TABLE III. SOUND ASSOCIATION, LABYRINTH H
Time required to reach food under different sound conditions
| I | II | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Order of gongs the same as when course was being learned. | Order of gongs reversed. | ||||||||||
| Animals | Animals | ||||||||||
| Trials | A | B | E | G | Trials | A | B | E | G | ||
| ˝ | ˝ | ˝ | ˝ | ˝ | ˝ | ˝ | ˝ | ||||
| (1) | 13 | 19 | 24 | 17 | (1) | 14 | 16 | 18 | 27 | ||
| (2) | 16 | 12 | 13 | 16 | (2) | 37 | 17 | 16 | 27 | ||
| (3) | 10 | 16 | 20 | 14 | (3) | 14 | 19 | 26 | 11 | ||
| (4) | 12 | 19 | 10 | 12 | (4) | 27 | 22 | 21 | 14 | ||
| Total, | 51 | 66 | 67 | 59 | ˝ | Total, | 92 | 74 | 81 | 79 | ˝ |
| 243 | 326 | ||||||||||
In the case of thirteen of the sixteen tests given with reversed acoustical conditions (see column II) the periods were longer than the corresponding ones given alternately with them for comparison, and there was an average time-lengthening of 5.2 seconds per trial, or 34.2%. The following is a short description of the animals' reactions to the changed conditions. It corresponds to the time-values expressed in column II of the table.
Bird A: test 1, animal undisturbed; test 2, drew back from S, turned to the left and went toward R, but later returned and passed S without pausing; test 3, paused at O for a short interval, but did not enter the blind alley; test 4, paused at O again, later drew back from S, turned to the left and entered blind alley 3; it soon escaped, and this time passed S without being disturbed, although it paused at T and U.
Bird B: tests 1 and 2, animal apparently undisturbed; test 3, a few slight pauses at openings; test 4, drew back from S, entered blind alley 3, but soon escaped and passed S without hesitation.
Bird E: test 1, undisturbed; tests 2 and 4, paused at openings; test 3, turned back from S, entered blind alley 3, and paused at several places later when passing toward F.
Bird G: test 1, many pauses; test 2, turned from S and entered blind alley 3; test 3, undisturbed; test 4, drew back from S, went toward R, but did not enter the blind alley.
As the animals gave little attention to the wooden gong, but were always sensitive to the metallic one, their observed movements probably must be accounted for chiefly on the basis of certain visual and organic sense-data now with, and now without, the ringing sound. The data governing the start (as already noticed) were probably sufficient for the avoidance of the first blind alley when the gongs were reversed. In case of the other two, however, the birds had come to depend upon acoustical data, and when these were lacking as they approached the openings O and Q, the left turn could not readily be initiated, hence certain hesitations and misdirections of movement frequently occurred. Experience with the blind alley in the first experiment assisted the animals in dealing with the second blind alley here, but mistakes were made. Visual data usually were sufficient to produce the proper turn at Q, but when the ringing sound was given just afterwards, it sometimes occasioned the left turn, thus bringing the animals toward the opening of the blind alley. While the tests given were not such as would indicate how far pigeons can discriminate sounds, they certainly show that these birds are capable of useful sound associations, although visual ones are evidently of greater importance to them.
D. Habits in Labyrinth O
I next made tests in which tactual and electrical sense-data could also be utilized. In one of the passages of a simple labyrinth was placed a board 8 in. square and 3/4 in. thick, over which were stretched copper wires which formed a series of interrupted electrical circuits. By closing a key a bird could be stimulated whenever it stepped upon the wire surface. A second key was connected with a metallic gong. When an animal on its way through the maze first stepped upon the wire surface, electrical and acoustical stimuli were given; later it was allowed to walk across the board without being thus stimulated; afterward acoustical stimuli were given it at various parts of the maze.
FIG. 5. Labyrinth H. For study of acoustical association. Passages 6 in. wide, as before.
Eight animals were used. All were found quite sensitive to the electrical shocks, and when next tested they avoided the board, especially if the gong sounded as they approached. Some would show signs of uneasiness anywhere in the maze on hearing the gong. When the board was so placed that they had to pass over it in reaching the food, when once on it they moved very leisurely, often lingering; and if they stepped upon the wire surface in the darkened maze, they showed no evidence of being frightened. Evidently no association had been formed between the peculiar tactual stimulus of touching the wires and the electrical shocks which had at first been given. But the tactual stimulus may have been below the threshold. Yerkes[203] saw evidence of association of this kind in the frog; this animal, however, is probably much more sensitive to tactual stimulation received from surfaces over which it passes than is the pigeon.
The results of these four experiments indicate that the pigeon easily acquires complicated labyrinth habits; that these remain fixed for some weeks at least; that acoustical, visual, and certain organic data are the most important sensory factors; and that the process of learning is one of "trial and error," in which the animal comes to form such a close connection between the sense-data of the interior of the box and those other sense-data arising from movements involved in reaching the food, that when the box impressions are again encountered the other sense-data are revived and readily condition the proper movements. How much memory of eating was involved in these tests cannot be told; but it was certainly not an essential part of the mental act.[204] Proper guidance throughout the course was the main thing, and this was determined by definite sense-data. That recognition, discrimination, and perhaps choice were to some extent present seems likely from the animal's hesitating movements at certain critical points. Thus it is highly probable that when the bird approached a blind alley which it had always entered before (see Fig. 2, stage 6), two alternatives were recognized, to enter, as before, or not to do so, as was usual thereafter, and that the pause had for its mental correlate a state closely bordering upon what in us would be deliberation.
2. Release experiments. Under this heading are included certain cage experiments in which some act, such as touching a lever, pecking, or stepping upon a platform, resulted in the opening of the door, and thus enabled the animal to escape and secure the food lying in view without. The animal was admitted to the cage through an entrance compartment as in the case of the maze trials. Before being tested it was allowed to become familiar with the cage and to reach the food directly by passing out through the open door. When first in the cage the animals did not seem to notice the release apparatus, and hence they probably did not begin learning the method of escape until later when they entered the cage and found the door closed, and the ordinary exit thus obstructed.
A. Latch Tests
The cage here employed was an 18-inch cubical box. The top was of chicken wire and the bottom and three sides of heavy boards; the fourth side was formed by narrow vertical bars and a wire door which opened inwardly and was held by a latch working on the outer surface of the bars. At first a long wooden latch was used, which the animals raised when reaching out for food. As this seemed an unnatural act, downward pressure was substituted by attaching to the latch, now made smaller and of brass, a string which ran over a pulley above the door and down into the cage. As nooses did not seem adapted to the birds, the end of the string was attached to a wooden lever which worked on the inner surface of the bars, about three inches from the floor. Eight animals were tested four times daily (twice in the forenoon and twice in the afternoon) for ten days. The time required to escape and the animals' behavior were recorded as in the case of the labyrinth tests.
FIG. 6. Curve of learning to operate latch, plotted from last column of Table IV. Each vertical division indicates 30 sec. The horizontal divisions represent successive days.
When they first entered the box (singly as in the other experiments) and found the usual exit closed they made various attempts to push through between the bars, springing and often flying about with great force and persistency. In course of their random movements they touched the bar and opened the door and thus escaped. Later the unnecessary movements were mostly dropped and the necessary ones became highly specialized. The first association was established between the box impressions received on entering and the movements involved in approaching the front of the box and depressing the lever; later a connection was formed between the sensations of touching the lever, of hearing the sound of the opening door, of feeling the jar, etc., and the movements of turning away from the lever and passing out. The sight of the opening door seemed to be of less service to the birds than the sound and jar. Each animal soon came to touch the bar at the point of least resistance, and usually with considerable precision. The time required by the several birds is shown in Table IV, next page. The daily average results are plotted in Fig. 6, above.