(Return to index)

THE experiments made on vision dealt with the following subjects:

    A. The threshold for light.
    B. Discriminative sensibility for brightness.
    D. Discrimination of color.
    E. Discrimination of visual areas.

The apparatus used in the experiments on the sensitiveness of the retina to light was a long wooden tube about eight inches square at the ends and four feet long. It was blackened on the inside and was made absolutely light-proof. At one end was a box-like cover under which the subject could sit, with his eyes on a level with the tube. When the subject was in position, the box was covered with a camera cloth in such a way that no light could reach his eyes. At the other end of the tube was a round opening one inch in diameter. It was found impossible to reduce white light to the threshold. The opening was therefore covered with violet glass. Since all light appears as mere brightness in its lowest intensities, the color of the glass was indifferent in the present case. The glass was held in place by a box-shaped cap which fitted over the end of the tube. The circular opening was closed by a round black disc. When the disc was in [p. 77] position no light whatever could reach the eye of the subject. The only possible way to stimulate the retina was to move the disc away from the opening in the tube.

The experiments were made in a completely dark room. The source of light was a Welsbach burner. One of the great difficulties in making experiments on the threshold of vision has been to find some way of reducing the light by measurable amounts. In this case no attempt was made to reduce the illumination itself. A source of light which would remain constant was all that was required, and other means presently to be described were taken to lessen the intensity. A year's experience with the apparatus previous to using it in this test was sufficient to convince us that the Welsbach burner, under full gas pressure, does furnish a constant source of light. The supply of gas is always sufficient to illuminate the mantel to its full extent, and that insures the maximum of light which the burner affords. It was found that the same threshold -- allowing for variations in temporary condition -- could be established for a given subject day after day with this apparatus. Assuming therefore that what variations of intensity there were in the source of light were beneath perception when applied in this way, the intensity was reduced to the amount required by shading the light in various ways and reducing reflection in the room. A few inches in front of the opening of the tube was placed a black cardboard screen at a given angle with the box and with the burner. The burner was placed opposite the black screen about five feet away, and shed its light on the screen, from which it was reflected into the box whenever the open-[p. 78]ing was exposed. The burner itself was placed in a case with a window opening toward the screen. This reduced reflection from the walls of the room. By this means the amount of light admitted to the box was lowered to a point near the threshold. The further diminution of intensity was accomplished by a series of cheese-cloth curtains which hung in front of the window of the case containing the burner. Every curtain lowered reduced the illumination of the black screen and consequently the amount of light entering the box. Since all the other factors in the situation remained constant the absolute sensitiveness of the retina could be measured by the number of curtains necessary to reduce the illumination to the least visible amount.

There were two time factors in the experiment which it was necessary to keep constant. One was the length of time the eyes were rested in the dark box before the experiment began, and the other the time of exposure of the light for each stimulation. The former was important because the sensitiveness of the retina increases fast on being completely protected from light. If the experiments were made a few minutes after the subject was put into the apparatus, the threshold found would be much higher than that found half an hour later. It was observed that after an hour the sensitiveness increased little if any. Consequently, the subjects were left entirely without stimulation of the retina for thirty or forty minutes. This time was used for determining the touch threshold (see chap. iii, sec. A, 1) and the discriminative sensibility for area on the skin (see chap. iii, sec. B, 2). At the end of that time the experiments on the eyes were begun, and were completed in twenty or twenty-[p. 79]five minutes. The sensitivity registered is that which obtains after protection of the retina from light -- except the minimal stimulations of the test -- for an hour.

The second time element -- the duration of the single stimulations -- is important because a very faint light may be visible when exposed for a longer time but not visible when exposed for a shorter time. This time interval was controlled by a mechanical contrivance for raising and lowering the disc covering the opening in the dark box. The disc was fastened by a projection at one side to the end of a vertical rod, in such a way that when the rod was moved up about half an inch, the disc was thrown down, uncovering the opening. The other end of the rod was joined by a pivot to the end of one arm of a lever which was mounted on the table. The fulcrum of this lever was a ball and socket bearing at its middle point. The downward pressure of the rod on one arm was balanced by a movable weight on the other arm. By means of this adjustment it was possible to bring the system into a state of equilibrium such that the lever arms would remain as they were placed. When the arm connected with the rod was moved up by pressing down on the free lever arm, the disc was thrown down and the box opening remained uncovered. It was closed again at the end of the required interval by means of a metal ball which rolled down a trough and into a second short trough which was fastened to the upright rod. A catch was arranged which held the balls in place until time to release them. When the free lever arm was pushed down, it removed the disc from the box opening, and at the same instant released [p. 80] the catch which held the ball. The opening remained uncovered until the bail rolled down the trough and into the short trough borne by the rod. The weight of the ball then carried the rod down and thus threw the disc up over the opening again. The duration of each stimulation was the time required for the ball to roll from the catch to the short trough employed and its inclination remained the same, the ball's time remained practically constant. Its variations were far within the limits of the time error for this experiment. The troughs were padded to make them noiseless. A padded inclined plane was arranged which received the ball as it left the trough and returned it to the operator.

The series of tests was begun with a light which was clearly above the threshold, to make sure that the subject was familiar with the stimulus. The stimulations were given in series of ten at irregular intervals. The subject was required to count aloud when he saw the light. As soon as the threshold was approached, two or three series were given for each grade of intensity. Rests of several minutes were allowed between series. When the subject was troubled with retinal activity which obscured the field of vision with clouds of gray or colored light, longer periods of rest were given for these to subside. As in other threshold tests, imaginary stimulations were inserted in the series [p. 81] by most subjects. If the tendency proved extreme, the subject was told of it, and in such cases it invariably decreased. The point taken as a threshold was the least intensity -- measured in terms of the number of curtains lowered -- at which three-fourths of the stimulations could be correctly counted, regardless of the imaginary ones inserted.

The results (Fig. 45) show an appreciably greater sensitiveness of retina, in the men than in the women. Men are decidedly more numerous in the region of greatest sensitiveness (six to seven curtains), and less numerous in the regions of slight sensitiveness (none to three curtains).

The Bradley color wheel, with two sizes of black and white discs, was employed for testing the discrimination of brightnesses. The smaller circle was made the standard. It remained half black and half white throughout the test. The proportion of black and white in the larger circle was shifted until the least amount of black necessary to make the outer ring appear darker than the inner circle was discovered. A disc with a circle divided into one hundred parts placed behind the large discs served to measure the amount of black added to the outer circle. The record was kept in terms of the percentage of black required in the outer ring to make it just perceptibly darker than the inner circle.

To insure a constant illumination, the tests were made in a dark room, and the light was furnished by a Welsbach burner placed at a fixed distance behind the subject in such a way that the light came over the [p. 82] left shoulder. The subject was seated facing the wheel. His eyes were kept closed during the shifting of the discs, and were not opened until the wheel was in full motion again. The only direction given the subject was that he was to tell each time whether the inner circle or the outer was darker. To avoid the error of having the outer circle constantly the darker, frequent tests in which the outer circle was lighter were inserted in the series. The test began with clearly distinguishable differences, and worked down, shifting back and forth, to the finest possible discrimination. Three out of four correct judgments served to fix the limit of discrimination. The discrimination was measured to fourths of 1 per cent.

The results (Fig. 46) show a better discrimination on the part of the men. Their curve is above that of the women in the region of fine discriminations (51-53 per cent.) and below it in the region of coarse discriminations (54-55 per cent.).

The apparatus used for testing keenness of vision consisted of cards, one black and one white, about four inches long and five inches wide, on each of which were pasted little squares (2 mm.), of red, blue, green, and yellow. The black card had also a white square [p. 83] and the white card a black square. The five squares were pasted at equal distances along the center of the cards. The subject was tested first with the white background, and second with the black background.

The tests were made in a dark room. The card was illuminated by the light of a Welsbach burner placed at a fixed distance from it. A screen behind the light protected the subject's eyes from it as he approached. The card was hung on the wall on a level with the subject's eyes. From the point below the middle of the card, a 5 m. line was marked across the room on the floor. The subject was placed at the end of this line, with one eye bandaged. The card was hung upon the wall, and the subject was asked how much he could see on it. He was then directed to approach slowly, telling at each step how much he could see, until all the squares and their colors were visible. The point at which each square and each color became visible was noted. With thirteen subjects of each sex the right eye was used first, and with twelve the left eye. The spots were, of course, visible much farther away than their colors. In order to make the subject careful in observing, he was asked about the appearance of the spots at each step, and particularly whether or not they all looked gray or black. After testing one eye, a pretense was made of changing the card for the other eye, in order that the subject might not be influenced by a knowledge of what was really on the card.

The test did not prove to be altogether a satisfactory one, because the maximum distance from the [p. 84] card, 5 m., was not sufficient to make the spots disappear to the normal eye, on either background. The blue spot on the black ground, and the yellow spot on the white ground, were not visible to most subjects at 5 m. The colors on the white card were very rarely visible at the end of the line. On the black card the red and green could be distinguished by a number of subjects.

The results of these tests bearing on the threshold for the perception of the spots appear in Table X. About all that can be gathered from this table is that there are more men than women with weak left eyes. The records for the blue on the black ground, and for the yellow on the white ground, do, however, afford some means of judging of the status of those who, in the other combinations, fall beyond the limits of the test, i. e., 5m. In these two instances, the right-eye test shows the men somewhat better than the women, and the left-eye test shows them very slightly inferior.

The results bearing on the threshold for the recognition of the colors of the spots are shown in Table XI. In this table the combinations, black on a white background and white on a black background, do not appear, because for them the threshold of perception was the only one obtainable. The difference between the sexes is more marked in the recognition of color than in the perception of the spots. In the latter respect the right eye is superior among the men; in the recognition of color it is superior among the women. There are but two instances, blue and yellow on white, in which it is superior in the men. In two more, green on white, and yellow on black,[p. 85]

The test regarding the discrimination of color was made with the larger series of Holgrem worsteds for testing color-blindness. The worsteds were given to the subject heaped upon a gray cloth. He was given a sample to serve as a standard, and was told to select all the worsteds in the pile which were of the same color as the sample. It was carefully explained that the worsteds might differ in shade from the sample, but must not differ in hue. When the required worsteds were selected, he was told to arrange them in order from lightest to darkest. The samples were given in the order -- green, blue, red, and yellow. After each series was selected it was mixed into the pile before the selection of the next one began. The subject was marked "color-blind" only in undoubted cases, where decided oranges were [p. 88] placed with yellows, lavenders with blue, or browns and grays in the color series. When bad mistakes were made, but not bad enough to rank the subject as unquestionably color-blind, he was marked "poor" in color discrimination. The series in which the mistakes were very slight were graded as "medium," and those which were perfect were recorded as "excellent."

The curve for the discrimination of color (Fig. 47) shows a strikingly better color discrimination in the women than in the men. The men predominate in the "color-blind" and "poor" sections, and the women in the "medium" and "excellent."

For the tests on discrimination of visual areas a series of small white squares mounted on large black cardboard squares was used. The size of the black squares, and the position of the white squares on them, were identical throughout the series. The standard square measured 20 mm. on a side. Those for comparison were 19.5, 19, 18.5, and 18 mm. The subject sat at a table facing a black screen. The experimenter stood behind the screen and placed the two squares to be compared in front of the screen successively. The directions were to report which of the two squares [p. 89] shown was the larger. The series began with the larger differences and worked down to the smaller. The point at which three-fourths of twelve judgments were correct was taken as the limit of discrimination.

The curves for the discrimination of visual area (Fig. 48) are very similar for the men and the women. What difference there is is in favor of the men. They are somewhat more numerous in the region of finest discrimination (19.5 and 19.5 mm. +).

There are no tests to compare with the present series on the absolute sensitiveness of the retina to light or on the discriminative sensibility for grays, and none exactly comparable with the test on keenness of vision. Those which are on record employed the method of reading type or numerals. Pearson (69) reports a series of measurements on men and women at Cambridge, in which the average is slightly higher for the men with the right eye and for the women with the left eye. The Columbia University tests (82), in which the same method was used, revealed no difference in sight. The present test happens to coincide with the Cambridge results in showing the right eye [p. 90] slightly better among the men, and the left eye among the women. In both cases, however, the differences are very slight. Krauskopf (43) found a greater percentage of eye defects among girls than among boys. The total mass of results does not warrant us in postulating any sex difference in keenness of vision.

Three distinct problems in color vision have been the objects of previous investigations: the absolute sensitiveness of the eye to color; the discriminative sensibility for color; and the presence of color-blindness. The first of these was investigated by Nichols (62). He prepared series of mixtures of white powder with colored pigments for the four colors, red, green, blue, and yellow. The series varied in intensity from mixtures indistinguishable from white, to clearly colored mixtures. The series for the four colors were placed in glass bottles, and the bottles were indiscriminately mixed. The subject was required to sort them according to hue and shade. Nichols found that men were able to distinguish smaller amounts of pigment than women in all colors except blue, in which women excelled. This result is not in accord with that of the present series of tests, which finds women more sensitive than men to all colors except yellow.

With regard to the second problem -- the ability to discriminate shades of a single color -- all observers agree with our result, i. e., that women are superior to men. Lombroso (51, chap. iii) reports that women are three times as keen as men in distinguishing colors with the Holgrem worsteds, a difference which he attributes to their practice in embroidery. Nichols, in the test reported above, finds that [p. 91] women are better than men in arranging the series of colors according to intensity. Gilbert (30) and MacDonald (55, p. 1106) both report that among school children girls are better than boys in distinguishing the shades of a single color. Luckey (53) Seems to be alone in his doubt about a sexual difference in this respect. He reports experiments on a good many children and a few adults, and says that he finds the color range and the power of discrimination in the primary colors equal for the two sexes.

The investigations which have been made regarding the third problem -- the presence of color-blindness -- may be said to have established without question the fact that this defect is more frequent among men than among women. Jeffries (40, 41, 42) reports tests on large numbers of persons both in this country and in Europe which show a decidedly larger percentage of color-blindness among males than among females. Mullen (59) collected the reports on tests for color-blindness made in the United States, France, England, Denmark, Sweden, Russia, Austria, China, and Japan between the years 1880 and 1897. In all of these reports in which a comparison of sex was made, the percentage of color-blindness was much higher among men than among women. Wissler (82) corroborates these findings in his report of the tests on Columbia University students, and it holds for the present series of tests on University of Chicago students.

No comparative test (other than that of the present series) on the judgment of visual areas is on record. MacDonald (55, p. 1104) reports a test on school children in the estimation of the length of a line.[p. 92] In that, as in the present test the males were somewhat more accurate than the females.

The generalization suggested by the experiments on vision is that on the whole men are somewhat better than women in brightness vision, while women are better than men in color vision. Although no difference between the two in keenness of vision has been established, men's eyes are shown to be absolutely more sensitive to light than women's, and they make finer discriminations of grays. The results as to the absolute sensitiveness of the eye to color are contradictory. Nichols finds it greater in men, and the present test shows it greater in women. There is general agreement, however, that women discriminate color better than men, and are less subject to colorblindness. The tests on visual discrimination of area and on estimation of length show that in this faculty males are somewhat superior to females.