 455 
FERREE and rand : efficiency of the Eye 
Horizontal 
Vertical 
45° 
Semi- 
Semi- 
Semi- 
Station 
Direct 
indirect 
Indirect 
Direct 
indirect 
Indirect 
Direct 
indirect 
Indirect 
32 
9.20 
5-5° 
5-00 
0-47 
0.28 
1-33 
5.20 
2.25 
3-40 
33 
6.20 
3.18 
3-70 
1-54 
0-75 
1.22 
4.60 
I.83 
2.60 
34 
5-75 
4-3° 
4.00 
2.85 
1.20 
1.46 
4-30 
2.92 
3.10 
35 
8.00 
6.90 
5-40 
3-70 
1.70 
1.65 
6.00 
4-40 
4-9° 
36 
5.60 
7.25 
5-30 
2.35 
1.91 
1-65 
4.20 
4.68 
4.00 
37 
5-45 
7.00 
5.80 
2.18 
2.15 
1.82 
3-78 
4-55 
4.00 
38 
8.25 
6.80 
5-4o 
3.60 
2.20 
1.72 
6.00 
4.60 
3.80 
39 
6-35 
3-7o 
4.00 
2.80 
1.40 
i-43 
4.60 
2.80 
3.00 
40 
3-°o 
2.05 
2-3° 
41 
2.70 
i-73 
2. IO 
42 
7-3° 
3-65 
3-5° 
2.50 
I.64 
1.36 
5-4o 
2-93 
2.80 
43 
9.80 
6.90 
5-00 
2.70 
2.08 
1.78 
7.20 
4-50 
3-9° 
44 
5-50 
7.10 
5-20 
2.42 
2.18 
1.88 
4-35 
5-io 
4-3° 
45 
5-45 
8.00 
5-20 
2.60 
2.00 
i-93 
4.80 
5-30 
4.20 
46 
10.00 
7-70 
5-20 
2-75 
1.90 
1.86 
8.00 
5-40 
4.10 
47 
6.60 
4.20 
3.60 
2-45 
1.56 
i-33 
5-30 
3-05 
2.90 
48 
5.80 
4-35 
3-70 
3-20 
I.69 
i-74 
5-°o 
3.60 
3-3° 
49 
8.40 
7.20 
4.80 
4-3° 
2-55 
2.10 
7.20 
5.80 
4.00 
50 
5-5° 
7-70 
4-9° 
3-35 
2.42 
2.10 
8.50 
5.80 
4.10 
5i 
5-40 
6.80 
5.00 
3-05 
2.68 
2.15 
4.60 
5-35 
4-35 
52 
8.00 
6.40 
4-70 
4.20 
2-55 
i-93 
6.50 
4.82 
4.00 
53 
6.60 
3.85 
3.60 
3.00 
1.77 
1.41 
5-oo 
3-20 
3-oo 
54 
6.95 
2.88 
2.80 
2.62 
1.80 
1.50 
5.80 
3.00 
2.90 
55 
9.00 
5.90 
3-90 
3-15 
2.40 
i-94 
8.00 
5-20 
3-75 
56 
4-95 
5-9° 
4.60 
3-i5 
2.50 
2.10 
5-30 
5.80 
440 
57 
4-65 
6.10 
4-50 
3.00 
2.60 
2.20 
4-65 
5.80 
440 
58 
9-75 
6-35 
4.00 
3-35 
2.58 
2.00 
8.50 
5.80 
4.00 
59 
5-85 
3-20 
2 90 
2.98 
1.90 
1.76 
5.60 
3-62 
3.10 
60 
3-85 
2.57 
2.60 
1.66 
2.90 
61 
5.20 
4.20 
3-i° 
4-45 
2.60 
1.90 
7.80 
540 
3-50 
62 
3-3° 
4.20 
3-20 
3-30 
2-95 
2.10 
4-95 
5-70 
3-70 
63 
3-52 
4.20 
3.00 
3.60 
2.80 
2.20 
5.60 
5-oo 
3-50 
64 
5-40 
3-70 
3.10 
4.60 
2-45 
i-93 
7-65 
4.60 
340 
65 
4-15 
2.40 
2.25 
4.00 
1.79 
i-54 
5-50 
2.82 
2.60 
66 
2.10 
1.42 
i-35 
- 
- 
- 
- 
- 
- 
- 
- 
- 
Average 5.0 
4-27 
3.61 
2.32 
i-59 
1.48 
4-77 
3-63 
3-30 
TABLB II.-Distribution Series.- ^Continued.') 
10 Reducedto equal wattages (800 watts) these installations give the following average 
illumination values in foot-candles for the receiving test plate in the positions specified 
above : Direct system : horizontal, 4.54; vertical, 2.2; 450, 4.33. Semi-indirect system : 
horizontal, 4.49; vertical, 1.67; 450, 3.82. Indirect: horizontal, 3.61; vertical, 1.48; 450, 3.3. 
It may not be out of place to suggest here that a careful study of the illuminating effi- 
ciency of different types of lighting units should be made under conditions that are strict- 
ly comparable for a wide range of variation. Such tests should be made under common 
supervision in a model room so constructed as readily to permit of the kind of variations 
needed; and should be, if possible, paralleled by tests for the efficiency of the eye. In 
working towards a reconstruction of lighting conditions, it is obvious that tests for the 
efficiency of the eye and for illuminating efficiency should go hand in hand. TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
456 
Figs. 2, 3 and 4 are taken from the series of 9 photographs (see 
Figs. 2-10, op. cit., pp. 4i6b-4i6d) showing the illumination effects 
produced by the three systems of lighting. In these figures are 
given the brightness measurements of all surfaces having very 
high or very low brilliancy. The spot measured is indicated by a 
cross and the numerical value of the brightness measurement in 
candlepower per square inch is printed nearby. These spots are 
also lettered for convenience of reference in the intensity series. 
That is, since several installations were used in the intensity 
series, it was found convenient to express these values in tabular 
form and to identify them with the surfaces measured by means 
of letters. These photographs were taken from a point in line 
with the four positions of the observer as near to the south wall 
of the room as was possible; but owing to the narrow field of the 
camera as compared with the binocular field, these views include, 
for example, only about one-half of the field of vision of the 
observer at the test station nearest to this wall of the room. The 
camera's field in this position corresponds in fact very closely to 
the field presented to the observer seated at the center of the 
room. While, therefore, not all of his field of view for all of the 
positions at which tests were made is covered by the brightness 
measurements shown in the photographs, still the order of bright- 
ness difference present in the field of view for the different sys- 
tems is well represented by these measurements, as can be seen by 
an inspection of the preceding photographs (see also Figs. 2-10, 
op. cit., pp. 4i6b-4i6d) and from the descriptions of the installa- 
tions used. In order to facilitate certain features of comparison 
such as, for example, the evenness of surface brightness for each 
system for all of the room; for all of the room but the sources of 
light; and for all of the room but the sources and the spots 
above the sources, the brightness measurements shown in Figs. 
2, 3 and 4 are also given in tabular form. These measurements 
and the letters identifying them with the surfaces measured, are 
given in Table III. In making the comparison it should be noted 
that the spots mentioned are not in all cases identical for the 
three systems. That is, owing to the different effects produced 
by the different reflectors, the same spots were not always con- 
spicuously light or dark for the three systems. The letters, E, 457 
FERREL and rand: EFFICIENCY of the Eye 
F, G, etc., may then refer to entirely different spots in case of 
the three systems. 
TABLE III.-Distribution Series. 
Showing the brightness measurements in candlepower per square inch for 
the surfaces A, B, C, D., etc., see Figs. 2, 3 and 4. 
Surface 
Direct 
Semi-iudirect 
Indirect 
measured 
system 
system 
system 
A 
1000.0000 
0.710 
O.I38 
B 
0.3816 
0.057 
0.0715 
C 
0.517 
0.093 
0.066 
D 
0.010 
0.059 
0.0022 
B 
0.00296 
O.OO29 
O.OO3O 
F 
0.0044 
O.OO33 
O.OOI23 
G 
0.0078 
O.OO53 
O.OO49 
H 
0.0077 
0.006 
0.0040 
I 
0.0075 
0.0062 
O.OO42 
J 
0.0014 
0.0010 
O.OOO95 
K 
0.0063 
0.0046 
0.00255 
L 
0.0042 
0.0027 
O.OO246 
M 
0.0065 
0.0051 
O.OO352 
N 
0.0047 
0.0027 
0.00272 
O 
0.0074 
0.0066 
O.OO343 
P 
0.006 
0.00484 
O.OO3O8 
Q 
0.00396 
TABLE IV.-Distribution Series. 
Showing the brightness measurements in candlepower per sq. in. of the test 
card, reading page horizontal, and reading page in the 45 deg. position for 
Positions I, II, III, and IV, for the direct, semi-indirect, and indirect systems. 
Position 
of observer 
Surface measured 
Direct 
system 
Semi-indirect 
system 
Indirect 
system 
I 
Test card 
• 0.00308 
0.0030 
O.OO299 
Reading page horizontal.... 
Reading page 450 position.. 
• 0.0057 
0.0058 
0.0088 
■ 0.004 
O.OO39 
0.00431 
II 
Test card 
• 0.00506 
O.OO453 
0.0046 
Reading page horizontal.... 
. 0.0088 
O.OIO7 
0.0088 
Reading page 450 position.. 
. 0.0068 
O 00726 
0.00792 
III 
Test card 
• 0.0055 
0.00462 
O.OO453 
Reading page horizontal. ... 
• 0.0092 
0.0087 
0.00814 
Reading page 450 position. . 
• 0.00704 
0.0077 
O.OO594 
IV 
Test card 
• 0.0066 
O.OO475 
O.OO453 
Reading page horizontal.. ■. 
• 0.00814 
O.OO572 
O.OO572 
Reading page 450 position. . 
. 0.0063 
O 00484 
0.00484 
In Table IV are given the brightness measurements in candle- 
power per square inch for the test card and the reading page for 
the four positions of the observer: I, II, III and IV, for the 
direct, semi-indirect and indirect systems. The measurements of 
the reading page were taken at the point of work for the four 
positions of the observer with the book in the horizontal and 45 
deg. position. During work the book was held in the 45 deg. 
position. 
In Tables V and VI are shown some prominent ratios of sur- TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
458 
face brightness for the three systems.11 (See also Table VIII, 
op. cit., p. 421.)12 
In compiling these ratios it has been considered important to 
make a comparative showing- for the three systems (a) of the 
extremes of surface brightness; and (b) of the relation of the 
brilliancy of objects in the surrounding field to the surface bright- 
ness at the point of work. The extremes of surface brightness 
11 In attempting to make comparisons of the effect of the different magnitudes of 
brightness ratios, one obviously must bear in mind that the surfaces between which 
the ratios are established are not in all cases in the same position in the field of 
vision for the three systems. For example, the brightest surfaces in case of the 
indirect system, namely, the spots on the ceiling directly above the reflectors, are 
farther removed from the direct line of vision of the observer in the working position 
than were the brightest surfaces in case of the direct and semi-indirect systems. 
The position of the surface in the field of vision would come into question, for ex- 
ample, in making a determination of the maximum value of brightness difference the 
eye is adapted to stand. While we have done a great deal of work on the effect of 
position of the brilliant surface in the field of vision in our investigation of the 
causes of discomfort, we have made no especial investigation of this point in relation 
to loss of efficiency. Doubtless what we shall all have to bear in mind is that, even 
in 'the end, we cannot hope to specify narrowly what is most favorable, etc. In 
lighting conditions. The factors that entei' into the concrete lighting situation are so 
complex or rather are so variable and so rarely duplicated that we can hope to make 
general specifications with regard to what is most favorable, for example, only within 
very broad limits. If one wishes to work the conditions down to a finer point than 
this, the particular installation must be tested in situ. We are at present working 
on a test which we hope will serve this purpose better than the test which has been 
used in the work described in the preceding papers. 
12 Table VIII, (op. cit., p. 421) was compiled from Tables IV-VII of that 
paper to show the mean variation in surface brightness for all the surfaces measured 
for the direct, semi-indirect, and indirect systems. In referring back to that paper 
it may not be out of place to call to mind again that the percentages given in Table 
VIII seem to indicate that the great advantage of the indirect over the other systems 
of lighting we have used with regard to the factor, evenness of surface brightness, 
comes, primarily at least, from its provisions for shielding the eye from the light 
source rather than from any conspicuously greater evenness of illumination given 
by it to the objects in the field of view. In fact, as may be seen from that table, all 
the systems give a fairly even distribution of surface brightness outside of the 
source and the surfaces immediately surrounding it. 
The need of keeping surface brightness within certain limits and the primary im- 
portance of properly shielding the eye from the source, to the accomplishment of this 
desideratum, are obvious. Doubtless many ways will be devised in course of time for 
cutting down useless and harmful brightness differences in lighting effects. For exam- 
ple, the possibility is here suggested of producing a still smaller brightness difference 
than is given by the indirect reflectors of the type we have employed, by using semi- 
indirect reflectors of such a density as to give a surface brilliancy equal to that of 
the spot of light cast upon the ceiling. The value of this brilliancy, because of the 
larger area of luminous surface presented, could then be made smaller than that of 
the ceiling spot cast by the indirect reflector and still give the same amount of light 
to the room. A similar effect may be obtained with the indirect reflector by using 
lamps of lesser wattage and adding the light needed to make up the deficiency by 
installing directly beneath the reflector lamps of low wattage in translucent en- 
closures of a density that gives a surface brilliancy equal to that of the ceiling spots. 
The effect of both of these devices would be to lower the surface brilliancy for a 
given light flux by increasing the area of the luminous surface. Whether either de- 
vice would be advisable from other standpoints we are not at present prepared to 
say. 459 
FErrEE and rand: efficiency of the Eye 
are shown by giving the ratios between surfaces of the first, 
second, third, etc., order of brilliancy and the surface of the 
lowest order of brilliancy; and the comparison of the brilliancy 
of objects in the surrounding field to the brightness at the point 
of work by giving the ratios of the surfaces of the first, second, 
and third order of brilliancy to the brightness of the test card 
and the reading page in the working position. The following 
points may be noted, (i) The illumination effects produced by 
the direct system are characterized by great extremes of surface 
brightness, and a high ratio of brilliancy of objects in the sur- 
rounding field to the surface brightness at the point of work. 
These effects are much less pronounced for the semi-indirect sys- 
tem and still less for the indirect. (2) A comparison of this 
table with the tables giving loss of efficiency as the result of work 
shows that while the extremes of surface brightness are enor- 
mously larger for the direct than for the semi-indirect system, 
the eye loses almost as much in efficiency for three hours of work 
under the semi-indirect as under the direct system. That is, the 
greatest ratio of brightness for the direct system is over one 
thousand times as much as the greatest ratio for the semi-indirect, 
while the difference in loss of efficiency for the two systems is 
comparatively insignificant. On the other hand, the greatest 
ratio of brightness for the semi-indirect system is only about five 
times as much as for the indirect; while the difference in loss of 
efficiency for three hours of work is very large, this loss of effi- 
ciency for three hours of work for the indirect system being, it 
will be noted, very small indeed. This seems to indicate (a) that 
for the scale of brightness magnitudes and the illumina- 
tion effects present in this series of experiments the gradation 
of surface brightness for the indirect system is very close 
to what the eye is adapted to stand without loss of efficiency; 
and (b) that an increase in difference in brightness above 
this point is followed at first by a rapid increase in loss of 
efficiency and later by a much slower increase. In the intensity 
series, in the work of the former paper, it will be remembered, 
the following points also came out. (1) The effect of size of 
ratio on loss of efficiency is different for different orders of mag- 
nitude of brightness. And (2) the size of the brilliant object, as TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
460 
well as its brilliancy, is of importance. That is, within certain 
limits, as yet undefined, an increase in the area of the brilliant 
surface causes an increase in loss of efficiency. 
In Table V the ratios were compiled from measurements show- 
ing the extremes of brightness of prominent surfaces in the room. 
In Table VI they were compiled to show the relation of the 
brilliancy of objects in the surrounding field to the surface bright- 
ness at the point of work for the positions of the observer, I, II, 
III and IV13 (see Fig. I, p. 452a). In general a falling off in 
the magnitude of brightness differences in the field of view 
will be noted in order from the Positions I to IV. This 
falling off is greatest for the direct system, next greatest for the 
semi-indirect, and least for the indirect. Thus there is not only 
a decrease in the number of surfaces in the field of view show- 
ing a high brilliancy from Positions I to IV, but also a decrease 
in the magnitude of brightness difference between the surfaces 
of high brilliancy and the test card, between these surfaces and 
the reading page, etc., especially for the direct and semi-indirect 
systems. An inspection of the table for loss of efficiency shows, 
roughly speaking, a correspondingly marked decrease in loss of 
efficiency from Positions I to IV for the systems which show the 
marked decrease in brightness difference, that is, for the direct 
and semi-indirect systems. The decrease in loss of efficiency, 
it will be noted, is practically nothing for the indirect system. 
Thus not only much less loss of efficiency is sustained by the 
eye for the indirect units used, but the results are much more 
independent of the position of the observer in the room. 
The loss of efficiency for the Positions I, II, III and IV for the 
three systems is shown in Table VII.14 
13 It may also be of interest to the reader to work out for these four positions the 
ratios: lightest to darkest, darkest to test card, darkest to reading page, etc. 
14 Obviously in the consideration of the effect of a given lighting situation on 
the ability of the eye to hold its efficiency for a period of work, the age of the ob- 
server and the condition of his eyes should be taken into account. All the observers 
that have been employed by us in this work were under 26 years of age. Following 
is a clinic report of the eyes of the observer whose results are given tn the following 
table, made by Dr. Wm. Campbell Posey of Philadelphia. 
Observer R. 
With glasses.-Vision of right eye = 20/25. Far muscle test = O esophoria. 
Vision of left eye = 20/20- Near muscle test = orthophoria. 
Ophthalmoscopic examination.-Right eye - mixed astigmatism, diopter. 
Left eye = hyperopic astigmatism, diopters. 
{Continued on next paged) 461 
FERREE AND rand: EFFICIENCY OF THE EyE 
Chart I gives a graphic representation of the results of this 
table. Loss of efficiency is plotted along the ordinate and time 
of work along the abscissa. Each of the large squares along the 
abscissa represents an hour of work and along the ordinate an 
integer of the ratio, time clear to time blurred. The effect on 
loss of efficiency of the number and magnitude of brightness of 
surfaces of high brilliancy, especially of primary sources, in the 
field of view is obvious from these charts. The chart for position 
IV, however, shows that there is still a considerable difference in 
the loss of efficiency produced by the three systems, even when 
there are no sources or other surfaces of high brilliancy in the field 
of view. The indirect system still gives the least loss of efficiency, 
the semi-indirect next, and the direct the most. As may be seen 
in Figs. 2, 3, and 4, and in Tables III and VI there was little dif- 
ference in the evenness of surface brightness in the field of view 
presented to the observer in this position, certainly none that could 
be considered of consequence in favor of the indirect system. The 
above results seem to indicate, therefore, that while the evenness of 
surface brightness is an important factor it is not the only factor 
in a lighting situation which may influence the amount of loss of 
efficiency sustained by the eye as the result of a period of work 
We wish to repeat in this paper what was very strongly empha- 
sized in our former paper, namely, that the units we have em- 
ployed were not selected as fully representative of the classes 
direct, semi-indirect, and indirect. Agreement in fact has not yet 
been reached with regard to what falls within each of these 
External condition.-Adduction good; eyes slightly divergent under cover; cornea 
clear; pupils, 2^2 mm.; irides respond equally and freely to 
light, accommodation, and convergence stimuli. 
Glasses worn during test.-Right eye = -S., 0.50 D.;-C., 0.37D., x 16010 
Reft eye = -C., 0.50 D., x 1800 
Rest the former paper has not appeared in print before this one is presented it 
may be well to make some mention here also of the reproducibility of results that 
may be obtained for dur test for loss of efficiency. The mean variation of the ratio, 
time clear to time blurred for the same observer working under conditions as nearly 
constant as possible, is very small indeed. The order of magnitude of the mean 
variation of the test for the fresh eye was obtained as follows. Beginning at 9 a. m. 
five 3 minute tests were run with a rest period of 20 minutes between each test 
This was done with all observers on several days under each system of lighting em- 
ployed. The rest period was taken in each case in a room lighted by daylight, with 
the observer facing a wall with an evenly lighted matt surface. For a single series 
of five tests the variation in the time seen clear in the 3 minute periods have 
always fallen within 1 per cent, for all of the observers we have used and for all 
systems of lighting. TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
462 
classes. The units employed were chosen rather to show the 
effect of varying the factors we have grouped under the heading 
of distribution on the ability of the eye to maintain its efficiency 
for a period of work. We hope ultimately to determine the 
CHART I.-Distribution Series. 
Showing the effect on loss of efficiency of varying the observer's position in 
the room, or the number of bright sources, primary and secondary, in the 
field of vision. 
POSITION I 
POSITION II 
POSITION IV 
POSITION III 
limits between which each of these factors may vary in a se- 
lected range of lighting situations, without damage to the eye, es- 
pecially the factor surface brightness. These most favorable 
conditions will then serve as a goal to be attained whatever prin- 
cipal of lighting is employed. 463 
FERREE AND RAND: EFFICIENCY OF THE EYE 
Ratio 
Direct system 
Semi-indirect system 
Indirect system 
Lightest to darkest 
IOOO/O.OOI4 = 714,285 
0.710 /o.ooi = 710 
0.138 ' /o. 00095 = 145 
2nd lightest to darkest...... 
0.3816/0.0014 = 123.9 
0.093 /o.ooi = 93.0 
0.0715 /o 00095 = 75.2 
3rd lightest to darkest 
O.O517/O.OOI4 = 37.O 
• 0.059 /o.ooi = 59.0 
0.066 /o. 00095 = 69.4 
4th lightest to darkest 
O.OI /0.0014 = 7.14 
0.057 /o.ooi = 57.0 
0.0049 0/.00095 = 5.I5 
5th lightest to darkest 
0.0078/0.0014 = 5.57 
0.0066 /o.ooi = 6.6 
0.0042 /o.00095 = 4-42 
6th lightest to darkest 
0.0077/0.0014 - 5.50 
0.0062 /o.ooi = 6.2 
0.0040 /o.00095 = 4.21 
7th lightest to darkest 
O.OO75/O.OOI4 = 5.35 
0.0060 /o.ooi = 6.0 
0.00352/0.00095= 3.70 
8th lightest to darkest 
0.0074/0.0014= 5.28 
0.0053 /o.ooi = 5.3 
0.00343/0.00095 = 3.61 
9th lightest to darkest 
0.0065/0.0014 = 4.64 
0.0051 /o.ooi = 5.1 
0.00308/0.00095 = 3.24 
10th lightest to darkest 
0.0063/0.0014 = 4.5 
0.00484/0.001 = 4.84 
0.0030 /o.00095 = 3.15 
1 ith lightest to darkest 
0.006 /0.0014 = 4.28 
0.0046 /o.ooi = 4.6 
0.00255/0.00095 = 2.68 
12th lightest to darkest ....... 
O.OO47/O.OOI4 = 3.36 
0.0033 /o.ooi = 3-3 
0.00246/0.00095 = 2.57 
13th lightest to darkest 
0.0404/0.0014 = 3.14 
0.0029 /o.ooi = 2.9 
0.0022 /o.00095 = 2.31 
14th lightest to darkest 
0.0042/0.0014 = 3.00 
0.0027 0/00.1 = 2.7 
0.00123/0.00095 - 1.29 
TABLE V.-Distribution Series. 
Ratios showing the extremes of surface brightness for the direct, semi-direct, and indirect systems used. 464 
transactions of ieeuminating engineering society 
TABLE VI.-"Distribution Series. 
Ratios showing the relation of the brilliancy of objects in the surrounding field to the surface brightness at the point of work 
for Positions I, II, III, and IV for the direct, semi-indirect, and indirect systems used. 
Position 
of observer 
Ratio 
Direct system 
Semi-indirect system 
Indirect system 
1 
Lightest to test card 
1000/0.00308 
= 
324,674 
0.7 IO 
/0.003 
= 
236.7 
0.138 
/o. 00299 
= 
46.0 
Lightest to reading page 
I000/0.004 
= 
250,000 
0.710 
/0.0039 
= 
182.0 
0.138 
/o. 00431 
= 
32.0 
2nd lightest to test card 
• - • 0.3816/0.00308 
= 
123.9 
0.093 
/0.003 
= 
31.0 
0.0715 
/0.00299 
= 
24.0 
2nd lightest to reading page • • 
... 0.3816/0.004 
= 
95-3 
O.O93 
/0.0039 
= 
24.0 
0.0715 
/o. 00431 
= 
16.5 
3rd lightest to test card 
•• 0.0517/0.00308 
- 
16.8 
0.059 
/0.003 
= 
29.7 
0.066 
/o.00299 
= 
22.0 
3rd lightest to reading page • • • 
.0.0517/0.004 
= 
12.9 
0.059 
/0.0039 
= 
15.0 
0.066 
/o. 00431 
= 
15-0 
II 
Lightest to test card 
1000/0.00506 
= 
197,628 
0.710 
/o. 00453 
= 
156.7 
0.138 
/0.0046 
- 
30.0 
Lightest to reading page 
I000/0.0068 
= 
147,059 
' 0.710 
/0.00726 
= 
97-8 
0.138 
/o. 00792 
= 
17.0 
2nd lightest to test card 
. .. 0.3816/0.00506 
- 
75-4 
0.093 
/o. 00453 
= 
20.5 
O.O715 
/0.0046 
= 
15-5 
2nd lightest to reading page • • 
... 0.3816/0.0068 
= 
56.0 
O.O93 
/o. 00726 
= 
12.8 
O.O715 
/0.00792 
= 
9.0 
3rd lightest to test card 
... 0.0517/0.00506 
- 
10.2 
O.O59 
/0.00453 
= 
13-0 
0.066 
/0.0046 
=• 
U-3 
3rd lightest to reading page . . • 
. .. 0.0517/0.0068 
= 
7-6 
0-059 
/o. 00726 
= 
8.0 
0.066 
/o. 00792 
= 
8.0 
III 
Lightest to test card 
I000/0.0055 
= 
181,818 
0.710 
/o. 00462 
= 
154.0 
0.138 
/0.00453 
= 
30-4 
Lightest to reading page 
1000/0.00704 
- 
142,055 
0.710 
/0.0077 
= 
92.0 
0.138 
/o. 00594 
= 
23-0 
2nd lightest to test card 
. 0.3816/0.0055 
= 
69.0 
0.093 
/o. 00462 
- 
26.0 
0.0715 
/o. 00453 
= 
15-8 
2nd lightest to reading page • • 
. 0.3816/0.00704 
- 
54-o 
0093 
/o 0077 
= 
12.0 
0.0715 
/o. 00594 
12.0 
3rd lightest to test card 
. .. 0.0517/0.0055 
9.0 
0.059 
/o.00462 
- 
12.8 
0.066 
/0.00453 
= 
12.0 
3rd lightest to reading page • • • • 
. . 0.0517/0.00704 
= 
0-7 
0.059 
/0.0077 
= 
7-7 
0.066 
/o. 00594 
= 
11.0 
IV 
Lightest to test card 
• • 0.0063/0.0066 
= 
0-954 
0.00484/0.00475 
= 
1.019 
0.00484/0.00453 
- 
1.068 
Lightest to reading page 
. . 0.0066/0.0063 
= 
1-047 
0.00475/0.00484 
= 
0.785 
0.00453/0.00484 
- 
0.936 
2nd lightest to test card 
■ • 0.0014/0.0066 
= 
0.212 
0.010 
/o. 00475 
= 
0.0210 
0.00095/0.00453 
= 
0.209 
2nd lightest to reading page . • 
.. O.OOI4/O.OO63 
0.222 
0.010 
/o. 00484 
= 
0.00207 
0.00095/0.00484 
= 
0.196 465 
FERREE and rand: EFFICIENCY OF the EYE 
TABLE VII.-Distribution Series. 
Showing the effect on loss of efficiency of varying the observer's position in the room, or the number of light sources 
primary and secondary, in the field of vision. At Position I, six light fixtures were in the field of vision ; at Position II, four 
light fixtures ; at Position III, two light fixtures ; and at Position IV, no light fixtures. 
Position 
Lighting 
Intensity foot- 
candles 
Maximal 
distance at 
which test 
object can 
be seen 
Work- 
ing 
Total 
time 
Total 
time 
Total 
time 
clear -j- 
total 
time 
Ratios 
reduced 
to comm- 
Hori- 
Verti- 
of observer system Watts Volts 
zontal 
cal 450 Time 
clear 
distance 
clear 
blurred 
blurred 
standard 
I. 
Indirect 800 107 
5-2 
1.36 3.5 9 A.M. 
84-5 
67-5 
135 
45 
3-0 
3-5 
12 M. 
84-5 
67.5 
132 
48 
2-75 
3-2 
Semi-indirect 760 107 
5-8 
I.45 4.0 9 A.M. 
80.5 
68.5 
142 
38 
3-73 
3-5 
. 12 M. 
79-5 
68.5 
92 
88 
1.04 
0-97 
Direct 880 107 
4.2 
1.41 2.6 9 A.M. 
81.0 
68.0 
139 
4i 
3-39 
3-5 
12 M. 
78.0 
68.0 
71 
109 
0.65 
0.67 
ii. 7 
Indirect 800 107 
5-i 
I.98 4.2 9 A.M. 
86.5 
74-0 
141 
39 
3-61 
3-5 
12 M. 
86.5 
74-o 
139 
4i 
3-39 
3-28 
Semi-indirect 760 107 
6.1 
2.5 4.7 9 A.M. 
87.0 
75-° 
138 
42 
3-28 
3-5 
12 M. 
87.0 
75-o 
123 
57 
2.19 
2.2 
Direct 880 107 
4-65 
2.75 4-4 9 a.m. 
87.0 
75-o 
134 
46 
2.91 
3-5 
12 M. 
84.0 
75-o 
99 
81 
1.22 
1.46 
in. 
Indirect 800 107 
3-9 
2.1 4.0 9 A.M. 
84.0 
70.0 
13° 
50 
2.6 
3-5 
12 M. 
84.0 
70.0 
127 
53 
2.4 
3-23 
Semi-indirect 760 107 
5-0 
2.6 5.4 9 A.M. 
84.0 
69.0 
131 
49 
2.67 
3-5 
12 M. 
84.0 
69.0 
122 
58 
2.1 
2-73 
Direct 880 107 
4-o 
2.9 4.6 9 A.M. 
83-5 
70.0 
148 
32 
4.6 
3-5 
12 M. 
83.0 
70.0 
137 
43 
3-8 
2.41 
IV. 
Indirect 800 107 
2-9 
2.1 3.6 9 A.M. 
84.0 
70.0 
139 
4[ 
3-39 
3-5 
12 M. 
84.0 
70.0 
137 
43 
3.18 
3-27 
Semi-indirect 760 107 
3-4 
3.0 4.4 9 A.M. 
83-5 
7o.5 
139 
4i 
3-39 
3-8 
12 M. 
83-5 
70-5 
132 
48 
2-75 
2.83 
Direct 880 107 
3-o 
3-4 4-5 9 A-M. 
82.0 
69.0 
119 
61 
i-95 
3-5 
' 12 M. 
81.0 
69.0 
108 
72 
i-5 
2.7 TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
466 
As was also stated in our former paper our next step in this 
division of the work will be to determine by using reflectors of 
different degrees of opacity the effect on loss of efficiency when 
the light is distributed to the plane of work both by the direct 
and indirect principles of lighting. That is, reflectors of differ- 
ent densities: prismatic, alba, opal lux, totally opaque, etc., will be 
used turned up and down. In each case the installation will be 
made with special reference to giving the best results obtainable 
for the particular type of unit employed; and the factors: even- 
ness of illumination, diffuseness of light, the angle at which the 
light falls on the work, and the evenness of surface brightness, 
will be varied separately in turn and the effect on loss of effi- 
ciency will be determined. Moreover, if it is found that the 
factors in question cannot be studied in sufficient detail in the 
concrete lighting situation, the work will be supplemented by 
more abstract investigations. The results of this series of tests 
should give us among other things, for example, a still better 
idea of what amount of brightness difference the eye is adapted 
to stand, and the comparative effect of different ratios of surface 
brightness on loss of efficiency. 
INTENSITY SERIES. 
In the work of the preceding paper we had undertaken to 
determine the most favorable intensities for the three types of 
artificial lighting we had used in the distribution series, and the 
effect of varying intensity with the particular grouping of distri- 
bution factors represented in each case. As was stated in the 
introduction of the present paper, this work was completed for 
the direct and semi-indirect systems but not for the indirect. For 
the semi-indirect installation it will be remembered that the eye 
fell off heavily in efficiency for all intensities with the exception 
of a very narrow range on either side of 2.2 foot-candles, meas- 
ured at the point of work with the receiving test plate of the 
photometer in the horizontal plane. For the direct installation 
no intensity could be found for which the eye did not lose a great 
deal in efficiency as the result of work. For the indirect installa- 
tion, however, as the following data will show, a comparatively 
wide range of intensity may be used without the eye suffering 467 
FERREk AND RAND.' EFFICIENCY OF THE EYE 
any considerable loss of efficiency as the result of three hours of 
continuous work. 
The tests were made in the same room, with the same fixtures, 
and in general, with the same conditions of installation and 
methods of working as were described in the work on the distri- 
bution factors. To secure the various degrees of intensity needed, 
lamps of different wattage were employed. These were selected 
from a series of tungsten lamps ranging from 15 to 100 watts. In 
order to keep the distribution factors as nearly constant as pos- 
sible for a given type of system, the lamps used in making the 
test for that type of system were all of one wattage, i. e., were 
all 15's, 25's, 40's, 6o's or 100's'. 
For the indirect system the total range of intensity employed 
is shown by the following figures. The series .was begun with 
25-watt lamps, and consisted of 25, 40, 60, and 100-watt lamps. 
For the 25 watt lamps the photometer reading at the point of 
work with the receiving test plate in the horizontal plane showed 
1.33 foot-candles of light; with the receiving test plate in the 
vertical plane, 0.39 foot-candle; and with the receiving test plate 
in the 45 deg. plane, 0.87 foot-candle. For the 100-watt lamps 5.2 
foot-candles were obtained with the receiving test plate in the 
horizontal plane; 1.36 foot-candles with the test plate vertical; 
and 3.5 foot-candles with the test plate inclined 45 deg. The tests 
for loss of efficiency15 showed probably a slight advantage for the 
25-watt lamps, although the difference in result for the different 
intensities is sufficiently near in value to the mean variation of 
the test as to be scarcely worthy of consideration. 
As was the case for the direct -and semi-indirect installations, 
the following specification was made of the illumination effects 
produced by the indirect installation. (1) Illumination measure- 
15 In conducting these tests it was found necessary to allow a period of adaptation 
without work, to the illumination of the room before the first test was taken. If this 
were not done, especially in case of the lower intensities of lights used, the changing 
sensitivity of the eye to the intensity of light employed, produced a noticeable change 
in the visual acuity between the times the tests before and after work were taken. 
Since the distance of the test card was kept the same for the two tests, this change in 
the visual acuity tended to influence the ratio, time clear to time blurred. To deter- 
mine the length of time needed under a given intensity of light to insure a constant 
acuity, so far as adaptation is concerned, preliminary tests were made as follows. 
The acuity of the observer was taken every 3 minutes until no noticeable change 
was found. This length of time was then always allowed for that observer as an 
adaptation period prior to the loss of efficiency test conducted for the given intensity 
of illumination. TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
468 
ments were made for the highest intensity employed at the 66 
stations in the test room. These measurements were made in 
the way described in the preceding section. For the other in- 
tensities used, measurements were made at nine representative 
stations to show in a general way the order of magnitude of 
reduction produced by using the lamps of lower wattage. (2) 
Brightness measurements were made of the prominent objects 
in the room, such as the test card, the book of the observer, and 
all surfaces showing very high or very low brilliancy for all of 
the intensities. 
In Table VIII are given the illumination measurements at the 
66 stations for the highest wattages used, made with the re- 
ceiving test plate of the photometer in the horizontal, vertical, 
and 45 deg. planes. Tables IX and X show the illumination meas- 
urements at the nine representative stations for the other watt- 
ages employed in the series. The order of magnitude of- reduc- 
tion of the illumination of the room produced by using the lamps 
of lower wattage conforms pretty closely in each case, it will be 
observed, to the simple ratio of the wattages employed. (See 
foot-note to Table XII, p. 472.) As was the case for the semi- 
indirect system, noted in the preceding paper, socket extenders 
had to be used with the 25 and 40-watt lamps. That is, without 
the extenders these lamps, owing to their smaller size, came so low 
in the reflector as to change the distribution effects given by 
reflectors. For example, without the socket extenders for these 
shorter lamps, the spot of light on the ceiling was made smaller 
and correspondingly more brilliant. It was thought advisable to 
determine whether this comparatively small change in illumina- 
tion effects would cause any difference in the eye's ability to hold 
its efficiency for a period of work. In the specification of illumi- 
nating effects, therefore, measurements have been made for the 
25 and 40-watt lamps both with and without socket extenders. 
In Table IX illumination measurements for the 25 and 40-watt 
lamps are given with socket extenders, and in Table X illumina- 
tion measurements for these lamps are given without socket 
extenders. In Table XI are given the brightness measurements 
for the indirect installation for the different intensities used, both 
with and without socket extenders for the 25 and 40-watt lamps. 469 
FERREE AND RANDI EFFICIENCY OF THE EYE 
The points at which the measurements were taken are indicated 
by the letters A, B, C, D, E, F, etc., see Fig. 4, p. 452b. In Table 
XII are given the prominent brightness ratios for the different 
intensities used. Obviously an important point of comparison 
for the purposes of this investigation is the ratios with and with- 
out socket extenders for the 25 and 40-watt lamps. 
TABLE VIII.-Intensity Series. 
Showing the illumination measurements in fpot-candles for each of the 66 
stations represented in Fig. 1 for the indirect system used. 
tatior 
1 Horizontal 
Vertical 
45° 
Station 
Horizontal 
Vertical 
45° 
I 
1.22 
- 
- 
34 
4.0 
1.46 
3-1 
2 
1.26 
- 
- 
35 
5-4 
1.65 
4-9 
3 
1-32 
- 
- 
36 
5-3 
1-65 
4.0 
4 
i-47 
- 
- 
37 
5-8 
1.82 
4.0 
5 
2.2 
- 
38 
5-4 
I.72 
3-8 
6 
2-95 
- 
- 
39 
4-o 
i-43 
3-o 
7 
2-9 
- 
. - 
4o . 
2.3 
- 
- 
8 
3-o 
- 
- 
41 
2.1 
- 
- 
9 
2.2 
- 
- 
42 
3-5 
1.36 
2 8 
IO 
i-35 
- 
- 
43 
5-o 
1.78 
3-9 
ii 
1.66 
- 
- 
44 
5-2 
1.88 
4-3 
12 
2.7 
0.47 
1.43 
45 
5-2 
1-93 
4.2 
13 
4.1 
0.42 
1.96 
46 
5-2 
1.86 
4-i 
U 
4-4 
0.47 
2.3 
47 
3-6 
i-33 
2-9 
15 
4-5 
0.47 
2.3 
48 
3-7 
i-74 
3 3 
16 
4-i 
0.48 
2.1 
49 
4.8 
2.1 
4.0 
17 
3-15 
0.46 
1.6 
50 
4-9 
2.1 
4.1 
18 
2.2 
0.47 
1.63 
5i 
5-o 
2.15 
4-35 
19' 
2-5 
- 
- 
52 
4-7 
i-93 
4.0 
20 
3-4 
0.86 
2.2 
53 
3-6 
1.41 
3-o 
21 
4.6 
0-94 
3-° 
54 
2.8 
i-5 
2-9 
22 
4.8 
1.07 
2-9 
55 
3-9 
1.94 
3-75 
23 
_ 5-i 
1.1 
2-9 
56 
4.6 
2.1 
4-4 
24 
5-o 
1.04 
3-0 
57 
4-5 
2.2 
4-4 
25 
3-5 
o-75 
2.1 
58 
4.0 
2.0 
4.0 
26 
2.2 
- 
- 
59 
2-9 
1.76 
3-i 
27 
2.4 
-• 
- 
60 
2.6 
1.66 
2-9 
28 
3-7 
1.12 
2-5 
61 
3-i 
i-9 
3-5 
29 
5-2 
1.48 
3-4 
62 
3-2 
2.1 
3-7 
30 
5-4 
1-4 
3-6 
63 
3-o 
2.2 
3-5 
31 
5-2 
1.24 
3-6 
64 
3-i 
i-93 
3-4 
32 
5-o 
1-33 
3-4 
65 
2.25 
i-54 
2.6 
33 
3-7 
1.22 
2.6 
66 
i-35 
- 
- 
Average 3.61 
1.48 
3- 3 TRANSACTIONS OR ILLUMINATING kNGINLERING SOCIETY 
470 
TABLE IX.-Intensity Series. 
Showing the illumination measurements in foot-candles at nine repre- 
sentative stations for the different intensities used for the indirect system. 
Socket extenders used with the 40 and 25-watt lamps. 
Station 
Horizontal 
Vertical 
45° 
800 
480 
320 
200 
800 
480 
320 
200 
800 
480 
320 
200 
Card 
5-2 
3-o 
i-7 
i-33 
I.36 
0.765 
0-49 
0.39 
3-5 
1.97 
1.08 
0.87 
12 
2.7 
1-63 
0-97 
0.65 
0-47 
0.265 
0.18 
0.12 
i-43 
0.83 
0.48 
0.44 
16 
4.1 
2.2 
1-32 
1.11 
0.52 
o-33 
0.24 
0.14 
2.1 
1.22 
0.66 
0.6 
3[ 
5-2 
2.7 
1.84 
i-45 
I.24 
0-77 
0.51 
0-47 
3-6 
r-95 
1.16 
1.01 
34 
4.0 
2.25 
1.21 
1.0 
1.46 
0-79 
0.52 
0-49 
3-1 
1-63 
0.89 
0.78 
39 
4.0 
2.2 
1.6 
0.83 
1-43 
0.725 
0.51 
0-37 
3-o 
i-57 
1.04 
0.64 
45 
5-2 
2-75 
i-94 
1.28 
i-93 
0-99 
0.58 
0-53 
4.2 
2.18 
1-43 
1.0 
54 
2.8 
1.48 
1.16 
0.68 
i-5 
0.82 
0.63 
0.41 
2-9 
i-5i 
1.23 
0.68 
58 
4.0 
2.1 
i-3 
1.09 
2.0 
0-94 
0.64 
0.52 
4.0 
2.2 
1-3 
0.98 
Ave. 
3.61 
2.16 
i-44 
0.9 
1-32 
0.89 
0-59 
0-37 
3-3 
1.98 
1-32 
0.83 
TABLE X.-Intensity Series. 
Showing the illumination measurements in foot-candles at nine represen- 
tative stations for the different intensities used for the indirect system. No 
socket extenders used with the 40 and 25-watt lamps. 
Station 
Horizontal 
Vertical 
45° 
800 
480 
320 
200 
800 
480 
320 
200 
8c 0 
480 
320 
200 
Card 
5-2 
3-o 
1.48 
1.16 
I.36 
0.765 
0.407 
0.37 
3-5 
i-97 
0-95 
0.76 
12 
2.7 
1-63 
0.84 
o-5 
0-47 
0.265 
o.i39 
0.99 
i-43 
0.83 
0-44 
0.282 
16 
4.1 
2.2 
1.01 
0.96 
O.52 
0-33 
0.143 
0.14 
2.1 
1.22 
0.5 
0.48 
31 
5-2 
2-7 
1.48 
i-3 
I.24 
0-77 
0.462 
0.39 
3-6 
i-95 
1.0 
0.86 
34 
4.0 
2.25 
°-99 
1.0 
I.46 
0-79 
0.5 
0.45 
3-i 
1-63 
0.84 
0.8 
39 
4.0 
2.2 
1.63 
0.78 
i-43 
0.725 
0-44 
0.36 
3-o 
i-57 
0.98 
0.6 
45 
5-2 
2-75 
1.62 
1.18 
i-93 
0.99 
0.52 
0.48 
4-2 
2.18 
I-3I 
0.98 
54 
2.8 
1.48 
1-03 
0.63 
i-5 
0.82 
0.61 
0.41 
2-9 
i-5i 
1.18 
0.65 
58 
4.0 
2.1 
1.11 
0.87 
2.0 
0-94 
o-54 , 
0.42 
4.0 
2.2 
1.11 
0.83 
The results of the tests 'for the intensity series for the indirect 
system are given in Table XIII. Three hours was selected as the 
period of work in all of these experiments. The tests were taken 
only as Position I (see Fig. i, p. 452a), the position, it will be re- 
membered, at which six of the fixtures were in the field of view. 
It will be noted that there is practically no difference in the loss 
of efficiency of the eye for the different intensities of illumination 
when socket extenders were used for the shorter lamps. When 
socket extenders were not used for these lamps, quite a little loss 
of efficiency was experienced. This loss, moreover, was consider- 
ably greater for the shorter 25-watt lamps than for the 40-watt 471 
FERREE and rand: efficiency of the eyf 
lamps. Since the prominent variable in this case was intrinsic 
brilliancy of the ceiling spot above the reflector, the increased loss 
of efficiency can probably be ascribed primarily to this cause; or 
more comprehensively stated perhaps, to the change in the magni- 
tude of the brightness differences that were present in the field of 
vision. For example, the ratio, lightest to darkest for the 100- 
watt lamps was 145; it was 133 for the 60-watt lamps; 142 for 
the 40-watt lamps with socket extenders; and 135 for the 25- 
watt lamps with socket extenders. For the 40-watt lamps with- 
out socket extenders, however, this ratio was raised to 326, and 
for the 25-watt lamps without socket extenders it was raised to 
374. Similar changes were also made in the other ratios: lightest 
to test card, lightest to reading page, etc., as may be seen by in- 
specting Table XII. 
TABLE XI.-Intensity Series. 
Showing the brightness measurements in candlepower per square inch 
for the different intensities used for the indirect system at points indicated 
by the letters A, B, C, D, etc., see Fig. 4. 
320 watts 
200 watts 
With 
Without 
' With 
Without 
Surface 
800 
480 
socket ex- 
socket ex- 
socket ex- 
socket ex- 
measured 
watts 
watts 
tenders 
tender? 
tenders 
tenders 
A 
.0.138 
0.0704 
O.O539 
0.088 
O.O352 
0.0748 
B 
0.0715 
O.O385 
0.0252 
0.0231 
0.0165 
O.OI87 
c 
0.066 
O.O352 
O.O244 
0.022 
O.OI59 
0.0165 
D 
.0.0022 
O.OOO97 
O.OOO79 
0.00059 
0.00064 
0.0004 
E 
.O.OO3O 
0.000163 
O.OOII9 
0.0007 
O.OOO84 
O.OOO57 
F 
O.OOI23 
0.00040I 
O.OOO35 
0.00022 
O.OOO32 
0.00018 
G 
.O.OO49 
O.OOI69 
O.OOI45 
O.OOIOI 
0.00128 
O.OOO84 
H 
0.0040 
0.00163 
O.OOI29 
0.00092 
O.OOII 
O.OOO72 
I 
.0.0042 
O.OOI58 
O.OOI27 
0.0009 
O.OOII 
O.OOO68 
J 
.O.OOO95 
O.OOO53 
O.OOO38 
0.00027 
0.00026 
0.0002 
K 
.0.00255 
O.OOI23 
0.00088 
0.00088 
0.00074 
0.00064 
L 
O.OO246 
O.OOI2I 
0.00085 
0.00079 
0.00066 
0.00046 
M 
.0.00352 
O.OOI58 
0.00106 
0.00097 
0.00052 
0.0007 
N 
.0.00272 
O.OOIOI 
O.OOO76 
0.00061 
0.00066 
O.OOO44 
0 
.O.OO343 
0.00128 
0.00076 
0.00028 
0.00055 
O.OOOI9 
P 
.0.00308 
O.OOII9 
0.00067 
0.00027 
0.00041 
0.00018 
x 
O.OO299 
O.OOI54 
O.OOIO9 
0.0008 
0.00074 
O.OOO59 
Reading page 
horizontal . 
.0.0088 
O.OO4O5 
O.OO281 
0.0022 
0.00198 
O.OOl6 
Reading page 
45° position 0.00431 
0.00273 
O.OOI67 
O.OOI54 
0.00117 
0.0009 transactions of illuminating ENGINEERING society 
472 
TABLE XII.-Intensity Series. 
Showing some prominent ratios of surface brightness for the different intensities used for the indirect system. An important 
point of comparison for the purpose of this investigation is the ratios with and without socket extenders.16 
Ratio 
800 watts 
480 watts 
320 watts 
With socket extenders 
Lightest to darkest 
Lightest to test card 
Lightest to reading page 
2nd lightest to darkest 
2nd lightest to test card 
2nd lightest to reading page • • • • 
3rd lightest to darkest 
3rd lightest to test card 
3rd lightest to reading page 
0.138 /0.00095 = 145.0 
0.138 /o.00299 = 46 ° 
0.138/0.00431= 32.0 
0.0715/0.00095 = 75.0 
0.0715/0.00299 = 24.0 
0.0715/0.00431 = 16.5 
0.066 /o.00095 = 66.3 
.... 0.066 /o.00299 = 22.0 
0.066 /O.OO43I = I5.0 
320 watts 
Without socket extenders 
0.088 /o.00027 = 326.0 
0.088 /0.0008 = IIO.O 
0.0704/0.00053 = 133.O 
0.0704/0.00154 = 45.O 
0.0704/0.00273 =260 
0.0385/0.00053 = 72.6 
0.0385/0.00154 = 25.0 
0.0385/0.00273 - I4-° 
0.0352/0.00053 = 66.0 
0.0352/0.00154 = 22.0 
0.0352/0.00273 = 12.9 
200 
°-°539/°- 00038 = 142.0 
0.0539/0.00109 = 49.0 
0.0539/0,00167 = 32.0 
0.0252/0.00038 = 66.0 
0.0252/0.00109 = 23 0 
0.0252/0.00167 = 15.0 
0.0244/0.00038 = 64.0 
0.0244/0.00109 = 22.0 
0.0244/0.00167 = 14.6 
watts 
Ratio 
Lightest to darkest 
Lightest to test card 
With socket extenders Without socket extenders 
0.0352/0.00026 = 135 O 0.0748/0.0002 = 374 0 
0.0352/0.00074 = 47.O 0.0748/0 00059 =• 127.0 
0.0352/0.00117= 30.0 0.0748/0.0009 = 83.0 
0.0165/0.00026 = 63.5 0.0187/C.0002 = 93.5 
0.0165/0.00074 = 22.3 0.0187/0.00059 = 31.7 
0.0165/0.00117= 14.0 0.0187/0.0009 = 20.8 
0.0159/0.00026 = 6l.O 0.0165/0.0002 = 82.5 
0.0159/0.00074 = 2I.5 0.0165/0.00059 = 28.0 
O.OI59/O.OOII7 = 13.6 0.0165/0.0009 = 18.3 
st to test card, lightest to reading page, etc., should be approximately the 
reflector, to the ceiling, etc. A certain unevenness in these ratios will 
d 40-watt lamps. This is no more than is to be expected because socket 
same relation of lamp to reflector for the 25 as for the 40-watt lamp, nor 
nd 100-watt lamps. Obviously too the 60-watt lamps did not sustain the 
2nd lightest to darkest 0.0231/0.00027 = 85.5 
2nd lightest to test card 0.0231/0.0008 = 28.9 
2nd lightest to reading page 0.0231/0.00154 = 15.0 
3rd lightest to darkest 0.022 /0.00027 = 81.5 
3rd lightest to test card 0.022 /0.0008 = 27.5 
3rd lightest to reading page 0.022 /o.00154 = I4-2 
16 Theoretically considered, the above ratios: lightest to darkest, lighte 
same for all of the wattages, if the lamps sustain the same relation to the 
be noted, however, even when socket extenders were used with the 25 an 
extenders of only one length could be obtained, and these did not give the : 
the same relation for either of these lamps as was obtained for the 60 a: 
same relation to the reflectors as did the 100-watt lamps. 473 
FERREE AND RAND.- EFFICIENCY OF THE eye 
TABLE XIII.-Intensity Series. 
Showing the effect on the efficiency of the eye of varying the intensity of light in an indirect system composed 
of 8 units, clear tungsten lamps. 
Foot-candles 
Maximal 
distance at 
which test 
object can 
be seen 
clear 
Total 
time 
clear 
Total 
time 
blurred 
Total 
time 
clear -t- 
total 
time 
blurred 
Ratios 
reduced to 
common 
standard 
Watts 
Horizon- 
tal 
Verti- 
cal 
45° 
Time 
ing 
distance 
800 
5-2 
I.36 
3-5 
9 A.M. 
84-5 
67-5 
135 
45 
3-0 
3-5 
12 M. 
84-5 
67-5 
132 
48 
2-75 
3-2 
480 
3-o 
O-765 
i-97 
9 A.M. 
82.5 
65.0 
150 
30 
5-o 
3-5 
12 M. 
82.0 
65.0 
148 
32 
4.6 
3-22 
320 
With socket extenders •• 
i-7 
0-49 
1.08 
9 A.M. 
8l.O 
66.0 
140 
40 
3-5 
3-5 
12 M. 
8l.O 
66.0 
137 
43 
3-18 
3.18 
320 
Without socket extenders 
1.48 
0.407 
0-95 
9 A.M. 
79-o 
64.0 
145 
35 
4-i 
3-5 
12 M. 
77-o 
64.0 
138 
42 
3-29 
2-7 
200 
With socket extenders •. 
i-33 
0-39 
0.87 
9 A.M. 
79-o 
64.0 
122 
58 
2.1 
3-5 
12 M. 
79-o 
64.0 
120 
60 
2.0 
3-3 
200 
Without socket extenders 
1.16 
o-37 
0.76 
9 A.M. 
78.0 
62.0 
149 
3i 
4.8 
3-5 
12 M. 
76.0 
62.0 
135 
45 
3-o 
2.1 TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
474 
CHART II.-Intensity Series. 
Showing a comparison of the effect on the efficiency of the eye of varying the intensity of light for the four installations of 
lighting used: the indirect, semi-indirect, and direct systems, 8 lamps; and the direct system, 16 lamps.17 
lighting system: Semi-indirect 
Lighting system: Indirect 
Lighting system: Direct (8 lamps) 
Lighting system: Direct (16 lamps) 
Foot-candles 
Foot-candles 
Foot-candles 
Foot-candles 
Hori- Verti- 
Hori- Verti- 
Hori- Verti- 
Hori- Verti- 
Watts Volts 
zontal cal 450 
Watts Volts 
zontal cal 450 
Watts Volts 
zontal cal 450 
Watts Volts zontal cal 450 
A 200 107 
1.6 0.45 1.15 
A 
200 107 
I.33 O.39 0.87 
A 120 107 
' 0.64 O.32 O.49 
A 240 107 1.23 0.54 0.935 
B 200 Iio 
I.72 0.484 I.29 
B 
320 107 
1.7 O.49 I.08 
B 200 107 
1.16 O.45 0.85 
B 365 107 1.6 0.6 1.33 
C 320 107 
2.2 O.58 I.52 
C 
480 107 
3.0 0.765 I.97 
C 320 107 
I.97 0.65 I.39 
C 400 107 1.86 0.8 1.46 
D 320 Iio 
B 480 107 
F 800 107 
X 760 107 
2.3I 0.62 I.6l 
3-3 °-94 2.4 
6.8 1.82 4.5 
5.8 1.45 4.0 
D 
800 107 
5-2 1.36 3'5 
D 480 107 
2.6 1.02 2.0 
X 880 107 4.2 1.41 2.6 
u 
<U 
^3 
o 
>. 
'O 
fl 
o 
p. 
J 
3 
tn" 
.2 
d 
tn 
tn 
<D 
fl 
2 
.y 
tn 
3 
a - 
<V 14 
fl Y 
tn oo 
fl ™ 
<u 
8 & 
a 
.2 
a '$ 
•8 
2 S 
a u 
S 
22 
d 
C P. 
rt tn 
tn 3 
8 ° 
-p s 
bi S' 
'u <u 
5 
tn qj 
o <U 
<D Qj 
fl 
o 5 
- £ 
8 s 
fl <u 
rp <u 
u 
5 
,., <u 
0 5 
.2 O 
d tn 
« s 
•P <u 
a I 
»J 
ZJ o 
fl 
ce -3 
u 0 
O 'T 
to 'p 
. 8 
rH <D 
tn 
fl 
fl 475 
FERREE AND RAND: EFFICIENCY OF THE EYE 
A graphic representation of the results for the indirect sys- 
tem with socket extenders is given in Chart II. In this chart 
loss of efficiency is plotted against time of work in the manner 
described in the preceding section. For the sake of comparison 
results are shown also on this chart for the direct and semi- 
indirect systems. A graphic representation has further been 
made of the results for the indirect, system with and without 
socket extenders. This is shown in Chart III. 
CHART III. - Intensity Series. 
Showing the effect on loss of efficiency of changing the height of the light 
source in the reflector of the indirect lighting fixtures. The effect on surface 
brightness is primarily to change the area and surface brilliancy of the spot of 
light thrown on the ceiling. Chart A shows the results when height of 
source in the reflector is changed; Chart B, the results when the height is 
kept approximately constant. 
CHART A 
CHARTS 
EYE SHADE SERIES. 
This series of experiments has been conducted for the follow- 
ing reasons, (i) In general two methods are used to protect 
the eye from the source of light, eye shades and lamp shades. 
It is desirable to know whether the eye is protected equally well 
by both; and if the eye shade can be substituted for the lamp 
shade, what type of shade would best serve the purpose. (2) 
And the statement has been made to us many times that with an 
eye shade the three systems of artificial lighting we have used 
should give equally good results; and results, moreover, as good TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
476 
as those given by the indirect system without an eye shade. 
There are in general two classes of eye shades, the translucent 
and opaque. Up to this time we have confined our work to the 
opaque shade. So far as we know it is customary to make the 
opaque shade with a dark lining. This kind of lining is em- 
ployed probably because of some notion that it is restful to the 
eye to darken as much of the field of vision as is possible.18 
The tests were begun with the opaque shade with the dark 
lining. What we found as the result of these tests was somewhat 
in contradiction to the predictions that had been made. The 
shade did give pretty nearly the same results for the three sys- 
tems; but it did this, contrary to prediction, by improving the 
direct and semi-indirect systems and making worse by almost 
an equal amount the indirect system. That is, protected by the 
opaque shade, the eye lost in efficiency for the three systems by 
an amount somewhere near the mean of the losses experienced 
by it for the three systems without a shade. Nor is this result 
surprising when one reflects upon the conditions imposed upon 
the eye by an opaque shade with a dark lining. While it pro- 
tects the eye from the sources of light, such a shade does not by 
any means eliminate harmful brightness differences in the field 
of vision. It in fact creates for the eye a very unnatural bright- 
ness relation, i. e., it renders the whole upper half of the field of 
vision dark in sharp contrast with the brightly lighted lower 
half. The direct effect of this is a strong brightness induction 
(physiological) over the lower half of the field of vision which 
manifests itself to the observer by causing glare in surfaces that 
have no glare, and by increasing the glare in surfaces in which 
glare is already present. This, it is scarcely necessary to point 
out, operates against the discrimination of detail and puts the 
eye under strain to see its objects clearly. Moreover, the unusual 
and strongly irregular character of the image formed on the ret- 
ina probably also sets up a warfare in the incentives given to 
the muscles which adjust the eye. That is, the upper half of 
the field of vision is dark and presents no detail. The effect of 
this is probably to exert a tendency to cause the muscular relax- 
18 Another popular view might be, so far as protection to the eye is concerned, to re- 
gard the opaque eye shade as the analogue of the opaque or perhaps the indirect lamp 
reflector and the translucent shade as the analogue of the semi-indirect reflector. 477 
FERRER AND RAND: FFFlClFNCV OF THp FY£ 
ation characteristic of the darkened field of vision. The lower 
half of the field is light and filled with detail. The incentive 
here is towards the- best possible adjustment of the eye for the 
discrimination of detail in the objects viewed, while the rim 
of the shade, the sharply marked boundary between the dark and 
light halves of the field of vision and much nearer to the eye 
than the objects viewed,19 serves as a constant and consciously 
annoying distraction to fixation and accommodation. These 
complex and somewhat contradictory impulses given to the mus- 
cles of the eye might very well, and doubtless do cause an exces- 
sive and unnatural loss of energy and efficiency in case of the 
prolonged adjustment of the eye needed for a period of work. 
Early in the course of the tests it occurred to us that we might 
render the brightness distribution in the field of view presented 
to the eye wearing a shade more natural, and thereby improve 
the effect of the shade on the eye, by employing a white instead 
of a dark lining. By using a matt white paper20 with a reflection 
coefficient of about 75 per cent, for this lining, the following 
effects were produced. The two halves of the field of vision 
were rendered much more nearly of equal brightness; the glare 
in the lower half of the field of vision was very noticeably 
lessened and the discrimination of detail was correspondingly 
improved; the upper half of the field of view no longer tended 
to give to the eye the reflexes of the darkened field of vision; 
and the rim of the shade did not stand out nearly so distinctly 
in the field of view to distract accommodation and fixation. The 
results of the test for loss of efficiency show, moreover, that our 
surmise with regard to the effect of this change on the eye was 
correct. The action of the white lining was greatly to improve 
the ability of the eye to maintain its efficiency for a period of 
work. As good results were not gotten, however, with the shade 
for any of the systems as were given by the indirect system 
without the shade. Since there was a still greater evenness of 
surface brightness in the field of view in case of the indirect 
system with the eye shade than without, the question arises why 
19 This rim is about three inches in front of the observer's eye when the shade is in 
position. 
20 Hering standard white paper was used for this lining. The reflection coefficient 
of the dark lining was about 6-8 per cent. transactions of irruminating FNGINFFRING society 
478 
at least as good results were not obtained with the shade as 
without. The answer, we believe, is to be found in terms of 
the distraction to fixation and accommodation caused by the eye 
shade even when a light lining was used. For the effect of a 
shade on the eye even when the most favorable lining is em- 
ployed is that of a constantly present distracting object with 
its lower margin not far removed from the center of the field of 
vision, and much nearer to the eye than are the objects which 
the observer is called upon to discriminate. It will be noticed 
also in Table XVII that the results were never so good for either 
kind of shade for the direct and semi-indirect systems as for the 
indirect. Since the evenness of surface brightness in the field 
of view was not very different for the three systems in both 
cases, this again probably indicates that the evenness of surface 
brightness is not the only one of the distribution factors that has 
to be taken into account in studying the effect of different con- 
ditions of lighting on the eye. 
These tests were made for the same installations that were used 
in the distribution series. Since the use of the eye shade did not 
affect the illumination of the room the reader is referred for the 
illumination measurements to the tables of the distribution series. 
The distribution of surface brightness in the field of vision, how- 
ever, was strongly affected. New measurements were made, 
therefore, of the brightness of the prominent surfaces in the field 
of vision. The tests were taken at Position I, see Fig. i, p. 452a 
The prominent surfaces in the observer's field of vision working 
in this position were J, K, and L (see Fig. 4, p. 452b) ; the top 
of the table carrying test and recording apparatus, immediately 
in front of the observer and below the level of his eyes; the test 
card; the reading page in the 45 deg. position; and the white and 
dark lining of the eye shade as seen by the observer when the 
shade was in position over his eyes. The measurements of the 
brightness of the lining of the eye shades as seen by the observer 
when the shades were in position were made as follows. A sur- 
face in front of the observer was made to match in brightness the 
lining of the shade as it was seen by him. The brightness of this 
surface was then measured by the method described on page 
452. In procuring the match between the comparison surface and 479 
FERREE AND rand: EFFICIENCY OF THE EYE 
the lining of the shade the series of Hering matt gray papers was 
employed. This series consists of 50 shades ranging from a white 
with a reflection coefficient of 75 per cent, to black. Sheets of 
these differing in brightness were placed in a vertical position at 
a given distance in front of the observer until an approximate 
match was made with the lining of the shade. The gradations 
needed to get the final match were secured by moving the sur- 
face to and from the observer and by tilting it at different angles 
with the line of sight. The former adjustment carried it into 
parts of the room having different intensities of illumination and 
the latter turned it so as to receive a greater or less amount of 
light. In making the brightness measurements, care was taken 
to have the receiving surface of the photometer arm normal at 
its central point to the line of sight taken by the observer when 
the match was made. The results of these measurements are 
shown in Table XIV. In Table XV are given some of the prom- 
inent ratios of surface brightness in the field of vision for the 
shade with the dark lining; and in Table XVI, some of the prom- 
inent ratios for the shade with the white lining. In Table XVII 
are shown the results for the test for loss of efficiency for the 
shade with the dark lining; and in Table XVIII for the shade 
with the white lining. For purposes of comparison the results 
of the three systems without a shade are repeated. These are 
given in Table XIX. A graphic representation results of all 
three tables is given in Chart IV. 
TABLE XIV.-Eye Shade Series. 
Showing the brightness measurements in candlepower per square inch 
for the various surfaces in the field of vision for the direct, semi-indirect and 
indirect systems used when the eyes were shielded in turn by an opaque 
eye shade with a dark lining, and an opaque eye shade with a white lining. 
Surface 
measured 
Direct 
system 
Semi-indi- 
rect system 
Indirect 
system 
J 
0.0014 
O.OOI 
O.OOO95 
K 
0.0063 
0.0046 
O.OO255 
L 
0.0042 
O.OO27 
O.OO246 
Table 
0.0029 
O 00255 
O.OO233 
Test card 
0.00308 
0.003 
O.OO299 
Reading page 45 ° position 
O 004 
0.0039 
O.OO431 
White lining of eye shade 
O.OOI97 
0.00204 
0.00207 
Dark lining of eye shade 
O.OOOO91 
0.000 11 
0.000126 TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
480 
TABLE XV.-Eye Shade Series. 
Showing some prominent ratios of surface brightness for the direct, semi-indirect, and indirect systems used when the eyes 
were shielded by an opaque eye shade with a dark lining. 
Ratio 
Direct system 
Semi-indirect system 
Indirect system 
Lightest to darkest 
- - - O.OO63/O.OOOO91 = 69.2 
O.OO46/O.OOOII = 42.O 
O.OO431/O.OOOI26 = 34.2 
Lightest to test card • 
... 0.0063/0.00308 = 2.04 
O.OO46/O.OO3 = I.53 
0.00431/0.00299 = 1.5 
Lightest to reading page 
... 0.0063/0.004 = I.56 
O.OO46/O.OO39 - 1.2 
0.00431/0.00431 = 1.0 
Lightest to lining of eye shade - - - - 
. • • 0.0063/0.000091 = 69.2 
O.OO46/O.OOOII = 42.O 
0.00431/0.000126 = 34.2 
TABLE XVI.-Eye Shade Series. 
Showing some prominent ratios of surface brightness for the direct, semi-indirect, and indirect systems used when the eyes 
were shielded by an opaque eye shade with a white lining. 
Ratio 
Direct system 
Semi-indirect system 
Indirect system 
Lightest to darkest 
0.0063/0.0014 = 3.9 
O.OO46/O.OOI =4.6 
0.00431/0.00095 = 45 
Lightest to test card 
0.0063/0.00308 = 2.04 
0.0046/0.003 = I.53 
O.OO431/O.OO299 = 1.0 
Lightest to reading page 
0.0063/0.004 = I.56 
0.0046/0.0039 =1.2 
0.00431/0.00431 = 1-0 
Lightest to lining of eye shade 
0.0063/0.00197 =3.2 
0.0046/0.00204 = 2.25 
0.00431/0.00207 = 2.5 481 
FERREE AND RAND: EFFICIENCY OF THE EYE 
» TABLE XVII.-Eye Shade Series. 
Showing the eye's loss in efficiency as the result of 3 hours of work under the direct, semi-indirect, and indirect systems 
of lighting employed. (With opaque eye shade with dark lining.) 
Total 
Hori- 
Foot-candles 
Maximal 
distance at 
which test 
object can 
be seen 
Working 
Total 
time 
Total 
time 
time 
clear 
total 
time 
Ratios 
reduced 
to common 
Verti- 
Lighting system 
Watts zontal 
cal 450 
Time clear 
distance 
clear 
blurred 
blurred 
standard 
Indirect 
... 800 5.2 
1-36 3-5 
9 A.M. 8o.O 
68 
143 
37 
3-86 
3-5 
12 M. 80.0 
68 
13° 
50 
2.6 
2-34 
Semi-indirect... 
• •• 760 5.8 
1-45 4-o 
9 A.M. 8o.O 
68 
126 
54 
2.33 
3-5 
12 M. 79.5 
68 
IIO 
70 
i-55 
2-33 
Direct 
.. . 880 4.2 
1.41 2.6 
9 A.M. 79.O 
67 
139 
4i 
3-39 
3-5 
12 M. 79.O 
67 
124 
56 
2.21 
2.27 
TABLE XVIII.-Eye Shade Series. 
Showing the eye's loss in efficiency as the result of 3 hours of work under the direct, semi-indirect, and indirect systems 
of lighting employed. (With opaque eye shade with white lining.) 
Foot-candles 
Maximal 
distance at 
which test 
object can 
be seen 
clear 
Total 
time 
blurred 
Total 
time 
clear 
Ratios 
reduced 
to common 
standard 
Lighting system 
Watts 
Hori- 
zontal 
Verti- 
cal 
45° 
Time 
Working 
distance 
time 
clear 
time 
blurred 
Indirect 
. •- 800 
5-2 
I-36 
3-5 
9 A.M. 
85-0 
68 
Il8 
62 
i-9 
3-5 
12 M. 
85.0 
68 
115 
65. 
i-77 
3-i8 
Semi-indirect ... 
... 760 
5-8 
i-45 
4.0 
9 A.M. 
85.0 
68 
128 
52 
2.46 
3-5 
12 M. 
85.0 
68 
124 
56 
2.21 
3-13 
Direct 
... 880 
4.2 
1.41 
2.6 
9 A.M. 
85-0 
70 
105 
75 
i-4 
3-5 
12 M. 
84-5 
70 
TOO 
80 
1.25 
3-07 transactions or illuminating engineering society 
482 
TABLE XIX.-Eye-Shads Series. 
Showing the eye's loss in efficiency as the result of 3 hours of work 
under the direct, semi-indirect, and indirect systems of lighting employed 
(No eye shade.) 
Maximal 
distance at 
Foot-candles 
which test 
, 
A 
X 
object can 
Lighting 
Hori- 
Verti- 
be seen 
system 
Watts zontal 
cal 
45° 
Time 
clear 
Indirect 
. 800 5.2 
1.36 
3-5 
9 A.M. 
84-5 
12 M. 
84-5 
Semi-indirect 
• 760 5.8 
1-45 
4.0 
9 A.M. 
80.5 
12 M. 
79-5 
Direct 
. 880 4.2 
I.41 
2.6 
9 A.M. 
8l.O 
12 M. 
78.0 
Total 
time 
clear 
Work- 
-A- 
Ratios 
ing 
Total 
Total 
total 
reduced 
dis- 
time 
time 
time 
to common 
tance 
clear 
blurred 
blurred 
standard 
Indirect 
67.5 
135 
45 
3-00 
3-5 
67.5 
I32 
48 
2-75 
3-2 
Semi-indirect • 
68.5 
142 
38 
3-73 
3-5 
68.5 
92 
88 
I.64 
o-97 
Direct 
68.0 
139 
4i 
3-39 
3-5 
68.0 • 
771 
109 
0.69 
0.671 
As yet we have not determined the effect of translucent shades 
on the eye. In attempting to deal in a general way with this 
class of shades we have the same type of difficulty to face that we 
have in case of the semi-indirect reflector. That is, we may have 
shades varying from transparent to opaque, and sharing in the 
merits and demerits of each extreme. Our judgment would be, 
however, that it would be very difficult to get a translucent shade 
that would give as good results as an opaque shade with a light 
lining; for the translucent shade when made sufficiently opaque 
to give the needed reduction to the image of the source will 
darken too much the upper half of the field of vision and thereby 
simulate too much the condition given by the opaque shade with 
the dark lining to give the best results for comfortable and effi- 
cient seeing. Moreover, from the results that have already been 
obtained with the opaque shade and from the principles it seems 
fair to infer from these results, it seems very probable to us that 
as good effects for seeing should not be expected from the use of 483 
FERREE AND RANDI EFFICIENCY OF THE EYE 
any kind of eye shade as may be gotten from lamp-shades. That 
is, if we are to secure the best results for seeing, the shade should 
be put on the lamp, not on the eye. 
THE ANGLE AT WHICH THE LIGHT FALLS ON 
THE WORK. 
The object of these experiments was to find out whether the 
CHART IV.-Eye Shade Series. 
Showing the effect on loss of efficiency of opaque eye shades with dark and 
with white lining for the installations direct, semi-indirect, and indirect 
with the same intensity of light at the point of work. Chart A shows 
results without shade ; Chart B, with shade having dark lining; Chart C' 
with shade having white lining. 
CHART A 
CHART B 
CHART C 
difference in the angle at which the light falls on the work pro- 
duces an effect on the eye that can be detected by the test we 
have used for loss of efficiency. For the purpose of this pre- 
liminary investigation it was decided to make the general illumi- 
nation of the room such as to cause the eye little loss of effi- TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
484 
ciency as the result of the period of work; and to add to that 
at the point of work a component of light which was less diffuse 
in order that the amount of light entering the eye would be more 
dependent upon the angle at which the reading page was held. 
The general illumination was obtained from the indirect system 
used in the work of the preceding sections with lamps totalling 
800 watts. The less diffuse component at the point of work 
was obtained from a 60-watt lamp with a porcelain reflector of 
the desk lamp type. This lamp was turned into the horizontal 
position and was placed behind the observer and to the left so 
that the light came over the left shoulder. When in the position 
for which the test was taken the tip of the lamp was slightly 
above the level of the observer's eye, at a distance of 1 meter 
from the left eye. 
The illumination and brightness measurements for the test 
room illuminated by the indirect system, 800 watts, are given on 
pp. 469 and 471. These measurements were not greatly changed 
by the addition of the 60-watt lamp behind the observer. Because 
of the presence of this lamp, however, the following measurements 
were added to those given on pp. 469 and 471 : the horizontal, 
vertical, and 45 deg. components of light at the point of work; the 
brightness of the test card in place for the test; and the brightness 
of the reading page when held respectively in the positions which 
gave the least and the greatest amounts of specular reflection. 
The illumination measurements at the point of work are given in 
Table XX. The brightness of the test card was 0.00365 cp. per 
sq. in.; of the reading page in the position that gave the least 
amount of specular reflection, 0.0059 cp. per sq. in.; and in the 
position that gave the greatest amount of specular reflection, 
0.0077 cp. per sq. in. A mirror surface was used as an aid in 
locating the position of least and greatest specular reflection. 
The results of the test for three hours of work done with the 
reading page in these two positions are also given in Table XX. 
A graphic representation of the results of this table is shown in 
Chart V. 
THE EFFECT OF DIFFERENT CONDITIONS OF LIGHTING 
ON THE FIXATION MUSCLES OF THE EYE. 
The test we have employed thus far in the conduct of our 485 
FERREE AND RANDI EFFICIENCY of THE Eye 
work is one designed to show the effect of different conditions of 
lighting on the ability of the eye to hold its efficiency for clear 
seeing for a period of three minutes. In itself this test is not 
TABLE XX.-The Angle at Which the Light Falls on the Work. 
Showing the effect on loss of efficiency of the angle at which the light 
falls on the work. 
Kind of 
reflection 
from read- 
ing page 
during 
work period 
Foot-candles at 
testcard 
Time 
Maximal 
distance 
at which 
test ob- 
ject can 
be seen 
clear 
Hori- 
zon- 
tal 
Verti- 
cal 
45° 
Diffuse .... 
•••• 5-3 
I.84 
3-9 
9 A.M. 
89 
12 M. 
89 
Specular ... 
•••• 5-3 
1.84 
3-9 
9 AM. 
89 
12 M. 
89 
Work- 
ing 
dis- 
tance 
Total 
time 
clear 
Total 
time 
blurred 
Total 
time 
clear 
total 
time 
blurred 
Ratios 
reduced 
to com- 
mon 
standard 
Diffuse 
•••73 
139 
41 
3-39 
3-5 
73 
137 
43 
3.18 
3-27 
Specular .... 
•••73 
137 
43 
3.18 
3-5 
73 
132 
48 
2-73 
3-o 
analytical in principle. The results, as is stated above, are ex- 
pressed in terms of an aggregate loss of function. The con- 
tributive factors may be inferred from the nature of the test, but 
CHART V.-The Angle at Which the Light 
Falls on the Work. 
Showing the effect on loss of efficiency of the angle 
at which the light falls on the work. 
the test is not in itself designed to separate them out. And 
indeed it is a question whether any practical good can accrue to 
the practise of lighting from a knowledge of just what part of TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
486 
the visual apparatus it is that falls off in function as the result 
of an unfavorable condition of lighting. Obviously the chief 
need is to find out what are the conditions that cause the eye to 
lose its ability to see clearly and to avoid these conditions in 
planning and installing a lighting system. From the beginning 
we have had in mind, however, an analysis of effect. Our tests 
for the sensitivity of the retina showed, for example, that very 
little, if any, of the difference in results we have gotten for the 
four types of lighting we have employed can be ascribed to a 
loss in the efficiency of the retina, or the light sensitive part of 
the visual apparatus. Three sets of factors are involved in clear 
seeing: (i) the sensitivity of the eye to colored and white light; 
(2) the ability to make fine space discriminations which is in 
part dependent upon our third factor; and (3) accurate fixation 
and accommodation. Both fixation and accommodation are the 
result of muscular action. When the muscles lose in tone because 
of excessive use or by sharing in a general condition or state of 
the body, the eye loses correspondingly in its power to sustain 
clear seeing. If, for example, the muscles of accommodation 
have fallen off in efficiency the lens is no longer held in the 
adjustment needed to bring the light to a sharp focus on the 
retina and loss of detail and blurring result; or, if it be the fixa- 
tion muscles that have suffered the loss, the eyes cannot be con- 
tinuously held in such a position that the images of the object 
viewed fall symmetrically on the fovea of each. When this latter 
condition is present loss of detail results from two causes. (1) 
The fovea and region immediately surrounding it are the most 
highly developed parts of the retina and the best fitted for the light 
and space discriminations needed for clear seeing. Moreover, the 
refracting media of the eye give the clearest images when the 
axis of the cone of rays from the object viewed deviates as little 
as possible, consistent with the mechanism of the eye, from the 
optic axis. And (2) if the images in the two eyes do not fall 
more or less symmetrically upon the fovea of each they are not 
accurately combined into one, and blurring and loss of detail 
results from the doubling of the objects seen. It is our purpose 
as fast as possible to isolate the effect of the three systems of 
lighting we have used on each of the above named factors. In 487 
FERREE AND RAND: EFFICIENCY OF THE EYE 
the work of the present section the effect of these systems on 
the fixation muscles has been studied. 
The doubling of the image seen when the fixation muscles lose 
their power of co-ordinated action furnishes us with our clue 
for a test for the loss of efficiency of these muscles. That is, 
just as blurring and the loss of ability to discriminate detail is 
taken as the criterion of the loss of acuity of vision, so will the 
doubling of the image seen be taken as our index of the loss of 
the co-ordinated action of the fixation muscles. If one were to 
stare continuously for an interval of time with natural vision at 
a simple test object, as, for example, a vertical line, doubling 
might be detected especially if there had been protracted strain 
or considerable loss of power to co-ordinate. For the purpose of 
our work, however, greater sensitivity than this would be 
needed. Obviously sensitivity can be added by putting the eyes 
under strain to combine their images. When this is done, even 
when the muscles are fresh, if the object is looked at or fixated 
for an interval of time it will be seen alternately a,s one and as 
two. The proportion or ratio of the time seen as one to the 
time seen as two can be regulated by the amount of initial strain 
under which the eyes are put to combine their images. The regu- 
lation of this ratio is empirical and of importance; for as is the 
case with the test for loss of efficiency for clear seeing, the sen- 
sitivity of the test depends to a considerable extent upon the 
initial value that is given to this ratio. The eyes may be put 
under strain to combine their images by interposing between 
them and the object viewed weak prisms and so adjusting them 
and regulating the distance of the object from the eye that with the 
maximum of effort to see it as one it is seen alternately as one and 
as two in the proportion desired.21 This result can be accom- 
21 It would seem that the above principle might be utilized to advantage by the 
opthalmologist in testing the extrinsic muscles of the eye. The abduction and adduc- 
tion tests, for example, determine only what the muscles are able to do by momen- 
tary effort. Obviously, however, it is not what the muscles are able to do by a 
momentary effort or jerk that measures their ability to hold the eyes continuously 
adjusted for work. It is rather their endurance or what they are able to accomplish 
in an interval of time. An expression may be had for this either for the eyes con- 
jointly or separately by the method described above. That is, the prisms may be put 
in front of either one or both eyes and the ratio be determined of the time the 
object is seen as one or as two for whatever interval of time the operator may select. 
Similarly, it seems to the writers that the time element might be introduced to ad- 
(Continued on next page.) TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
488 
plished still more conveniently, however, by using an adaptation 
of the Brewster stereoscope. In this case a stereograph consisting 
of two vertical lines exactly alike may be used as the test object. 
In the stereograph employed in our test the vertical lines were 
2.5 cm. long and were printed on the card 4.5 cm. apart or at 
2.25 cm. from the center of the card. When this was put in a 
sliding carrier and was made to approach the eyes, a position 
was reached at which with the maximum of effort the observer 
was no longer able to see the two vertical lines as one. They 
were seen alternately as one and as two. In making the test the 
hood was removed from the stereoscope so that the eyes were 
fully exposed to the conditions of illumination that were being 
tested. . The stereoscope was mounted in front of the eyes of 
the observer in position at the point of work. The distance of 
the carrier containing the test object from the observer's eyes 
was adjusted until the proper ratio of time seen as one and time 
seen as two was obtained. Having determined this position a 
record was made of the time seen as one and the time seen as 
two for three minutes at the beginning and the close of work. The 
ratio of the sum of these intervals may in either case be taken 
as a measure at that time of the power of the fixation muscles 
to act in co-ordination for three minutes of continuous effort; 
and the decrease in this ratio from the beginning to the close of 
work may be taken as a measure of the loss in that power, sus- 
tained as the result of work. In making this test the same re- 
cording apparatus was used as was employed in the test for loss 
of efficiency for clear seeing. That is, the record was traced 
on a kymograph by means of an electro-magnetic marker and a 
vantage into the visual acuity test used by the ophthalmologist when the cycloplegic 
is not employed or in cases of post-cycloplegic refraction. Is it, for example, enough 
to know that the eye has 20/20 acuity or can discriminate a certain standard visual 
angle by momentary effort? Would it not give a more complete repreentation of 
the functional condition of the eye to know what it can discriminate clearly through 
an interval of time; or better still perhaps, for what proportion of an interval of 
time it can discriminate a certain detail or standard visual angle clearly? For ex- 
ample, just as a fatigued eye may for the moment under the spur of the test over- 
come the functional results of fatigue, so might small errors of refraction be overcome 
for the moment by muscular effort, especially in the cases in which the muscles of 
the eye are unusually strong. But just as the fatigued muscle can not do this through 
an interval of time, so it would seem that a residual error of refraction might not 
be so easily masked through an interval of time by means of muscular effort. In 
short, this form of test is suggested as affording possibly a closer approximation to 
the conditions and demands imposed upon the eye during a period of work than is 
afforded by the acuity test based upon the momentary judgment. 489 
FERREE AND rand: EFFICIENCY OF THE EYE 
telegraph key, and a time line was run beneath the record by 
means of a Jacquet chronograph registering seconds. 
The test for the effect on the fixation muscles of a period of 
work was made under the same installations, conditions of 
work, and with the same observers that were, used in the dis- 
tribution series. The test, however, was made at only one of the 
positions used in that series, namely, the position at which the 
greatest loss of efficiency was obtained. (See Position I, Fig. I, 
p. 452a.) At this point, it will be remembered, six of the lighting 
TABLE XXI.-Fixation Muscles Series. 
Showing the loss of efficiency of the fixation muscles as the result of 3 
hours of work under the direct, semi-indirect, and indirect 
systems of lighting- employed. 
Foot-candles 
Distance 
at which 
test ob- 
ject is 
normally 
Hori- 
zon- 
Verti- 
Lighting system 
Watts tai 
cal 
45° 
Time 
seen single 
Indirect .... 
800 
4.2 
0-99 
2-5 
9 A.M. 
12 M. 
18 
l8 
Semi-indirect .... 
760 
4.8 
0.98 
2.6 
9 A.M. 
12 M. 
18 
18 
Direct 
880 
3-9 
Work- 
ing 
dis- 
tance 
1.0 
Total 
time 
single 
1.99 
Total 
time 
double 
9 A.M. 
12 M. 
Total 
time 
single 
total 
time 
double 
18 
18 
Ratios 
reduced to 
common 
standard 
Indirect 
22 
22 
142 
140 
38 
40 
3-7 
3-5 
3-5 
3-31 
Semi-indirect 
22 
141 
39 
3-6 
3-5 
22 
138 
42 
3-28 
3-24 
Direct 
20 
20 
153 
151 
27 
29 
5-66 
5-21 
3-5 
3-21 
units were in the field of view. The specification of the lighting 
effects produced by these installations are given on pp. 452a-459 
Nothing need be added at this point to these specifications but 
the brightness of the stereograph or the test object in position 
for the three systems of lighting, and the illumination meas- 
urements at the test card. The brightness measurements are as 
follows. The brightness of the card, corrected for the ab- 
sorption of the prisms of the stereoscope was for the direct 
system 0.00172 cp. per sq. in.; for the semi-indirect system, TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
490 
0.00163 cp. per sq. in.; and for the indirect system, 0.00167 cp. 
per sq. in. New illumination measurements were needed at the 
test card because the card had to be moved closer to the eyes 
than was the case in the tests for loss of efficiency for clear 
seeing, which brought it into a region of different illumination. 
These measurements are given in Table XXI. The results of 
our tests for loss of efficiency of the fixation muscles for the 
three systems of lighting are given also in this table. These re- 
sults show (a) that very little loss of co-ordination is suffered by 
the fixation muscles as the result of three hours of work under the 
systems selected; and (b) that there is very little difference in 
CHART VI.--Fixation Muscle Series. 
Showing the loss of efficiency of the fixation muscles 
as the result of 3 hours of work under the direct, 
semi-indirect, and indirect systems of lighting em- 
ployed. 
Foot-candles 
Lighting system 
Watts 
Horizontal 
Vertical 
45° 
Indirect 
. 800 
4.2 
0-99 
2.5 
Semi-indirect ■ 
• 760 
4.8 
0,98 
2.6 
Direct 
S80 
3-9 
1.0 
1.99 
the effect for the three systems. Since there is no reason for think- 
ing that the test has not as great sensitivity as the test for loss of 
efficiency for clear seeing, and since the same observers, condi- 
tions of lighting and working were used as in the former tests, it 
does not seem to us at this time that the loss of efficiency for clear 
seeing that is sustained under these conditions, shown by the 
former tests, can be ascribed to any great extent to an effect on 
the muscles of fixation. In a later report experiments will be 
described in which the effect on the muscles of accommodation 
has been studied. 
A graphic representation of the results of Table XXI is shown 
in Chart VI. 491 
FWEE AND rand: EFFICIENCY OF THE eye 
THE EFFECT OF MOTION PICTURES ON THE EFFICIENCY 
OF THE EYE. 
The belief that motion pictures subject the eyes to undue 
strain is too prevalent to need more than mention in passing. 
All are familiar with the conditions,-the initially dark-adapted 
and highly sensitized eye, the comparatively brilliant screen with 
its dark surrounding field, the flickering light, and the shifting 
and very often unsteady pictures. We have already seen that 
differences in surface brightness of considerable magnitude in 
the field of vision cause loss of efficiency and produce discom- 
fort, and we have discussed the causes for these effects. We have 
nothing further to add to that discussion here. We are, how- 
ever, facing for the first time in our work the question of the 
effect upon the eye of a flickering light and lack of steadiness in 
the object viewed. The following reason is suggested why a 
flickering or unsteady picture may cause loss of efficiency. The 
eye is so constituted that when its images lose in clearness or 
distinctness it is incited to a muscular readjustment to bring 
about the clearness needed. Ordinarily in seeing, the conditions 
for loss in clearness come about primarily through the difference 
in the distance or direction from the eye of the objects which 
are successively viewed. In motion pictures, however, the chang- 
ing clearness of the objects viewed is not due to any change in 
their distance or direction from the eye; nor to anything in fact 
which the readjustment of the eye can remedy to any consider- 
able degree. The effort expended, therefore, is of little avail for 
seeing, if, indeed, the new setting of the parts is not a detriment 
to clear seeing and a condition which in turn must be corrected. 
This should, and doubtless does, lead to muscular strain and 
loss of efficiency. It was decided, therefore, to make an explora- 
tive investigation to determine whether there is an effect of motion 
pictures on the eye which can be detected by our test for loss of 
efficiency. The tests were conducted in a local theater, selected 
primarily because of the favorable conditions that prevailed. The 
definition at the screen was good and the pictures were unusually 
steady and free from flicker. The conditions were, we think, fairly 
representative of what is found in the better class of motion 
picture houses. TRANSACTIONS OF IEEUMINATING FNGINFFRING SOCIETY 
492 
The tests were taken immediately before and after two hours 
of observation of the pictures. During the exhibition the ob- 
server sat directly in front of the center of the screen. The 
observation was made at successive times at three distances from 
the screen,-in the front, middle, and the back of the house. 
These positions were respectively 25, 48, and 71 ft. (7.62, 14.6, 
and 21.6 m.) from the screen. The room in which the pictures 
were shown was 78 ft. (23.7 m.) long and 48 ft. (14.6 m.) wide. 
The tests were taken in a room 14 ft. (4.2 m.) long, 9 ft. (2.74 
m.) wide, 11 ft. (3.35 m.) high, adjoining the stage. The walls 
and ceiling of this room were of rough plaster, painted a flat white. 
When taking the test the observer sat facing one of the side walls 
of the room, 1.5 m. distant. The room was lighted for the pur- 
pose of the test by one 100-watt and one 60-watt clear tungsten 
lamp suspended behind and slightly to the right of the observer 
when in position for the test, at about 2 ft. (0.6 m.) above the 
level of his eyes. The source of light was thus entirely out of 
the field of view and the light fell evenly and without shadow on 
the test card and the wall in front of the observer. At the point 
of the test card, the illumination measured with the receiving test 
plate of the photometer in the horizontal plane was 1.3 foot- 
candles; in the vertical plane, 1.9 foot-candles; and in the 45 deg. 
plane, 2.3 foot-candles. The surface brightness of the test card 
was 0.003256 cp. per sq. in., and that of the wall directly behind 
the card was 0.002288 cp. per sq. in. The distribution of surface 
brightness on the wall which the observer faced was very even. 
At the point of maximum brightness to the right of the observer, 
as nearly as that point could be located, the brilliancy was 0.00308 
cp. per sq. in.; and to the left of the observer, 0.002024 cp. per 
sq. in. 
In order that there might be no intermission between the pic- 
tures for changing the films, two projection machines were used. 
The following is the specification of the apparatus employed as 
given by the operator. 
Type of machine, Powers 6-A Projector. 
Lens equipment, 1 pair pearl white condensers, 6% in. F. L. 
1 Bausch and Lomb objective combination, 
424 in. E. F. FgRREF AND rand: EFFICIENCY OF THE EYE 
493 
Lamp, i io,ooo-cp. adjustable arc. 
Carbons, 24 in. cored bio's. 
Current, 22 volt a. c. through Halberg transformer. 
Line current, 28-30 amperes. 
Arc voltage, 45-50 volts. 
Length of throw or distance from objective to screen, 72 ft. 
(21.9 m.) 
Screen, sheet muslin sized and coated with flat white alabastine. 
Speed of film through machine, 66 ft. 8 in. (20.3 m.) per min. 
Number of pictures per 1 ft. (0.3 m.) of film, 16. 
Size of picture on film, 24 in. (1.9 cm.) high by 15/16 in. (2.38 
cm.) wide. 
Size of picture on screen, 11 ft. (3.35 m.) high by 14 ft. (4.26 
m.) wide. 
Approximate brightness of screen with film removed from pro- 
jector, 3.47 cp. per sq. in. 
Exceptional steadiness, it may be said, is given to the move- 
ment of the film and, therefore, to the picture in this type of pro- 
jector by the special type of intermittent movement that is em- 
ployed. Details of this movement need not be given here. As 
has already been stated, our reason for making the test in this 
particular theater was the comparative steadiness of the pictures 
and the comparative freedom from flicker, that was obtained. 
The results of the tests are shown in Table XXII. Quite a 
great deal of loss of efficiency is shown as the result of two 
hours of observation. The nearer the observer was to the screen, 
the greater was this loss found to be. The loss, however, so far 
as we can tell, is no greater than is caused by steady work under 
the direct and semi-indirect installations of lighting used in our 
distribution series. Unfortunately, we have not for the pur- 
poses of comparison, results for the same observer for the same 
length of time of exposure for the two sets of condition. The 
loss for observer R for two hours observation of the motion 
pictures was not nearly so great as for three hours of reading 
from good print and paper, under the direct and semi-indirect 
systems of lighting. But comparing the results for observer G for 
two hours of reading from the same type and paper with those 
for observer R for two hours observation of the pictures the transactions of illuminating FNGINFFRING society 
494 
loss seems to be about the same. That is, our results indicate 
that while the eyes are strained a great deal by the observation 
TABLE XXII.-Motion Picture Series. 
Showing the loss of efficiency of the eye caused by two hours' observation 
of motion pictures. 
Position 
Time 
Maximal 
distance 
at which 
test ob- 
ject can 
be seen 
clear 
Work- 
ing 
dis- 
tance 
Total 
time 
clear 
Total 
time 
blurred 
Total 
time 
clear 
total 
time 
blurred 
Ratios 
reduced 
to com- 
mon 
standard 
25 ft. (7.62 tn.) 
from projec- 
8 P.M. 
86.2 
tion screen.. 
70.5 
123 
57 
2.14 
3-5 
IO P.M. 
86.1 
70.5 
95 
§5 
1.12 
i-79 
48 ft. (14.63 tn.) 
from projec- 
tion screen.. 
8 P.M. 
85.8 
71.0 
128 
52 
2.46 
3-5 
71 ft. (21.64m.) 
IO P.M. 
85.6 
71.0 
108 
72 
i-5 
2.13 
from projec- 
tion screen.. 
8 P.M. 
86.0 
69.0 
137 
43 
3-19 
3-5 
IO P.M. 
86.0 
69.0 
124 
56 
2.2 
2.42 
CHART VII. Motion Picture Series. 
Showing the loss of efficiency of the eye caused by two 
hours observation of motion pictures. 
Position A 25 ft. from projection screen 
Position B 48 ft. from projection screen 
Position C 71 ft. from projection screen 
of moving pictures, even in the better moving picture houses, they 
are damaged little more by that in all probability than they are by 495 
FERREE AND RAND: EFFICIENCY OF THE EYE 
CHART VIII. 
Distribution Series (Observer R) Motion Picture Series Distribution Series (Observer G) 
Showing the loss of efficiency of the eye as (Observer R) Showing the loss of efficiency of the 
the result of three hours of reading under Showing the loss of efficiency of the eye as the result of two hours reading 
the systems of direct, semi-indirect, and . , , , , , • under the systems of direct, semi-in- 
indirect lighting used, and daylight. ere caused b>' two hours direct, and indirect lighting used, 
UgbUn, ^_Foo^„al„ of motion pictures. and daylight. 
system Watts Horizontal Vertical 
45° 
4-2 
3-5 
Position A, 25 ft. from projection screen 
Position B, 48 ft. from projection screen 
Position C, 71 ft. from projection screen 
A 
B 
Lighting 
system Watts 
Daylight.... - 
Indirect .... 800 
Volts 
107 
Foot-candles 
Daylight -• 
Indirect 800 
5-5 
5-2 
1.32 
1-36 
Hor. 
5-5 
5-2 
Ver. 
1.32 
I.36 
45° 
4-2 
3-5 
Semi-indirect . 760 
5-8 
i-45 
4-o 
C 
Semi-indir't 760 
107 
5-8 
1-45 
4.0 
Direct 880 
4.2 
1.41 
2.6 
D 
Direct 880 
107 
4.2 
I.41 
2.6 TRANSACTIONS OF ILLUMINATING FNGINKERING SOCIETY 
496 
reading steadily the same length of time under the greater part of 
the lighting that is now in actual use. 
A graphic representation of the results of Table XXII is given 
in Chart VII. For the sake of comparing the effect of motion 
pictures on the eyes with the effect of reading steadily under the 
direct, semi-indirect, and indirect systems of lighting we have 
employed, Chart VIII has been prepared. 
THE TENDENCY OF DIFFERENT LIGHTING CONDITIONS TO 
PRODUCE DISCOMFORT, AND A COMPARISON OF THE 
TENDENCY OF THESE CONDITIONS TO CAUSE 
LOSS OF EFFICIENCY AND TO PRODUCE 
DISCOMFORT. 
In the former papers we have held that the general level or 
scale of efficiency of the fresh eye, loss of efficiency as the result 
of work, and the tendency to produce discomfort are all separate 
aspects of the problem of lighting in its relation to the eye, and 
that our knowledge of each must be obtained by different methods 
of investigation. A correlation between these three moments is 
doubtless possible, but that correlation should be founded upon 
the results of careful investigation; it should not be assumed. 
It is our purpose in this section of the paper to show the relative 
tendency of the different conditions of lighting we have used to 
produce discomfort, and to make a rough comparison of each con- 
dition to cause loss of efficiency and to produce discomfort. Any 
comparative study of the conditions producing discomfort neces- 
sitates a means of estimating discomfort. It is obvious that the 
core of the experience of discomfort is either a sensation or a com- 
plex of sensations. As such it should have a limen or threshold 
just as other sensations have; and just as we are able in general to 
estimate sensitivity in terms of the threshold value so should we in 
this case be able to use the threshold value in estimating the eye's 
sensitivity or liability to discomfort under a given lighting con- 
dition. Threshold values are usually determined by finding how 
much energy or intensity of a given stimulus, applied for a short 
interval of time, is required to arouse a just noticeable sensation. 
This form of procedure, however, is not adapted to the needs of 
our problem. It is much better to reverse the process and find 
how long the eye has to be exposed to a stimulus of a given in- 
tensity to arouse just noticeable discomfort. Our threshold thus 497 
FFRRFX AND rand: EFFICIENCY OF THF FYE 
becomes a time threshold and is measured in units of time instead 
of units of intensity. In order to determine whether the judg- 
ment of the threshold of discomfort can be made with certainty 
and to perfect the method and to test in general its feasibility, an 
abstract investigation was undertaken first, running through an 
entire year, in which a better and more convenient control of 
conditions could be secured than is possible in the investigation 
of a concrete lighting situation. That is, we undertook to de- 
termine the comparative sensitivity of the eye to discomfort when 
a single source of light was exposed in different parts of the 
field of vision. In order to carry out this investigation a lamp 
house with a circular opening in one side 3 cm. in diameter was 
attached to the arm of a perimeter in such a way that the opening 
was always directed towards the observer's eye. In the lamp 
house could be placed a lamp of whatever candlepower was de- 
sired. The arm of the perimeter could be shifted to any meridian 
in which it was desired to work and the lamp house could be 
moved at will along this arm. It was thus possible to expose 
the light for any length of time in any part of the field of vision 
that was desired. Working in this way we have not only investi- 
gated the effect of many types of variation of the position of the 
light in the field of view, the effect of intensity of light, etc.; but 
we have studied and standardized the factors that influence the 
sensitivity and reproducibility of the judgment and have given 
our observers the training that was needed for the concrete in- 
vestigation. In making the concrete investigation we have used 
every variation of the conditions of lighting described in this 
and the preceding paper. That is, the tendency to produce dis- 
comfort, measured in terms of the value of the time threshold, 
has been determined for all the conditions of lighting we have 
used in the tests for loss of efficiency. Two cases may be made 
of the investigation,-a determination of the tendency to cause 
discomfort when the eye is at rest, and a determination of this 
tendency when the eye is at work. Both of these cases were 
included in our investigation. The following determinations were 
made, (a) The time threshold of discomfort was gotten when 
the observer was sitting with the accommodation muscles relaxed 
and with the fixation muscles as nearly relaxed as was practica- transactions of illuminating Hnginllring society 
498 
TABLE XXIII.-Distribution Series. 
Showing a comparison of the tendency of the direct, semi-indirect, and indirect installations of lighting used in the distri- 
bution series to cause loss of efficiency and to produce discomfort. The loss of efficiency is the result of three hours of work. 
The tendency to produce discomfort is estimated by the time required for just noticeable discomfort to be set up. 
Position 
of observer 
Foot-candles 
Per cent, 
loss of 
efficiency 
Time limen 
of discomfort 
in seconds 
(not reading) 
Time limen 
of discomfort 
in seconds 
(reading) 
Lighting system 
Watts 
Horizontal 
Vertical 
45° 
I. 
Indirect 
8co 
5-2 
I-36 
3-5 
8.6 
263 
IOO 
Semi-indirect 
760 
5-8 
1-45 
4.0 
72.0 
15 
8 
Direct 
880 
4-2 
1.41 
2.6 
81.0 
IO 
9 
II. 
Indirect 
800 
51 
1.98 
4-2 
6-3 
259 
103 
Semi-indirect 
760 
6.1 
2-5 
4.7 
37-o 
26 
14 
Direct 
880 
4-65 
2-75 
4-4 
58.3 
20 
13 
III. 
Indirect 
800 
3-9 
2.1 
4.0 
7-7 
255 
99 
Semi-indirect 
760 
5-o 
2.6 
5-4 
22.0 
120 
35 
Direct 
. ' 880 
4.0 
2-9 
4.6 
31.0 
55 
24 
IV. 
Indirect 
800 
2-9 
2.1 
3-6 
6.6 
265 
101 
Semi-indirect 
760 
3-4 
3-0 
4-4 
19.0 
240 
87 
Direct 
880 
3-o 
3-4 
4-5 
23-0 
235 
57 499 
FERREE and rand: ffficifncy of thf Fyf 
TABLE XXIV.-Intensity Series. 
Showing a comparison of the tendency of the direct, semi-indirect and indirect installations of lighting for the different in- 
tensities used in the intensity series to cause loss of efficiency and to produce discomfort. The loss of efficiency is the result 
of three hours of work. The tendency to produce discomfort is estimated by the time required for just noticeable discomfort 
to be set up. 
Time limen 
Time limen 
Foot-candles 
Per cent. 
of discomfort 
of discomfort 
Lighting system Watts Horizontal 
Vertical 
45° 
efficiency 
(not reading) 
(reading) 
Indirect 800 5.2 
1.36 
3-5 
8.6 
263.0 
TOO 
480 3.0 
0.765 
i-97 
8.0 
265.0 
103 
320 (with socket extenders) 1.7 
0-49 
1.08 
9-1 
256.0 
98 
200 (with socket extenders) 1.48 
0.407 
0-95 
5-7 
251.O 
104 
320 (without socket extenders) 1.33 
0-39 
0.87 
23-0 
50.0 
33 
200 (without socket extenders) 1.16 
o-37 
0.76 
40.0 
20.0 
14 
Semi-indirect 320 2.2 
0.58 
1.52 
11.4 
102.0 
35 
200 1.6 
0-45 
1.15 
40.9 
62.0 
16 
480 3.3 
0-94 
2.4 
50.0 
50.0 
15 
760 5.8 
i-45 
4.0 
72.0 
15-0 
8 
800 6.8 
1.82 
4-5 
78.0 
14.0 
3 
Direct (16 lamps). 240 1.23 
o.54 
0-935 
57-4 
23-5 
17 
365 i-6 
0.6 
i-33 
62.0 
14.0 
11 
400 1.86 
0.8 
1.46 
65.0 
12.0 
11 
880 4.2 
1.41 
2.6 
81.0 
10.0 
9 
Direct (8 lamps).. 200 1.16 
0-45 
0.85 
34-3 
56.0 
27 
120 0.64 
0.32 
0-49 
45-5 
52.0 
15 
320 1.97 
0.65 
i-39 
55-5 
23-0 
13 
480 2.6 
1.02 
2.00 
67.0 
20.0 
12 TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
500 
ble under the conditions. That is, the observer sat in the positions 
shown in Fig. I, p. 452a, and took an easy fixation of an area at the 
level of the eye on the opposite wall of the room. The fixation dis- 
tance, for example, for Position I, Fig. 1, p. 452a, was 22 ft. Since 
blinking was found to be one of the variable factors which influ- 
TABLE XXV.-Eye Shade Series. 
Showing a comparison of the tendency of the direct, semi-indirect, and 
indirect installations of lighting used in the distribution series to cause loss 
of efficiency and to produce discomfort when the eye was protected by an 
opaque eye shade with a dark lining and by an opaque eye shade with a 
white lining. The loss of efficiency is the result of three hours of work. 
The tendency to produce discomfort is estimated by the time required for 
just noticeable discomfort to be set up. 
Foot-candles 
Per- 
cent. 
Time 
limen 
of 
discom- 
fort 
in sec- 
onds 
Time 
limen 
of 
discom- 
fort 
in sec- 
Lining 
of eye Lighting- 
shade system Watts 
Hori- Verti- 
zontal cal 450 
loss 
of effi- 
ciency 
(not 
read- 
ing) 
onds 
(read- 
ing) 
White Indirect...... 800 
5-2 1.36 3-5 
9-i 
85 
50 
Semi-indirect . 760 
5.8 1.45 4.0 
10.6 
8l 
48 
Direct 880 
4.2 1 41 2.6 
12.0 
75 
45 
Dark Indirect 800 
5-2 1.36 3-5 
33-0 
23 
19 
Semi-indirect . 760 
5.8 1.45 4.0 
33-4 
19 
15 
Direct 880 
4.2 1.41 2.6 
35-o 
16 
13 
TABLE XXVI.-The Angee at which the Light Faees on the Work. 
Showing a comparison of the tendency to cause loss of efficiency and to 
produce discomfort of the angle at which the light falls on the work. The 
loss of efficiency is the result of three hours of work. The tendency to pro- 
duce discomfort is estimated by the time required for just noticeable dis- 
comfort to be set up. 
Kind of 
reflection from 
Foot-candles 
Per cent, loss 
Time 
limen of 
discomfort 
in seconds 
Hori- 
Verti- 
reading page 
zontal 
cal 
45° 
of efficiency 
(reading) 
Diffuse• • •■ 
• 5-3 
I.84 
3-9 
6.6 
95 
Specular... 
• 5-3 
I.84 
3-9 
14-3 
30 
ence the tendency to produce discomfort, the amount of blinking 
was made constant from test to test. This was accomplished by 
having the observer blink at equal intervals during the test, timing 
himself by means of the stroke of a metronome. The interval 
most natural and suitable for this purpose was determined for 501 
FERREE and rand : efficiency of the EYE 
each observer separately. In the results given in the follow- 
ing table a three-second interval was used. And (b) the time 
threshold of discomfort was determined when the observer was 
reading from print and paper similar to that used in the loss 
of efficiency tests. In these tests all the conditions were kept as 
nearly the same as they were in the work on loss of efficiency as 
was possible. The results of both of these sets of experiments 
on the tendency to produce discomfort are shown in Tables 
XXIII-XXVI. The tendency to produce discomfort should be 
estimated, roughly speaking, probably as inversely proportional 
to the time it was required for discomfort to be set up. The time 
required for discomfort to be set up is given in the tables. In 
order to make convenient a comparison of the tendency of the 
various conditions of lighting to cause loss of efficiency and to 
produce discomfort the percentage loss of efficiency caused by 
the given lighting conditions is given in a parallel column in each 
table. The percentage loss of efficiency was computed by divid- 
ing the loss in the ratio of time seen clear to time seen blurred 
sustained as a result of work by 3.5, the standard ratio to which 
all the ratios at the beginning of work were reduced. A rough 
correspondence of the tendency to produce discomfort and to 
cause loss of efficiency will be noted in every case. This cor- 
respondence by no means amounts to a t : 1 correlation, however. 
In Table XXIII is given the comparison of the tendency to cause 
loss of efficiency and to produce discomfort for the distribution 
series; in Table XXIV, for the intensity series; in Table XXV, 
for the eye shade series; and in Table XXVI, for the series 
showing the effect of the angle at which the light falls on the 
work. 
In conclusion we wish to state that in this work, and the work 
reported in the former papers, the purpose has been primarily 
to procure methods of working and to find out, as broadly as 
one may, the applicability of these methods to the problems sur- 
rounding the hygiene of the eye. While in many places attention 
has been called to results that seemed to have general significance, 
the intention has been, in general, to limit all comments and 
conclusions strictly to the conditions under which the work was 
done. Reprinted from The Journal of Philosophy, Psychology and Scientific Methods, 
Vol. XII., No. 24, Nov. 25,1915. 
A RESUME OF EXPERIMENTS ON THE PROBLEM OF 
LIGHTING IN ITS RELATION TO THE EYE 
THE work of which, this paper is a brief outline was done under 
the auspices of the American Medical Association. The ob- 
ject of the work has been to compare the effect of different lighting 
conditions on the eye and to find the factors in a lighting situation 
which cause the eye to lose in efficiency and to experience discom- 
fort. 
Confronting the problem of the effect of different lighting condi- 
tions on the eye, it is obvious that the first step towards systematic 
work is to obtain some means of estimating effect. The prominent 
effects of bad lighting systems are loss of efficiency, temporary and 
progressive, and eye discomfort. Three classes of effect, however, 
may be investigated: (1) the effect on the general level or scale of 
efficiency for the fresh eye; (2) loss of efficiency as the result of a 
period of work; and (3) the tendency to produce discomfort. A 
description of tests designed especially for the investigation of these 658 
THE JOURNAL OF PHILOSOPHY 
effects has already appeared in print.1 The test for the second 
effect, it may be mentioned, is not analytical in nature. Its results 
express an aggregate loss of function. Supplementary tests have 
been devised, therefore, by means of which the factors eontributive 
to a given result may be separated out. A description of these 
tests is also included in the papers referred to above. 
The following aspects of lighting sustain an important relation 
to the eye: the evenness of illumination, the diffuseness of light, 
the angle at which the light falls on the object viewed, the evenness 
of surface brightness, intensity, and quality. The first four of 
these factors, which may be grouped together as distribution factors, 
will be discussed briefly with reference to types of lighting now in 
common use. 
The ideal condition with regard to the distribution factor is to 
have the field of vision uniformly illuminated with light well dif- 
fused and no extremes of surface brightness. When this condi- 
tion is attained the illumination of the retina will shade off more or 
less gradually from center to periphery, which gradation is neces- 
sary for accurate and comfortable fixation and accommodation. In 
the proper illumination of a room by daylight, we have been able 
thus far to get the best conditions of distribution. Before it reaches 
our windows or skylights, daylight has been rendered widely dif- 
fuse by innumerable reflections; and the windows and skylights 
themselves acting as sources have a broad area and low intrinsic 
brilliancy, all of which features contribute towards giving the ideal 
conditions of distribution stated above. Of the systems of artificial 
lighting the best distribution effects, speaking in general terms, 
are given by the indirect systems, and the semi-indirect systems with 
a small direct component of light. In the indirect systems the 
i Ferree, 0. E. " Tests for the Efficiency of the Eye under Different Sys- 
tems of Illumination and a Preliminary Study of the Causes of Discomfort," 
Trans. III. Eng. Soc. 1913, 8, pp. 40-60; " Untersuchungsmethoden fur die 
Leistungsfahigkeit des Auges bei verschiedenen Beleuchtungssystemen, und 
eine vorlaufige Untersuchung uber die Ursachen unangenehmer optischer Em- 
pfindungen," Z. f. Sinnesphysiol., 1915, 49, pp. 59-78; "The Efficiency of the 
Eye under Different Systems of Lighting," Fourth Intern. Congress on 
School Hygiene, Buffalo, 1913, 5, pp. 351-364; "Ophthalmology," July, 1914, 
pp. 1-16; "Mind and Body," 1913, 20, pp. 280-286, 345-353; "The Problem 
of Lighting in Its Relation to the Efficiency of the Eye," Science, July 17, 
1914, N. S., 15, pp. 84-91; Ferree, C. E. and Rand, G. " The Efficiency of th® 
Eye under Different Conditions of Lighting: The Effect of Varying Distribu- 
tion and Intensity," Trans. Ilium. Eng. Soc., August, 1915, 10, pp. 407-447; 
"Further Experiments on the Efficiency of the Eye under Conditions of Light- 
ing," Trans. III. Eng. Soc., August, 1915, 15, pp. 448-504. See also J. R. Cra- 
vath. "Some Experiments with the Ferree Test for Eye Fatigue," Trans. III. 
Eng. Soc., 1914, 9, pp. 1033-1047; also C. E. Ferree, "Discussion of Mr. Cravath's 
Paper, Some Experiments," etc., ibid., pp. 1050-1059. PSYCHOLOGY AND SCIENTIFIC METHODS 
659 
source is concealed from the eye and the light is thrown against 
the ceiling or some other diffusely reflecting surface in such a way 
that it suffers one or more reflections before it reaches the eye. In 
some of the respects most important to the eye, this and the semi- 
indirect systems with a small direct component of light give the best 
approximation of the distribution effects characteristic of daylight 
of any that have yet been devised. The direct lighting systems are 
designed to send the light directly to the plane of work. There is 
in general in the use of these systems a tendency to concentrate the 
light on the working plane or object viewed rather than to diffuse 
it, and to emphasize brightness extremes rather than to level them 
down. Too often, too, the eye is not properly shielded from the 
light source and frequently no attempt at all is made to do this. 
The semi-indirect systems are intended to represent a compromise 
between the direct and indirect systems. A part of the light is 
transmitted directly to the plane of work through the translucent 
reflector placed beneath the source of light, and a part is reflected 
to the ceiling. Thus, depending upon the density of the reflector, 
this type of system may vary between the totally direct and the 
totally indirect as extremes and share in the relative merits and 
demerits of each in proportion to its place in the scale. By giving 
better distribution effects this type of system is supposed also to 
be a concession to the welfare of the eye, but our tests show that 
the concession, at least for the reflectors of low and medium den- 
sities, is not so great as it is supposed to be. In fact, installed 
at the intensity of illumination ordinarily used or at an intensity 
great enough for all kinds of work, little advantage seems to be 
gained for the eye in this type of lighting with reflectors of low 
or medium densities; for with these intensities of light and den- 
sities of reflector, the brightness of the source has not been suffi- 
ciently reduced to give much relief to the suffering eye. Until 
this is done in home, office, and public lighting, we can not hope to 
get rid of eye strain with its complex train of physical and mental 
disturbances. 
In the experimental work the following points are covered: the 
effect of varying the distribution factors on the ability of the eye 
to maintain its maximal efficiency for a period of work; the effect 
of varying the intensity of light with various groupings of distri- 
bution factors; and certain miscellaneous experiments relating to 
the hygienic employment of the eye. These latter experiments in- 
clude the effect of varying the area and conversely the intrinsic 
brightness of the ceiling spots above the reflectors in an indirect 
system of lighting; the effect of varying the angle at which the 
light falls on the work in a given lighting situation; the effect of 660 
THE JOURNAL OF PHILOSOPHY 
using an opaque eye-shade with light and dark linings with each 
of the lighting installations used in the distribution and intensity 
series; the effect on the efficiency of the fixation muscles of a period 
of work under each of these installations; the effect of motion pic- 
tures on the eye at different distances from the projection screen; 
and a determination of the tendency of all the conditions of light- 
ing employed to produce discomfort and to cause loss of efficiency. 
The investigations are not abstract in character. All the varia- 
tions obtained were gotten in actual concrete lighting situations by 
employing lighting installations in common use. In order that a 
correlation might be made between lighting conditions and the 
effect on the eye, the following specification of illumination effects 
was made in each case. (1) A determination was made of the aver- 
age illumination of the room under each of the installations of light- 
ing used. The room was laid out in 3-ft. squares and illumination 
measurements were made at 66 of the intersections of these squares 
and at the point of work. Readings were taken in a plane 122 cm. 
above the floor with the receiving test-plate of the illuminometer 
in the horizontal, the 45 deg., and the 90 deg. positions, measuring 
respectively the vertical, the horizontal, and the 45 deg. compo- 
nents of illumination. The 122 cm. plane was chosen because that 
was the height of the test object. In the work on the distribution 
series the illumination was made as nearly as possible equal at the 
point of work. (2) A determination was made in candle-power per 
sq. in. of the brightness of prominent objects in the room, such as 
the test surface; the reflectors for the semi-indirect installation; the 
reflectors and filament for the direct installation, etc.; the reading- 
page; the specular reflection from surfaces; etc. The brightness 
measurements were made by means of a Sharp-Millar illuminometer 
with the test-plate removed. The instrument was calibrated against 
a magnesium oxide surface obtained by depositing the oxide from 
the burning metal. By this method the reflecting surfaces were 
used as detached test-plates. The readings were converted into 
candle-power per sq. in. by the following formula: 
_ . , , Foot-candles 
Brightness = -- 
7F x 144 
(3) Photographs were made of the room from three positions under 
each system of illumination. 
The tests for the effect on the eye were made at four represen- 
tative positions in the room. The observers used were all under 26 
years of age. A clinic record was made of the eyes of each ob- 
server. The following results were obtained. 
1. Of the lighting factors that influence the welfare of the eye, PSYCHOLOGY AND SCIENTIFIC METHODS 
661 
those we have grouped under the heading of distribution are ap- 
parently fundamental. They seem to be the most important we 
have yet to deal with in our search for the conditions that give us 
the minimum loss of efficiency and the maximum comfort in seeing. 
If, for example, the light is well distributed in the field of vision, 
and diffuse, and there are no extremes of surface brightness, our 
tests indicate that the eye, so far as the problem of lighting is con- 
cerned, is practically independent of intensity. That is, when the 
proper distribution effects are obtained, intensities high enough to 
give the maximum discrimination of detail may be employed with- 
out causing appreciable damage or discomfort to the eye. 
2. For the kind of distribution effects given by the direct and 
semi-indirect reflectors of low or medium densities, our results show 
that unquestionably too much light is being used in ordinary work 
for the comfort and welfare of the eye. 
3. The angle at which the light falls on the object viewed is an 
important factor, but not nearly so important, for example, as even- 
ness of surface brightness in the field of vision. Extremes of sur- 
face brightness in the field of vision seem to be the most important 
cause of the eye's discomfort and loss of efficiency in lighting sys- 
tems as we have them at the present time. In lighting from exposed 
sources it is not infrequent to find the brightest surface from 
1,000,000 to 2,500,000 times as brilliant as the darkest; and from 
300,000 to 600,000 times as brilliant as the reading-page. These 
extremes of brightness in the field of vision are, our tests show, 
very damaging to the eye. 
4. Of the systems of artificial lighting tested thus far, the best 
results have been obtained for the indirect system, and the semi- 
indirect systems with reflectors having a high density. By means 
of these reflectors the light is well distributed in the field of vision 
and extremes of surface brilliancy are kept within the limits which 
the eyes are prepared to stand. Considerable loss of efficiency has 
been found to result from the use of direct reflectors and semi- 
indirect reflectors of low or medium density. 
5. The loss of efficiency sustained by the eye under an unfavor- 
able lighting situation is found to be muscular, not retinal. The 
retina has been found to lose little if any more in efficiency under 
one than under another of the lighting systems employed. 
6. Loss of efficiency of the fixation muscles is, according to our 
tests, a very small part of the eye's aggregate loss in muscular effi- 
ciency as the result of work under an unfavorable lighting sys- 
tem. The chief loss seems to be sustained by the accommodation 
muscles or the muscles which adjust the lens of the eye. 
7. Eye-shades are apparently not an adequate substitute for 662 
THE JOURNAL OF PHILOSOPHY 
lamp-shades for the protection of the eye from the sources of light. 
The best results are obtained by means of an opaque eye-shade with 
a light lining. The usual opaque eye-shades with dark linings, 
while they shield the eye from the source of light, do not by any 
means eliminate harmful brightness differences in the field of vision. 
They in fact create for the eye a very unnatural brightness rela- 
tion; i. e., they make the whole upper half of the field of vision 
dark in sharp contrast with the brightly lighted lower half. The 
direct effect of this is a strong brightness contrast (physiological) 
over the lower half of the field of vision which causes glare in sur- 
faces which have no glare and increases the glare in surfaces in 
which glare is already present. Moreover, the unusual and strongly 
irregular character of the image formed on the retina probably also 
sets up warfare in the incentives given to the muscles which adjust 
the eye. That is, the upper half of the field of vision is dark and 
presents no detail. The effect of this is probably to exert a tendency 
to cause the muscular relaxation characteristic of the darkened 
field of vision. The lower half of the field of vision is light and 
filled with detail. The incentive here is for the best possible ad- 
justment of the eye for the discrimination of detail in the objects 
viewed, while the rim of the shade, the sharply marked boundary 
between the light and dark halves of the field of vision and much 
nearer to the eye than the objects viewed, serves as a constant and 
consciously annoying distraction to fixation and accommodation. 
These complex and somewhat contradictory impulses given to the 
muscles of the eye might very well and doubtless do cause an ex- 
cessive and unnatural loss of energy and efficiency in case of the 
prolonged adjustment of the eye needed for a period of work. 
Translucent shades when made sufficiently opaque to give the neces- 
sary reduction to the image of the source, darken too much the 
upper half of the field of vision and simulate thereby too much the 
effect given by the opaque shade with the dark lining to give the 
best results for efficient and comfortable seeing. 
8. The observation of motion pictures for two or more hours 
causes the eye to lose heavily in efficiency. The loss decreases 
rather regularly with the increase of distance from the projection 
screen. It seems little if any greater, however, than the loss caused 
by an equal period of working under much of the artificial light- 
ing now in actual use. In making these tests care was taken to 
choose a projection apparatus which gave a picture comparatively 
steady and free from flicker. 
9. In all the conditions tested a rather close correlation is found PSYCHOLOGY AND SCIENTIFIC METHODS 
663 
to obtain between the tendency of a given lighting condition to 
cause loss of efficiency and to produce discomfort. 
C. E. Ferree, 
Gertrude Rand. 
Bryn Mato College. [Reprinted from the Transactions of the Ieeuminating 
Engineering Society, No. 9, 1915.] 
SOME EXPERIMENTS ON THE EYE WITH INVERTED 
REFLECTORS OF DIFFERENT DENSITIES* 
C. E. FERREE AND G. RAND. 
Synopsis: In previous papers read before this society by the present 
writers, the gradation of surface brightness and its distribution in the 
field of vision were shown to be important factors in the effect of lighting 
conditions on the eye. In the work described in the present paper, grada- 
tion of surface brightness is made the chief variable. Inverted reflectors 
of six degrees of density are employed, and a correlation is made between 
the illuminating effects obtained and the tendency to cause loss of power 
to sustain clear seeing and to produce ocular discomfort. 
This paper is the fourth in a series in which the effect of differ- 
ent conditions of lighting on the eye is investigated. In the 
first paper, two tests were described-one designed to be used as 
a general test for detecting the comparative tendencies of different 
lighting conditions to cause a loss in the eye's power to sustain 
clear seeing for a period of work; the other for detecting the 
tendency to produce ocular discomfort. In the second paper, the 
application of the first of these tests to various lighting conditions 
was begun. Two purposes were had in making this application: 
(1) the studying and perfecting of the test itself for use in 
lighting work, which it is obvious could not be done effectively 
under one set or type of lighting conditions;1 and (2), the inves- 
tigation of pertinent lighting effects, the results of which could be 
made both to serve as a guide for further work, and to provide 
cumulative data from which conclusions may be drawn as the con- 
ditions and stage of advancement of the work may warrant.. 
This paper was divided into two sections. In the first the test 
was applied to the determination of the effect on the eye of three 
lighting installations, direct, semi-indirect and indirect, so se- 
lected as to give wide differences in illuminating effects. In the 
second section the effect of six variations in intensity for the 
direct and semi-indirect installations was determined. In both 
INTRODUCTION. 
* A paper presented at the ninth annual convention of the Illuminating Engineer- 
ing Society, Washington, D. C., September 20-23, 1915. 
The Illuminating Engineering Society is not responsible for the statements or 
opinions advanced by contributors. 1098 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
of these cases the tests were all made at one position in the room, 
the point marked as the position of the observer in Fig. i of the 
present paper. Obviously, however, the effect of an unfavorable 
installation on the eye will vary with the position of the observer 
in the room. In the third paper, therefore, the tests were re- 
peated for these installations at four positions in the room: the 
first with six reflectors in the field of view; the second with four; 
the third with two, and the fourth with none. The following 
features were also included: the work of the intensity series was 
completed, i. e., six intensities of light were used with the indi- 
rect reflectors; a test was described for determining the effect on 
the fixation muscles of the eye; and a series of miscellaneous 
experiments was conducted pertaining to the hygienic employment 
of the eye. In these experiments the following points were taken 
up: the effect of varying the area, and conversely the intrinsic 
brilliancy of the ceiling spots above the reflectors of the indirect 
system of lighting used; the effect of varying the angle at which 
the light falls on the work in a given lighting situation; the effect 
of using an opaque eye-shade with dark and light linings with a 
number of lighting installations; the effect on the efficiency of 
the fixation muscles of three hours of work under each of these 
installations; the effect of motion pictures on the eye for different 
distances of the observer from the projection screen; and a deter- 
mination of the tendency of the different conditions of lighting 
used in these experiments to produce ocular discomfort, and a 
comparison of the tendency to produce discomfort and to cause 
loss of efficiency. 
Time cannot be taken here even for a brief statement of the 
results obtained in these experiments. For the purpose of this 
paper, it will be sufficient to say that gradation of surface bright- 
ness and its distribution in the field of vision were shown to be 
important factors in the effect on the eye. In the work to be 
described in the present paper, gradation of surface brightness 
has been made the chief variable. Inverted opal glass reflectors 
of six degrees of density have been employed and a correlation 
has been obtained between the illuminating effects produced and 
their tendency to cause loss of efficiency and to produce ocular 
discomfort. As the work progresses, an attempt will be made FERREE AND RAND: EXPERIMENTS ON THE EYE 
1099 
not only to investigate this factor further in some of its more 
important relations to lighting practise, but to take up in turn, so 
far as is practicable, each of the other factors mentioned in the 
former papers.2 
CONDITIONS TESTED 
An effort has been made to get a series or reflectors similar in 
size and shape and differing only in density. It is our ultimate 
purpose to use these reflectors both in accord with the principles 
of direct and indirect lighting, and by employing additional trans- 
lucent and opaque reflectors, differing if need be in size and 
shape, to vary first one and then the other of the distribution 
factors mentioned in the former papers. So far, however, we 
have been able to use only six of the number of reflectors needed 
to carry out this plan, and these in accord with the principle of 
indirect lighting. They were all turned towards the ceiling and 
were installed the same distance from it. So installed, as the- 
photometric measurements will show, the chief variables have 
been the brightness of the reflectors and the ceiling spots above- 
the reflectors,-more especially, the brightness of the reflectors. 
The reflectors used will be designated here by the numerals,. 
I, II, III, IV, V and VI; and will be described in greater detail 
in an appendix to the paper. They were all installed 30 in. 
(0.76 m.) from the ceiling3 and were held by Plume and Atwood 
semi-indirect holders attached to cords dropped from the eight 
outlets shown in Fig. 1. 
It has been our wish to conduct this investigation, as has been 
the case in all our work on the distribution factors, with the 
quality and intensity of the light made approximately the same. 
Unfortunately, with the material available, the quality of the 
light could not be made in all cases uniformly alike. Clear tung- 
sten lamps were used as light sources with each installation, but 
two of the reflectors, I and II, were not free from color. The 
density of these reflectors had been secured in part, by giving 
them a brownish tone. Just how much effect this would have, 
if any, on the results of the tests we are not prepared at this time 
to say. The fact should be borne in mind, however, in considering 
the results obtained. It was decided to make the intensity of 
light as nearly equal as possible at the test object and to give a 1100 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
supplementary specification of the lighting effects in the remainder 
of the room. 
At the test object the light was photometered in several direc- 
tions. It was made approximately equal in the plane of the test 
object and as nearly as possible equal in the other directions. 
The specification of the lighting effects in the remainder of the 
room was accomplished as follows. (i) A determination was 
made of the average illumination of the room under each set of 
reflectors. The room was laid out in 3 ft. (0.90 m.) squares and 
Fig. 1.-Plan of test room 
illumination measurements were made at 66 of the intersections 
of the sides of these squares. Readings were taken in a plane 
122 cm. above the floor with the receiving test-plate of the illum- 
inometer in the horizontal, the 450 and 900 positions, measuring 
respectively, the vertical, the 450 and horizontal components of 
illumination. The 122 cm. plane was chosen because that was the 
height of the test object. (2) A determination was made of the 
brightness of prominent objects in the room, such as the test 
card, the reflectors, the reading page, the specular reflection from FERREE AND RAND: EXPERIMENTS ON THE EYE 
1101 
surfaces, etc. The brightness measurements were made by means 
of a Sharp-Millar photometer with the receiving test-plate re- 
moved. The instrument was calibrated against a magnesium 
oxide surface obtained by depositing the oxide from the burning 
metal on a white card. By this method the reflecting surfaces 
were used as detached test-plates. The readings were converted 
into candlepower per square inch by the following formula: 
brightness - foot-candles/7r X 144- (3) Photographs were made 
of the room for each set of reflectors employed. They will not all 
be included in this paper, however, because too little difference 
in illuminating effects is shown for the different reflectors to 
warrant so extensive a use of the photographic method of speci- 
fication. 
The tests were conducted in a room 30.5 ft. (9.29 m.) long, 
22.3 ft. (6.797 m-) wide, and 9.5 ft. (2.895 m-) high- In Fig. 1, 
this room is shown drawn to scale: plan of room, north, south, 
east and west elevations. In the plan of room are shown the 66 
stations at which the illumination measurements were made; and 
the positions of the outlets for the lighting fixtures, A, B, C, D, 
E, F, G and H. In the drawing, east elevation, the position of 
the observer at which the tests were taken is represented.4 So 
far in the work with these reflectors the tests have been made at 
only one point in the room. 
Table I gives the illumination measurements for each of the 
66 stations represented in Fig. I. These measurements were 
made with the receiving test-plate of the illuminometer in the 
horizontal, the vertical and the 450 planes. Tables II and III 
have been compiled to supplement Table I for the purpose of 
making a comparative showing of the evenness of illumination at 
the 122-cm. level given by the six sets of reflectors. Two cases 
may be made of this : (1) a comparison may be made of a given 
component from station to station; or (2) the difference between 
the components may be compared. To facilitate these compari- 
sons (a) the mean variation from the average of each of the 
components has been computed; and (&) the difference in the 
average of the three components has been determined. Results 
for the first of these points are shown in Table II; and for the 
second in Table III. 1102 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
TABLE I. 
Showing the illumination measurements in foot-candles for each of the 
66 stations represented in Fig. 1 for the six types of reflectors used. 
Division A. 
Horizontal, reflector type 
Vertical, reflector type 
450, reflector type 
Station 
I 
11 
Hi 
I 
11 
hi 
I 
11 
in 
I 
I.40 
i-35 
i-3° 
2 
I-5O 
1.28 
1.20 
3 
1-49 
1-52 
1.27 
4 
1.85 
1.46 
1.47 
5 
2.40 
2.20 
2.0 
6 
2.20 
2.40 
2.10 
7 
2.60 
2-5° 
2.30 
8 
2.9O 
3.10 
2.90 
9 
2.70 
2.60 
2.50 
IO 
1-54 
1.41 
I-32 
ii 
2. IO 
1.88 
1.78 
12 
3-9° 
4-3° 
3-70 
0-55 
o-53 
0.50 
2. IO 
2.0 
1.72 
13 
4-50 
4-9° 
4-50 
0.58 
o-55 
o-49 
2-3° 
2.70 
2.40 
14 
3-20 
3-40 
3-50 
0.50 
0.46 
0.48 
I.65 
1.70 
1.90 
15 
3.10 
3.10 
3-20 
0.52 
0.40 
0.42 
I.62 
1.70 
1.69 
16 
4-5° 
4-40 
43° 
0.50 
0.41 
0.50 
2-5° 
2.20 
2.10 
17 
3.80 
3-i° 
3.10 
0-54 
0.40 
o-43 
2.20 
1.48 
1.62 
18 
2.60 
1.86 
1.90 
0-57 
0.46 
o-45 
1-50 
1.10 
o-97 
i9 
3-30 
2.40 
2.50 
1-25 
0.87 
o-93 
2.40 
1.60 
1.90 
20 
4.10 
3-7° 
3-70 
1.17 
1.0 
0-94 
2.70 
2.20 
2.20 
21 
5-20 
4.60 
4-5o 
1.30 
i-3° 
1.22 
3-10 
2.90 
2.70 
22 
4.0 
3-50 
3.80 
1.16 
1.0 
0-94 
2.50 
2.20 
2.30 
23 
4.0 
3-7° 
3 70 
1.10 
1.06 
1.20 
2.60 
2-3° 
2.10 
24 
4-9° 
4-9° 
4-7° 
1.1 
1.20 
0-97 
3-o 
2.70 
2.80 
25 
3-90 
4.10 
4.20 
1-03 
0.91 
o-95 
2.40 
2.30 
3-30 
26 
2.50 
2.10 
1.78 
27 
2.80 
2.80 
2.40 
28 
4.80 
5.85 
4-70 
i-i5 
1.05 
1.20 
3-4° 
3-o 
3-20 
29 
5.80 
6.0 
6.20 
i-34 
i-35 
1-25 
3-80 
4.0 
4.0 
30 
4-50 
3.80 
4-50 
1.42 
1.11 
i-37 
3-4° 
2,70 
3-io 
3i 
4-5° 
3-9° 
4.60 
1.42 
I-I5 
1.14 
3-30 
2.50 
3-o 
32 
5.60 
5-30 
5.60 
1.42 
i-33 
r-32 
4,0 
3-70 
3-4° 
33 
5-o 
3.60 
4-2,5 
1-52 
1.10 
i-i5 
3-3° 
2.60 
2.80 
34 
4-7o 
3.80 
3.80 
1.86 
i-54 
1.42 
3-70 
2.90 
2.80 
35 
5-20 
4-90 
5-o 
2.10 
1.64 
1.60 
4-40 
3-5° 
3.80 
36 
4-50 
4.10 
4-50 
1.80 
1.61 
1.48 
3.60 
2.80 
3-30 
37 
4.60 
4.0 
4.60 
1.99 
i-54 
1.66 
3-7° 
3-o 
3-5o 
38 
4-9° 
5-40 
5-4° 
2.0 
1.90 
1.80 
4.0 
3.80 
4.20 
39 
4.10 
4.10 
4.0 
1.72 
1.68 
1.50 
3-20 
3.10 
3-20 
4o 
2.60 
2.40 
2.0 
4i 
2.0 
1.67 
1.62 TERREL AND rand: EXPERIMENTS ON THE EYE 
1103 
TABLE I.-(Continued.} 
Station 
42 
Horizontal, reflector type 
Vertical, reflector type 
450, reflector type 
I 
3-9° 
11 
4-40 
hi 
44o 
I 
I.80 
II 
I.60 
hi 
1.64 
I 
340 
11 
3-20 
Hl 
340 
43 
5-40 
5-4o 
5.60 
2.20 
1.86 
1.84 
4-70 
4.0 
440 
44 
4-40 
4.10 
3-70 
2.10 
1.76 
1.50 
4.0 
340 
3-20 
45 
4.10 
4-3° 
4.20 
2.10 
1.71 
1.68 
3.80 
340 
3-30 
46 
5-20 
5.60 
5-7° 
2.10 
1.60 
r.68 
4.60 
4.10 
440 
47 
4-50 
4.20 
4-5° 
I.76 
1.58 
1-52 
3.60 
3-30 
3-3° 
48 
3-90 
3-70 
3.80 
2.0 
i-94 
1.76 
3.60 
3-30 
3-30 
49 
4-9° 
4.80 
5-1° 
2-3° 
2.10 
2.0 
3-90 
4.20 
3.80 
50 
4.10 
3.60 
4.0 
2.3° 
2.0 
2.10 
3.80 
3-50 
340 
5i 
3-90 
3-70 
3-9° 
2.30 
1.90 
1.98 
3-70 
3-30 
3-50 
52 
4-40 
4-50 
4.60 
2.3° 
i-95 
2.0 
4.10 
3-70 
3-90 
53 
3.60 
3-70 
3.80 
2.0 
1.58 
1.80 
3-3° 
3-20 
3-20 
54 
3-io 
3-50 
340 
I.70 
1.48 
1.46 
3-20 
2.90 
3.10 
55 
4.10 
4-30 
4.10 
2.3° 
1.70 
1.80 
4.20 
3.80 
3-70 
56 
3-6° 
3-o 
3-30 
2.10 
1.80 
1.86 
3-50 
3.10 
3.60 
57 
cp .60 
3-o 
3.80 
2.3O 
1.82 
2.0 
3-50 
3-o 
3-70 
58 
4-40 
440 
540 
2. IO 
2.10 
2.10 
4.20 
4.10 
440 
59 
3-30 
3.60 
3.60 
I-63 
1.85 
1.66 
3-o 
3-20 
3-20 
60 
3-o 
2.60 
2.90 
2.0 
1.90 
1.66 
3-5° 
3.10 
3-20 
61 
3.10 
2.90 
3-20 
2.50 
2.0 
2.0 
3-9° 
3.60 
3-90 
62 
2.60 
2.60 
2.50 
2.20 
2.10 
1.92 
3-50 
340 
3-20 
63 
2.50 
2.50 
2.10 
2.20 
2.15 
2.60 
34o 
340 
3.10 
64 
3-io 
2.30 
3-0 
2.40 
2.0 
2.10 
4.0 
3-30 
3.60 
65 
2.40 
2.40 
2-3° 
I.98 
1-65 
i-59 
3.10 
2.70 
2.80 
66 
1-23 
1-25 
1.20 
Average 3.61 
345 
349 
1-65 
144 
143 
3-31 
2.98 
3-°5 
Division B. 
Horizontal, reflector type 
Vertical, reflector type 
450 reflector type 
Station 
' IV 
V 
VI 
' IV 
V 
VI 
IV 
V 
VI 
I 
I-5O 
1-45 
i-37 
2 
1-32 
1-36 
1-30 
3 
I.42 
1.40 
i-35 
4 
I-5O 
i-53 
1.58 
5 
2-3° 
2.50 
2.40 
6 
2.70 
2-3° 
2.40 
7 
2.60 
2.60 
2.50 
8 
3-40 
3-6° 
3-30 
9 
2.80 
2.80 
2.80 
IO 
i-55 
1.48 
1-56 
ii 
2.00 
i-94 
2.00 
12 
4.20 
4-30 
4.10 
0.65 
O.51 
0.53 
2.60 
2.40 
2.10 
13 
4-7° 
5-4o 
4-90 
0-59 
0.58 
0.60 
2.90 
3-io 
2.90 
14 
3-6° 
3-4o 
3.60 
0.56 
0-49 
0.50 
1.88 
1.88 
I.92 1104 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
TABLE I.-{Continued.} 
Horizontal, reflector type 
Vertical, reflector type 
450, reflector type 
Station 
' IV 
V 
VI 
'iv 
V 
VI 
IV 
V 
VI 
15 
3.10 
340 
3-30 
0.58 
049 
0.43 
I.72 
1.80 
1.80 
16 
4-40 
4-70 
4.80 
0-55 
0.40 
0.56 
2.30 
2.70 
2.60 
17 
3.10 
3.60 
3-7o 
044 
043 
0.50 
I.64 
1.85 
2.0 
18 
1.88 
1.96 
2.0 
049 
0.51 
0.57 
1.14 
1.15 
1.15 
19 
2-75 
2.60 
2.80 
1.25 
I.IO 
1.13 
2.10 
1.88 
1.84 
20 
4.20 
3.80 
4-30 
1.20 
1.20 
1.18 
2.70 
2.30 
2.60 
21 
4.60 
5-o 
5-o 
1.22 
I.48 
1-36 
2.80 
2.50 
3.20 
22 
3-70 
3-80 
4.10 
1.18 
1.05 
1.04 
2.50 
2.40 
2.50 
23 
3.80 
440 
4.20 
I-I5 
1.26 
1.06 
2.30 
2.70 
2.50 
24 
5-20 
5-9o 
5-70 
147 
1.52 
1.27 
3-20 
3-30 
3-o 
25 
4-90 
4-50 
4-3° 
1.25 
i-3i 
1.07 
2.9O 
2.80 
2.40 
26 
2.40 
2.50 
2.20 
27 
2.70 
2.80 
2-75 
28 
5-ro 
4-9° 
5-20 
145 
1.24 
1.14 
3.80 
34o 
3.60 
29 
5.60 
6.10 
6.40 
i-54 
i-34 
i-35 
4-3° 
440 
4.60 
30 
4-30 
4-3° 
4-50 
i-37 
1.27 
146 
3-20 
3-o 
3-50 
31 
4.20 
4.0 
440 
1-30 
1.26 
1.30 
3-o 
2.90 
3-3° 
32 
5.60 
5.80 
6.20 
148 
1.40 
1-53 
3-7o 
4-25 
4-5 
33 
4.10 
4-50 
4.80 
1.42 
1.28 
148 
3.10 
340 
3.80 
34 
4.0 
4.0 
4.20 
1.80 
2.0 
1.88 
3-o 
3-30 
3-5o 
35 
5-40 
5.60 
5.80 
2.10 
2.20 
2.40 
4.0 
4.0 
440 
36 
440 
4-30 
4-50 
1.70 
2.0 
1.98 
340 
3-50 
3.60 
37 
4-30 
440 
4-30 
1.88 
1.96 
1.92 
34o 
3-30 
3-50 
38 
5-20 
5-o 
5-50 
2.3° 
2.30 
2.20 
4.60 
4.10 
440 
39 
4-3° 
4.20 
4-50 
2.20 
1.68 
1.90 
3.80 
3-30 
3-7o 
40 
2.60 
2.40 
2.60 
4i 
1.80 
1.81 
1.92 
42 
4-50 
4.20 
440 
1.82 
1.90 
1.85 
3-5° 
3.60 
3.80 
43 
540 
5-5° 
5.80 
2.10 
2.10 
2.10 
4-50 
4-3° 
4.80 
44 
3.80 
3-70 
4-5o 
1.90 
2.10 
2.0 
3-3° 
2.70 
3-90 
45 
4.20 
440 
4.60 
1.90 
1.90 
1.98 
3.60 
3-9° 
3-90 
46 
540 
5.80 
6.20 
1.90 
1.85 
1.88 
4-3° 
4-7o 
4.60 
47 
3-9° 
4.0 
4-50 
1.80 
1.78 
1.72 
3.60 
3-30 
3.80 
48 
3.60 
3-70 
4-50 
1.91 
1-94 
2.30 
3-20 
3-30 
4.0 
49 
5-o 
4-90 
5-30 
2.20 
2.60 
2.60 
4-50 
4.60 
4.80 
50 
3-90 
4.0 
4.20 
2.40 
2.20 
2.80 
3-7° 
3-70 
4.20 
5i 
3-90 
3.80 
4.20 
2.40 
2.20 
2.60 
3-70 
3-60 
4.10 
52 
4-65 
4-7o 
5-o 
2.50 
2.50 
2.50 
4.10 
4.20 
4.20 
53 
4.0 
3-50 
4.0 
2.10 
2.10 
2.10 
3-5° 
3-20 
3.60 
54 
3-70 
3-50 
3.80 
i-74 
1.70 
1.82 
3-20 
3-3° 
3-60 
55 
4.20 
4-70 
5-o 
2.40 
2.0 
2.20 
4.0 
4-3° 
4-9° 
56 
3-20 
3-3° 
3-6° 
2.-20 
2.20 
2.10 
3-50 
3.60 
3-90 
57 
340 
3-50 
3-60 
2.20 
2-3° 
2.20 
3-60 
3-80 
3-7o 
58 
4.60 
5.10 
5-20 
2.20 
2-3° 
2.40 
4.20 
4-90 
4.60 FERREE AND RAND: EXPERIMENTS ON THE EYE 
1105 
table I.- 
(Continued.) 
Horizontal, reflector type 
Vertical, reflector type 
450, reflector type 
Station 
IV 
V 
VI 
' IV 
V 
VI 
IV 
V 
VI 
59 
3-90 
3-9° 
4-30 
1.78 
1.81 
2.40 
3-2° 
3-5° 
4-o 
60 
2.50 
2.50 
3-o 
2.0 
2.20 
2.40 
3-3° 
3-70 
3.80 
61 
3-2° 
3-20 
3.8° 
2.30 
2.60 
2.40 
4-40 
4.20 
4.6 
62 
2-3° 
2.60 
2.50 
2. IO 
2.30 
2.40 
3-20 
3-o 
3-4o 
63 
2.50 
2.40 
2.60 
2.60 
2.80 
2.10 
3.60 
3.80 
3-40 
64 
3-io 
2.90 
3-° 
2.30 
2.40 
2.40 
4.0 
4.0 
4.0 
65 
2.30 
2.40 
2.40 
2.0 
2.0 
2.10 
3.10 
3.00 
3.10 
66 
1.14 
1.16 
1.42 
Average 3.80 
3-70 
4.20 
1-675 
1.68 
1.71 
3-3° 
3-3i 
3-49 
table II. 
Compiled from Table I to show a comparison of the evenness of the illu- 
mination at the 122-cm. level given by the six types of reflector used. 
Mean variation of components 
Percentage of 
mean variation of components; 
A J pt tix 
reflector 
Vertical 
Horizontal 
45° 
Vertical 
Horizontal 
45° 
I 
0.976 
0.516 
0.582 
27.0 
3i-3 
17.6 
II 
0-999 
0.487 
0.576 
29.O 
33-8 
19-3 
III 
1.066 
0.430 
0.562 
30.5 
30.1 
18.4 
IV 
1.21 
0.498 
0.601 
31-8 
29-7 
18.2 
V 
1.10 
0-539 
0.628 
29.8 
32-i 
19.0 
VI 
1-47 
0-574 
0.677 
35-0 
31.2 
19-4 
TABLE III. 
Compiled from Table I to show the difference in the average values of the 
three components of illumination for the six types of reflector used. 
Difference between components 
Percentage of 
difference between components 
Type of 
Vertical and 
Vertical 
450 and 
Vertical and 
Vertical 
450 and 
reflector 
horizontal 
and 450 
horizontal 
horizontal 
and 450 
horizontal 
I 
I.96 
0.30 
1.66 
54-3 
8-3 
50-2 
II 
2.OI 
°-47 
i-54 
58.3 
13-6 
51-7 
III 
2.06 
0-44 
1.62 
59-o 
12.6 
53-1 
IV 
2.125 
0.50 
1-625 
559 
13-2 
49-2 
V 
2.02 
0-39 
1-63 
54-6 
10.5 
49-2 
VI 
2-49 
0.71 
1.78 
59-3 
16.9 
51-0 
Figs. 2-5 are taken from the series of photographs showing the 
illumination effects produced by the six types of reflector used.5 
As was stated earlier in the paper, not so much use has been 
made of the photographic method of specification in this as in 
the former papers. In the former papers three photographs were 
given for each set of reflectors. One of these was taken from 
the south end of the room at a point 4 ft. (1.22 m.) from the 1106 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
west wall. This photograph was taken so as to comprehend as 
much of the room as was possible in one view. It included the 
greater part of the ceiling, floor, and north wall, six of the fixtures 
and about one-half of the east wall. Another was taken to show 
the illumination effects in the west half of the room. This photo- 
graph represents the distribution of light and shade on the greater 
part of the west wall and adjacent ceiling and includes two of the 
fixtures. A third was taken primarily for showing the brightness 
measurements of all surfaces having a very high or very low 
brilliancy in the field of view of the observer. To have carried 
out this program in full in the present work would have required 
the insertion here of eighteen photographs. The amount of 
difference in the distribution of light and shade for the different 
reflectors was much too small to warrant this. It has in fact been 
deemed sufficient to include in this paper photographs for only the 
second and third of these positions and for only two of the sets 
of reflectors used,-the most opaque and the least opaque. The 
photographs for the second position are shown in Figs. 2 and 3; 
for the third, in 4 and 5. In representing the brightness measure- 
ments in Figs. 4 and 5, the spot measured is marked by a letter 
and the numerical value of the brightness measurement in candle- 
power per square inch is printed near by. The spots are lettered 
for convenience of reference in the tables of brightness measure- 
ments. The photographs were taken from a point directly be- 
hind the position of the observer as near to the south wall of the 
room as was possible; and although not all of the observer's field 
of view is covered by the brightness measurements made, owing 
to the narrow field of the camera as compared with the binocular 
field, still the order of magnitude of brightness differences present 
in the field of view is well represented by these measurements. 
In Tables IV and V are given the brightness measurements of 
the room for the six sets of reflectors. These tables also include 
the letters identifying the measurements with the spots measured 
as shown in Figs. 4 and 5. The distribution of light and shade 
in the room was so similar for the different sets of reflectors that 
the spots measured have approximately the same location for 
each set of reflectors. Two sets of measurements were made 
of the brightness of the reflectors,-one with the opening of the Fig. 2.-Showing the illumination of the west wall of the room, Reflector I. 
Fig. 3.-Showing the illumination of the west wall of the room, Reflector VI. Fig. 4.-Showing the illumination effects in the north end of the room, Reflector I; 
and the brightness measurements of all surfaces having a very high or a very low 
brilliancy. This photograph was taken from a point directly behind the observer 
as near to the south wall of the room as was possible, and comprehends as much 
of the observer's field of view as could be included in the field of the camera. 
Fig. 5.-Showing the illumination effects in the north end of the room, Reflector VI; 
and the brightness measurements of all surfaces having a very high or a very low 
brilliancy. This photograph was taken from a point directly behind the observer 
as near to the south wall of the room as was possible, and comprehends as much 
of the observer's field of view as could be included in the field of the camera. EERREE AND rand: EXPERIMENTS ON THE EYE 
1107 
illuminometer close to the reflector and the other with the opening 
as nearly as possible in the position of the observer when making 
the test. In the former case the receiving arm was turned normal 
to the surface measured and the instrument was supported in 
such a position that the opening was about 4 in. (10.16 cm.) from 
this surface. The surfaces of some of the reflectors presented so 
much unevenness of brightness that overlapping measurements 
were made and an average taken. These average values are given 
in Table IV. In Table V is given the brightness of the reflectors 
as measured from the position of the observer. These measure- 
ments were taken of the reflectors at outlets A, B and C (Fig. 1) 
for each of the six installations. A comparison of these measure- 
ments will show that reflector B has in each case a higher value 
than reflector A, and C a higher value than B. Whether or not 
this can be wholly accounted for because the reflectors were not 
perfect diffusers we are not prepared to say. That is, the angle 
subtended by reflector A at the point of observation was less than 
that subtended by B, and by B less than that subtended by C; 
so that at the distance at which these reflectors was viewed ap- 
proximately all of A occupied the field of the illuminometer in 
making the brightness match, while only the brighter central 
portions of B and C were comprehended in this field, still less of 
the duller periphery being included for C than for B. 
In Tables VI and VII, are shown some prominent ratios of 
surface brightness for the six sets of reflectors. In compiling 
these ratios it has been considered important to make a compara- 
tive showing for the different types of reflectors (a) of the 
extremes of surface brightness and (&) of the relation of the 
brilliancy of objects in the surrounding field to the surface bright- 
ness at the point of work. Extremes of surface brightness are 
shown by giving the ratios between surfaces of the first, second, 
third, etc., order of brilliancy and the lowest order of brilliancy; 
and the comparison of the brilliancy of objects in the surrounding 
field to the brightness at the point of work by giving the ratios 
of the surfaces of the first, second, and third order of brilliancy 
to the brightness of the test card and the reading page in the 
working position. 1108 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
TABLE IV. 
Showing the brightness measurements in candlepower per square inch for 
the surfaces A, B, C, D, etc. (Figs. 4 and 5), the test card and reading 
page. These measurements were taken with the illuminometer close to the 
surface measured and with its receiving arm normal to this surface. 
Surface 
Reflector 
Reflector 
Reflector 
Reflector 
Reflector 
Reflector 
measured 
type I. 
type II 
type III 
type IV 
type V 
type VI 
A 
0.264 
0.361 
O.392 
0.614 
0.848 
0.920 
B 
0.030 
O.OI985 
0.024 
O.OIOI 
O.OI37 
O.OI93 
c 
0.029 
O.O2I 
0.021 
0.0123 
0.0166 
0.0156 
D 
O.OI93 
0.0106 
O.OO75 
0.0070 
0.00767 
0.0107 
E 
0.00238 
O.OO246 
0.00229 
0.00282 
0.00255 
0.0026 
F 
O.OO34 
O.OO394 
O.OO34 
0.00396 
0.00396 
0.00396 
G 
0.0040 
O.OO392 
O.OO42 
0.00497 
0.00418 
0.00458 
H 
0.00414 
O.OO396 
O.OO44 
0.00506 
O.OO43 
0.00466 
I 
0.0044 
O.OO4O2 
O.OO453 
0.00528 
O.OO42 
0.00484 
J 
0.00163 
O.OCII 
0.00128 
0.00141 
O.OOI23 
O.OO163 
K 
0.0036 
0.00387 
O.OO414 
0.0044 
O.OO425 
0.00414 
L 
0.0023 
0.00224 
0.00282 
0.00299 
O.OO273 
O.OO299 
M 
0.00458 
0.00405 
O.OO484 
0.0052 
O.OO427 
0.00506 
N 
O.OO277 
0.00216 
0.00216 
0.00334 
0.00268 
0.00268 
0 
0.00348 
0.00299 
O.OO462 
0.00361 
0.00361 
O.OO365 
P 
O.OO37 
0.00312 
O.OO5O6 
0.00409 
O.OO37 
O.OO397 
Q 
O.OOO97 
0.00083 
0.00106 
0.00099 
O.OOO924 
0.00106 
R--" .... 
O.OOI99 
0.0029 
0.00207 
0.00220 
0.00246 
0.00238 
Test card. 
O.OO312 
0.00308 
O.OO3O8 
0.00317 
0.00312 
0.00317 
Reading 
page 
hori- 
zontal. • 
0.00528 
0.00497 
O.OO5O6 
0.0052 
0.00484 
0.00484 
Reading 
page 
45° po- 
sition •. 
O.OO352 
0.00348 
O.OO352 
0.00348 
O.OO334 
O.OO339 
TABLE V. 
Showing the brightness measurements in candlepower per square inch of 
the reflectors used when the measurements are made from the position oc- 
cupied by the observer during the test. In these measurements the receiving 
arm of the illuminometer was placed as nearly as possible in the position of 
the observer's eye during the test, and was pointed at the reflector. The 
position of the reflector in each case is shown by the letters A, B and C in 
Fig- I- 
Position of 
Reflector 
Reflector 
Reflector 
Reflector 
Reflector 
Reflector 
reflector 
type I 
type II 
type III 
type IV 
type V 
type VI 
A 
0.119 
0.156 
0.180 
O.2325 
0.327 
0.382 
B 
0.1755 
O.1913 
0.2025 
0.2535 
0.338 
0.405 
c 
0.2025 
0.338 
0-397 
0-544 
O.722 
0.830 EERREK AND rand: EXPERIMENTS ON THE EYE 
1109 
Supplementary to Tables IV, VI and VII we have computed 
for the six types of reflector the mean variation of the several 
brightness values from their average values. While important 
from the standpoint of showing the variations from the mean for 
the different types of reflector, such a comparison is, however, 
probably not so important from the standpoint of the eye as are 
the comparisons given in Tables IV to VIL That is, from the 
standpoint of the effect on the eye it is probably more important 
to give a representation of the brightness of individual surfaces, 
more especially of surfaces showing extremes of brightness, than 
it is to give the mean variation from the average brightness of all 
the surfaces. In order to make possible the comparison with and 
without the reflector and the spot above the reflector, the table is 
made to show separately the mean variation of the following 
measurements: (a) for all: (b) for all but the reflector; and 
(c) for all but the reflector and the spot above the reflector. 
Results are given in Table VIII. 
As was stated earlier in the paper the effect of a harmful in- 
stallation on the ability of the eye to maintain its efficiency for a 
period of work varies with the position of the observer in the 
room. In the former work the tests were made at four positions, 
one in which six fixtures were in the field of view; one in which 
four were in the field of view; one in which two were in the field 
of view; and one in which none was in the field of view. This 
variation of the position in which the observation is made accom- 
plishes two purposes : (i) it gives us a more representative idea 
of the difference of the effect on the eye of the six types of light- 
ing used; and (2) it shows the effect of varying the number of 
surfaces in the field of view showing brightness differences, par- 
ticularly the number of primary sources. So far we have been 
able to conduct the tests for the reflectors used in this work at 
only one of these positions, namely, the one with six reflectors 
in the field of view.6 Later we expect to repeat the tests for 
at least a part of these reflectors at the other three positions. 
The results for the effect on the eye are given in Table IX.7 
The values given in this table are averaged in each case from 
the results of 6 three-hour tests and are typical of the results 
obtained for all of our observers. In order to show the repro- TRANSACTIONS OF ILLUMINATING FNGINUFRING SOCIETY 
1110 
Division A. 
Ratio 
Reflector type I 
Reflector type II 
Reflector type III 
Lightest to darkest 
. ••• 0.264 /0.00097 = 272.0 
0.361 /o. 00083 = 435-0 
0.392 /o.00106 = 370.0 
2nd lightest to darkest 
• 0.030 /o.00097 = 31.0 
0.021 /o.00083 = 25-3 
0.024 /o.00106 = 22.6 
3rd lightest to darkest 
.... 0.029 /o. 00097 = 29.9 
0.01985/0.00083 = 23.9 
0.021 /o.00106 = 19.8 
4th lightest to darkest 
. • • 0.0193 /o.00097 = 20.0 
0.0106 /o.00083 = 12.8 
0.0075 /o.00106 = 7.08 
5th lightest to darkest 
0.00458/0.00097= 4.72 
0.00405/0.00083 = 4.88 
0.00506/0.00106 = 4.77 
6th lightest to darkest •... 
... 0.0044 /o.00097 = 4.54 
0.00402/0.00083 = 4.84 
0.00484/0.00106 = 4.57 
7th lightest to darkest 
... 0.00414/0.00097 = 4.27 
0.00396/0.00083 = 4.77 
0.00462/0.00106 = 4.36 
Sth lightest to darkest 
• •• 0.0040 /0.00097 = 4.12 
0-00394/0.00083 =t 4.75 
0.00453/0.00106 = 4.27 
9th lightest to darkest 
... 0.0037 /0.00097 = 3.81 
0.00392/0.00083 = 4.72 
0.0044 /o.00106 = 4.15 
10th lightest to darkest .... 
... 0.0036 /0.00097 = 3.71 
0.00387/0.00083 = 4.66 
0.0042 /o.00106 = 3.96 
nth lightest to darkest . ... 
• • • 0.00348/0.00097 = 3.59 
0.00312/0.00083 = 3.76 
0.00414/0.00106 = 3.91 
12th lightest to darkest .... 
... 0.0034 /o. 00097 = 3-51 
0.00299/0.00083 = 3.60 
0.0034 /0.00106 = 3.21 
13th lightest to darkest . ... 
... 0.00277/0.00097= 2.86 
0.0029 /o.00083 = 3.49 
0.00282/0.00106 = 2.66 
14th lightest to darkest -... 
... 0.00238/0.00097= 2.45 
0.00246/0.00083 = 2.96 
0.00229/0.00106= 2.16 
15th lightest to darkest .... 
.... 0.0023 /o. 00097 = 2.37 
0.00224/0.00083 = 2.70 
0.00216/0.00106 = 2.04 
16th lightest to darkest .... 
.... 0.00199/0.00097= 2.05 
0.00216/0.00083 = 2.60 
0.00207/0.00106 = 1.95 
17th lightest to darkest - ... 
... 0.00163/0.00097= 1.68 
0.0011 /o.00083 - i-33 
0.00128/0.00106 = 1.21 
TABLE VI * 
Ratios showing the extremes of surface brightness for the six types of reflectors used. FERREE AND RAND: EXPERIMENTS ON THE EYE 
1111 
Ratio 
Reflector type IV 
Reflector type V 
Reflector type VI 
Lightest to darkest 
•... 0.614 /0.00099 = 620.0 
0.848 /O.OOO924 = 918.0 
0.92 /0.00106 = 868.0 
2nd lightest to darkest 
.... 0.0123 /o. 00099 = 12.4 
O.OI66 /0.000924 = 18.O 
O.OI93 /O.OOIO6 = 18.2 
3rd lightest to darkest 
. ... o.oioi /o.00099 = 10.2 
0.0137 /o.000924 = 14.8 
O.OI56 /0.00106 = 14.7 
4th lightest to darkest • • 
.... 0.007 /o.00099 = 7.70 
0.00767/0.000924 = 8.30 
0.0107 /0.00106 = IO. I 
5th lightest to darkest 
• ... 0.00528/0.00099 = 5.33 
O.OO43 /O.OOO924 = 4.65 
0.00506/0.00106 = 4.77 
6th lightest to darkest 
.... 0.0052 /o. 00099 " 5-25 
O.OO427/O.OOO924 = 4.62 
O.OO484/O.OOIO6 = 4.57 
7th lightest to darkest 
.... 0.00506/0.00099= 5.11 
O.OO425/O.OOO924 = 4.60 
0.00466/0.00106 = 4.40 
8th lightest to darkest 
.... 0.00497/0.00099= 5.02 
0.0042 /0.000924 = 4.55 
0.00458/0.00106 = 4.32 
9th lightest to darkest 
.... 0.0044 /o. 00099 = 4-44 
0.00418/0.000924 = 4.52 
0.00414/0.00106 = 3.9I 
10th lightest to darkest .... 
... 0.00409/0.00099= 4.13 
O.OO396/O.0OO924 = 4.29 
0.00397/0.00106 = 3.74 
1 ith lightest to darkest .... 
.... 0.00396/0.00099 = 4.0 
0.0037 /o.000924 = 4.0 
0.00396/0.00106= 3.73 
12th lightest to darkest • • •. 
.... 0.00361/0.00099= 3.65 
0.00361/0.000924 = 3.91 
0.00365/0.00106 = 3.44 
13th lightest to darkest .... 
.... 0.00334/0.00099= 3.37 
0.00273/0.000924 = 2.95 
O.OO299/O.OOIO6 = 2.82 
14th lightest to darkest • • • • 
.... 0.00299/0.00099= 3.02 
O.OO268/O.OOO924 = 2.90 
0.00268/0.00106 = 2.53 
15th lightest to darkest ... 
.... 0.00282/0.00099= 2.85 
O.OO255/O.OOO924 = 2.76 
0.0026 /o.00106 = 2.45 
16th lightest to darkest .... 
.... 0.0022 /o.00099 = 2.22 
0.00246/0.000924 = 2.66 
O.OO238/O.OOIO6 = 2.24 
17th lightest to darkest .... 
.... 0.00141/0.00099= 1.42 
0.00123/0.000924 = I.33 
0.00163/0.00106 = I.54 
* It will be noted in Tables IV and VI that while the reflectors grade in brightness from I to VI in an unbroken series, the ratio lightest 
to darkest for Reflector III is less than for Reflector II, and for Reflector VI than for Reflector V. This in all probability is because of a 
difference in distribution effects given by these reflectors due to a difference in power to diffuse the light. That is, while the brightest spots 
(the reflectors) grade in an unbroken series, the darkest spots do not, 
TABLE VI. 
Division B. 1112 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
Ratio 
Lightest to test card 
Lightest to reading page 
2nd lightest to test card 
2nd lightest to reading page . .. 
3rd lightest to test card 
3rd lightest so reading page • • • 
Reflector type I 
.. 0.264 /0.00312 = 84.7 
. . O.264 /O.OO352 = 75.O 
• • 0.030 /O.OO312 = 9.61 
• • 0.030 /O.OO352 = 8.5 
. . 0.029 /O.OO312 = 9.29 
.. 0.029 /0.00352 = 8.2 
Reflector type II 
0.361 /o. 00308 = 117.2 
0.361 /0.00348 = 103.8 
0.021 /o.00308 = 6.8 
0.021 /o.00348 = 6.3 
0.01985/0.00308 = 6.46 
0.01985/0.00348 - 5.70 
Reflector type III 
0.392 /0.00317 = 123.7 
0.392 /o.00348 = I I 2.6 
0.024 /o.00317 = 7.57 
0.024 /o.00348 = 6.9 
0.021 /0.00317 = 6.62 
0.021 /0.00348 = 6.03 
Ratio 
Lightest to test card 
Lightest to reading page 
2nd lightest to test card 
2nd lightest to reading page • .. 
3rd lightest to test card 
3rd lightest to reading page • • • 
Division 
Reflector type IV 
.. 0.614 /0.00317 = 193.7 
• • 0.614 /0.00348 = 176.4 
• . 0.0123/0.00317 = 3.88 
.. 0.0123/0.00348= 3.53 
.. 0.0101/0.00317= 3.19 
.. 0.0101/0.00348= 2.90 
B. 
Reflector type V 
0.848 /o.00312 = 272.0 
0.848 /o.00334 = 248.0 
0.0166 /o.00312 = 5.32 
0.0166 /0.00334 = 4.97 
0.00137/0.00312 = 4.39 
0.00137/0.00334 = 4.1 
Reflector type VI 
0.92 /o.00317-= 290.2 
0.92 /o.00339 - 271.0 
0.0193/0.00317= 6.09 
0.0193/0.00339= 5.7 
0.0156/0.00317 = 4.92 
0.0156/0.00339 = 4.6 
Table vil. 
Ratios showing the relation of the brilliancy of objects in the surrounding field to the surface brightness 
at the point of work for the six types of reflector used. 
Division A. EERREE AND RAND: EXPERIMENTS ON THE EYE 
1113 
ducibility of the results obtained and that the variations produced 
by the changes in lighting effects are much greater than the vari- 
ations in the test itself, subject to all the variable factors which 
may influence it, the mean variation from the average result has 
been computed in each case. The value of this in per cent, is 
given in column 15.8 
TABLE VIII. 
Compiled from Table IV to show the mean variations in surface bright- 
ness for the six types of reflector used. 
Division A. 
Mean variation 
for the three reflectors 
Percentage 
of mean variation 
for the three reflectors 
Measurements considered 
Reflector type 
I II 
III 
Reflector type 
I II III 
All 
0.02885 
0.0373 
0.0405 
134.8 
148.0 148.4 
All but the reflector 
All but the reflector and 
0.00667 
0.00412 
0.0041I 
93-2 
75-3 70.3 
the spots above the re- 
flector 
0.000917 
0.000884 
0.0012 
29-5 
29-7 35-8 
Division B. 
Mean variation 
for the three reflectors 
Percentage 
of mean variation 
for the three reflectors 
Measurements considered 
Reflector type 
IV V 
VI 
Reflector type 
IV V VI 
All 
0.06494 
0.08852 
0.09597 
168.O 
170.8 170.5 
All but the reflector 
All but the reflector and 
O.OO2O 
0.00274 
0.00342 
42.4 
56.0 62.0 
the spots above the re- 
flector 
O.OOIII 
0.000964 
0.00104 
30.9 
30.0 30.2 
In Chart i a graphic representation is made of the results of 
this table. In constructing this chart, the total length of the test 
period is plotted along the abscissa, and the ratio of the time the 
test object is seen clear to the time it is seen blurred in the three- 
minute records before and after work is plotted along the ordin- 
ate. Each one of the large squares along the abscissa represents 
one hour of the test period, and along the ordinate an integer of 
the ratio. 
So far in all our work we have shown for the sake of complete- 
ness of representation the gradation of surface brightness in 
three ways.-(i) Brightness measurements of prominent sur- 
faces have been made. (2) Ratios have been given between 
surfaces of the first, second, third, etc., order of brilliancy and 
surfaces of the lowest order of brilliancy; and between sur- 1114 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
Reflector 
type 
Watts 
Volts 
Intensity foot-candles 
Hori- Ver- 
zontal tical 450 
Time 
Maximal 
distance 
at which 
test object 
can be 
seen Working 
clear distance 
Total 
time 
clear 
Total 
time 
blurred 
Total 
time 
clear 
total 
time 
blurred 
Ratios re- 
duced to 
common 
standard 
Toss of 
efficiency 
expressed 
in per- 
centage Mean 
drop m variation 
ratio (per cent.)9 
I 
800 
III.O 
4-1 
1.14 
2.7 
9 A.M. 
80.2 
62.2 
138.5 
41-5 
3-34 
3-5 
- 
- 
12 M. 
80.0 
62.2 
133-0 
47-0 
2.83 
2-97 
15 
0.86 
II 
800 
IIO.O 
3-7 
1.13 
2.6 
9 A.M. 
80.0 
64.0 
139.O 
41.0 
3-39 
3-5 
- 
- 
12 M. 
79-5 
64.0 
115-0 
65-0 
i-77 
1.82 
48 
0-77 
III 
800 
IO7-5 
4.2 
1.16 
2.6 
9 A.M. 
80.0 
64.0 
145-0 
35-o 
4-14 
3-5 
- 
- 
12 M. 
79-5 
64.0 
Il8.O 
62.0 
i-9 
1.60 
54 
0.70 
IV 
800 
IO5-5 
3-8 
1.15 
2-5 
9 A.M. 
79-5 
63-6 
145-0 
35-o 
4-i4 
3-5 
- 
- 
12 M. 
79-o 
63.6 
112.0 
68.0 
1.65 
i-39 
60 
2.0 
V 
800 
105.5 
3-7 
I-I5 
2.6 
9 A.M. 
79-5 
63-6 
140.1 
39-9 
3-51 
3-5 
- 
- 
12 M. 
78.5 
63-6 
91.0 
89.0 
1.02 
1.017 
7i 
0.86 
VI 
800 
107.5 
4.2 
1.16 
2-7 
9 A.M. 
79-5 
63-6 
141.0 
39-o 
3-62 
3-5 
- 
- 
12 M. 
78.5 
63.6 
89.O 
91.0 
0.978 
0-945 
73 
1.0 
TABLE IX. 
Showing the tendency of the six types of reflector to cause loss of visual efficiency, or power to sustain clear seeing. FERREE AND RAND: EXPERIMENTS ON THE EYE 
1115 
TABLE X 
Showing a comparison of the tendency of the six types of reflector used to cause loss of efficiency and to produce ocular dis- 
comfort. The tendency to produce discomfort is estimated by the time required for just noticeable discomfort to be set up. 
Reflectors 
Volts 
Foot-candles 
Per cent, 
loss of 
efficiency 
Mean 
variation 
(Per cent.) 
Time 
limen of 
discomfort 
in seconds 
(not read- 
ing) 
Mean 
variation 
(Per cent.) 
Change 
produced 
by chang- 
ing type of 
reflector 
(Per cent.) 
Time 
limen of 
discomfort 
in seconds 
(reading) 
Mean 
variation 
(Per cent.) 
Change 
produced 
by chang- 
ing type of 
reflector 
(Per cent.) 
Hori- 
zontal 
Vertical 
45° 
Type I .. 
III 
4-1 
1.14 
2.7 
15 
0.86 
80 
1.1 
25 
1.6 
Type II.. 
IIO 
3-7 
I-I3 
2.6 
48 
0-77 
45 
1-5 
43-8 
.17 
i-5 
32.0 
Type III. 
107-5 
4.2 
1.16 
2.6 
54 
0.70 
34 
1.8 
244 
14 
2.1 
17-7 
Type IV. 
105.5 
3-8 
i.i5 
2.5 
60 
2.0 
21 
2-4 
38.2 
10 
2.0 
28.6 
Type V• • 
105.5 
3-7 
1.15 
2.6 
7i 
0.86 
17 
2.2 
19.O 
8 
2-3 
20.0 
Type VI. 
107.5 
4.2 
1.16 
2-7 
73 
1.0 
15 
2-7 
11.8 
6-5 
3-o 
18.8 1116 
transactions of illuminating engineering society 
faces of the first, second and third order of brilliancy and the 
brightness at the point of work. And (3) the mean variation 
from the average and the percentage of mean variation have been 
shown. In the consideration of these specifications, a number of 
CHART I. 
Showing the tendency of the six types of reflectors to cause loss of 
visual efficiency, or power to sustain clear seeing. Ratio time clear to 
time blurred is plotted against length of test period. 
Foot-candles 
Reflector 
Volts 
Horizontal 
Vertical 
45° 
Type I 
III 
4-1 
I.I4 
2.7 
Type II 
IIO 
3-7 
I.I3 
2.6 
Type III ; 
107-5 
4.2 
1.16 
2.6 
Type IV 
105.5 
3-8 
1.15 
2.5 
Type V 
105.5 
3-7 
i.i5 
2.6 
Type VI 
IO7.5 
4.2 
1.16 
2.7 
single items might be selected as of possible significance in rela- 
tion to the effect on the eye. Among these may be mentioned the 
order of magnitude of the highest brilliancies; the average bril- 
liancy; the ratio of the highest to the lowest order of brilliancy; 
the ratio of the highest order of brilliancy to the average bril- FERREE AND RAND: EXPERIMENTS ON THE EYE 
1117 
liancy; the ratio of the average to the lowest order of brilliancy; 
the ratio of the highest order of brilliancy to the brilliancy 
at the point of work, (brightness of test card and reading page) ; 
etc. In order to see which of these correlate most closely with 
the results of the test for tendency to cause loss of efficiency, 
CHART II. 
Showing the tendency of the six types of reflectors to cause loss of 
visual efficiency or power to sustain clear seeing. Percentage drop in ratio 
time clear to time blurred is plotted against brightness of reflector in 
candlepower per square inch. 
Reflector 
Volts 
Foot-candles 
Cp. per 
sq. in. 
Horizontal 
Vertical 
45° 
Type I 
III 
4.1 
1.14 
2.7 
0.264 
Type II . .. . 
... IIO 
3-7 
I-I3 
2.6 
0.361 
Type III ... 
... 107.5 
4.2 
1.16 
2.6 
0.392 
Type IV ... 
... 105.5 
3-8 
I.I5 
2.5 
0.614 
Type V .... 
... 105.5 
3-7 
I-I5 
2.6 
0.848 
Type VI ... 
... 107.5 
4-2 
1.16 
2.7 
0.920 
curves are constructed in which some of these features are plotted 
against the results of the test. These curves are given in Charts 
II to IV. In Chart II per cent, loss of efficiency is plotted against 
the highest order of brilliancy, namely, the brightness of the re- 
flectors. In Chart III and IV are grouped the remainder of the 
curves. 
Another method of evaluating the results of our test was briefly 1118 
TRANSACTIONS OF IIAUMINATING ENGINEERING SOCIETY 
treated of in a discussion of Mr. Cravath's paper by one of the 
writers. (The Transactions, 1914, IX, pp. 1051-1053.) In 
this method the ratio of the time seen clear to the total time of 
the observation is taken as the measure of the ability of the eye 
CHART III. 
Showing the tendency of the six types of reflectors to cause loss of 
visual efficiency or power to sustain clear seeing. In curve A percentage 
drop in ratio time clear to time blurred is plotted against ratio of average 
brightness to brightness at point of work; in B, against ratio of lightest 
surface to brightness at point of work; and in C, against average 
brightness. 
to sustain clear seeing at the time the test is taken. For the sake 
of comparing this method of evaluation with the one we have used 
in the rest of the paper, Charts V and VI have been constructed. 
In Chart V length of test period is plotted along the abscissa, and 
the ratio of time clear to total time of observation is plotted along 
the ordinate. In plotting these lines, one of the larger squares FERRE® AND rand: EXPERIMENTS ON THE eye 
1119 
along the abscissa represents one hour of the test period, and 
along the ordinate, o.i ratio, time seen clear to the total time of 
the observation. That is, in this method of treating the results, 
since the ratios, or the quantities to be plotted along the abscissa, 
are much smaller than they are in the former method, the scale 
CHART IV. 
Showing the tendency of the six types of reflector to cause loss of 
visual efficiency or power to sustain clear seeing. In curve D percentage 
drop in ratio time clear to time blurred is plotted against ratio of lightest 
surface to average brightness; in E, against ratio of lightest surface to 
darkest surface; and in F, against ratio of average brightness to darkest 
surface. 
has been multiplied by io for convenience of representation. In 
order that the lines may all start at a common point, the initial 
ratios are reduced to i as a common standard. In Chart VI, per 
cent, loss of efficiency as evaluated by this method is plotted 
against intrinsic brilliancy of reflector. As before, intrinsic 
brilliancy of reflector is plotted along the abscissa, and per cent. 1120 
transactions of illuminating engineering society 
loss of efficiency along the ordinate. A comparison of these re- 
sults with the former will show the same order of rating of the 
reflectors but a slight change in the position in the scale given to 
some of the reflectors. For the purpose of discovering what is 
CHART V. 
Showing the tendency of the six types of reflectors to cause loss of 
visual efficiency or power to sustain clear seeing. Ratio of time clear to 
total time of observation is plotted against length of test period. 
Foot-candles 
Reflector 
Volts 
Horizontal 
Vertical 
45° 
Type I 
III 
4-1 
I.I4 
2.7 
Type II 
IIO 
3-7 
I.I3 
2.6 
Type III 
107-5 
4.2 
1.16 
2.6 
Type IV 
105.5 
3-8 
I.I5 
2.5 
Type V 
105.5 
3-7 
I.I5 
2.6 
Type VI 
107.5 
4.2 
1.16 
2.7 
the best way of treating the results of the tests, several methods 
have been employed. Up to and including the present paper, 
however, only three of them have been given in print: ratio of 
time clear to time blurred, ratio of time clear to total time of FERREE AND rand: EXPERIMENTS ON THE EYE 
1121 
observation, and the per cent, of drop in the ratio time clear to 
time blurred. An ultimate decision with regard to what is the 
best method of treatment of the results can come, we believe, only 
with the consideration of a larger number of cases. 
The work was concluded by determining for the six types of 
CHART VI. 
Showing the tendency of the six types of reflectors to cause loss of 
visual efficiency or power to sustain clear seeing. Percentage drop in ratio 
time clear to total time of observation is plotted against brightness of 
reflector in candlepower per square inch. 
Reflector 
Volts 
Foot-candles 
Cp. per 
sq. in. 
Horizontal 
Vertical 
45° 
Type I 
III 
4.1 
I.I4 
2.7 
0.264 
Type II .... 
... IIO 
3-7 
I.I3 
2.6 
0.361 
Type III ... 
... 107.5 
4.2 
1.16 
2.6 
0.392 
Type IV ... 
.. • 105.5 
3-8 
I.I5 
2.5 
0.614 
Type V .... 
... 105.5 
3-7 
I.I5 
2.6 
0.848 
Type VI ... 
... 107.5 
4.2 
1.16 
2.7 
0.920 
installations the relative tendencies to produce ocular discomfort. 
As before, two cases were made of this determination,-one when 
the eye was at rest, the other when it was at work. For a de- 
scription of how the determination is made, and a discussion of 
the method that is used, see the Transactions of the I. E .S., 
1913, VIII, pp. 54-58; and 1915, X, pp. 496-501. Space will be 
taken here only for presentation of the results. These are given 1122 
TRANSACTIONS OE IRRUMINATING ENGINEERING SOCIETY 
in Table X. In this table are given also, for the sake of com- 
parison, results expressing the tendency of the six types of re- 
flectors to cause loss of ability to sustain clear seeing. 
APPENDIX. 
The reflectors used in this work were supplied by the Holo- 
phane Works of the General Electric Co., and are opal glass of 
light, medium, and heavy densities. They are all of the bowl 
type and of the same size, 8 in. Reflector I is a pressed Sudan 
toned brown; reflector II, a blown white glass, toned brown 
(experimental); reflector III, a pressed Sudan; reflector IV, a 
pressed Druid; reflector V, a blown Veluria; and reflector VI, a 
blown white glass (experimental). Reflectors I, III, IV and V 
are commercial products, but II and VI are special, inserted in 
the series to give gradations in density. As was stated in the 
text these reflectors presented considerable unevenness of surface 
brightness. This was especially true of the pressed reflectors, 
which are smooth on the inside and grooved on the outside. The 
glass in these grooves being thinner than in the spaces between, 
a very uneven surface brilliancy is given to the reflector. Further, 
reflector IV, because of its imperfect diffusion, was quite a little 
brighter in the center, at the location of the filament, than at the 
top and bottom. In determining the brightness of these reflectors, 
overlapping readings were taken and an average obtained. 
NOTES. 
1 The truth of this should be obvious to any methodological critic. It is in 
fact the logical corollary of the application of a new test to a new field. Until a 
range of application is made which is reasonably representative of the work for 
which the test is designed to be used, a complete description of the test itself, in- 
cluding a statement of the factors which may influence its results and full directions 
how to use it, cannot possibly be given without more presumption than we care to 
exercise. While an attempt to do this might afford a certain amount of specious satis- 
faction to the practicing engineer, it would be superficial and incomplete and calculated 
to produce trouble in the work of others. When in the opinion of the authors a 
sufficient range of work has been covered, a separate paper will again be devoted to 
the test method itself in which data collected from all the work will be submitted, 
and the adaptability and application of the method to different kinds of work will be 
discussed. It is clear, we think, to anyone who has had experience in developing and 
applying a new test that this can be done more safely and effectively at the close 
of a section of the work which is sufficiently comprehensive to be representative of 
the accomplishment of the test, than at its beginning or while the work is yet in 
progress. In this later paper data will be submitted also on four types of test devised 
by us to detect changes in the functional condition of the retina as the result of 
working under different conditions of lighting. 
Two points keep coming up, however, with a degree of persistence which may EERREE and rand: experiments on the Eye 
1123 
justify a somewhat detailed discussion at this time. The first pertains to the sensi- 
tivity of the test to factors extraneous to the conditions that are being tested. The 
point was briefly discussed in the original paper on the test and again in the two suc- 
ceeding papers. It was brought out more especially in Mr. Cravath's paper and in 
the discussions following it. Among other things it was shown in this paper that by 
purposely varying these factors in some extreme way they could be made to influence 
the results of the test. The more crucial point was not shown however; namely, 
that they operate against the usefulness of the test when the work is done under the 
conditions that ordinarily obtain in a well conducted experiment; nor does the paper 
contain any evidence that Mr. Cravath thinks this is the case. In our own work a 
different plan has been pursued with regard to this point. Instead of trying to find 
out what effects could be produced by means of procedures that would not be per- 
mitted in making a test, every care has been taken from the beginning to eliminate or 
hold as constant as possible all extraneous factors which might influence the general 
and muscular efficiency of the eye, and to check *up the effectiveness of this control 
by carefully determining the mean variation in the results for each set of lighting 
conditions. This we have considered to be the most direct and feasible plan of con- 
ducting the work. In any event, it is obvious that there is no need of futile spec- 
ulation concerning the possibilities of influence of these factors, nor of any in- 
definiteness either in the discussion or investigation of the point, so long as the 
actual value of the influence can be measured by determining the mean variation and 
its relative value be estimated by comparing the mean variation with the variations 
produced by changing the conditions to be tested. That is, a measure of the absolute 
and relative value of these factors is readily available and this measure has been 
carefully used at every step in the work. We need scarcely to point out that it is a 
well recognized principle of experimentation in comparative work such as we are 
doing that as long as the mean variation is safely within the experimental variation, 
the method is considered satisfactory for the purpose for which it is being used. 
In this connection it may not be out of place to give here a more detailed account 
than has yet been given of the method that has been used in selecting and training 
observers. Care is exercised in the first place to choose one who has shown a satis- 
factory degree of precision in threshold and equality judgments in other optical work 
in the laboratory, and whose clinical record shows no uncorrected defects of conse- 
quence The observer is then practised on the three minute record under a lightin 
condition selected and maintained for the purpose, until a satisfactory degree of 
reproducibility is shown. These records are usually run in series of five with a twenty 
minute rest interval between each record. So far we have not published the results 
of an observer who has not been able to attain a reproducibility in the time seen 
clear of I per cent, for a series of five records in these preliminary experiments, al- 
though this degree of precision is unnecessary unless the observer is being trained for 
work in which there are very small differences in the conditions to be tested. Since 
these records are run with no change in the lighting conditions and with rest intervals 
to prevent general or optical fatigue, they serve primarily as a training in making the 
judgment and as a check on the precision of the judgment. In the second stage of 
preparation the observer makes a number of three hour tests with records before and 
after work for two or more lighting installations, and the mean variation of the 
results from the average is determined. Again, if a sufficiently small mean variation 
is not shown where there has been no change in the lighting conditions, the observer 
is not allowed to take part in the actual work of testing. This last mean variation is 
the final preliminary check upon all the factors that may vary under the control im- 
posed,-lack of reproducibility in the judgment, variable physical and mental fatigue, 
etc. The final check is had in the course of the work itself. That is, a number of 
tests are made for each lighting condition of the series to be investigated, and the 
mean variation is determined for each and compared with the variations that are 
produced by the changes in the conditions to be tested, to find out to what extent 
these variations may be ascribed to the changes made and to what extent to the 
normal variation of the test. How much larger is the variation which is produced by 
changing the lighting conditions than is the normal variation for each condition may 1124 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
be seen by comparing Columns 14 and 15 of Table IX. In the work of the preceding 
papers the excess of the experimental variation over the mean variation was much 
greater still as might be expected from the greater differences that were present in 
the lighting conditions tested. For example, in five three hour tests for the indirect 
system for Position I (see this Transactions 1915, X, pp. 413-426) the mean variation 
in per cent, was 1.1; for the semi-indirect system it was 1.4; and for the direct system, 
1.2; while the percentage drop in the ratio from beginning to close of work for these 
systems was respectively, 8.5, 72.3, and 80.9. Similar citations may be made for the 
other conditions tested. When one compares in these cases the mean variation with 
the magnitude of change of ratio produced by changing the lighting system, it be- 
comes obvious how unnecessary has been the concern about the influence of extraneous 
factors in case of the work that has as yet been done. In fact, the mean variation 
has been so safely within the experimental variation that the writers have not felt 
it necessary heretofore to make the numerical comparison in print. It is so well 
recognized as an experimental principle that the experimental variation shall safely 
exceed the mean variation that it has been their custom to give the comparison only 
when there exists some grounds for doubt. Heretofore, we have, as a general case, 
been working with conditions that produced a large difference in results. 
As bearing on another phase of the question of reproducibility, namely, where 
a long interval has elapsed between two series of tests, we may cite one example 
where two series were taken under the same lighting conditions a year apart, and 
the variation in the average per cent, loss of efficiency was only 0.3. In this case a 
favorable lighting system was used, the initial ratios were closely the same, and the 
control in general good, although no especial care was taken to make it so more than 
what is ordinarily exercised. It is not presented, however, as a typical instance. It 
happens to be the only case of which we have a record, where a long interval has 
elapsed between two tests. 
Moreover, there is nothing in the nature of the test other than its superior sen- 
sitivity that should make it more susceptible to the influence of extraneous factors 
than any other test of acuity. The principle of the test will be remembered from the 
earlier papers. It is merely the conventional acuity test subjected to certain features 
of standardization for the sake of greater reproducibility, and made into an endurance 
test to give it additional sensitivity. The older test had not been found to be suf- 
ficiently sensitive to fatigue conditions to warrant adoption in our work. This test 
is in fact not meant to be a fatigue test. It was designed to test the dioptric condition 
of the eye, and may be used with more or less success as a test of how far a given 
lighting condition is conducive to clear seeing with a maximum of momentary effort-, 
but it has not the essentials of a fatiglue test nor of its converse, the ease with which 
clearness of seeing is attained,-which is what is needed primarily for the selection 
of lighting conditions for the greater part of the work that we are ordinarily called 
upon o do. Almost if not quite as good resultsl for example, may be gotten with it 
after work as before, when there is every other reason to believe the eye has suf- 
fered considerable depression in functional power. The reason for this is obvious. 
Although greatly fatigued, the eye can, under the spur of the test, be whipped up to 
give almost if not quite as good results as the non-fatigued organ when only a 
momentary effort is required. (See Column 8, Table IX, and former papers.) If 
fatigued, however, it can not be expected to sustain this extra effort for a period of 
time. The demonstration of this fact led early in our work to the introduction of 
the time element into the test. The principle involved is not a new one. It is 
merely the application of a very old and well known one to the work of testing for 
optical fatigue. If, for example, a sensitive test is wanted for the detection of 
fatigue in a muscle, as good results can not be expected if the test requires only a 
momentary effort on the part of the muscle as would be attained if the endurance of 
the muscle were taken into account. For our purpose, therefore, the old acuity test 
has been made into an endurance test, in which the fatigue or loss of functional 
efficiency of the eye is measured by its power to sustain clear seeing for a period 
of time. As such it should and does show a sensitivity for detecting fatigue far be- 
yond what can be attained by the older and more established test when it is used Ferree and rand: EXPERIMENTS on the eye 
1125 
for that purpose. And being a test which is more sensitive to functional changes in 
the eye, it doubtless does show in some proportion to its greater sensitivity more 
effect of the indirect as well as of the direct factors that influence acuity; but since 
the indirect factors can be subjected to control, while the direct factors are varied, 
there is in proportion to the sensitivity of the test and the control exercised a gain for 
the purpose for which the test is used. That this gain is great is shown in all our 
work by a comparison of the size of the mean variation with that of the variation 
produced by the change in the conditions to be tested. 
The second point we wish to discuss here refers to the part played in our ex- 
periments by a factor known among psychophysicists as the error of expectation. The 
belief that there is a need to take account of this error in sense judgments arises 
from the difficulty in keeping the observer in ignorance of the test material and of 
a certain amount of the experimental procedure. In our experiments there are just 
two points on which the observer has knowledge: namely, the test object and the 
lighting conditions or system under which the work is done. All the rest is kept 
concealed from him unless the experimenter should in turn serve as observer in which 
case his results are checked up by those of observers who have not served as experi- 
menter. We will consider this factor first in relation to the test object. The ob- 
server knows what the test object is (the letters li in 8 point type) and is told to 
record, for example, when the dot is seen separate from the vertical line in the 
letter i. The question at issue then is whether proper account is taken in our ex- 
perimental procedure of the influence of expectation on this judgment. The question 
can be discussed the most comprehensively perhaps by first considering rather broadly 
the status and development of experimental method with regard to this factor. As 
we have already intimated, the probable influence of expectation is an inherent diffi- 
culty in all sense judgments,-photometric, acuity, threshold, etc. That it can not be 
entirely eliminated is, we think, generally conceded as axiomatic. Psychophysicists 
have, therefore, turned their attention to attempts to compensate for it, and a need has 
been felt to do this in most cases only when the work requires that the determination 
be made with a great deal of precision. Different methods may be employed for this 
purpose all of which are more or less open to question. The one most frequently 
used perhaps, is the method of ascending and descending series. From a consideration 
of this method an idea may be had in a general way of all the methods of its class. 
Rather than to eliminate or even to lessen the operation of the factor, the purpose of 
this method is to control its direction and to plan the experiment in opposing series, so 
as to compensate for its influence in the final result. That is, in making a threshold 
determination, for example, the series in one case is begun below the threshold, 
and the observer is told that the stimulus will be increased until the threshold is 
reached; in the other case the procedure is reversed. For the final result an average 
is taken of the values so determined on the assumption that expectation in the two 
cases will influence the determination by equal amounts in opposite directions. Much 
has been said in the literature of psychophysics with regard to whether this method 
accomplishes what it is intended to accomplish, and more might be said; but it is 
immaterial for our purpose whether it does or not, for it is obvious that it could 
not be applied to our 3 minute records, for here the image to be judged rises to the 
threshold of discrimination independently of the control of either the observer or the 
experimenter. The individual judgments, therefore, could not be arranged in op- 
posing series for the purpose of compensation. An entirely different type of method 
is to use an objective check on the judgment of the observer, and by this means en- 
deavor to weed out from the results the influence of subjective factors. We tried 
for several months to devise a means of changing the stimulus in such a way that 
an objective check could be had on the registration of the observer without sac- 
rificing the principle of the test. Such a change, however, could not be made in 
the test object which did not at the same time permit the eye to relax its strain at 
the instant of change, which it is obvious destroys the very feature which gives the 
test its superior sensitivity. The attempt to get an objective check, however, was 
made more for the sake of offsetting possible criticism than it was because of any 
belief that it was necessary for the purpose for which the test has so far been used; 1126 
TRANSACTIONS OF ireuminating engineering society 
for, as we have already stated, a determination of the mean variation for the 3 minute 
record, each one of which consists of a number of separate judgments, had shown us 
that the influence of expectation as a source of variable error is of negligible con- 
sequence. That is, the mean variation is the measure of the aggregate effect of all 
the variable factors including expectation, if indeed it be a source of error in the 
case under consideration, and it was found to be too small as compared with the 
variations produced by the changes in the conditions tested to be the cause for any 
concern for the purpose of the work. Moreover, it will be remembered that a knowl- 
edge of the test object is given to the observer as one of three changes that were 
made in the conventional acuity test to minimize very obvious sources of variable 
error, among waich were memory and expectation, and to give a greater reproduci- 
bility to the judgment. We can do no better probably than quote from the original 
discussion. "Visual acuity tests of the Snellen type, especially when used in work 
in which it is required to make successive tests on the same person, are open to the 
following objections, (a) The judgment is in terms of recognition. A letter may be 
recognized when it is not seen clearly. In any judgment based on the recognition of 
even a single letter, memory plays an important role. It is, so far as the writer 
knows, impossible to standardize this memory factor and to obtain results strictly in 
terms of acuteness of vision, (b) The test card is made up of quite a long series 
of letters. As the fest progresses the letters are memorized more and more completely. 
It is practically impossible to eliminate this progressive error when a number of 
successive judgments have to be made as is the case before a final result is reached 
in any single visual acuity test and as is especially the case when a number of suc- 
cessive tests have to be given to the same person, which happens in much of the 
work involved in the solution of the problem here proposed (c) The 
Snellen series contains quite a large number of letters. The eye is found to fatigue 
and vision to blur before the series ,is completed. This introduces an error which it 
is practically impossible to render constant." All of the above errors were elimin- 
ated, or at least minimized, in the tests finally adopted by us by changing the type 
of judgment and by adopting a simple test object, made up of only two characters, 
the letters li in 8 pt. type. In this test the observer's acuity of vision is determined 
by the distance at which he can just clearly distinguish the two test objects. In 
practise it has come to be a matter of distinguishing whether or not the dot is 
separated from the vertical line in the image of the letter i. The results are thus 
rendered directly in terms of acuity of vision and the progressive errors due to 
memory and expectation are minimized. In this regard the significance of the change 
in the type of judgment from recognition to the judgment of the separateness of two 
simple objects, e. g., the dot and the line in the letter i, should not be overlooked. 
When the criterion is recognition and the task set for the observer is merely to 
identify the test object with its name or some memory of it from past experience, as 
is the case in the old form of the test, memory and expectation play their maximum 
role. Any extraneous clue or a partial discrimination of the object may in fact serve 
as a basis for all that is required in the judgment. When, however, the task set 
for the observer is a different one and he is required to judge the presence or ab- 
sence of a space between the dot and line in the letter i, the role of these factors 
is reduced to a minimum, and the task is narrowed down to the judgment of a space 
threshold, one of the simplest and most reproducible types of sense judgment. In 
short then, a knowledge of the test object is given to the observer as a part of the 
modification of the conventional acuity test to minimize the effect of variable factors, 
among which memory and expectation play the chief role. And that it has accom- 
plished its purpose is abundantly attested by a comparison of the size of the mean 
variation given by the test so revised as compared with that given by the older form. 
We may add that the letter 1 is used in connection with the letter i for two reasons. 
(1) A steadier fixation is given than can be attained by so small an object as the 
letter i; and (2) a standard is afforded (an unbroken vertical line) in terms of which 
to judge the separateness of the dot from the vertical line -in the letter i. 
The only other way in which expectation can come into the experiment through 
knowledge on the part of the observer is, as we have already stated, through an Ferree and rand: experiments on the eye 
1127 
awareness of the conditions or lighting system tested. The observer can not work 
for three hours under a given lighting installation without being more or less aware 
that the same installation is being used as was used before, or a different one. More- 
over, we do not see how this unfortunate factor can be completely eliminated unless 
imbeciles be used for observers. We wish to point out, however, that there is no 
greater liability to harmful influences from this factor in our test than in the older 
acuity test or any other that could be applied to the same type of work. We grant 
that, in any test that could be used, if observers of strong commercial or other bias 
should in two isolated trials get better results for one type of lighting than another, 
there might be grounds for suspecting that prejudiced observations were made: but 
if each condition were tested a number of times, as has been the case in all of our 
work, and a small mean variation were obtained for each series of tests, the result 
would look much more like the response of an organism to a constant set of condi- 
tions in obedience to physiological law than it would like a voluntary reproduction 
guided by prejudice, however strong and constant that prejudice might be. Here 
again the size of the mean variation is the check upon the validity of the results, 
for it is obvious, we think, even to a novice, that records taken at intervals of from 
one to five days could not show a close reproduction if the fidelity of the registration 
were in any way interfered with by the wishes or prejudice of the observer. Further- 
more, it is only fair to say that it would be difficult to find a group of observers 
freer from a direct interest in lighting conditions or a knowledge of their significance 
than is the group from which the greater number of our observers are selected. 
2 These factors are the evenness of illumination, the evenness of surface bright- 
ness, the diffuseness of light, the angle at which the light falls on the work, intensity, 
and quality. 
3 The problem of installing is probably not the same for the inverted translucent 
as for the inverted opaque reflectors. In the latter case the height should be so ad- 
justed as to give as nearly as possible an even distribution of surface brightness on the 
ceiling, and evenness of illumination on the working plane. In case the inverted 
translucent reflectors, however, if the distance from the ceiling is made great enough 
in all cases to produce these effects, it may throw the bright reflectors too low in 
the field of vision for the highest efficiency and the greatest comfort to the eye. In 
this regard the opaque reflectors have. the advantage that it is always easier with 
them to get the brightest surface in the room out of the zone of most harmful in- 
fluence in the field of vision. In later work we expect to conduct a series of ex- 
periments with the above reflectors in which the height from the ceiling is the 
factor varied. 
4 The track along which the test card was moved was parallel to the east and 
west walls of the room. When taking the test the observer faced the north wall in 
such a position that when the eyes were in the primary position the lines of regard 
were parallel with the east and west walls of the room, and approximately normal 
to the north and south walls. That is, the head was erect and held in such a position 
that the objects in the room, reflectors, etc., fell as symmetrically as was possible 
within the field of view. During the three hours of reading which intervened be- 
tween the two three minute records, the observer moved just far enough back from 
the upright supporting the mouth board to give room for the book to be held, and 
to permit of a comfortable reading position. Care was taken to have the eyes sustain 
as nearly as was possible the same general relations to the objects of the room 
as were sustained when the three minute records were taken. This could be 
done either by holding the head erect, etc., or by tilting slightly backward in the 
swivel chair used by the observer and allowing the head to relax a compensating 
amount. So far as the direct optical effects are concerned, it would seem to be 
immaterial which of these positions is chosen, so long as approximately the same field 
of vision is obtained. The latter is usually preferred by the observer as causing less 
general fatigue. When taking this position, the book is elevated and held at ap- 1128 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
proximately an angle of 45° (a little nearer to the vertical than this perhaps). The 
brightness measurements of the book at this angle and in the horizontal are not 
taken, however, so much because of this as to give the brightness of the book in 
two fixed representative positions at the point of work. Care is taken to have print 
of uniform size and distinctness for use with the three systems and to have a page 
which gives a comparatively small amount of specular reflection. Uniformity in these 
regards can usually be secured by using numbers of the same journal. 
5 It should not be needful to mention that the recording apparatus is screened 
from the observer's view while the test is being made. Before photographing, the 
screen was removed and the apparatus regrouped. 
6 This is the test station shown in Fig. 1, and of the four used in the former 
work is the one nearest to the south wall of the room. 
7 As has been stated in our former papers, in the consideration of the effect of 
a given lighting situation on the ability of the eye to hold its efficiency for a period 
of work, the age of the observer and the condition of his eyes should be taken into 
account. All of the observers who have been employed by us in this work have been 
under 28 years of age. Following is the clinic report of the eyes of the observer 
whose results are given in Tables IX and X, made by Dr. Wm, Campbell Posey of 
Philadelphia. 
Observer R. 
With glasses.-Vision of right eye = 20/25. Far muscle test = Q % esophoria. 
Vision of left eye = 20/20. Near muscle test = orthophoria. 
Ophthalmoscopic examination.-Right eye = mixed astigmatism, % diopter. 
Deft eye = hyperopic astigmatism, 1% diopters. 
External condition.-Adduction good; eyes slightly divergent under cover; 
cornea clear; pupils, 2)4 mm.; irides respond equally and freely 
to light, accommodation, and convergence stimuli. 
Glasses worn during test.-Right eye = -S., 0.50 D.; -C., 0.37D., x 1600 
Deft eye = -C., 0.50 D., x 1800 
Early in our work the problem arose whether the three minute records before and 
after work should be taken in the same room in which the work was done or in 
a separate room reserved solely for that purpose. To test this point, work was 
done in both ways. It was found that the effects of smaller differences in lighting 
conditions could be detected when both the three minute records and the work were 
done under the lighting conditions to be tested. That is, the total test procedure, 
which includes both the three minute records and the reading, is more sensitive when 
it is all done under the conditions to be tested, than when a part of it is done under 
these conditions and a part in a separate room. Since the method is more sensi- 
tive when the whole procedure is conducted under the lighting conditions to be 
tested, we can see no reason why even the purist should demand that a part of it be 
done under the conditions to be tested and a part somewhere else, so long as the 
results are recognized to be the consequence both of the three minute records and of 
the reading. Our purpose, it will be remembered, has been to get a sensitive means 
of detecting the relative tendencies of different lighting conditions to cause a loss 
in the power of the eye to sustain its ability to see clearly; and the method is more 
sensitive when the three minute records, also, are made under the conditions to be 
tested. This, we may say, is our chief reason for the practice. A justification, we 
believe, is not logically needed. Moreover, the method so conducted is just as 
amenable to control and to checks upon its reproducibility, as if it were used in the 
less sensitive form. It is, in fact, considerably more amenable to control, for if a 
separate room were used for the three minute records, very great care would have 
to be exercised to see that it was always illuminated with exactly the same intensity 
of light that was used in the room in which the reading was done. If the illumina- 
tion were not accurately the same, a period of adaptation would have to be allowed 
before the three minute record could be made, which, in case of the record taken 
after work, would give the eye opportunity to recover from the fatigue induced by FERREE AND RANDI EXPERIMENTS ON THE EYE 
1129 
the work. It is obvious that a great deal of difficulty would be encountered in ac- 
curately maintaining this control; and, if it were not so maintained, an error of 
considerable consequence would be introduced into the work. In getting control not 
only the illumination of the test object must be taken into account, but the bright- 
ness of the whole field of vision with its complex distribution of light and shade, for 
this conditions the state of adaptation of the paracentral and peripheral portions of 
the retina which in turn exerts an influence on the part of the retina that receives 
the image of the test object. It may be added also that adaptation effects in the 
paracentral and peripheral portions of the retina are stronger and more rapid than in 
the central portions. 
In connection with the fact that the three minute records add sensitivity to the 
method when they are also taken under the conditions to be tested, we may say 
that we are now working on a short method in which three minute records with 
proper rest intervals are used. This test is rougher and less sensitive than the 
longer method, but if it can be made satisfactory, it might be more adaptable to 
practical work. 
8 It will be noted in this table that there is very little variation in the value of 
the initial ratios. We noted in each of our preceding papers and again in our dis- 
cussion of Mr. Cravath's paper that the sensitivity of the test varies with the ratio of 
the working distance of the test object from the observer to the acuity distance. 
After considerable investigation of the point, we adopted, as a standard to be at- 
tained approximately, a ratio of distances that would give for the initial record a 
ratio of time clear to time blurred of 3.5. As might be expected, it is impossible to 
get this ratio of 3.5 exactly from any single ratio of working distance to acuity 
distance that can be determined in advance of the actual record. But with care a 
close approximation may be attained, and since the loss of efficiency is judged from 
the amount this ratio is changed from the beginning to the close of work, and not 
from the ratio itself, the failure to obtain it does not affect a comparison of the 
favorableness of different lighting conditions for the eye, any more than is represented 
by its effect on the sensitivity of the test. In short, the variations in this ratio from 
test to test form merely one of the group of variable factors, the check upon the effect 
of which on the results of the test, is the size of the mean variation; and, so long 
as this mean variation is safely within the amount of variation produced by changing 
the conditions to be tested, the control may be considered as satisfactory for the 
purpose of the work that is being done. That is, when this check is properly exer- 
cised, the influence of a variation in this ratio can not possibly be mistaken for the 
effect of the condition which is being tested. However, in the course of the deter- 
mination of what value of initial ratio should be used, considerable study was made 
of the effect of varying the ratio. While space will not permit us to quote largely 
from these results here, still an idea may be given in the space at our command, of 
the order of magnitude of the effect that is produced. That is,, we will take three 
cases including a range of differences amply great to cover what is ever apt to occui 
in actual work. In the first, the initial ratios were 2. and 3. (difference, 1); in the, 
second, 2.67 and 5 (difference, 2.33); and in the third, 1.93 and 7.57 (difference, 
5.64). Tne difference in the percentage loss of efficiency for the first case was 1.4; 
in the second, 2; and in the third, 1.7. The effect shown in these cases, it will be 
observed, is about of the same order as the normal mean variation of th® test. 
8 In order to make a fair comparison between the drop in ratio time clear to time 
blurred caused by working under a given lighting condition and the mean variation 
of the drop, the percentage drop and the percentage mean variation are both esti- 
mated in the above table, also in the citation made in Note 8, p. 1129, on the same 
base, 3.5. If this comparison had not been wanted especially to show that the mean 
variation produced by changing the type of reflector, it would have been more in 
accord with custom perhaps to have expressed the mean variation as a percent, of 
the mean value of the drop. Computed in this way the value of the mean variation 
for Reflectors I-VI would have been in order 5.6 per cent., 1.6 per cent., 1.3 per 
cent., 3.3 per cent., 1.2 per cent., 1.4 per cent.; and for the citation in Note 8, they 1130 
TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
would have been for the indirect system 4.5 per cent., for the semi-indirect system 
0.7 per cent., and for the direct system 0.5 per cent. This method of estimating 
the mean variation gives, it will be noted, the largest per cent, variation for the most 
favorable lighting condition because the drop in ratio, the base on which the per- 
centage is estimated, is the smallest for this condition. The actual variation we have 
found as a rule is, as might be expected, the least for the most favorable condition. [Reprinted from the Transactions of the Illuminating 
Engineering Society, No. 9, 1915.] 
DISCUSSION. 
Mr. J. R. Cravath : Dr. Ferree calls his test a test "loss of 
efficiency of the eye." I think the term "eye-fatigue" is much 
briefer and more expressive. The work reported in a previous 
paper of Dr. Ferree covered conditions rather widely varied. 
The paper we have before us now covers conditions which 
come within fairly narrow limits of visible source brightness. 
The results have been especially interesting to me as a member 
of the Committee on Glare because we have, during the past year, 
attempted to formulate or to express certain limits of good prac- 
tise which are least conducive to glare. In ordinary interior 
illumination, we state in our report, which is soon to be published, 
that contrasts of brightness of adjacent surfaces (I mean by 
adjacent surfaces, those which are adjacent within the visual 
field) should not be over a ratio of one to two hundred, and 
preferably not over one to one hundred. That ratio was taken 
as the result of an examination of a good deal of data, some of 
Fig. i. 
them in previous papers of Dr. Ferree. It was therefore of con- 
siderable interest to me to see how the results in the present paper 1131 
EXPERIMENTS ON THE EYE 
conform with these limits and in order to do that, I have taken 
the ratio between the brightest spot, which of course, is the re- 
flector, and the point N on the room photograph a little to the 
right of the reflector, and plotted a curve, Fig. I, corresponding to 
Chart II but, instead of using the brightness of the reflector, I 
have used the ratio of the brightness of the reflector to the darkest 
point along side of it; that is the point N. The ratios of A to N 
under 200 seem to give notably less fatigue than those above 
1 to 200, which would confirm the judgment of our committee 
as well as the previous results obtained from various sources. 
Dr. J. W. Schereschewsky : I want to congratulate Dr. 
Ferree on the extreme care which is evident in all these series of 
tests to obtain small mean deviations and secure reproducibility of 
results. I wish to call attention, however, to one factor, which 
in all probability, will have considerable effect on all tests which 
involve muscular action and that is the question of weather con- 
ditions. It is plain to all who have read this paper that tests of 
this kind such as Dr. Ferree has here published are labors requir- 
ing a great deal of care in arranging and carrying out, and it 
seems to me that we ought, by all means, in working tests of this 
character, to insure, as much as possible that the work shall not 
be thrown away because of great variations in the results due to 
extraneous factors. Now it seems to me that tests of this kind 
ought never to be conducted in very hot weather. The effect of 
high degrees of heat and humidity is to reduce the endurance of 
muscle. That seems to be plain from Prof. F. Lee's experiments 
which were done in the investigations of the New York Ventila- 
tion Commission, in which it was found that sections of muscles 
removed from animals which had been subjected to high degrees 
of temperature and humidity furnished on the average 40 to 50 
per cent, less contractions than muscles of animals which had 
not been so exposed. Therefore, it seems to me that persons who 
conduct tests of this kind in very hot weather would find a great 
loss of efficiency of the eye simply from exposure to weather con- 
ditions ; so in the future, when we are endeavoring to corroborate 
the results of these tests by similar tests, we must take the pre- 
caution never to undertake such tests except when the atmos- 
pheric conditions are distinctly comfortable. TRANSACTIONS OF IEEUMINATING ENGINEERING SOCIETY 
1132 
Mr. T. W. Roeph : The data given in this paper are very 
valuable to us who have lighting systems to design. 
I should like to call attention to the ratio between the brilliancy 
of the ceiling and brilliancy of the reflectors as obtained in these 
measurements. Reflectors 4 and 5 are typical reflectors of the 
class which is most commonly used to-day for semi-indirect light- 
ing, and in this particular installation, which does not represent 
all installations, but is possibly a typical installation, the ratio 
between the brilliancy of the reflector and the brilliancy of the 
ceiling is one to fifty, and with reflector 5, one to sixty-one, 
taking the brightest point on the ceiling. Those are the re- 
flectors which show the greatest eye fatigue, and this shows 
us how far we will still have to go in reducing intrinsic brilliancy 
in order to get semi-indirect lighting systems which are correct 
from the engineering standpoint. Now, reflector No. 3 is a 
commercial reflector which is being used to a certain extent for 
semi-indirect lighting. It is denser glass and it is harder to sell 
than reflectors 4 and 5, so that we who are working for better 
lighting in the commercial field have that to contend with. Re- 
flector 3 shows a ratio between the brilliancy of the ceiling and 
the brilliancy of the reflector of one to sixteen. The densest re- 
flector tested has a ratio of one to nine. Three years ago, in a 
paper on the "Engineering Principles of Semi-Indirect Lighting," 
I argued that, from the engineering standpoint, the brilliancy of 
the reflector in an installation should be approximately the same 
as the brilliancy of the ceiling. That was not particularly from 
the standpoint of eye protection, but from the standpoint of ob- 
taining the maximum diffusion of illumination, arguing that if 
we are going to sacrifice the efficiency of direct lighting by in- 
stalling semi-indirect systems, we should try to get the maximum 
engineering value of the semi-indirect systems, by obtaining max- 
imum diffusion, and that this would be obtained when the 
brilliancy of the bowl is approximately the same as the brilliancy 
of the ceiling. This paper indicates that to obtain good eye- 
protection in semi-indirect lighting, we should work to very much 
denser glassware or very much lower brilliancy of reflector bowl 
than is generally practised to-day. Even a ratio of one to nine, 
where the reflector is nine times as bright as the ceiling, shows a 
considerable degree of eye fatigue. 1133 
EXPERIMENTS ON THE EYE 
There is one point I should like to bring out in connection with 
the measurement of the intrinsic brilliancy of these reflectors, 
and that is merely a suggestion that possibly a good way to obtain 
the intrinsic brilliance of a reflector of this character would be to 
take the candlepower as determined on the photometer, and the 
area of the reflector as determined on a drawing board, and thus 
find the candlepower per square inch, rather than to take lumin- 
osity measurements of the reflector at various points and average 
them. I believe it would be more accurate to take the candle- 
power of the reflector and simply divide it by the projected area 
in the direction of view. 
Mr. J. R. Cravath : This question of what shall be taken as 
the criterion of brightness is something that Dr. Ferree evidently 
is not sure of, and I don't think any of the rest of us are-I mean, 
what particular contrast shall be taken. Mr. Rolph has just 
mentioned the contrast of the brightness of the reflector with the 
brightness of the ceiling above. Dr. Ferree has given us results 
showing the highest brilliancy, that is, the brilliancy of the re- 
flector, and the average brilliancy, and the ratio of the highest to 
the lowest and the ratio of the highest to the average, the ratio of 
the average to the lowest and the ratio of the highest to the bril- 
liancy of the point of work. There is such a great deal to be said 
in favor of his objection that, possibly, in a case where the subject 
is working continuously on desk work or reading that the ratio of 
the brightest object in view to the work, that is, to the paper on 
which the eye is working, should be the criterion; because in that 
case, the brightest objects in view appear simply on the edge of 
the retina most of the time, while the paper is on the center; but 
for most practical purposes, I think perhaps the criterion adopted 
by the Committee on Glare of the ratio between the nearest adja- 
cent surfaces would answer all practical purposes for the present. 
I also want to express the debt that I feel the practical men of the 
society owe to the investigators who bring out this kind of data; 
it is exactly what we need to make progress in our work. 
Dr. C. E. Ferrer (In reply) : I suggested in a former paper 
that, theoretically considered, better results should be gotten with 
the semi-indirect reflector of such a density as to give a surface 
brilliancy equal to that of the ceiling spot than are obtained with TRANSACTIONS OF IREUMINATlNG ENGINEERING SOCIETY 
1134 
the totally indirect reflector. That is, if the reflector is made of 
the same brightness as the ceiling spot, the same light flux can be 
obtained with a lower intrinsic brilliancy of the brightest surface 
than if the light all comes from the ceiling spot because of the 
increase of luminous area. This is in agreement, I believe, with 
the general tenor of Mr. Rolph's discussion. Unfortunately, 
however, I have not yet been able to obtain reflectors of sufficient 
density to test the point directly. However, in the work that we 
have done, an increase in the density of the reflector, so far as 
we have been able to carry the increase with the reflectors sup- 
plied us for the purpose, has been accompanied by a consistent 
improvement in the effect on the eye. There is one thing to be 
claimed, however, in favor of the indirect reflector when all is 
said and done. It is easier with it to remove the brightest spot in 
the field of vision from the zone of harmful influence to the eye, 
especially in rooms of the height ordinarily found in dwelling 
houses, because with this type of reflector the brightest spot is 
always on the ceiling. With reference to the effect of position or 
rather height of the brightest spot in the field of vision, it may not 
be out of place to anticipate here in slight measure the content of 
a future paper. In the work of the present paper the reflectors 
were installed 30 inches from the ceiling. This is in accord with 
general practise for the installation of totally indirect reflectors 
in rooms of the height of our test room and is considered to give 
a favorable distribution of light and shade on the ceiling and a 
comparatively even distribution of light on the working plane. 
So installed, however, the brightest spot (the reflector) is dropped 
well into the field of view, especially at the outlets most removed 
from the observer. The question arises, therefore, whether semi- 
indirect reflectors should be installed according to the principles 
of indirect lighting, direct lighting, or whether some compromise 
should be made between the two. We have begun, therefore, 
a series of tests in which the distance of the reflector from the 
ceiling is varied. So far we have been able to finish the compari- 
son for the reflectors of least and greatest density at distances of 
30 in. and 15 in. from the ceiling. The 15-in. distance gave 
quite considerable improvement in the effect on the eye for the 
reflector of least density, but not nearly so much for the reflector 
of greatest density. This result suggests that a more careful 1135 
EXPERIMENTS ON THE EYE 
study should be made of the method of installing semi-indirect re- 
flectors differing in density. It would seem that the denser they 
are the more nearly they can afford to be installed as indirect re- 
flectors and the less dense they are the more nearly they should be 
installed as direct reflectors so far as eye effects of the kind re- 
vealed by our tests are concerned. 
I have no doubt that Dr. Schereschewsky is right about the 
probable effect of excessive temperature on the results of tests 
such as ours. I am very frank to confess, however, that I never 
do anything on a hot day if I can help it; and I certainly would 
not conduct a test when the temperature is excessively high. 
Through the greater part of the year the temperature of our test 
room is kept within a small variation by thermostat control. If 
it is necessary to work on warm days electric fans are used; but 
on no account are tests ever made on hot, humid days. In fact 
nearly as much care has been taken, I should say, to secure uni- 
formity in temperature control in our work as has been taken to 
secure a uniform control of illumination and brightness effects. 
I am confident, therefore, that our results so far have not suf- 
fered from temperature as a variable factor. If I may digress 
here for a moment, I should like to say, with no reference what- 
ever to Dr. Schereschewsky, whose discussions I have always 
found to be most considerate and intelligently liberal in tone, that 
I am becoming somewhat tired of the subject of extraneous 
factors. To speculate about their probable influence may be of 
considerable cultural value to those who have heretofore thought 
little about the subject, but there is no need to worry about their 
influence or to stand in the way of reasonable progress when a 
gauge on the amount of their influence may be and has been had 
at every step in the work. In this latter connection I refer to a care- 
ful determination of the mean or average variation. If this is done 
as it has been done at every step in the training of the observer; 
if moreover it is done for each condition tested and a comparison 
made of its amount with the amount of variation produced by 
changing the condition tested; exact knowledge is had in every 
case whether or not the results obtained are significant. The 
subject of gauging the influence of variable factors is too old and 
has been too carefully worked out to justify the raising in any TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
1136 
scientific body of as much elementary discussion as has been 
raised with regard to it in this Society. The procedure in general 
is very simple and straightforward. Train the observer on every 
feature of the test method with careful attention to the size of 
the mean variation. In the actual work determine the mean varia- 
tion for each set of conditions tested and compare it with the 
variation produced by changing the conditions to be tested. If 
its sum for any two sets of conditions is not less than the dif- 
ference between the average results obtained for the two con- 
ditions these results, it is usually considered, can not be claimed 
as significant. I have spent months, for example, in the training 
of an observer on the different features of the test method, only 
to discard him at the end of that time because a sufficient degree 
of precision of record could not be obtained under a constant set 
of lighting conditions. Those who have shown in such a course 
of training an unsatisfactory degree of precision usually reveal on 
examination, I may say, some uncorrected optical defect. Muscle 
imbalance more often than any other seems to have been the 
defect in the cases which have so far given us trouble. This 
may mean that the extrinsic more often than the intrinsic muscles 
are the cause of a variable performance on the part of the eye 
in tasks such as we have set for it in our tests, but it is also 
probable that the occurrence is due to a considerable extent to the 
fact that in ophthalmological practise small defects in muscle 
balance are more often left uncorrected than are, for example, 
refraction defects. 
In concluding my comments on this point I think I may be 
justified in mentioning that I have spent a greater number of 
years than I like to recall in trying to get control of the variable 
factors that influence the response of the eye; and that I have 
added considerably to the precision of its performance under 
experimental conditions, I can only call upon my published work 
to testify. It is not likely, therefore, that in the course of de- 
veloping a new test I would show such a degree of incaution with 
regard to the most elementary and well-known principles of ex- 
perimentation as was made the subject of serious and somewhat 
pretentious inquiry in the discussions of the paper preceding the 
present one, and in the discussions aroused by Mr. Cravath's 
paper. 1137 
EXPERIMENTS ON THE EYE 
I am glad Mr. Cravath has given us still another way of plotting 
the results of the tests against brightness effects. It has not 
occurred to me, however, to attach any especial importance as a 
separate factor to the ratio of the brilliancy of the brightest area 
to that of its immediately contiguous surroundings. There are, 
for example, only two possible effects that I could conceive to be 
due to this relation, neither one of which would seem to me to 
warrant making of it a separate factor, (i) It would enhance 
by physiological induction in some proportion to the difference 
in physical brightness, the brightness of the sensation aroused by 
the reflector and thus increase its power to set up muscular strain 
by distracting the eye from the adjustment needed for the work 
in hand. So considered, however, its action would merely be 
that of an auxiliary factor, supplementary to the actual bright- 
ness of the reflector. As such it is of course of a great deal of 
importance, greater perhaps, for example, that the relation of 
lightest to darkest surface, as brightnesses are graded in a room 
ordinarily well illuminated. In short it would seem to me that the 
point of reference in determining the relations that are of im- 
portance to the eye is the brightness at the point of work. Any 
extreme deviation above or below this brightness, especially 
above, or anything that would make these deviations conspicuous 
to vision would seem to me to be of prime importance. I would,, 
therefore, consider it an important addition to our present 
method of specifying brightness effects to give more detailed 
measurements, so far as is practicable, of the surroundings im- 
mediately contiguous to the brightest spot, because the effect of 
that spot upon sensation is to an important degree dependent 
upon the immediate surroundings; but I would by no means be 
willing to make these measurements and that of the brightest spot 
the sole specification of brightness effects, as Mr. Cravath sug- 
gests might be sufficient for our present needs. Moreover, it must 
not be overlooked in this connection that a curve plotted on a 
basis of the ratio of A, brightness of the reflector, to N for the 
different conditions tested must give a curve very similar to that 
plotted on the basis of the brightness of the reflectors alone, for 
N does not vary greatly from a constant value for the six sets of 
reflectors we have used. Obviously, therefore, cognizance should 
be taken of this fact before too much general importance is attri- TRANSACTIONS OF ILLUMINATING ENGINEERING SOCIETY 
1138 
buted to this ratio as a separate factor from the shape of the 
curve plotted by Mr. Cravath for this particular set of conditions. 
(2) There might, it is conceivable, be some unknown effect on 
the retina which directly depresses its functional power or in- 
directly disturbs the adjustment of the eye. I have, however, 
already tested quite extensively the tendency of different lighting 
conditions to depress the functional power of the retina for as 
much as ten hours of continuous work and have found reason to 
believe that very little indeed of our results for the tests for loss 
of efficiency could be ascribed to a depression of retinal function. 
There are four ways, I may say, in which a change in the func- 
tional activity of the retina may be manifested: (a) a change in 
sensitivity to color and brightness; (b) a change in lag or the 
time required for the sensation to reach its maximum; (c) a 
change in the susceptibility to fatigue or exhaustion, measured by 
rate of exhaustion; and (d) a change in the power of recovery, 
measured by rate of recovery. All of these points were covered 
in the tests mentioned above. Short of such an investigation a 
complete record can not be given of the functional state of the 
retina at any time or as the result of any condition or set of 
conditions to which it may be subjected. But when such tests 
have been conducted for a period of exposure of the eye to the 
conditions tested more than three times as long as was used in the 
tests for loss of efficiency, it would seem reasonable to conclude 
that the results of this latter test could not be ascribed to any 
considerable extent to a depression of retinal activity. 
Mr. Cravath has always quarreled with me over what the test 
should be called and perhaps on good grounds. If "fatigue" is a 
more palatable term to the engineer than "loss of efficiency," I am 
quite willing that the test shall be called a fatigue test. I have 
in fact called it that part of the time myself. My reason for calling 
it something else in the beginning was primarily one of pro- 
fession. Among men in physiological and psychological optics 
the term fatigue as applied to the eye has been, since the days of 
Fechner and Helmholtz, a technical term connotating a retinal 
condition. It was chiefly to avoid the chance of confusion with 
the narrower usage of fatigue that I chose the broader term loss 
of efficiency as a brief designation of what is really tested, namely, 
the loss in the power of the eye to sustain clear seeing. [Reprinted from Journal of Experimental Psychology, Vol. I, No. 1, Feb., 1916.] 
A NEW METHOD OF HETEROCHROMATIC 
PHOTOMETRY 
BY C. E. FERREE and GERTRUDE RAND 
Bryn Maur College 
In a former paper1 it was stated that a satisfactory method 
of photometry should combine the following features. (1) It 
should enable us to detect small differences in luminosity 
and to reproduce our results for a given observer with a 
small mean variation and for a number of observers with a 
comparatively small mean variation. That is, the method 
should possess an adequate degree of sensitivity, and should 
give results with a satisfactory degree of reproducibility.8 
(2) It should be known either to possess of itself logical 
sureness of principle or its results must agree in the average 
with those of some method which can be shown to possess 
this sureness of principle. Methods having these features 
have been developed for the photometry of colorless light. 
The problem for the photometry of colored light, however, 
has presented great difficulty. 
The methods for the photometry of colored light may be 
grouped under two headings: the direct and the indirect 
methods. In the former class we have the method of direct 
1 Ferree, C. E. and Rand, Gertrude, 'A Preliminary Study of the Deficiencies of 
the Method of Flicker for the Photometry of Lights of Different Color,' Psychol. Rev., 
1915, 22, pp. 110-162. 
* It is scarcely needful to point out in this connection that the method should 
give results with a satisfactory degree of reproducibility for observations separated 
both by short and by long intervals of time. 
1 2 
FERREE AND RAND 
comparison or as it is sometimes called the equality of bright- 
ness method. This method is generally accredited with 
sureness of principle but it is deficient on the side of sensitivity 
and reproducibility when lights differing in color value are 
compared.1 Of the latter class the method of flicker has 
received the greatest amount of attention and has been the 
most favored. But in our former paper it was shown (1) that 
the method of flicker, so far as it has been developed up to 
the present time, does not possess of itself the sureness of 
principle needed to meet the requirements of a satisfactory 
method; and (2) that as yet its results have not been found 
to agree on the average with those of any method which can 
be shown to have this sureness of principle. It was also 
stated that in a later paper a method of photometry would 
be described which possesses approximately as high a degree 
of sensitivity and of reproducibility as the method of flicker 
and gives results which agree very closely on the average 
with those obtained by the equality of brightness method, much 
more closely than results obtained by the method of flicker. 
It will be the purpose of this paper to give a brief description 
of this method. It was first used by the writers in connection 
with work in color sensitivity for the purpose of detecting 
small changes on the illumination of a room by daylight.2 
Although not so convenient perhaps as the equality of bright- 
ness method for many of the purposes for which photometry 
may be used, for this work the method did not present any 
great amount of difficulty, and it proved to be much more 
sensitive than the equality of brightness method. The equality 
of brightness photometers of the Sharp-Millar type, for ex- 
ample, are very insensitive for the determination of the illumi- 
nation of a room by daylight, because the standard field 
1 Loc. cit., pp. 114-115. 
1 See Ferree, C. E. and Rand, G., 'An Optics-Room and a Method of Standardizing 
Its Illumination,' Psychol. Rev., 1912, 19, pp. 364-373; Rand, G., 'The Effect of Changes 
in the General Illumination of the Retina upon Its Sensitivity to Color,' ibid., pp. 
463-490; 'The Factors that Influence the Sensitivity of the Retina to Color: A Quanti- 
tative Study and Methods of Standardizing,' Psychol. Rev. Monog., 1913, 15, No. 62, 
pp. 79-166; see also Ferree, C. E., 'Tests for the Efficiency of the Eye Under Different 
Systems of Illumination and a Preliminary Study of the Causes of Discomfort,' Trans. 
III. Eng. Soc., 1913, 8, p. 51. METHOD OF PHOTOMETRY 
3 
illuminated by the tungsten lamp is deep orange in color, 
while the comparison field illuminated by daylight is clear 
white. This difference in color tone makes the judgment 
of equality of brightness difficult to make, and renders the 
instrument extremely insensitive for daylight work. 
So far as the type of judgment is concerned, the method we 
are about to describe is in reality an equality of brightness 
method but it has the important advantage that the observer 
is never required to judge impressions differing in color 
quality. In making a photometric balance by this method, 
the feature is thus eliminated which has from the beginning 
of the work of measuring light intensities rendered the 
equality of brightness method difficult of application to hetero- 
chromatic photometry. Because of this factor we have at 
the suggestion of certain photometrists undertaken to work 
out the principle on which the method is based in a way that 
will be of general service to the work in heterochromatic 
photometry and to secure data on the points in terms of 
which a verdict is rendered for every method of photometry, 
namely, the sensitivity and precision of the method, and the 
agreement of its results with those obtained by the equality 
of brightness method. 
The method is based upon the extreme sensitivity of the 
peripheral retina to brightness contrast, especially to the 
induction of black by a white screen. Not only is the pe- 
ripheral retina extremely sensitive to brightness contrast, 
but, which is the crucial point of the method, very small 
changes in the illumination of the contrast screen produce a 
change in the amount of the contrast. That is, for any part 
of the retina, for example, the amount of contrast induced 
by the surrounding field upon a contrast stimulus increases 
with decrease of illumination. In the peripheral retina we 
have an extreme case of this. In this part of the retina very 
small changes of illumination indeed are required to produce 
a noticeable change in the amount of contrast induced, especi- 
ally when the surrounding field is white and the stimulus is a 
gray of the proper brightness. 
The apparatus needed for the method consists of a vertical 4 
FERREE AND RAND 
screen 60 X 50 cm. with an opening 15 mm. in diameter mid- 
way between top and bottom and 16 cm. from one edge, a 
motor to carry the measuring discs, and a photometer bar. The 
surface of the screen is covered with Hering mat white paper, 
or preferably with a paper overlaid with magnesium oxide de- 
posited from the burning metal. This screen is graduated in 
the horizontal meridian from the center of the 15 mm. opening. 
Over the opening between the white covering and the 
screen is placed the gray which is to serve as the con- 
trast-stimulus. Nos. 3-5 of the Hering series of standard 
gray papers were found to serve this purpose best for 
the intensities of light we used in the work the results 
of which are given in Table I. The measuring discs were 
made up of a sector of the same gray as was used for the 
contrast-stimulus and a sector of black, the varying pro- 
portions of which produce the match which is needed between 
the stimulus gray darkened by contrast and the measuring 
disc. The measuring disc was carried by a motor the shank 
of which is at one end of the screen and normal to it. When 
in position, the inner edge of the disc extends approximately 
to the 200 point of the graduations, and is about 3 cm. in front 
of the screen.1 Immediately in front of the disc and 250 from 
the center of the stimulus-opening to the temporal side is placed 
the fixation-point.2 The photometer bar was placed in a 
1 The measuring disc was placed this distance in front of the screen partly to 
give freedom of rotation but primarily to eliminate any induction by the screen on 
the disc. That induction is not present in any considerable amount under the con- 
ditions described above, is evident from the following considerations, (a) It is not 
directly detectable by the eye; and (b) a very high sensitivity is attainable by the 
method which could not be the case were induction present to any degree, for if it 
occurred in amounts great enough to influence sensation, the effect would obviously 
be to lower the sensitivity of the method. That there has been no lowering of sensi- 
tivity from this cause is rendered probable by the fact that approximately as small 
differences can be detected for lights differing in color quality as can be detected by 
Lummer-Brodhun photometers of the most improved type when the lights compared 
are of the same color quality. 
2 Care should be taken not to place the fixation-point so far from the stimulus- 
opening that the color used is seen of a different quality at this point than at the 
stimulus-opening. If this should occur for any color when the fixation is at the 25 ° 
point, a lesser excentricity should be chosen. The sensitivity of the contrast-effect 
to change of illumination is not lessened to any considerable extent, for example, by 
taking a 200 instead of a 250 fixation-point. We have found the 250 point to be very METHOD OF PHOTOMETRY 
5 
plane normal to the screen midway between the stimulus- 
opening and the fixation-point. The contrast-stimulus and 
the measuring disc thus receive equal amounts of light from 
the source to be measured. 
In Fig. 1 is given a picture of the apparatus in the rough 
form in which it is at present used. At S is shown the screen 
which is illuminated in turn by the standard and comparison 
lights; at B the photometer bar; at D the measuring disc; 
at L the lamp to be photometered; at M the mouthboard; 
Fig. i 
and at 0 the opening filled with gray which is darkened in 
graded amounts by induction from the screen S as the 
amount of light received by it from the lamp L is varied. 
The screen S is supported by short rods screwed into heavy 
tripod bases. In order to provide for adjusting the height 
of the screen above the table the frame of the screen is 
fastened to the supporting rods by means of a collar and set 
screw. On the upper edge of this frame is attached a device 
for lining up the eye with the stimulus-opening 0. This 
device consists of the following parts: A vertical arm carrying 
at its lower end a small circle 15 mm. in diameter extends 
satisfactory, however. For a fixation-point we used a black knot on a gray thread, 
stretched from a point on the screen S to a rod in front and on the opposite side of 
the measuring disc. 6 
FERREE AND RAND 
down to the level of the stimulus-opening 0. Attached to 
the vertical arm just above this circle is a horizontal arm 
15 cm. long which carries at its outer end and at right angles 
to it a disc 22.3 mm. in diameter. The size and position of 
the disc, the circle, and the opening 0 and their distances 
from each other sustain such relations that when the eye is 
in position 25 cm. in front of the opening 0, the line of regard 
contains the centers of the opening 0, of the circle, and of 
the disc; and the inner edge of the circle is just contained 
within the stimulus-opening; and the edge of the disc is just 
contained within the circle. That is, in effect the device is a 
peep-sight arrangement, and the alignment described above 
is possible only when the eye is in such a position that the 
line of regard is perpendicular to the stimulus-opening, the 
circle, and the disc at their centers. The device is attached 
to the metal frame by means of a screw with a milled head, 
so that after the alignment is made it can be readily turned 
out of the road and clamped. In order that the distance of 
the eye may be accurately adjusted at the same time its 
alignment is made with the stimulus-opening, a measuring 
device is also provided. This device consists of a slender 
brass rod 25 cm. long fitted at either end with two short 
right-angled arms 5 mm. in length. On the end of one of 
these arms is a ring which is just larger than the stimulus- 
opening, and on the other is a brass disc of the same diameter 
as the ring, provided at its center with a large pupillary 
aperture. In adjusting the distance of the eye this disc is 
rested lightly against the forward surface of the eyeball and 
the ring against the screen S concentric to the stimulus- 
opening. When the position of the eye is once determined, 
a mouthboard is adjusted and clamped in position so that 
the observer's teeth fit into impressions previously made and 
hardened in wax. This fixes the relation of the observer's 
eye to the stimulus-opening 0. All that is needed, there- 
fore, at subsequent sittings to bring the eye again into this 
relation, is to fit the teeth into the impressions on the mouth- 
board. As stated above, the photometer bar when adjusted 
stands in a plane normal to the screen S and midway between METHOD OF PHOTOMETRY 
7 
the stimulus-opening and the fixation-point. This guarantees 
that the stimulus-opening and the measuring-disc at the 
fixation-point shall receive equal amounts of light from the 
lamp on the bar. The nearer end of the bar is placed at 
a distance below the level of the center of the stimulus-opening 
and fixation-point equal to the height of the center of the fila- 
ment of the lamp above the bar. Owing to the fact that the 
head may cast a shadow on the stimulus-opening when the 
bar is horizontal, the rear end of the bar is raised until the 
stimulus-opening and the screen for some distance around are 
free from shadow for any position of the lamp on the bar that 
may be needed to obtain the photometric balance. In order to 
provide for this tilting, the bar is made adjustable in height. 
It is supported on two heavy tripod bases each carrying a 
vertical stem made of hollow tubing split at the upper end and 
provided with collar and set screw. Into these tubes fit rods 
attached to the photometer bar by means of hinge joints. This 
arrangement permits of an independent adjustment of the 
height of the two ends of the bar. In order that a given 
amount of slant may be reproduced at any subsequent 
observation, an attachment is provided to indicate the angle 
which the bar makes with the vertical. This consists of a 
graduated half circle attached just above the hinge joint to 
the supporting rod furthest from the screen. On this rod 
just below the hinge joint is fastened a pointer. When the 
further end of the rod is raised this half circle is rotated past 
the pointer and the amount of rotation may be directly read 
from its graduated limb. Since the angle at which the bar 
is tilted is kept the same for the comparison and standard 
lights, the tilting interferes in no way with the accuracy of 
the photometric balance or the application of the law of 
squares to the determination of the relative luminosities of 
the standard and comparison lights.1 
1 We hope soon to use the principle upon which the method is based in connection 
with apparatus which will be more conveniently applicable to the general needs of 
the photometrist than is the rough device which is described above. One of the 
improvements which we expect to make, for example, is to eliminate the need of a 
slanting bar. For while the slanting bar does not interfere with the establishment of 
a correct photometric balance at the screen, it is not the most favorable condition under 8 
FERREE AND RAND 
The method is applied as follows: The observer's eye is 
lined up with the stimulus-opening in the manner described 
above and the contrast stimulus is put in position. The stand- 
ard lamp is placed on the photometer bar at a given distance 
from the screen, the eye is turned out to the fixation-point and 
the sectors on the measuring disc situated just back of the fixa- 
tion-point are adjusted by means of a number of separate judg- 
ments to match the gray at the stimulus-opening darkened by 
contrast.1 The relative value of these sectors becomes the index 
of the illumination of the screen for the standard lamp at the 
given distance. The standard lamp is then removed, the com- 
parison lamp is placed on the bar, and its position is adjusted 
which to operate all types of lamps. This should not be difficult to do by moving the 
observer farther from the screen and using a correspondingly larger stimulus-opening 
and a greater linear distance from this opening to the fixation-point. 
1 That is, with the change in the amount of light falling on the photometer head 
made up of gray stimulus and surrounding screen of white, there is a change in the 
amount by which the gray stimulus is darkened by contrast from the surrounding 
screen. This amount is recorded on the measuring disc directly behind the fixation- 
point, by adding a sector of black to the gray sector, which was chosen, it will be 
remembered, of the brightness of the gray stimulus uninfluenced by contrast. The 
value of this black sector thus becomes the fixed index of the amount of light falling 
on the screen in any given case. In short, with a given brightness relation of contrast 
stimulus to surrounding screen, there is a finely graduated scale of contrast values 
inversely correlative with the amounts of light falling on the screen. That is, with a 
large amount of light or high illumination of the screen, the amount of contrast in- 
duced is low; with a small amount of light falling on the screen the amount of contrast 
induced is high. The amount of change of the illumination of the screen needed to 
produce a noticeable change in the amount of contrast induced is very small indeed 
and because of this the method possesses a high degree of sensitivity. So far as we 
are able to determine, the amount of change in this illumination required to produce 
a just noticeable change in the contrast sensation is no greater than is needed to 
produce a just noticeable change in the positive sensation under the best conditions 
that have been as yet devised in photometric work for making the judgment, namely, 
the conditions presented by the Lummer-Brodhun photometer heads of the most 
improved type when the lights compared are the same in color quality. If this were 
not true, the method could not show, as it does, as great sensitivity for lights either 
the same or differing in color quality as does the equality of brightness method for 
lights of the same color quality. Also so far as we have been able to determine there 
is no greater variability in the eye's sensitivity to contrast from day to day than there 
is in its sensitivity to brightness itself. That is, photometric determinations by this 
method of a given light density on the screen show a very small mean variation from 
day to day, quite comparable in fact, with the mean variation found in judgments of 
brightness equality, when there is no difference in color quality to confuse the judg- 
ment. There is, moreover, so far as we have applied the method, a comparatively 
small variation for different observers. METHOD OF PHOTOMETRY 
9 
until the contrast-stimulus again just matches the measuring 
disc. Since the proportion of sectors has not been changed from 
that which was determined for the standard lamp, it can be 
assumed that the illumination of the contrast screen is again 
of the same value it was for the standard lamp at the given 
distance. The comparative luminosities of the two light 
Table I 
Showing a Comparison of the Results Obtained by the Contrast and Equality 
of Brightness Methods 
Distance of 
White Light Giv- 
ing Equality of 
Illumination 
Sensitivity or 
Amount of Change 
that Can be De- 
tected by 
Precision or the 
Mean Deviation 
from the Average 
Source of Colored 
Light1 
Color 
Equal- 
ity of 
Bright- 
ness 
Method 
New 
Method 
Equality 
of Bright- 
ness 
Method 
New 
Method 
Equalityj 
of Bright- 
ness 
Method 
New 
Method 
E 
O 
E 
O 
E 
O 
•centage 
dle-power) 
E 
O 
•centage 
dle-power) 
E 
u 
-centage 
dle-power) 
e' 
-centage 
dle-power) 
C a 
u fl 
Ph ri 
0 
<D fl 
<D C 
87 cp. 41 cm. 
Red 
66.1 
66.6 
2.0 
5-6 
o-5 
i-3 
1.1 
3-2 
0.20 
o-55 
distant from 
photometric screen 
Blue-green 
59-5 
59-5 
2.5 
7-85 
o-5 
1.65 
i-5 
4-85 
0.22 
o-97 
52 cp. 38 cm. 
Red 
82.2 
82.2 
2.1 
5-o 
o-5 
1.1 
1.1 
2-5 
0.20 
0.41 
distant from 
photometric screen 
Blue-green 
70.9 
70-5 
4-0 
10.4 
o-5 
1.4 
2.0 
5-5 
0.25 
1.1 
13 cp. 38 cm. 
Red 
160.6 
160 
3-o 
3-7 
o-5 
0.65 
1.6 
i-9 
0.23 
o-3 
distant from 
photometric screen 
Blue-green 
134-5 
134-9 
4.0 
o-5 
0.76 
2.2 
3-2 
O.24 
0.46 
sources, thus, can be computed directly by the law of squares 
from distance readings on the photometer bar. The advan- 
tage of the method for color photometry is, as has already 
been pointed out, that in the judgment in case of both the 
standard and comparison lights, the surfaces, the brightness 
values of which are being compared, are both illuminated 
by the same light. The eye is thus never compelled to 
make its judgment of two surfaces differing in color quality. 
The method is in fact almost if not quite as sensitive for the 
1 The colored lights were obtained in these experiments by means of the Wratten 
and Wainwright filters inserted in openings in the front of lamp-houses. The red 
filter used transmits the spectrum from .768 p to .65 mJ and the blue-green from .52 m 
o .465 p. 10 
FERREE AND RAND 
photometry of lights differing in color quality as it is for the 
photometry of lights of the same color quality or for colorless 
light; and its sensitivity for lights of the same color quality 
or for colorless light compares very favorably, by test, for 
the same observer with photometers of the best Lummer- 
Brodhun type. The indirect vision judgment, it may be 
said, however, presents more difficulty to the unpractised 
observer than is afforded by the flicker judgment, but this 
difficulty yields rather readily to practice. The method has 
been in almost daily use in our laboratory for four years. 
It has, for example, been employed exclusively for the 
standardization of the illumination of our optics-room in the 
determination of color sensitivity. 
Table I. has been prepared to show a sample of a compari- 
son of the results of this method and the equality of bright- 
ness method. These results are typical. They report the 
average of twenty-five determinations. Sensitivity and the 
average deviation or precision are expressed both in divisions 
of the photometer bar and in percentage candle-power. 
The lamps used for this work were standardized by the 
New York Electrical Testing Laboratories. They were 
operated on a storage battery circuit. 
With some modifications the contrast principle described 
above can be used for the determination in terms of pigment 
standards of the white-black value of colored papers for a 
given degree of illumination of the papers. The method 
of doing this can perhaps best be explained by comparing 
it with the method just described. In this method, it will 
be remembered, we determined how much contrast would be 
induced by one surface upon another of a lower coefficient 
of reflection when a given amount of light was incident upon 
the two surfaces, and used this amount of contrast as an 
index of the amount of light falling upon the two surfaces in 
establishingthephotometric balance, or equality of illumination 
at the screen. Here the relation of coefficient of reflection of 
the screen to the stimulus is fixed and the amount of light falling 
on the screen from the comparison lamp is varied until the right 
amount is obtained to give the photometric balance with the METHOD OF PHOTOMETRY 
11 
standard lamp. In the determination of the white-black value 
of a colored paper under a given illumination, the situation is a 
little different. The amount of light falling on the screen 
is kept constant and the relation to the screen of the pigment 
standard which is to express the white-black value of the 
color, must be varied until the right one is obtained. Here 
again we find out how much contrast is induced by one 
surface upon another of a lower reflection coefficient, in this 
case the colored paper; and this amount of contrast is taken 
as an index of the brightness relation of the colored paper to 
the surface of higher reflecting power, the screen S in Fig. I, 
for the given intensity of illumination. The problem thus 
becomes to determine what gray sustains the same brightness 
relation to the screen S as does the colored paper. This gray 
should represent the white-black value of the color. 
In detail the method is as follows: The stimulus-opening 0 
is filled with the colored paper for which the determination 
is to be made and a disc of the same color is made a part 
of the measuring disc. The fixation is taken and black is 
added to the color on the measuring disc until it is sufficiently 
reduced in brightness to match the stimulus color darkened 
by contrast. The amount of black that has to be added 
becomes the index of the amount of contrast induced and 
therefore of the brightness difference between the screen S 
and the colored paper for the degree of illumination used, 
and is kept constant during the remainder of the work. In 
order to be able to find out what gray differs as much in 
brightness from the screen as does the colored paper, a disc 
made up of variable proportions of white and black is substi- 
tuted for the colored stimulus behind the screen; and a sector 
made up of variable proportions of the same white and black, 
for the colored sector on the measuring disc. The white and 
black are then varied in proportions which are carefully kept 
equal for stimulus and sector until a match is obtained. When 
this match is gotten the relative amounts of white and black 
either in the stimulus disc or the substitute sector on the 
measuring disc should give the gray value of the color for the 
illumination employed. That is, there is obviously only one 12 
FERREE AND RAND 
gray that could be used in place of the contrast stimulus and 
in the substitute sector in the measuring disc which would 
satisfy the conditions of the experiment, and that is a gray 
having the luminosity or white-black value of the color origi- 
nally used. It is scarcely necessary to caution that throughout 
the determination the light falling on the screen, the stimulus 
and the measuring disc must be kept constant. Care must be 
taken also in this case not to take the fixation so far from 
the stimulus-opening that a color difference is present be- 
tween stimulus and measuring disc. 
If the determinations are wanted in photometric units 
they can be gotten by the equality of brightness method 
from the grays which have been equated with the colors. 
A convenient instrument to use for this purpose is the Sharp- 
Millar portable photometer with its scale calibrated to give 
results in terms of candle-power per sq. in. The calibra- 
tion can be readily made by the experimenter from a mag- 
nesium oxide surface with a known amount of light falling 
on it. If the light incident on the grays is daylight, the same 
difficulty will be experienced with the Sharp-Millar photom- 
eter, however, which was mentioned earlier in the paper, 
namely, the two halves of the photometer field will differ in 
color quality due to the fact that the carbon standard in this 
instrument gives a light rich in the long wave-lengths. 
1 In discussing the application of this method to the determination of the bright- 
ness of colored papers, reference should probably be given to the discussion of the 
Talbot-Plateau law in the paper preceding this. (See Ferree and Rand, 'A Preliminary 
Discussion of the Deficiencies of the Method of Flicker for the Photometry of Light* 
of Different Color,' Psychol. Rev., 1915, 22, pp. 149-150.) [Reprinted from Journal of Experimental Psychology, Vol. I, No. 3, June, 1916.] 
A SPECTROSCOPIC APPARATUS FOR THE IN- 
VESTIGATION OF THE COLOR SENSITIVITY 
OF THE RETINA, CENTRAL AND 
PERIPHERAL 
BY C. E. FERREE AND GERTRUDE RAND 
Bryn Mawr College 
The need has long been felt in physiological and psycho- 
logical optics for better methods of specifying and standard- 
izing the stimulus. In an adequate specification and stand- 
ardization of the stimulus for work in these fields, three points 
are involved. Means must be had (1) of securing a small 
range of wave-lengths and of stimulating with them any 
part of the retina; (2) of determining accurately the values 
of the wave-lengths employed; and (3) of specifying the 
amounts of light used. In the present series of papers ap- 
paratus will be described with which it is, comparatively 
speaking, easily and conveniently possible (1) to stimulate 
any part of the retina with the light of the spectrum and to 
control as desired the conditions of preexposure and sur- 
rounding field; (2) to control the amounts of light used within 
the small gradations needed for threshold and just noticeable 
difference determinations; and (3) to specify in C.G.S. units 
in any case in which it is wished the amount of light used. 
This work was begun and announced three years ago as the 
logical completion of work published at that time1 and has 
1 See Rand, G., 'The Factors that Influence the Sensitivity of the Retina to Color: 
A Quantitative Study and Methods of Standardizing,' Psychol. Rev. Monog., 1913, 15, 
No. 1, 166 pp.; and Ferree, C. E., and Rand, G., 'A Note on the Determination of the 
Retina's Sensitivity to Colored Light in Terms of Radiometric Units,' Amer. Journ. 
of Psychol., 1912, XXIIL, pp. 328-332. 
If more time has been taken to accomplish what we have here to present than 
seems justifiable, we beg to point out that the work has been in many respects pioneer 
in character and that the difficulties on the radiometric side had not yet been over- 
come by the physicist at the time we began our work. We have also been greatly 
diverted by the pressure of other work. 
The apparatus has been greatly simplified from its original construction. As it 
247 248 
C. E. FERREE AND GERTRUDE RAND 
involved the construction of a new type of campimeter, 
spectroscope, apparatus for the regulation of the amount of 
light used applicable to the spectroscope employed, and 
radiometric apparatus non-selective in its response to wave- 
length and sufficiently sensitive for the purpose in hand. 
The apparatus included under the first three of the above 
classes was constructed under direction by our department 
mechanician and should be within the technical capabilities 
of any good college mechanician. The description of this 
apparatus will form the subject matter of the present paper. 
For the radiometric apparatus we are indebted to Dr. W. W. 
Coblentz, radiometric specialist for the Bureau of Standards. 
Believing that the sensitivity of the thermopile could be 
increased by enlarging the area of the receiving surface and 
wanting to measure for at least one color all the light energy 
falling on the stimulus-opening in our campimeter screen 
rather than attempt to calculate it from one or more linear 
elements in this surface, we took up with Dr. Coblentz the 
feasibility of constructing surface thermopiles. He was asked 
to construct for us a surface thermopile having a receiving 
surface 16 mm. square, a trifle larger than the opening in our 
campimeter screen.1 By means of this thermopile and one 
of the linear type we have been able to measure the light at 
three places-at the analyzing slit, at the campimeter open- 
ing, and at the eye. The use that is made of these measure- 
ments in the color investigation will be stated in a later paper. 
Description of Apparatus 
For the sake of convenience the apparatus may be de- 
scribed under three headings: («) an apparatus for getting 
now stands it is comparatively simple in form; it is not difficult to operate, and is, 
we believe, within the financial and technical possibilities of any laboratory in which 
serious work is being done in the optics of color. There is no work known to us in 
the investigation of color sensitivity in which light of a standard quality and intensity 
is needed to which this or a similar apparatus can not be adapted. On the spectro- 
scopic side at least we hope it will prove sufficiently feasible and practicable to render 
undesirable the use of colored papers and filters in much of the work that is now 
being done in the optics of color. 
1 See Coblentz, W. W., 'Instruments and Methods Used in Radiometry'-IL, 
Bulletin of the Bureau of Standards, 1912, IX., p. 22, ff. SPECTROSCOPIC APPARATUS 
249 
the light stimulus of spectrum purity and of exposing different 
parts of the retina to this stimulus under any conditions of 
surrounding field, preexposure, etc., that may be desired; 
(b) an apparatus for varying the intensity of the stimulus 
within the limits needed for threshold and j.n.d. work; and 
(c) an apparatus for measuring the intensity of the stimulus 
in terms of common or C.G.S. units. 
1. An apparatus for getting a light stimulus of spectrum 
purity and of exposing the different parts of the retina to this 
stimulus under any conditions of surrounding field, preexposure, 
etc., that may be desired. 
To design an apparatus which will combine all the above 
features and which can besides be easily and conveniently 
operated is a task which has presented considerable difficulty.1 
Dreher2 and Abney,3 for example, have each described ap- 
paratus devised to accomplish a part of what is outlined 
above, but the apparatus of neither would at all serve the 
purpose that has impelled us to take up anew the problem of 
the construction of apparatus. 
1 A statement of the points needed by an apparatus by means of which all the 
factors influencing the sensitivity of the retina to colored light may be controlled and 
which will have besides a wide range of serviceability is, in the belief of the writers, as 
follows, (i) It must provide an accurate means of separating out the desired ranges 
of wave-lengths throughout the spectrum. (2) Means must be provided for directing 
these wave-lengths to any part of the retina that is desired. (3) Provision must be 
had for controlling the brightness of the field surrounding the color and the brightness 
of the preexposure. (4) The apparatus must be available for use in a light as well 
as in a dark room, else (a) the influence of the brightness of the surrounding field and 
of the preexposure upon the response of the retina to the different colors can not be 
eliminated from the investigation when wanted; and (b) a large part of the work in 
the optics of color can not be done, namely, the investigations of response under 
different degrees of illumination. (5) A method of presenting the light to the eye 
must be had which will give the effect of a surface or field variable in shape and size 
that may be imaged on the retina. (6) The method of presentation must be such 
as to allow of no admixture of light from the room with the range of wave-lengths 
needed for the stimulus, and it must be as little as possible wasteful of light, else a 
sufficient range of intensity can not be had. (7) A beam of colored light should be 
provided of such intensity that its energy value per unit of cross section can be speci- 
fied when desired at the point of entrance of the light into the eye. 
2 Dreher, E., 'Methodische Untersuchung der Farbentonanderungen homogener 
Lichter bei zunehmend indirektem Sehen und veranderter Intensitat,' Z. /. Sinnes- 
physiol., 1912, 46, pp. 1-82. 
3 Abney, W. deW., 'The Sensitiveness of the Retina to Light and Color,' Philos. 
Trans, of the Roy. Soc. of London, 1897, Ser. A, 190, pp. 155-195- 250 
C. E. FERREE AND GERTRUDE RAND 
A schematic representation of the apparatus under this 
heading is shown in the diagram in Fig. I. It consists in 
Fig. i. Showing the Path of the Light from the Source to the Eye. 
general of light source (N), spectroscope (Sp), micrometer 
slit (S2) for separating out the wave-lengths needed for the 
colored stimulus, rotary campimeter (RC), collimating lens 
(ZJ for rendering parallel the beam of light emerging from 
the analyzing slit (S2), and focusing lens (Z2) for bringing 
this beam to a focus on the pupil and for shifting the image 
formed to follow the pupil as the eye takes fixation at different 
degrees of excentricity. 
It is scarcely needful to call to mind that one of the chief 
difficulties in devising a peripheral vision spectroscope has 
been to secure a means of throwing the colored light on the 
excentric portions of the retina. Dreher, for example, has 
tried to do this with limited success by holding the eye 
stationary and by reflecting the light to the different portions 
of the retina by a system of mirrors. By this means so much 
light is lost by reflection that the device would be useless for 
our purpose because we wish to be able to radiometer the 
light which finally reaches the eye. Moreover the difficulty 
in making the adjustments and the impossibility of making 
fine adjustments even for the different portions of the retina 
in the same meridian render the apparatus practically useless 
for extended work. Also the ordinary campimeter pre- 
cautions with regard to the control of the brightness of the SPECTROSCOPIC APPARATUS 
251 
surrounding field and preexposure can not be taken nor can 
the apparatus be used in an illuminated room without the 
admixture of a great deal of light from the room with the 
wave-lengths selected for the stimulus. So far as we know, 
no attempt has yet been made to adapt the spectroscope to 
the campimeter, at least not in such a way that any meridian 
of the retina can be worked in easily and conveniently from 
center to periphery and an accurate control be had at every 
step of the brightness of the field surrounding the stimulus 
and the preexposure.1 
1 The hollow hemisphere, for example, used by Abney can scarcely be considered 
a campimeter for it was not designed for exercising any control over the brightness of 
the surrounding field. The apparatus was designed solely for dark-room work and 
was called by him a perimeter. The practical impossibility of uniformly controlling 
the brightness of the field surrounding the stimulus by means of an opaque campimeter 
screen hemispherical in shape was shown by one of the writers (Ferree) in a previous 
article. ('A Note on the Rotary Campimeter,' Psych. Rev., 1913, 20, pp. 373-378.) 
Abney describes two devices for exposing the peripheral retina to the light of the 
spectrum. Both are for use in the dark room. The first consists of a perimeter, the 
graduated arm of which is painted white, and the following auxiliary apparatus. 
Below the eye is placed "a small mirror on a ball and socket joint which by means of 
an arm causes a beam of light falling on it to be cast in any direction that is desired." 
The monochromatic light for the stimulus is gotten by means of an instrument which 
he calls his "color patch apparatus." The light passes from this apparatus to the 
mirror on the ball and socket joint which reflects it to the different points on the arm 
of the perimeter. In this way patches of color which serve as the stimuli for the eye 
in the color investigation are formed at different degrees of excentricity on the peri- 
meter arm. That is, with regard to the method of securing the stimulation of the 
peripheral retina, this apparatus is of the usual type. The eye is held stationary and 
the retina is stimulated from center to periphery in a given direction by causing the 
colored stimulus to pass from the center to the periphery of the field of vision in the 
opposite direction. In his second apparatus, also called a perimeter by him, a hollow 
white hemisphere made of papier mache is employed instead of a rotating arm. In 
the center of this surface is a circular aperture l| in. in diameter to admit the stimulus 
light. To give the effect of a colored surface to this aperture when illuminated from 
behind with colored light, it was filled in with glass ground on both sides. In order 
that the shape and size of the stimulus may be varied, additional special apertures 
are provided which are placed as desired in the path of light immediately behind the 
fundamental aperture. As before, the color patch apparatus is used as the source 
of the stimulus light. The method of securing the excentric stimulation of the retina, 
however, is just the reverse of that described before. That is, the colored stimulus 
is kept in a fixed position and the eye follows a phosphorescent point which is moved 
from the center to the periphery of the hemisphere in the meridian in which the in- 
vestigation is being made. 
Neither of Abney's devices would have answered our purposes for the following 
reasons. (l) Neither is adapted for use in a light room. (2) Provisions for the 252 
C. E. FERREE AND GERTRUDE RAND 
One of the difficulties encountered in attempting to do 
this is that when the eye is turned to take an excentric fixa- 
tion, the pupil is rotated out of the beam of light. To over- 
come this, we have taken advantage of the simple fact that 
when the axis of a double convex lens is displaced from the 
axis of the beam of light, the image of the light is displaced 
in the same direction in which the lens is moved. There- 
fore, by moving the focusing lens (Z2) an appropriate amount 
in the direction the eye is moved in taking its excentric fixa- 
tion, the beam of light can always be kept focused on the 
pupil, while no light energy is lost in the operation, at least 
not enough to be detected either by the eye or by the thermo- 
piles, linear or surface, that we are now using. That is, in 
the apparatus shown in the diagram, the lens (Z2) which 
focuses the light on the eye, is mounted on a rack and pinion 
operating in the line of the fixation-arm of the campimeter. 
When a given meridian is to be investigated, the fixation-arm 
is turned into this meridian and the light is always kept 
focused on the pupil of the eye by means of fine changes in 
the position of the lens made possible by the rack and pinion 
adjustment.1 
control of surrounding field and preexposure are lacking. And (3) the loss of light is 
so great and the method of presentation is such that the intensity of the light entering 
the eye could not be measured. 
1 One of the objects in devising this means of presenting the stimulus light to 
the eye has been to get an apparatus that could be used in a well illuminated room, for 
only under these conditions, as has been shown in previous articles (see 'A Note on 
the Determination of the Retina's Sensitivity to Colored Light in Terms of Radio- 
metric Units,' Amer. Jour, of Psychol., 1912, 23, pp. 331-332; 'The Effect of Changes 
in the General Illumination of the Retina upon its Sensitivity to Color,' Psychol. 
Rev., 1912, 19, pp. 463-490; 'The Factors that Influence the Sensitivity of the Retina 
to Color; A Quantitative Study and Methods of Standardizing,' Psychol. Rev. Monog., 
March, 1913, pp. 135-166) can the influence of the brightness of the surrounding field 
and of the preexposure be eliminated from the investigation. Obviously a serious 
difficulty in working in an illuminated room is to get a method of presenting the 
stimulus to the eye that will give the effect of a colored field and prevent the admixture 
of light from the room with the wave-lengths which are needed for the investigation. 
The following methods may be used in dark-room work to get the effect of a colored 
field or surface to image on the retina. (1) Light of the desired range of wave-lengths 
is allowed to fall on a diffusely reflecting surface which is non-selective in reflection of 
wave-length. The standard reflecting surface for purposes of this kind is prepared 
from magnesium oxide deposited from the burning metal. Such a surface properly 
prepared has a high reflection coefficient approximately 90 per cent., and possesses SPECTROSCOPIC APPARATUS 
253 
The Source of Light.-The requirements of a source for our 
purpose are (a) that it should give a light in which all the 
wave-lengths are represented in sufficient strength both for 
the needs of the investigation of color sensitivity and for 
radiometric standardization; and (b) that the light emitted 
shall be as constant as possible in intensity and spectro- 
radiometric composition. After trying many light sources 
we have finally adopted as best for our purpose the Nernst 
filament. This filament, after having been properly seasoned, 
the additional advantage that it can easily be kept fresh. And (2) light of the desired 
composition is allowed to fall on a diffusely transmitting surface also non-selective, 
or approximately so, in its transmission of wave-length. These methods of getting the 
effect of a colored surface or field while serviceable within limits for dark-room work, 
have the following objections for our purpose, (a) Both kinds of surface would reflect 
the white light of the room-the magnesium oxide surface, for example, would reflect 
approximately 90 per cent, of the light falling on it. This effect, however, is reduced 
to a minimum by using a double convex lens in the manner employed by us. For 
when a parallel beam of light falls on the lens and the eye is placed at the focus, the 
lens fills uniformly with light, and, so far as we are able to determine, white light is 
not reflected to the eye from the surface of the lens in appreciable amounts under the 
conditions obtaining in our work. That is, the lens is placed 2.5 cm. behind the 
stimulus-opening and the whole spectroscope and lens system are enclosed up to the 
campimeter screen in a light-tight compartment. Under these conditions by exer- 
cising a little care in the selection of the position of the apparatus with reference to 
the distribution of light in the room, so little light is reflected from the lens to the 
eve that the presence of the lens can not be detected when the eye is in position before 
the stimulus-opening and no colored light is coming from the spectroscope. In other 
words, not enough white light is reflected from the lens to render it visible to the eye. 
(i) Both of these methods are very wasteful of light. That is, there is not only a 
comparatively high percentage of absorption of light t)y the types of diffusing media, 
but only a small percentage of the light coming from any point or unit of the surface 
of these media enters the pupil of the eye. Thus there is not the possibility of getting 
the wide range of intensities of light needed to give the apparatus maximum service- 
ability for the investigation of color sensitivity. Furthermore, we are desirous of 
being able to specify in radiometric units the amount of light per unit area in the cross 
section of the beam of stimulus light at its entrance into the eye. This would be a 
very difficult task indeed with either of the two methods of presenting the stimulus 
light to the eye mentioned above. Its accomplishment, however, is not at all difficult 
by the method we have adopted. That is, instead of the light spreading from the 
stimulus-opening as if emanating from a source, it is concentrated into an image on 
the pupil of the eye, an image in this case of the analyzing slit. (So, see pp. 250, 261 and 
Figs. 2 and 3.) The amount of energy concentrated into this image can easily be deter- 
mined by direct measurement with our radiometric apparatus for any of the wave- 
lengths used by us as stimuli, and the amount per unit area be estimated; further, if 
an artificial pupil were used, or the substitute to be described in a later paper, see 
Psychological Review, 1916, September number, the total amount of light entering the 
eye could be determined in C.G.S. units. 254 
C. E. FERREE AND GERTRUDE RAND 
gives a light which changes comparatively little through long 
intervals of time. It has the advantage, moreover, that its 
shape well adapts it for use with the slit of the spectroscope, 
i. e., the shape is such as to make it possible to utilize for the 
illumination of the face of the prism a relatively large pro- 
portion of the light emitted. Also by increasing the length 
of the filament and collimator slit, it is possible to increase 
in direct ratio the amount of light falling on the face of the 
prism. When in use, in order that no light shall be lost in 
passing through one or more condensing lenses, the filament 
Fig. 2. Showing the Apparatus Grouped for Work-light source, spectroscope, 
thermopile, and rotary campimeter with its auxiliary lens system. In order that all 
stray light either from the general illumination of the room or by reflections from 
the Nernst glower be eliminated, the spectroscope and lens system are, when in use, 
enclosed from just in front of the Nernst to the campimeter screen in a light-tight 
compartment. 
is mounted as closely as possible to the jaws of the slit. It is 
shown in position for work at (TV) in Figs. 2 and 3. The 
filament is mounted in a lamp socket fastened to an upright SPECTROSCOPIC APPARATUS 
255 
(^1). In order that the height of the filament shall be ad- 
justable this upright consists of a short rod sliding in a tube 
fitted with collar and set screw (5) which permits of a move- 
ment of the socket up and down, also to right and left. The 
upright (^1) is fastened to the horizontal support (Di) ex- 
Fig. 3. Showing a Second View of the Apparatus Grouped for Work. 
tending out from the collimator arm of the spectroscope by a 
clamp and set screw (Ci) which permits of the motion of the 
socket to right and left and to and from the collimator slit. 
The Nernst filament used by us is designed to be operated 
at no volts on a D.C. current, and has a carrying capacity 
rated at 1.2 amperes. When operated at its maximum 
capacity on a direct current, we have found, however, that 
the life of the filament is short. Satisfactory results have 
been obtained by us only when it is operated at or below 0.6 
amperes. In series with the ballast, therefore, which is 
ordinarily used for the reduction of the current from the line 
and to compensate for the change in the resistance of the 256 
C. E. FERREE AND GERTRUDE RAND 
Nernst material as its temperature varies, we have found 
it necessary to use additional resistance. This resistance is 
needed not only to cut down the current to the desired value, 
but to correct for fluctuations in the line. Two coils are 
employed, one coarse and one fine. The former is used to 
cut down the current to approximately the desired value, 
and the latter to correct for the fluctuations in the line. Both 
are in the form of adjustable rheostats. The second is of 
special construction to give the small changes needed. It 
consists of a cylindrical coil of wire wound on an insulated 
core of brass tubing and is operated by a screw motion in 
such a way as to give the effect of a slider on a single wire. 
This rheostat is described in greater detail on p. 275. It 
is not on the market but can readily be constructed to 
order by a laboratory mechanician. The finely graduated 
control afforded by it not only makes it possible to correct 
for fluctuations in the line, as is stated above, but also to 
compensate for the slow decrease in the carrying capacity 
of the Nernst material with use. The current consumed is 
measured by a Weston ammeter graduated to 0.02 amperes. 
Operated with this type of control the light flux obtained from 
the Nernst may be kept constant within the limit of change 
that can be detected either by the radiometer or the eye. 
The Spectroscope.-In addition to the usual features 
attaching to a good spectroscope, our instrument was de- 
signed especially to meet the following needs. (1) To answer 
all the purposes for which a source of colored light may be 
used in the investigation of color sensitivity, a wide range of 
intensity is needed. In order too that an adequate radio- 
metric standardization be made, it is especially desirable 
that light of high intensity be available. (2) If the spectro- 
scope is to be used in conjunction with a campimeter, it is 
necessary that the objective arm remain in a fixed relation 
to the stimulus-opening of the campimeter screen and that 
some convenient and accurate means be had of changing the 
wave-lengths without changing the position of the objective. 
The first of these needs was met by employing a collimator 
slit long enough to admit the amount of light needed, and a SPECTROSCOPIC APPARATUS 
257 
prism and lenses large enough to take care properly of the 
amount of light admitted. The second point has been met 
by us in the two following ways, one of which is technically 
more correct perhaps than the other. («) The spectroscope 
was mounted on a stationary base supported by levelling 
screws. In fixed relation to this base a slit (S2) was perma- 
nently mounted to separate out the wave-lengths which are 
to fall on the stimulus-opening in the campimeter screen. 
In order to be able to change the wave-length of the light 
falling on the slit (S2) the stationary base carries a track 
along which the whole spectroscope may be shifted by very 
small amounts. This movement is made by means of a 
screw of fine thread turned by a wheel 4I inches in diameter. 
This track and the base of the spectroscope carry a Vernier 
scale graduated to 1/50 of a mm. by means of which any previ- 
ous setting may be accurately reproduced, and in terms of 
the readings of which the spectroscope may be calibrated in 
wave-length. This method of changing wave-length not 
only provides abundantly for small changes in wave-length 
but it obviates any necessity for readjustment of the colli- 
mator arm which would be the case were the wave-lengths 
changed, for example, by rotating the prism. In case the 
wave-length is changed by this device, the prism is set for 
minimum deviation for the D-line, and this adjustment is 
kept throughout. An adjustment which gives minimum 
deviation for the D-line alone is generally considered by 
spectroscopists to be adequate for the work in the visible 
spectrum when the light is passed through only one prism. 
Kayser, for example, says:1 "Bei den Apparaten mit nur 
einem Prisma verzichtet man fur gewohnlich darauf, alle 
Wellenlangen unter dem Minimum zu beobachten, sondern 
stellt das Prisma fest auf, so dass etwa die D-Linien unter 
dem Minimum durchgehen. Die Dispersion ist bei einem 
Prisma so gering, dass dies gewohnlich geniigt. Das ist 
aber nicht mehr der Fall, wenn man fiber die Grenzen des 
sichtbaren Spectrums hinausgeht, und so hat vielleicht zuerst 
Langley bei seiner Untersuchung des ultrarothen Theiles 
1 'Handbuch der Spectroscopic,' Bd. I., p. 510. 258 
C. E. FERREE AND GERTRUDE RAND 
des Sonnenspectrums an einem Apparat mit einem Prisma 
eine Vorrichtung beschrieben, welche automatisch das Mini- 
mum erhalt, und welche wegen ihrer Einfachheit seitdem 
oft verwandt wird." (P) In order that minimum deviation 
may be had automatically for all wave-lengths falling on our 
analyzing slit, one of our spectroscopes was built with a 
minimum deviation attachment.1 A schematic representa- 
tion of this attachment and the prism in position for use is 
shown in Fig. 4. In this figure, P represents the prism so 
placed on the prism table that its refracting angle is bisected 
by one of the radii of the table; PB and PA represent respec- 
Fig. 4. Showing the Prism Table with Its Auxiliary Minimum Deviation 
Attachment. 
1 This attachment is a modified form of a device apparently used first by A. 
Cornu in 1873. He does not describe the apparatus, however, until ten years later 
('Sur un spectroscope a grande dispersion,' J. de Phys., 1883 (2), 2, pp. 53-57). The 
first suggestion of such a device seems to have been made by Mascart ('Recherches sur 
le spectre solaire ultra-violet, et sur la determination des longueurs d'onde,' Ann. 
ecole norm., 1864, 1, pp. 219-272). See also S. P. Langley, 'The Selective Absorption 
of Solar Energy,' Amer. J. of Sc., 1883 (3), 25, pp. 169-176, Philos. Mag., 1883 (5), 
15, pp. 153-183, Ann. Chim. et Phys., 1883 (5), 29, pp. 497-542, Wied. Ann., 1883, 
19, pp. 226-244, 384-400; B. B. Donath, ' Bolometrische Untersuchungen fiber Ab- 
sorptensspectra fluorescerender Substanzen und atherischer Oele,' Wied. Ann., 1896, 
58, pp. 608-661; and F. L. O. Wadsworth, 'Fixed-arm Spectroscopes,' Philos. Mag., 
1894 (5), 38, PP- 337-351- SPECTROSCOPIC APPARATUS 
259 
tively the collimator and objective arms which are fastened 
to the stem of the spectroscope independent of the prism 
table; TD represents an arm fastened to the prism table in 
such a position as to be continuous with the radius of the 
table which bisects the refracting angle of the prism; AE 
and BE represent two rods of equal length which connect 
PA and PB at points equidistant from the center of the 
table to a collar which is free to play back and forth along 
the arm TD. M is a micrometer screw with a graduated 
cylindrical head, which is used to move the collimator arm 
through the small angles needed to give the change of wave- 
length. Opposed to this screw is a plunger p working against 
the spring in the case C. When the screw advances it moves 
the collimator arm forward and when it recedes the collimator 
arm is made to follow it by the push of the plunger p. The 
screw and plunger are supported by a curved arm coming off 
from the stem of the spectroscope, which can be clamped 
rigidly in any position which may be desired. This arm 
ends in two right-angled extensions, one of which carries the 
screw and the other the plunger. Between X and Y, the 
vertical arms of these extensions, the collimator arm is moved 
to give the change of wave-length. The micrometer screw 
passes through a fixed sleeve S graduated in fortieths of an 
inch. As the screw advances the cylindrical head telescopes 
on the sleeve S, one of the graduated spaces being traversed 
with each complete turn of the screw. The forward end of 
the cylindrical head is bevelled and on the circumference of 
the bevelled edge is a scale of equal divisions graduated in 
twenty-fifths of an inch. By means of this scale and the 
scale on the sleeve S, the advancement of the screw can be 
read in thousandths of an inch. 
That minimum deviation is given automatically by this 
device to all the wave-lengths falling on the analyzing slit 
may be understood from the following considerations. In 
attaining minimum deviation the incident and emergent rays 
make equal angles with the normal to the refracting faces of 
the prism, therefore equal angles with the refracting faces 
themselves. When an adjustment is made for minimum 260 
C. E. FERREE AND GERTRUDE RAND 
deviation for the D-line of the spectrum, for example, and 
the wave-length is changed by moving either the collimator 
or objective arms, the prism must be turned through half 
the angle through which the collimator or objective arm is 
moved if the wave-length traversing the axis of the objective 
tube is to be deviated the minimum amount. That is, for 
the prism to be set for minimum deviation the line bisecting 
the refracting angle of the prism will also bisect the angle 
made by the incident and emergent rays; hence if in changing 
the wave-length, the angle between the incident and emergent 
rays be changed a given amount by a movement of the 
collimator or objective arm, the prism must be moved through 
half that angle in order that the line which bisects its refracting 
angle will also bisect the angle made by the incident and 
emergent rays. The special purpose of the attachment 
described above, therefore, is to turn the prism through half 
the angle traversed by the collimator arm in changing the 
wave-length. This is attained by placing the prism on the 
table so that the radial arm PD bisects the refracting angle, 
and by connecting the moveable collar on the arm PD with 
the collimator and objective arms at points equidistant from 
the center of the prism table by rods of equal length (AE and 
BE of Fig. 4). Then when an adjustment is made for the 
D-line, and the collimator arm is moved through the angle 
needed to change the wave-length which falls on the analyzing 
slit, the arm {PD} turns through half that angle and takes 
up a position midway between the collimator and objective 
arms, and the part of the spectrum which falls on the ana- 
lyzing slit is deviated the minimum amount in passing 
through the prism.1 
1 A constant deviation prism may be used also for getting automatically minimum 
deviation for all parts of the spectrum. Such a prism may be constructed by setting 
two 300 prisms against the faces of a right-angled totally reflecting prism; or the 
prism may be made in one piece in such a way that the four vertical faces enclose four 
angles of 900, 75°, 1350, and 6o° respectively. In using this type of prism the colli- 
mator and objective tubes are set permanently at right-angles to each other. When 
this is done and the prism is rotated about a vertical axis, the different portions of 
the spectrum are thrown on to the analyzing slit and the portion that enters it at 
any moment, will traverse the system at minimum deviation. 
In the early part of our work we used a constant deviation prism of the Cassie SPECTROSCOPIC APPARATUS 
261 
Following is a description of the parts and mounting of 
the spectroscope. The collimator slit (Si) is 12 mm. long. 
Its width is adjusted by means of a micrometer screw fitted 
with a head graduated to read thousandths of a mm. The 
collimator lens (CZ) is a Zeiss triple achromat, 180 mm. 
focal length, 60 mm. diameter; the objective lens (OL) is 
also a Zeiss triple achromat, 240 mm. focal length, 60 mm. 
in diameter. The prism (P) is 100 mm. high and has a 
refracting base of 85 mm. and a refracting angle of 6o°. 
Owing to the large size required, a liquid (CS2) prism has thus 
far been used. With the exercise of a reasonable amount of 
caution to keep the CS2 free from impurities, this prism 
has given very good satisfaction. At present we see no good 
reason why its use should not be continued. The analyzing 
slit (S2) is made adjustable in length. The range attainable 
varies from o mm. to 12 mm. A slit adjustable in length is em- 
ployed in order that the amount of light entering the eye may be 
made independent of the natural pupillary aperture. (See 'A 
Substitute for an Artificial Pupil,' Psy. Review, 1916, -, Sep- 
tember number.) So far it has been mounted on an independ- 
ent base screwed to the table in a fixed relation to the base of 
the spectroscope and the stimulus-opening in the campimeter. 
If desired this base might be made a continuation of the base 
of the spectroscope. In order that the distance of this slit 
from the objective lens (OL) may be adjusted for the different 
focal distances for the different wave-lengths, the frame in 
which the slit is mounted is furnished with a rack and pinion 
adjustment. 
In Figs. 2 and 3 is shown the mounting of the spectroscope. 
At E may be seen the platform and track on which moves 
the carriage bearing the spectroscope. The platform is 
supported on four upright somewhat pointed screws (Fi) 
which serve the triple purpose of leveling the apparatus, of 
type which not only gives automatically minimum deviation but also the effect of a 
train of prisms; i. e., the arrangement is such that the light is passed back and forth 
through the prism a number of times before it finally enters the objective tube. 
(Cassie, W., 'Multiple Transmission Fixed-arm Spectroscopes,' Philos. Mag., 1902, 
Ser. 6, 5, 449-457.) This prism was constructed to special order of the dimensions 
needed for our apparatus. 262 
C. E. FERREE AND GERTRUDE RAND 
allowing an adjustment of 4 inches in height, and of pre- 
venting chance shifting of the position of the apparatus on 
the table.1 To this movable platform is securely fastened 
the heavy stem and tripod base on which the spectroscope 
is mounted. The prism table of the spectroscope is shown 
at W. It has two adjustments, (a) With its supporting 
table it can be rotated in the horizontal plane to provide for 
the adjustment of the collimator and objective systems at 
the angle of minimum deviation and to allow for change of 
wave-length by the rotation of the prism if that should be 
desired. And (b} it is furnished with three leveling screws 
(F2) by means of which it can be accurately leveled. The 
collimator and objective tubes are mounted on two horizontal 
arms (D2 and D3) which rotate about the upright stem on 
specially prepared collars. D2 furnishes the support for the 
light source and the collimator tube, and Ds for the objective 
tube. In order that these arms when adjusted shall be 
1 In the construction of the carriage and track great care was taken that there 
should be no play or free motion of the parts. This was necessary in order that for a 
given reading of the Vernier scale on the carriage and platform the axis of the objective 
should always sustain the same relation to the slit (S2)-in other words that for a 
given reading the same wave-lengths should always fall on the slit. 
The calibration of this Vernier scale was accomplished by means of a Hilger direct 
vision spectrometer which has a scale reflected across the upper half of the spectrum. 
Since wave-lengths can not be read directly from this type of instrument, a supple- 
mentary chart must be made in which are given the values of the different scale divi- 
sions in terms of wave-length. This was done as follows. Twenty-six points in this 
scale were identified with the bright lines given by potassium, strontium and cadmium 
arcs and by the mercury tube. The wave-lengths of these bright lines were obtained 
from Hagenbach and Konen's 'Atlas of Emission Spectra,' and from these values the 
curve of wave-lengths was plotted for the entire spectrum. For the calibration of the 
Vernier scale the Hilger spectrometer was then mounted behind the stimulus-opening 
in the campimeter screen, and the wave-lengths coming through for any given reading 
of the Vernier scale were determined. 
The calibration of the graduated scale on the fixed sleeve and cylinder head of 
the minimum deviation attachment (p. 258-260) maybe accomplished in a similar way. 
In case the minimum deviation attachment is used and the changes of wave- 
length are produced by it, such precision in the construction of the carriage and track 
as is described above is of course not necessary. That is, the carriage and track, 
while devised in our particular apparatus primarily for changing the wave-length, is 
very serviceable for other purposes which do not require such careful construction. 
For example, we have found this feature to be very useful in lining up the instrument 
with other apparatus, especially when the work requires the objective arm to be fixed 
as is needed in our adaptation of the spectroscope to the campimeter. SPECTROSCOPIC APPARATUS 
263 
held rigidly in place, they are firmly clamped to the sup- 
porting table by means of the metal pieces (Gi and G2) one 
end of each of which is milled down to fit in a groove in this 
table (FT) and the other is clamped respectively to the arms 
(D2 and Ds). The collimator and objective lenses are each 
mounted in brass telescope tubing provided with a rack and 
pinion for the adjustment of its length. In order that the 
axes of these tubes may be in the same horizontal plane the 
following provisions for leveling are made. The tubes fitted 
about one-fourth of their length from the larger end with 
collars (Zi and Z2), are swung on trunions (/x and /2) in 
U-shaped housings (TTi and A"2) supported by small vertical 
pillars coming up from the horizontal arms (D2 and D3). 
The smaller end of each tube rests on a leveling screw (T3 
and F^ threaded in a short horizontal stem coming out from 
the housings (Ki and Zf2), by means of the adjustment of 
which the axis of the tube may be brought into the proper 
position. When this position is attained provisions are 
made for clamping the tube firmly in place. In order that 
room may be allowed between the prism and the collimator 
and objective tubes for the introduction of apparatus for 
reducing the intensity of light or for any other purpose that 
may be desired, the vertical pillars carrying the tubes run 
in slots io cm. long cut in the horizontal arms (D2 and D^. 
When adjusted to the proper position the pillars are firmly 
clamped to the horizontal arms by means of the milled nuts 
(Mi and M2). The analyzing slit to separate out the wave- 
lengths to be used for the colored stimulus is shown at S2. 
The length of this slit may be made to vary from o mm. to 
12 mm. Its width is adjustable by means of a micrometer 
screw fitted with a head graduated to read to thousandths 
of a mm. This slit is mounted vertically in an oblong brass 
frame, 12 cm. long and 6.5 cm. broad, carrying on each side a 
groove. In this groove slides a narrow brass plate on which 
is mounted a linear thermopile (T) with its receiving surface 
facing in the direction of the slit through a circular opening 
in the plate. During the color observation the plate and 
thermopile are raised out of the path of the beam of light 264 
C. E. FERREE AND GERTRUDE RAND 
and clamped. For the energy measurements they are lowered 
to the level of the slit. In order to focus the different wave- 
lengths of the spectrum on the analyzing slit, a micrometer 
adjustment is provided to regulate the distance from the 
objective lens. The frame for thermopile and slit and the 
micrometer attachment are carried by an upright (^2) which 
is mounted to one side of the path of the beam of light on a 
heavy independent tripod base screwed to the table. In 
order that there may be an adjustment of height, the upright 
consists of a tube 34 cm. long fitted at its upper end with a 
collar and set screw into which slides a steel rod (Ah) 38 cm. 
in length. The micrometer adjustment consists of a tube 
30 cm. long and 15 mm. in diameter along the axis of which 
runs a finely threaded screw fitted with a milled head. In 
this tube is a beveled slot 5 mm. wide and 22.5 cm. long 
exposing the threaded screw. Telescoping the tube are two 
sections of tubing of larger bore (^i and ^2), 4-5 cm. in length, 
to the inner surfaces of which are screwed threaded pieces 
which extend down into the slot and engage the micrometer 
screw. As the micrometer screw is turned, these tubular 
sections move slowly along the screw. They are each fitted 
with collar and set screw which hold the horizontal rods sup- 
porting respectively the framework for the thermopile and 
the holder for the collimating lens (Zi). The collimating 
lens (Zi) is inserted in the path of the beam of light in order 
that the light emerging from the analyzing slit (S2) may enter 
parallel the focusing lens (Z2) in front of the stimulus-opening 
in the campimeter screen. This lens has a diameter of 40 
mm. and a focal length of 140 mm. Obviously if it is to 
act as a collimator for all the waves of light emerging from 
the slit (S2), its distance from the slit must always be kept 
equal to its focal length. That is, when the distance of the 
slit from the objective lens is changed to accommodate for 
the difference in focus for the different wave-lengths of light, 
the distance of the lens (Lx) must also be changed by an equal 
amount and in the same direction. This is accomplished 
by the micrometer adjustment described above. That is, 
the tubular sections which carry respectively the framework SPECTROSCOPIC APPARATUS 
265 
for the slit and the holder for the lens are both operated by 
the same micrometer screw, hence any movement of the 
slit is accompanied by a similar movement of the lens. An 
adjustment of the distance of the lens for the slit, once made, 
need not, therefore, ever be disturbed in the process of 
accommodating for the difference in focus for the different 
wave-lengths of the spectrum. 
In order that all stray light either from the general 
illumination of the room or by reflections from the Nernst 
glower be eliminated, the spectroscope and lens system are 
enclosed from just in front of the Nernst to the campimeter 
screen in a light-tight compartment. So far we have found 
it most convenient to make this compartment of light- 
proofed fabric. That is, the compartment must be made 
large enough to enclose all of the auxiliary reduction and 
adjusting apparatus, the thermopile, etc., and must permit of 
easy entrance for the purpose of making the adjustments and 
settings required. The construction of a compartment having 
these requirements we found could be most simply and feasibly 
accomplished by employing, as stated above, light-proofed 
fabric. There are so many ways in which such a compartment 
can be constructed that space will not be taken here for a de- 
scription of the compartment we have used. It will be suffi- 
cient to say that adequate care has been taken to exclude 
any stray light that might find its way into the stimulus by 
means of reflections and refractions within the system. 
It is well known to spectroscopists that if some portion of 
the spectrum obtained by a single spectroscope be examined 
by a second instrument of good resolving power, frequently 
more than one band will be obtained. For example, if a 
spectroscope is so adjusted that only a narrow region of the 
red falls on the objective slit and the light emerging from this 
slit is passed through a second spectroscope, it may be found 
that the second resolution gives one or more comparatively 
faint bands in some other part of the spectrum. In most 
cases where a good instrument is used this degree of impurity 
would probably not appreciably change either the radio- 
metric or optical results. However, we have endeavored to 266 
C. E. FERREE AND GERTRUDE RAND 
devise means whereby light of such a degree of purity may 
be obtained that the second resolution shows only the one 
band. Our first step in this direction was (tz) to eliminate 
as far as possible all stray light and internal reflections. 
Stray light was eliminated by carefully light-proofing the 
housing of the apparatus. Also the base of the prism and 
every other surface which might either admit or reflect ex- 
traneous light into the path of the refracted beam was 
blacked. Some of the harmful sources of internal reflection 
were found to be the bounding surfaces of the objective lens, 
the prism faces and the analyzing slit, principally the surfaces 
of the lens.1 For example, in looking into the forward end 
of the objective tube a number of small images of the spectrum 
can be seen graded in size and nearly in line but projected to 
different distances. These images show of course that light 
is being reflected back and forth in the optical system. The 
amount of this reflection can be lessened to a considerable 
extent by reducing the area (cutting out the edges) of the 
lens and prism surface exposed to the beam of light. How- 
ever, since this method of reducing the amount of internal 
reflection lessens also the intensity of light, only a limited 
use was made of it. If care is not taken to prevent it, light 
from neighboring regions of the spectrum will also be reflected 
back into the objective tube from the surfaces on either side 
of the analyzing slit. These surfaces should, therefore, be 
very carefully blacked. Some good perhaps is accomplished 
by slanting them slightly so that the edges of the slit point 
in the direction from which the light is coming. The effect 
of this is to reflect out of the system all light but that which 
passes through the slit. The edges of the slit also require 
careful attention if reflections are to be avoided. The knife- 
edge should not be obtained by a steeply pitched bevel on 
either side. We have found that very satisfactory edges 
may be made from sections of Gillette razor blades. (Z>) A 
minimum deviation device was attached to our spectroscope 
(see pp. 258-260). This was found to add a great deal to the 
1 The surfaces of this lens being concave on the side opposite to which the light 
enters the lens tend to reflect the parallel waves toward the axis of the beam, and 
thus again to mix the light which had been resolved by the prism. SPECTROSCOPIC APPARATUS 
267 
purity of the spectrum.1 But (c) the simplest and most 
effective device proved to be the use of thin gelatine filters 
especially selected with reference to the bands that were to 
be eliminated. The use of these filters in fact is enough to 
give the desired purity even when no other precautions are 
taken. Moreover, we were able to select these filters so that 
there was so little absorption of the light that we wished to 
use as to be of no consequence for the purposes of our work. 
They were mounted in an especially constructed holder be- 
tween the objective tube of the spectroscope and the ana- 
lyzing slit. A similar result could be accomplished by using 
two spectroscopes in series.1 But the expense and incon- 
venience of doing this make it undesirable in work of the type 
we are doing, more especially when the desired purity can 
be obtained by the simple means described above with 
probably a negligible reduction of the intensity of the useful 
light. 
Apparatus for Varying the Intensity of the Stimulus 
Light 
In designing an apparatus which will be at all broadly 
serviceable in the investigation of color sensitivity, due 
1 Our attention was called to the effect of the minimum deviation by the obser- 
vation that when the prism was adjusted for minimum deviation for the D-line and 
the wave-length was changed by shifting the spectroscope by means of the driving 
screw shown in Fig. 2, the farther the D-line was from the analyzing-slit, the greater 
were the number of bands which appeared in the field when the light emerging from 
the slit was examined by means of a second spectroscope. Without the use of the 
gelatines or a second spectroscope, however, we have not been able to get rid entirely 
of the bands mentioned above. 
1 It is obvious that an advantage is gained here for purity over and above the 
increase in resolving power given by two prisms. Increase in resolving power is 
usually gotten by using a train of prisms in a single spectroscope or a single prism 
through which the light is passed more than once. (The Cassie prism used earlier in out 
work is of this latter type, see this paper, footnote, p. 260.) But while increasing the 
resolving power increases the purity of the spectrum (Schuster, 'Theory of Optics,' 
p. 163 expresses the dependence of purity on resolving power by the formula P = pR 
where P represents purity; R. resolving power; and p, a factor which is a function 
of the slit-width), the impurities caused by internal reflections are clearly of a kind 
that can never be eliminated entirely by an increase of the resolving power of the 
prism or train of prisms. This, so far as we know, can be done only in ways similar 
to those described above. The use of a second spectroscope, however, besides being 
much more expensive and more inconvenient to manipulate, causes a greater decrease 268 
C. E. FERREE AND GERTRUDE RAND 
attention should be paid to adequate means of varying 
intensity. This has become of especial importance because 
of the growing recognition of the usefulness of the threshold 
and just noticeable difference determinations as the basis 
of comparative work. We have designed, therefore, ap- 
paratus to produce both gross and fine changes in the amount 
of light employed, more especially the small gradations 
Fig. 5. Showing Auxiliary Apparatus Designed for the Reduction of the In- 
tensity of Light-specially constructed resistance coils, coarse grating, aluminum 
sectored discs with Vernier protractor, and collimator slits including just noticeable 
difference slit.1 
needed for threshold and just noticeable difference deter- 
minations. This apparatus is shown in Fig. 5. It consists 
in the intensity of light used as stimulus than do the filters mentioned above, which 
when properly selected may produce practically no change in a given region of the 
spectrum and absorb heavily in other regions. 
1 In III. b the disc made up of a single 150 sector (v) is^shown bright side before to 
make it distinguishable from the remainder of the compounded disc. When in use 
it is placed behind the other discs and the black side is turned towards the focusing 
lens just behind the stimulus-opening. 269 
SPECTROSCOPIC APPARATUS 
of (i) two types of collimator slit, one of which is especially 
devised for just noticeable difference work; (2) coarse gratings 
designed to give gross variations in the intensity of light; 
(3) sectored discs suitable for threshold and just noticeable 
difference work in which all of the changes from o°-348.75° 
total aperture may be made, and a protractor with Vernier 
scale to permit of close reading of the discs especially when 
the aperture is small; and (4) special resistance coils designed 
to vary the intensity of light at the source. 
Collimator Slits.-One of these slits is of the usual type 
having carefully beveled knife-edged jaws 12 mm. long, 
with a micrometer adjustment of width graduated to thou- 
sandths of a mm. Such a slit may be used (a) to make a 
reduction of the spectrum as a whole or of any part, and the 
reduction may be computed directly from the slit width 
provided the source is uniformly luminous over the surface 
exposed; or (b) one part of the spectrum may be made in 
turn to sustain, within limits and under the conditions men- 
tioned above, any ratio that may be desired to any other 
part of the spectrum, providing the original intensities from 
which the reductions are made, are known. This slit, how- 
ever, is not adapted to just noticeable difference determina- 
tions for a given color or range of wave-lengths. The second 
slit (shown at I, Fig. 5) is especially designed for just notice- 
able difference work. This slit is constructed so that its 
upper and lower halves are independently variable in width. 
It was designed especially for some new methods we are using 
for a quantitative comparative determination of the retina's 
inertia to the different wave-lengths of light in which just 
noticeable differences are employed. In the short exposures 
used in tracing the sensation from the threshold to its maxi- 
mum, it is obvious that the sectored disc could not be em- 
ployed in making the variations needed for just noticeable 
differences. While designed to meet this special need we 
have found it to be a very convenient means of making the 
variations needed for much of the general work in just 
noticeable difference determinations. The slit is formed by 
three knife-edged jaws. That is, one jaw of the slit is made 270 
C. E. FERREE AND GERTRUDE RAND 
in one piece and is 12 mm. long; the other jaw is made in 
two pieces, each 6 mm. long. One edge, the upper for the 
upper jaw and the lower for the lower jaw, is beveled to fit 
into a dovetailed guide cut in the frame. The other edge of 
each jaw is held in place by a slender close fitting key of 
appropriate length. The jaw, 12 mm. long, is stationary 
and the other two jaws are made to move away from it by 
two independent micrometer screws operated by drum heads 
graduated to thousandths of a millimeter. In operation the 
source is adjusted so that one edge of its equally luminous 
surface is flush with the stationary jaw and the other jaws 
are pulled away from it exposing as desired different widths 
of this surface. In a just noticeable difference series one half 
of the slit is held constant and the other is varied to give the 
just noticeable difference. When the width of slit needed 
for this is obtained, that half of the slit is held constant and 
the other is varied, and so on until the series is completed. 
The Coarse Grating.-This device is serviceable for gross 
reductions of the spectrum as a whole. It is shown at II. 
in Figs. 5 and 2. The grating consists of an exposed photo- 
graphic plate ruled on a dividing engine with lines 60 to the 
centimeter. The gelatine side of the plate is covered with a 
thin glass plate and the whole is mounted in a brass frame 
supported by a slender rod running parallel to the lines of 
the grating. This rod fits into a collar furnished with a set 
screw so that the grating may be rotated any amount that 
is desired about the axis of the rod. When interposed in the 
path of the light this grating allows the maximum amount 
of light to pass through when it is perpendicular to the path. 
As it is rotated, less and less light gets through the open spaces 
afforded by the lines, the amount depending upon the angle 
of rotation. If it were wanted to use this grating as a more 
precise instrument of reduction, it would be comparatively 
simple to add a graduated scale so that any previous setting 
might be reproduced, and to calibrate the scale so that the 
amount of reduction might be read directly from it. So far 
in our work we have not felt the need to do this as the grating 
has been used only in making gross reductions in the amount SPECTROSCOPIC APPARATUS 
271 
of light employed, the finer changes being made either by 
the slit or by the sectored discs. This grating may be 
mounted anywhere in the path of the light from the source 
to the campimeter screen. Thus far we have found it most 
convenient to insert it between the collimator and the prism 
(see Fig. 2). 
The Sectored Discs and Vernier Protractor.-Probably 
the most convenient and widely applicable apparatus for 
varying the intensity of light by known amounts is the 
sectored disc. A strong objection to the use of the sectored 
disc when fine changes are needed such as are required, for 
example, in threshold and just noticeable difference work, is 
the difficulty of getting and of measuring accurately suffi- 
ciently small amounts of change. Such discs are ordinarily 
constructed with two or more open sectors, and a change in 
one is multiplied as many times as there are open sectors. 
Moreover, an error made in the measurement of one sector 
is multiplied by the number of open sectors. This latter 
difficulty becomes especially significant in working with 
intensities at or near the threshold when a small error may 
represent a high percentage of the total open sector. We 
have sought to overcome these difficulties in two ways. 
(1) Our discs for a low total aperture are so constructed that 
one sector may be varied at a time. And (2) a special 
protractor has been designed fitted with a movable arm 
carrying a knife edge and Vernier scale graduated to read 
to 1/60 of a degree. 
Two sets of discs were made in all-one for simple reduc- 
tions and threshold work, and the other for just noticeable 
difference determinations. The discs were cut from hard 
sheet aluminum No. 20 B. & S. gauge, 0.9 mm. thick; and 
are for the first set 19.5 cm. in radius. In order to give a 
wide range of change a number of discs are required. For 
example, a variation of total range of open sector from o°- 
348.75° is obtained in the first set of discs by using four pairs 
of two-sector discs and one pair of one-sector discs with a 
counter-balancing weight. In the first pair the breadth of 
each of the two open sectors is 90°, and the range of variation 272 
C. E. FERREE AND GERTRUDE RAND 
of total open sector is from o°-i8o°; in the second pair the 
breadth of each of the two sectors is 45° and the range of 
variation of total open aperture is from i8o°-27O°; in the 
third pair the breadth of each of the two open sectors is 
22.50 and the range of total aperture is 27O°-3I5°; in the 
fourth pair the breadth of each of the two sectors is 11.25° 
and the range of total aperture is 3i5°-337-5°; and in the 
fifth pair the breadth of the single sector is 11.25° and the 
range of total aperture is from 337.5°-348.75°. In order 
that very small apertures may be obtained or that small 
variations of aperture may be had, when each open sector 
is 15° or less, a small single sector furnished with counter- 
balancing weight was provided in addition to the five pairs 
of discs. When this is used with the pair of discs having 90° 
sectors to cover one of the open sectors, the total aperture 
may be varied from o° through 15°. This single sector may 
also be used in connection with any of the other pairs of 
discs to aid in making smaller variations in the total aperture 
than is obtainable with the pair of discs alone. That is, 
as one of the open sectors is opened, the other may be closed 
by any desired amount less than 15° by means of this single 
sector. Following this principle broader single sector? may 
be constructed to permit of smaller variations when the total 
aperture is still larger. We have found, however, that the 
15° sector satisfies the need for the values of total aperture 
for which small changes are significant.1 
In order to make the first set of discs serve also for just 
noticeable difference determinations, it was necessary to 
supplement each of the pairs of discs in this set with single 
discs of the same breadth of sector and of lesser radius. 
1 In the construction of these discs a solid disc with a radius of 19.5 cm. was 
first cut from a sheet of aluminum. The open sectors were then cut into these discs 
of the breadth desired to a depth of 13.5 cm. This left a small solid disc of 6 cm. 
radius to support the pairs of sectors. As was stated above in order to guarantee 
symmetry of rotation in case of the single sectors, a counterbalancing weight was 
fastened on the opposite side from the sector 6 cm. from the center of rotation. This 
weight was in the form of a small lead disc soldered to the inside supporting disc in 
line with the radius which just bisects the sector. It is scarcely needful to say that 
great accuracy is demanded in the cutting of the discs. This accuracy was obtained 
by means of a special cutter designed for cutting with accuracy straight and curved 
edges in metal. SPECTROSCOPIC APPARATUS 
273 
These discs were 17 cm. in radius and the open sectors were 
cut in from the outer edge to a depth of II cm. A margin 
of 2.5 cm. was thus left between the edges of the two sets of 
discs. In making the just noticeable difference determina- 
tion the two sets of discs are adjusted so that the edge of the 
inner disc just bisects the stimulus-opening. The openings 
of the two sets are then varied independently as may be 
required, it being necessary of course always to make the 
open sectors of the outer discs the larger. The sectored discs 
are shown at III. a in Fig. 5. The method of using them is 
further illustrated at III. b. Here the discs are shown 
mounted on an electric color mixer and are set for a just 
noticeable difference determination low in the intensity scale. 
Two of the set of larger discs each having two 900 sectors are 
mounted to give two open sectors one of which is closed by 
means of the single 150 sector shown at x, leaving a total 
aperture of io°. In front of these is mounted the disc of 
smaller radius with the edge of one of its sectors shown at y 
projecting 40 into the io° opening reducing it to 6°. When 
the disc so compounded is adjusted in front of the stimulus- 
opening so that the edge of the disc of shorter radius just 
bisects this opening and the disc is rotated at the fusion rate, 
the two halves of the opening are illuminated with light of 
intensities proportional respectively to 10 and 6. By using 
the discs of different breadths of sector, similar variations 
can be achieved over a range o°~247-7S° open sector. 
The special protractor by means of which the width of 
the open sectors may be read to 1/60 of a degree is shown at Z. 
This protractor is provided with a 1800 arc of 10 cm. inside 
radius, and an arm 27 cm. long which rotates about a central 
collar closely fitting the chuck of the motor. The 1800 arc 
is graduated to 1/4 degrees and the movable arm carries a 
Vernier scale graduated to 1/60 degrees. The movable arm 
carries a beveled slot also of a 10 cm. radius of curvature 
into which the 1800 arc fits. To insure accuracy of setting 
this movable arm is provided with finely beveled straight 
edges. When making a measurement of open sector, one 
of these edges is set flush with one of the edges of the sector 274 
C. E. FERREE AND GERTRUDE RAND 
and the reading made. It is then rotated until the same edge 
is flush with the other edge of the sector, and the difference 
between the readings is taken as the value of the sector. We 
have thought it necessary to stress the accuracy with which 
these measurements must be made because the threshold 
value of sensation is frequently obtained for the intensities 
of light employed by us with a total aperture of i/i 50. With 
such small apertures it is obvious that accuracy of measure- 
ment becomes of prime importance. When used in connec- 
tion with the spectroscopic apparatus described in a pre- 
ceding section to determine the threshold and just noticeable 
differences in sensation, the sectored discs are interposed in 
the path of the parallel beam of light just behind the lens 
Z2 (see Figs. 2 and 3). As is shown in this figure in order 
to eliminate as far as possible all vibrations and conse- 
quent displacements of the edges of the discs from the de- 
sired alignment with the beam of light, the discs are mounted 
on a motor (S) suspended by springs. With discs so designed 
and used, and with the proper standardization of the factors 
which influence the response of the eye, determinations may 
be made having a very high degree of reproducibility. 
Special Resistance Coils.-Special resistance coils have 
been devised which serve the following purposes: (1) to give 
the fine changes of resistance needed to compensate for 
fluctuations of voltage in the lighting circuit which otherwise 
might produce troublesome variations in the flux of light 
from the Nernst filament; (2) to produce changes in the 
intensity of the spectrum given by the filament;1 and (3) to 
make possible the fine changes in the speed of rotation of the 
discs that are so frequently required in work in the optics of 
color. These coils are of two general types, {a} Coils which 
give fine changes over a narrow range; and (b) coils which 
will permit of fine changes over a wide range. Coils of the 
first type are shown in Fig. 5 at IV. a and b\ a coil of the 
second type at IV. c. The resistances of the first type con- 
sist of one coil and are constructed to give the effect of a 
1 It was recognized of course in using the resistance for this purpose that a change 
in the amount of current by which the Nernst is operated changes the spectro-radio- 
metric composition of the light. SPECTROSCOPIC APPARATUS 
275 
contact sliding along a single wire. The effect is produced 
by winding the coil of wire of the desired size and resistance 
on a hollow brass cylinder insulated with micanite, and by 
turning this coil by means of a screw motion under a U-shaped 
contact. In this way the contact is made to travel along the 
entire length of the wire, and the fineness of change is limited 
only by the size and coefficient of resistance of the wire. 
In one example of this general type of resistance shown in 
Fig. 5 (IV. a), the contact is kept stationary and the cylinder 
is mounted on a rod as its axis, threaded at both ends. As 
this rod operated by a milled head turns, the cylinder rotates 
and slowly advances, so that the contact travels continu- 
ously over the whole length of the wire. In a second example 
of this type, shown at IV. b, the contact slowly advances as 
the cylinder is rotated, and passes continuously over the 
whole length of the wire. This effect is accomplished as 
follows. The cylinder is mounted on a horizontal rod turned 
by a geared wheel 5 cm. in diameter. The contact is mounted 
on a threaded rod which is turned by a geared wheel of the 
same diameter as the wheel which turns the cylinder. As the 
latter wheel turns, its teeth engage the teeth of the wheel 
which rotates the rod on which the anchor piece of the con- 
tact is threaded, and the contact advances along the rod at 
a rate which keeps it continuously in touch with the wire 
throughout its whole length. The coil which we use to con- 
trol the amount of current operating the Nernst filament is 
made of No. 26 wire, 47.8 ft. long, and has a resistance of 55 
ohms. 
The resistance of the second general type consists of two 
coils in series, one designed to make gross changes and one 
to add fine changes to it for any given setting of the contact. 
The wire of the first coil is wound on a long section of brass 
tubing insulated by micanite. This tube was given a station- 
ary mounting and the changes in resistance are produced by 
a sliding contact shown at v. The supplementary coil was 
wound on a brass drum shown at w, insulated by micanite 
and mounted on a vertical threaded rod operated by a 
small wheel. As this wheel is turned the drum rotates and 276 
C. E. FERREE AND GERTRUDE RAND 
slowly changes its level bringing every point on the wire 
successively in touch with a stationary contact. The con- 
nections are so made in this rheostat that by means of a 
double-pole switch shown at t, the poles of the large coil 
can be reversed and a high rate of speed can be instantly 
changed to a low rate of speed, and vice versa. In the rheostat 
shown in Fig. 5, the large coil is made of Advance wire No. 26, 
139 ft. long, and has a resistance of 160 ohms. The small 
supplementary coil is made of wire of the same composition, 
No. 26, 17.4 ft. long, and has a resistance of 20 ohms. By 
means of these two coils changes of resistance amounting to 
very small fractions of an ohm can be made and the speed 
of rotation of motors of the type constructed by the C. H. 
Stoelting Co., for example, can be controlled to fractions of a 
revolution per second. We have found such control to be 
very useful in the general field of work in which impressions 
are to be g ven to the eye in succession, and especially neces- 
sary in the studies that we have made of the factors that 
influence the results obtained by the method of fl cker for 
the photometry of lights of different color. It is obvious 
that some such control must be had if the phenomena pro- 
duced by changing the rate at which impressions are given 
to the eye are to be studied in satisfactory detail. 
The Rotary Campimeter 
In a previous paper1 a rotary campimeter was described 
especially devised for use with pigment papers. At that time 
it was stated that this apparatus had also been adapted for 
use with a spectroscope, and that a description of it would 
be given in a 'ater paper. The apparatus we shall here 
describe has been in use, therefore, for three years and its 
feasibility has been tested for that length of time both in 
the research and drill work of the laboratory. 
The object of the rotary campimeter is to add to the verti- 
cal campimeter the rotary features of the perimeter, and thus 
to allow investigation of every possible meridian of the retina 
1 C. E. Ferree, 'Description of a Rotary Campimeter,' American Jour, of Psychol., 
1912, 23, pp. 449-453. SPECTROSCOPIC APPARATUS 
277 
with as much ease and precision as was possible with the old 
form of campimeter in the nasal meridian only, or at most 
in the nasal and temporal merid'ans. As designed for use 
with the spectroscope this apparatus consists of two parts 
with appropriate supports and accessories: camp meter screen 
and lens to focus the light on the pupil of the eye and to shift 
the image to follow the pupil as it takes an excentric fixation; 
and attachment to line up the eye with the stimulus-opening. 
The campimeter screen rotates on a brass collar around a 
circular support. The stimulus is exposed through an 
opening in the center of the campimeter screen. Behind 
this opening mounted in a smaller brass collar is the focusing 
lens (£2)- This lens is carried on a rack and pinion which 
moves its center back and forth along a line which passes 
through the center of the stimulus opening and which con- 
tains all the fixation points on the arm {I-I'). 
Fig. 6 shows the skeleton apparatus. It consists of the 
following parts: supporting base, frame for campimeter screen, 
Fig. 6. Showing the Rotary Campimeter, the Lens for Focusing the Light on 
the Eye, and the Rack and Pinion used to shift the focal point as the eye changes its 
fixation. 
rack and pinion adjustment and support for the focusing 
lens, and attachment for lining up the eye with the stimulus- 
opening. The supporting base consists of a horizontal steel 278 
C. E. FERREE AND GERTRUDE RAND 
bar, 83 cm. long, supported by two tripod rests (B and Bf). 
To this bar is clamped an upright (c) which serves as a sup- 
port for the framework of the campimeter screen. In order 
that the distance of the upright of the screen above the table 
may be adjustable, this upright consists of a steel tube 27 cm. 
long and 15 cm. in diameter furnished at its upper end with a 
collar and set screw into which fits a rod 20 cm. long, to which 
the framework of the campimeter screen is attached. The 
framework of the campimeter screen consists of a stationary 
brass ring about which rotates a larger brass collar (//), 
20 cm. in diameter.1 
The back circumference of collar (7/) is graduated from 
o°-36o°. To this collar are fastened the radiating arms. 
There are eight of these arms, one for each 450 mark on the 
graduated collar. They are made of steel and are 2 cm. broad 
and 40 cm. long. The eighth arm (I-I') differs from the 
other seven. It forms a right angle, one side of which lies 
in the plane of the background, and the other in front of this 
plane. The part in the plane of the background is 60 cm. 
long and the part at right angles to this plane is 28 cm. long. 
The arm is graduated from i8°-57° along the section that 
lies in the plane of the background, and from 57°-92° along 
the section at right angles. The graduations are based on 
the arc of a circle of 25 cm. radius. The rack and pinion 
adjustment (R') which carries the focusing lens is attached 
to the rotating collar. Thus when the arm carrying the 
fixation points is rotated into any given meridian, the rack 
and pinion adjustment is also rotated so that the line of 
motion of the center of the lens always contains the fixation 
points. The focusing lens (Z2) is a double convex lens 50 
1 This ring was made large in diameter for two reasons, (a) The ring had to be 
made thick in order to give sufficient rigidity to support the campimeter screen and to 
furnish the proper attachment for the rotary collar. Had the circumferance been 
made small, the effect of the ring would have been that of a short tube. If the stimulus 
were viewed through a short tube, an induction factor would have been involved which 
would have been difficult if not impossible to standardize. The opening of the ring 
was, therefore, made considerably larger than any stimulus we wished to use in order 
to avoid the introduction of this factor, (b) The large circumference of the ring makes 
the apparatus available for investigating the effect upon sensitivity of varying the 
size of the stimulus. SPECTROSCOPIC APPARATUS 
279 
mm. in diameter and with a focal length of 275 mm. It 
moves in a plane 2.5 cm. behind the stimulus-opening, hence 
the parallel rays of light entering it from the collimating 
lens (£1) are brought to a focus on the pupil of the eye when 
in position 25 cm. behind the stimulus-opening. To the eye 
at this point, the lens, or as much of it as is visible through 
the stimulus-opening, is seen uniformly filled with light. 
That is, it is a well-known fact in physiological optics that 
when parallel rays of light are focused on the pupil of the 
eye, or more accurately perhaps, at the optic center of the 
refracting mechanism, by means of a double convex lens, 
the lens is seen by the eye as if uniformly filled with light. 
As the eye takes the different fixation points on the arm 
(I-I') the light is kept focused on the pupil by slightly 
displacing by means of the rack and pinion adjustment the 
center of the lens in the direction in which the eye is turned 
in taking the new fixation.1 The adjustment is quickly and 
easily made. In fact the rotary campimeter adapted to the 
spectroscope in the manner described in this paper presents 
little if any more difficulty of operation than it does when 
pigment papers are used as stimuli. 
The device for lining up the eye with the stimulus-opening 
is also attached to the rotary collar (77) and is constructed 
as follows. A cross piece (S) 16 cm. long and 1.8 cm. wide 
is fastened by a screw with a milled head to one side of the 
collar (77) and is supported by a pin on the opposite side. 
A pin is used instead of a second fastening in order that the 
device may be conveniently and quickly turned out of the 
path of light when not in use. In the center of the cross 
piece (S) is a circular opening (0) 15 mm. wide. When in 
position immediately behind the opening in the campimeter 
screen which admits the stimulus light, the center of this 
opening lies in a line perpendicular to the opening in the 
screen at its central point. To the cross piece (S) 3 cm. to 
1 In taking a new fixation the pupil is seen to turn from under the colored image 
and to the observer the stimulus-opening is no longer filled with colored light. As 
the center of the lens is shifted in the appropriate direction, however, the image is 
seen to travel towards the pupil, and when it falls full upon it, the observer again 
sees the stimulus-opening filled with light. 280 
C. E. FERREE AND GERTRUDE RAND 
the side of the opening (0) is fastened at right angles an 
arm 12 cm. long and 1 cm. wide, terminating in a disc (P) 
2 cm. in diameter. Three centimeters from its outer end this 
arm is bent at right angles so that the disc lies directly behind 
the opening (0) and in a plane parallel to that opening. 
The size and position of the stimulus-opening, the opening 
(0), and the disc (P), and their distances from each other 
sustain such relations that when the eye is in position 25 cm. 
behind the stimulus-opening with the center of the image of 
that opening on the fovea and the line of regard normal to 
the plane of the opening, the edge of the opening (0) is just 
contained within the stimulus-opening and the edge of the 
disc is just contained within the opening 0. That is, in effect 
the device is a peep sight arrangement, and the alignment 
described above is possible only when the eye is at the center 
of curvature from which the fixation points on the arm 
(I-P) are determined. As stated above, this attachment is 
fastened to the collar (IP) by means of a screw with a milled 
head, so that after the alignment is made it can be readily 
turned out of the road and clamped. In order that the 
distance of the eye may be adjusted at the same time as its 
alignment is made with the stimulus-opening, a measuring 
device 25 cm. long is provided. This device consists of a 
slender brass rod fitted at either end with two short right 
angled arms 5 mm. in length. On the end of one of these 
arms is a ring which is just larger than the stimulus-opening, 
and on the other is a brass disc of the same diameter as the 
ring, provided at its center with a pupillary aperture. In 
adjusting the distance of the eye the disc is rested lightly 
against the forward surface of the eyeball and the ring 
against the campimeter screen concentric with the stimulus- 
opening. When the position of the eye is once determined, 
a mouthboard is adjusted and clamped in position so that 
the observer's teeth fit into impressions previously made and 
hardened in wax. This fixes the relation of the observer's 
eye to the campimeter system. All that is needed, therefore, 
at subsequent sittings to bring the eye into this relation is 
again to fit the teeth into the impressions on the mouthboard. SPECTROSCOPIC APPARATUS 
281 
In order to facilitate excentric fixation in the nasal and 
temporal meridians, the head should be turned in adjusting 
the mouthboard 450 nasalwards or temporalwards as the case 
may be. With the head so placed, the eye can swing easily 
from the stimulus-opening to a fixation point whose ex- 
centricity exceeds 900. 
The front view of the campimeter in readiness for use 
may be seen in Fig. 2 of our former paper; a back view is 
given in Fig. 3 of this paper. A cardboard background has 
been fastened to the steel arms by means of metal fasteners 
pushed through holes in the steel arms and clinched. Since 
the background is fastened to the arms attached to the brass 
collar (H), a circular gap is left at its center. This gap is 
filled by a disc (A) shown in Fig. 3, which has been fastened 
to the arms just outside of the collar (//). The disc is 27 cm. 
in diameter and contains the stimulus-opening (0), the size 
of which may be varied to accord with the purpose of the 
investigation. In order to complete the graduations on the 
fixation arm to the stimulus-opening, disc (A) is graduated 
from o°-i8°. A background, 40 cm. in height, is fastened to 
the extension arms (/'). This background for screen and 
extension arm may be covered with whatever standard paper 
or surface that is desired.1 The graduations from oo~92° 
are pricked in this covering at points determined by the 
markings on the back of the disc (A) and the arm (/-/'). 
These constitute the fixation points. 
The method of using this apparatus is as follows. The 
eye of the observer is lined up with the stimulus-opening by 
means of the attachment described above at a point 25 cm. 
from the campimeter screen; and the position of his mouth- 
board is adjusted. With his eye in this position the lens 
should fill uniformly with light whenever the image of the 
analyzing slit falls on the pupil. Before the color observa- 
tions are begun this is tested out by taking a number of 
1 In all tests of the relative and absolute sensitivity of the retina this screen 
should be made of a gray of the brightness of the color to be used. No departure from 
this rule should be permitted in tests of sensitivity unless it is for the purpose of de- 
termining the effect of different screens on sensitivity, or of using this effect as a means 
of varying sensitivity. 282 
C. E. FERREE AND GERTRUDE RAND 
fixation points from o° to the periphery of the field of vision 
and adjusting the focusing lens in each case to bring the 
image of the slit full upon the pupil of the eye. In making 
the color observation the unused eye is covered with a 
bandage. The arm (I-I') is turned into the meridian to be 
investigated, the position being determined by the gradua- 
tions on the collar (H). The experimenter inserts a card 
which we shall call the preexposure card, between the eye 
and the stimulus-opening as near to the opening as possible 
while the observer takes the fixation required. In all in- 
vestigations of relative and absolute sensitivity, this card 
should be made of a gray of the brightness of the color to be 
used. At a signal given by the observer the preexposure 
card is withdrawn, the eye is exposed to the stimulus for the 
required length of time and the card is replaced in the path 
of the stimulus light. The observer is required to rest the 
eye after each observation. Further provisions against fatigue 
are made by frequent and regular intervals of rest. 
It is often desirable to have an equation representing 
what is sensed at a given point in the peripheral retina in 
terms of what is sensed in the central retina. In this way a 
representation may be had for comparative purposes of the 
color tone, brightness and saturation for light of a given 
intensity and range of wave-length.1 In the campimeter 
devised for use with pigment papers provisions were made 
for this. That is, it was possible to rotate a small disc on 
which the representation might be made, just behind the 
fixation point for all positions of the point from o°-9O° in 
any meridian. This disc was rotated by a small motor the 
shank of which protruded through a slit 8 mm. wide running 
the full length of the fixation arm (I-I'). For a further 
description of this motor, its supports, adjustments, etc., 
and the method of making the match of the sensations 
aroused in central and peripheral retina, see the former 
1 In making a comparative study of the sensitivity of the different parts of the 
retina, such an investigation is a valuable supplement to a survey of sensitivity made 
on the basis of threshold and just noticeable difference determinations. The two 
methods are needed in fact to give a complete representation of the sensitivity of the 
retina in its quantitative and qualitative aspects. SPECTROSCOPIC APPARATUS 
283 
article, p. 452. In the rotary campimeter devised for use 
with the light of the spectrum no similar provision has as 
yet been made. The difficulties appending the attempt to 
get a spectrum light variable at will in intensity, color, and 
brightness which can be easily and conveniently presented 
to the eye at any point of any meridian of the field of vision 
are obvious at a glance. Dreher, for example,1 was able to 
make a match of the sensations aroused at the center and 
periphery of the retina without very inconvenient changes 
in his apparatus for one point only in the peripheral retina. 
An attempt is being made to adapt our present apparatus so 
that this match can be accomplished conveniently for any 
point in the peripheral retina, but as yet nothing definite can 
be promised. 
1 Dreher, loc. cit. • f r /"T 'V 1 J 
.. . . 
'V- ■ 
' ' : ."A fe'- - '..< - - ' '". 
'■ '- ■• T . ; >r-...-■■<'■■' - ~' - '-J..-' .'■■■ > - ■-• • :> 
• ■ • ■ -''M 
■ " fWi- ■■ - ■ ... 
' BRYN MAWR COLLEGE MONOGRAPHS 
The Bryn Mawr College Monographs are issued in two series: the 
Monograph Series, containing studios that appear here for the first time, 
and the Reprint Series, containing reprints of article^ that have appeared 
in other journals. 
Of the 'Reprint Series, volumes I, II, III, V, VI, VII, IX, and X contt}in 
contributions from the biological laboratory; volumes IV and VIIt, contri- 
butions from. the mathematical and physical departments, and volume XJ, 
from the psychological laboratory.