TABLE VI. ASSOCIATION OF POSITION: DISTRIBUTION OF RIGHT CHOICES

The method of learning in these position tests was the same as that noticed in previous experiments, namely, building upon chance successes. When first admitted to the large box containing the row of small ones at the farther end, the animal accidentally found the receptacle containing the food, and later associated the movements involved in reaching that position with various sense-impressions of the box, especially those experienced upon entering—certain tactual impressions of the small entrance compartment, sound of the lifting door and sight given of the interior of the large box.

While the results clearly indicate that pigeons readily learn the position of objects, nothing is proved as to "counting." Some experimenters speak of similar trials as "number-tests," just as they do of "form-tests," but this is probably going too far. To investigate counting in animals, experiments should be arranged which minimize spatial responses. These tests certainly show that pigeons can discriminate positions readily, especially toward the ends of the group, but little more is certainly indicated. Porter[210] says: "If we do not find in birds the power to count, we have in their nice sense for the location of a member of a series ... something of that preliminary number-sense which Ribot describes as belonging to children and savages."

B. Color Tests

To investigate the animals' ability to utilize colors[211] in finding their food, I employed the same apparatus as before, except that six boxes were used throughout and each was covered with paper of a different color: red, yellow, green, blue (Bradley's standards, except red, RO being substituted), gray, and black. The boxes covered with black and gray paper were employed merely to complete the group of six. The same method as before was employed, except that the board to which the boxes were attached was left stationary at the end of the large box, and also that the position of all six boxes was changed irregularly for each test.

The general behavior of the animals at the beginning of these tests was quite similar to that shown in the preceding experiment; but it was soon evident that colors occasioned them far more difficulty than positions. The general distribution of choices is given in Table VII. It will be seen that the proper box was usually chosen more often than any one of the empty ones, but never oftener than the other five combined, as occurred in the position tests; also that in the case of each color there were instances in which another color was as often, or more often selected. Yet it is clear that colors may serve as valuable sense-data for these animals. In the first series of six tests (see Table VIII) there were few right choices or none, but in each succeeding series the number increased. The learning process was evidently of the same type as before observed (selection, in this case gradual, of chance but useful movements), and involved visual data largely.

TABLE VII. COLOR ASSOCIATION. GENERAL DISTRIBUTION OF CHOICES

TABLE VIII. COLOR ASSOCIATION. DISTRIBUTION OF RIGHT CHOICES

There is no evidence that the color-preference of the animals assisted them in choosing correctly, in fact, they were rather less successful in dealing with those colors for which they had previously shown decided preference,[212] since the whole number of right choices was less in the case of the green and blue boxes (85) than in the case of the red and yellow ones (92), and since there was a relative diminution in the rate of learning toward the last in case of the former boxes.[213]

To test the animals' ability to discriminate shades of colors in finding their food, two birds were used, with four boxes, each covered with a different shade of red paper, and two with the boxes covered with green paper. The brightness of the different shades was not measured, but to the eye it seemed to be equal in each of the cases. The food was placed in the box having the most nearly saturated color, and twenty-four trials in series of six, as before, were given each bird. The results were quite similar to those secured with different colors. With the red shades there were twenty-two choices of the best saturated shade to eight, ten, and eight, respectively, of the other three; and with green, twenty-one to nine, ten, and eight. The 43 correct choices were distributed from series 1 to series 4 as follows: 7, 11, 12, and 13, which shows learning as before. The relatively large number of right choices was probably due, partially to the fact that fewer alternative choices were possible since only four boxes were used, instead of six, and partially to the fact that the box containing the food may have been slightly brighter than the others.

Throughout these trials the position-element was a decidedly disturbing factor. When the animals were first learning to choose a box of a definite color, some would show a marked tendency to approach a receptacle occupying a certain position, and would persist in this from series to series. Others at first showed no special preference for certain positions, but, after happening to make a correct choice, they would return to that same place the next time, and thus miss the right box which had been changed for the new test.

C. Form Tests.

In this experiment the six food-boxes were each of different form: triangular, square, oblong, hexagonal, circular, and elliptical. They were of the same capacity, and were covered with light-brown paper. As in the preceding experiment, the birds were tested for only four of the boxes, and were given thirty trials each. Six animals were used, and as they were not the same as those previously employed, the square box (which had always been used before) had no advantage over the others in attracting the birds at the beginning of the trials. The tests were given as in the preceding experiment, except that it did not seem necessary to change the position of each of the six forms before giving each test; it was thought sufficient to move the food-box, and, if a wrong choice had been made in the preceding test, also the box wrongly chosen. The results are shown in Tables IX and X.

TABLE IX. FORM ASSOCIATION. GENERAL DISTRIBUTION OF CHOICES

TABLE X. FORM ASSOCIATION. DISTRIBUTION OF RIGHT CHOICES

It will be seen that each animal chose the right box oftener than any other one box, but not oftener than all of them; also that there was a small increase in the number of right choices from series to series. No one of the four forms seemed better discriminated than the others if we may judge from the practical equality of right choices made in each case (55, 53, 55, 52) or from the similar increase in number of right choices from series to series; the hexagonal and cylindrical boxes received fewer choices in the first series than did the triangular and square, but this was exactly counterbalanced in the last series. The triangular box was more often confused with hexagonal and square, and the square with triangular and oblong, than with the others. For the hexagonal box the cylindrical was more frequently mistaken than were the other forms, especially the oblong; and with the cylindrical the elliptical was more frequently confused than were the others, especially the oblong. In this series of tests nothing new as regards general behavior or method of learning was observed.

TABLE XI. POSITION, COLOR AND FORM ASSOCIATION

If we compare the results obtained in these three experiments (see Table XI and Fig. 7), we shall see that the pigeons were governed much more by the position of the food-box than by either its color or its form, and that color was better associated than form. Position was a most important factor throughout, as was observed also by Porter[216] in the case of the English sparrow. Porter[217] also found that his sparrows could associate color better than form. In the position-tests the pigeons showed very little improvement from series to series (see table); almost all that the animals could learn was acquired at the beginning. The more difficult color- and form-trials, however, showed almost constant improvement, although we should have expected this to be greater in the latter case than it was. When judged entirely by the actual number of right choices in a given kind of tests, some of the birds made a very poor showing; but from the standpoint of increasing number of right choices they appeared in a wholly different light.

Thus, for example, bird F (Table VII) made only 33 right choices in a possible 120, yet their arrangement is significant, being, from series 1 to series 5, respectively, 1, 5, 7, 10, 10. It is probable that there would have been still greater improvement had the tests been continued; perhaps the animal would have become as proficient in finding its food by depending upon the color of the receptacle usually containing it, as by relying upon the position of the box in the group.

V. SUMMARY

1. Respiration in pigeons is sensitive to various stimuli, and since its alterations of rate, amplitude, etc., can be easily recorded pneumo-graphically without frightening the animals, it may well serve as a process through which to study their mental life.

2. By repetition meaningless stimuli, for example, pistol-shots, quickly lose their disturbing influence; whereas the breathing remains sensitive to those of a significant character, such as the noises made by other birds.

3. Reaction to light of moderate intensity consists principally in an immediate quickening, the amount varying with the color; since a direct correspondence was found between color-preference and breathing-rate, it would seem that here agreeable feeling involves increased breathing activity.

4. Visual, acoustical, probably tactual, and certainly organic data, are the principal sensory factors of the associations of pigeons.

5. The animals readily form useful associations by a method of "trial and error," or the selection of successful movements which were at first accidental.

6. Apparently a pigeon does not learn by merely seeing a new act performed by another pigeon; yet there are instances of simple ("instinctive") imitation, and "trial and error" learning is not wholly independent of social conditions, since it proceeds much more satisfactorily if the animal is trained at least within hearing distance of other pigeons.

7. When a habit is being formed, the "period" required for the first test is usually very long, but learning proceeds quite rapidly during the next few trials; later it is more gradual, but it continues till the act becomes thoroughly familiar.

8. Associations are fairly permanent, and some remain practically unaltered for at least six weeks. Modification is easily accomplished, however, on the basis of new experience.

9. Pigeons differ widely both as to the ease with which they acquire associations and also as to their permanence. Difference in activity seems the chief reason for this.

10. While these birds seem mentally inferior to English sparrows and to various mammals which have been tested in a similar manner, they are capable of numerous ready adjustments. They discover circuitous labyrinth passages, they learn to manipulate latch apparatus when adapted to their natural habits and conveniently placed, and they easily reach their food by depending upon the position, color, or form of the box containing it. But the process is apparently simple association throughout. There is no evidence of higher mental activity—no looking the situation over and acting accordingly, no "reasoning" in the proper sense of the word, but only blind movements, some of which are retained and become highly specialized, merely because successful.

REACTIONS OF THE CRAYFISH

Huxley[218] states that crayfish avoid direct sunlight, hiding under stones during the day, and becoming active in the evening. On the other hand, they are attracted like moths to fires lighted on the bank at night, and may be scooped out by hand. Abbott,[219] giving an account of the burrowing crayfish, Cambarus diogenes, states that it is very difficult to observe the animals at work, since all their digging is done at night. It would seem from the account of Miss Hoppin, quoted by Garman,[220] that the blind crayfish, Cambarus pellucidus, is not altogether insensitive to light, for, reporting on the fauna of the caves of Missouri, she says that the crayfish are all found near the entrance to the cave, where there is considerable light. In the dark recesses there are only little white fishes. Blind fish and crayfish are also taken from the wells in the neighborhood, where the crayfish are found only in wells that are rather shallow and light; the fish, on the other hand, are only obtained from deep, dark wells.

According to the above accounts it would appear that the crayfish is negatively phototactic to direct sunlight or diffuse daylight, but positively phototactic to a light at night, and moreover, that light may influence the behavior of the animal even when the eyes have ceased to function.

The directive influence of light upon the movements of the crayfish has never been experimentally studied to my knowledge. Dearborn[221] thinks that light has no effect upon the animals. Yerkes[222] and Towle[223] have shown that Daphnia move toward the light. Bethe[224] finds that Carcinus is negatively phototactic, and also shows a tendency to hunt out corners. When the eyes are varnished with lampblack, the phototaxis disappears, but the tendency to seek out corners still remains. Bethe says that he has observed the same phenomenon in the crayfish. Keeble and Gamble[225] discovered that Hippolyte varians responds positively to light under all conditions, and Palæmon is just as markedly negative. Macromysis, however, reacted now positively now negatively, depending on the background. A black (absorbing) background called forth a positive response, while a white (scattering) background produced a negative reaction. Spaulding,[226] in studying the habits of the Hermit Crab (Eupagurus), found that it is strikingly positively phototactic. When animals are placed in an aquarium, one half of which is shaded, none of them are ever noticed inside of the dark line. Herrick[227] notes that lobsters are nocturnal, and avoid the light when placed in a tank, and Bateson[228] says that prawns and shrimps lie hidden during the day, and are active only at night. Parker,[229] in a study of Copepods, finds that the females have a strong positive phototaxis for light of a low intensity, while males show a weak negative phototaxis. To light of over 100-candle power at a distance of 10 cm. or to direct sunlight the female Copepods are negative, while the reaction of the males does not seem to be altered.

In his work on Carcinus, Bethe obtained retraction of the eye-stalks by suddenly throwing a strong light on the eye by means of a mirror. "Usually the eyes were quickly drawn in and protruded again, sometimes several times in rapid succession, like a man blinking under a sudden, strong light." When a dark object, the size of the hand, was moved just over the water, the eyes were seldom retracted, but the antennules were usually drawn in. Lemoine[230] observed that in Astacus retraction was due to touch alone, and that no light, however strong, was able to bring about such a reaction. Gulland[231] takes just the opposite view with reference to Astacus, stating that there are no setæ of any sort on the eye-stalk, and therefore it is insensitive to touch, but is withdrawn only because the animal sees the object by which the stimulus is given. If a curved needle is used, and the stimulus is applied from behind, no retraction follows. Dearborn,[232] however, working with Cambarus, agrees with Lemoine in saying, "Withdrawal of the ophthalmites into their sockets occurs only on contact with some hard object,—not from any light-stimulus of an ordinary sort." I may say in passing that in none of the following experiments on Cambarus was there ever a sign of retraction due to stimulation by light, the retraction always taking place in response to a touch-stimulus.

Lyon,[233] in his study of compensatory movements of the eye-stalks, found that when the eyes were painted with lampblack, the crayfish showed a reduction of about 10% in the compensatory movements when rotated in vertical planes, but the compensation remained the same for rotation about the dorsi-ventral axis. On rotation in the dark the compensatory movement of the eyes was found to be from 5° to 8° less than in the light.

EXPERIMENTAL

In the investigations to be described, 58 crayfish of the species Cambarus affinis were made use of, and for identification the animals were marked on the back with white enamel paint, the males receiving the even numbers from 2 to 64, the females the odd numbers from 1 to 51.

1. Reactions to White Light

The questions proposed for investigation were, (a) How does the crayfish react to diffuse daylight; (b) to reflected sunlight; (c) to direct sunlight; (d) to artificial light of different intensities? (e) What is the influence of previous conditions of exposure to light upon the reactions of the animal? (f) Do changes of temperature affect the reactions?

A wooden box, 80 cm. long, 25 cm. wide, and 20 cm. high, painted black on the inside, and constructed so as to hold water, was covered with a heavy black cloth to exclude the light from above. The front end of the box was of glass, thus admitting the light from the end. In all the experiments except those with direct sunlight, this glass end was covered with black cardboard in which a hole 10 cm. long and 5 cm. high had been so cut that the light was admitted at the middle of the bottom of the glass. The direct sunlight was admitted through the whole of the glass end. At the rear of the box a piece of black cardboard was so arranged that an aperture was afforded for observing the animals without admitting any appreciable amount of light, and this aperture could be readily closed by a slide when not in use.

The method of experimentation was to place the animal in the box about 20 cm. from the glass end, and observe whether it went toward or away from the source of light. The animals were experimented on in two groups of five each, and one hundred observations were made on the individuals of each group with each intensity of light, that is, twenty observations on each animal. In order to check the influence of the orientation of the animal at the time of exposure to the stimulus, the following four positions for placing the animal were chosen: (1) Head toward the light; (2) Head away from the light; (3) At right angles to the light with right side toward it; (4) At right angles with the left side toward the light. Thus five observations were made on each animal of each group in each position, exposed to each of the different intensities of light.

Seven different intensities of light were employed, and the results have been arranged in eight sets, as follows: I. Diffuse daylight in dry box, i. e., the animals were taken out of their ordinary medium, water, and were exposed to the stimulus of diffuse daylight in the air. The reactions under these conditions, however, were so slow and so unsatisfactory that the test was abandoned after the first group, and thus the second group has nothing to show for itself under this head. The remaining seven sets of observations were made on animals placed in 10 cm. of water at 15° C. II. Diffuse daylight. III. Reflected sunlight. The box was placed near a window on a clear day, and the sunlight was thrown in horizontally by means of a mirror. IV. Direct sunlight. On a clear day the box was placed in such a position that the sun shone in directly and illuminated the front half of it. V. 9-candle-power incandescent electric light. This lamp was marked 16 c., but it had been used a great deal, and on being tested with a Lummer-Brodhun photometer showed only 9 c. VI. An incandescent electric light of about 50 c. This lamp was marked 100 c., but had been used considerably and was slightly smoked. Unfortunately it was broken before there was any opportunity to test it. Judging from the fact that another 100 c. lamp of the same manufacture, in slightly better condition, measured 64 c., the estimate of 50 c. seemed a safe one. VII. The incandescent electric light alluded to above, which measured 64 c. VIII. An arc light which varied in intensity from 150 c. to 250 c.

In intensities V and VI the lamp was placed 5 cm. from the glass end of the box to allow the interposition of a heat-screen consisting of an alum solution in a flat glass jar 5 cm. thick. Reckoned in candle-metres, therefore, the intensity of the illumination at the surface of the animal in V was 144 c. m., and that in VI was about 800 c. m. In VII two heat-screens were used, and between these was placed a lens of considerable but not accurately determined curvature, so that it is impossible to express the intensity in candle-metres. In VIII the light was so variable that such an expression would mean nothing.

Unfortunately it was impossible to keep the two groups constant throughout the whole series, owing to the death of two individuals in each group during the experimentation. Group 1 was composed of nos. 1, 3, 4, 8, and 9, of which 1 and 8 were replaced by nos. 13 and 42 respectively. Group 2 was begun with nos. 23, 27, 32, 34, and 38, and the vacancies caused by the death of 23 and 32 were filled by nos. 21 and 36. The following table exhibits the reactions to the different intensities of light, + indicating an orientation toward the source of light, - an orientation away from the light, and ± an indifferent orientation, which usually means no movement at all.

TABLE I. SUMMARY OF REACTIONS TO WHITE LIGHT