On November 20, Nos. 4 and 5 were placed in the wooden box and left there for half an hour. As they had failed to escape at the end of this interval, they were taken out of the box by the experimenter and returned to the nest-box. November 21 and 22 this test of their ability to learn to climb the ladder was repeated with the same result. On November 23 they were placed in the box with the three mice which had previously been trained to climb the ladder. The latter escaped at once. Apparently the attention of Nos. 4 and 5 was drawn to the ladder by the disappearance of their companions, for they approached its foot and No. 5 climbed up a short distance. Neither succeeded in escaping, however, and they made no further efforts that day. On the 24th, and daily thereafter until the 29th, these two dancers were placed in the box for half an hour, with negative results. At the end of the half hour on the 29th, Nos. 2 and 6 were placed in the box and permitted to go back and forth from one box to the other repeatedly within sight of Nos. 4 and 5. The latter made no attempts to follow them, although at times they seemed to be watching their movements as they ascended the ladder.

To render the results of this test of imitation still more conclusive No.

5 was given further opportunity to learn from No. 1000. Beginning December 2, the following method of experimentation was employed with these two individuals. They were placed in the wooden box together. No. 1000 usually climbed out almost immediately. Sometimes No. 5 apparently saw him disappear up the ladder; sometimes she paid no attention whatever either to the presence or absence of her companion. After he had been in the nest-box for a few seconds, No. 1000 was returned to the wooden box by the experimenter and again permitted to climb out for the benefit of No. 5.

This mode of procedure was kept up until No. 1000 had made from three to ten trips. No. 5 was left in the box for half an hour each day. This test was repeated on 18 days within a period of 3 weeks. No. 5 showed no signs of an imitative tendency, and she did not learn to climb the ladder.

To this evidence of a lack of an imitative tendency in the dancer I may here add the results of my observations in other experiments. In the discrimination tests and in the labyrinth tests I purposely so arranged conditions, in certain instances, that one individual should have an opportunity to imitate another. In no case did this occur. Seldom indeed did the animals so much as follow one another with any considerable degree of persistence. They did not profit by one another's acts.

Excellent evidence in support of this conclusion was furnished by the behavior of the mice in the discrimination experiments. Some individuals learned to pull as well as to push the swinging wire doors of the apparatus and were thus enabled to pa.s.s through the doorways in either direction; other individuals learned only to pa.s.s through in the direction in which the doors could be pushed open. Naturally I was interested to discover whether those which knew only the trick of opening the doors by pushing would learn to pull the doors or would be stimulated to try by seeing other individuals do so. At first I arranged special tests of imitation in the discrimination box; later I observed the influence of the behavior of one mouse upon that of its companion in connection with visual discrimination experiments. This was made possible by the fact that usually a pair of individuals was placed in the discrimination box and the tests given alternately to the male and to the female. Both individuals had the freedom of the nest-box and each frequently saw the other pa.s.s through the doorway between the nest-box, _A_, and the entrance chamber, _B_ (Figure 14), either from _A_ to _B_ by pushing the swing door or from _B_ to _A_ by pulling the door.

Although abundant opportunity for imitation in connection with the opening of the doors in the discrimination box was given to twenty-five individuals, I obtained no evidence of ability to learn by imitation. The dancers did not watch the acts which were performed by their companions, and in most instances they did not attempt to follow a mate from nest-box to entrance chamber.

These problem tests, simple as they are, have revealed two important facts concerning the educability of the dancer. First, that it does not learn by imitation to any considerable extent, and, second, that it is aided by being put through an act. Our general conclusion from the results of the experiments which have been described in this chapter, if any general conclusion is to be drawn thus prematurely, must be that the dancing mouse in its methods of learning differs markedly from other mice and from rats.

CHAPTER XIII

HABIT FORMATION: THE LABYRINTH HABIT

The problem method, of which the ladder and door-opening tests of the preceding chapter are examples, has yielded interesting results concerning the individual initiative, ingenuity, motor ability, and ways of learning of the dancer; but it has not furnished us with accurate measurements of the rapidity of learning or of the permanency of the effects of training.

In this chapter I shall therefore present the results of labyrinth experiments which were planned as means of measuring the intelligence of the dancer.

The four labyrinths which have been used in the investigation may be designated as _A, B, C,_ and _D_. They differ from one another in the character of their errors, as well as in the number of wrong choices of a path which the animal might make on its way from entrance to exit. In the use of the labyrinth method, as in the case of the discrimination method of earlier chapters, the steps by which a satisfactory form of labyrinth for testing the dancer was discovered are quite as interesting and important for those who have an intelligent appreciation of the problems and methods of animal psychology as are the particular results which were obtained. For this reason, I shall describe the various forms of labyrinth in the order in which they were used, whether they proved satisfactory or not. At the outset of this part of my investigation, it was my purpose to compare directly the capacity for habit formation in the dancer with that of the common mouse. This proved impracticable because the same labyrinth is not suited to the motor tendencies of both kinds of mice.

[Ill.u.s.tration: FIGURE 25.--Labyrinth A. _I_. entrance; _O_, exit; 1, 2, 3, 4, blind alleys.]

The first of the four labyrinths, A, appears in ground plan in Figure 25.

It was constructed of wood, as were the other labyrinths also, and measured 60 cm. in length and width, and 10 cm. in depth. The outside alleys were 5 cm. wide. In the figure, _I_ marks the starting point or entrance to the maze, and _O_ the exit through which the mouse was permitted to pa.s.s into its nest-box. Any turn in the wrong direction which the animal made in its progress from entrance to exit was recorded as an error. The four errors, exclusive of the mistake of turning back, which were possible in this labyrinth, are indicated in the figure by the numerals 1, 2, 3, and 4. By retracing its steps a mouse might repeat any one or all of these errors, and add to them the error of turning back.

In the experiments a mouse was permitted to enter the maze from a small box which had been placed by the experimenter at _I_, and an accurate record was kept of the number of errors which it made in finding its way from entrance to exit, and of the time occupied. Each of five dancers was given 31 tests in this labyrinth. The number of tests per day varied, as is indicated in Table 36, from 1 to 4. The results of the tests, so far as errors and times are in question, appear in the table. _T_ at the head of a column is an abbreviation for time, _E_ for errors.

The dancers did not learn to escape from this labyrinth easily and quickly. In fact, the average time of the thirty-first test (198") is considerably longer than that of the first (130"). The number of errors decreased, it is true, but even for the last test it was 6.6 as compared with only a little more than twice that number for the first test. The last column of the table furnishes convincing proof of the truth of the statement that the animals did not acquire a perfect labyrinth-A habit.

Was this due to inability to learn so complex a path, or to the fact that the method is not adapted to their nature? Observation of the behavior of the mice in the experiments enables me to say with certainty that there was no motive for escape sufficiently strong to establish a habit of following the direct path. Often, especially after a few experiences in the maze, a dancer would wander back and forth in the alleys and central courts, dancing much of the time and apparently exploring its surroundings instead of persistently trying to escape. This behavior, and the time and error results of the accompanying table, lead me to conclude that the labyrinth method, as it has been employed in the study of the intelligence of several other mammals, is not a satisfactory test of the ability of the dancer to profit by experience. That the fault is not in the labyrinth itself is proved by the results which I obtained with common mice.

TABLE 36

RESULTS OF LABYRINTH A TESTS WITH DANCERS

AVERAGE TEST DATE No. 1000 No. 2 No. 6 No. 4 No. 5 FOR ALL 1905 T E T E T E T E T E T E

1 Nov 23 130" 14 100" 8 170" 13 60" 6 190" 26 130" 13.4 2 24 140 19 78 7 60 8 149 6 211 25 128 13.0 3 25 392 31 87 1 98 5 185 13 120 9 176 11.8 4 26 448 38 38 3 47 2 50 3 121 12 141 11.3 5 27 142 8 21 2 27 3 27 2 17 1 47 3.2 6 28 45 2 61 7 63 5 102 8 33 4 61 5.2 7 29 303 17 64 7 36 3 42 2 57 4 100 6.6 8 30 222 15 26 2 37 5 42 3 7 0 67 5.0 9 Dec 1 185 9 36 5 48 3 63 3 94 8 85 5.6 10 2 52 2 71 4 19 0 196 5 95 11 87 4.4 11 3 180 8 32 2 107 4 52 3 38 4 82 4.2 12 4 310 10 133 11 65 3 242 6 125 6 175 7.2 13 4 153 9 335 55 130 10 195 15 154 18 193 21.4 14 5 330 7 69 2 42 2 201 6 130 10 154 5.4 15 5 287 7 34 4 61 4 136 7 25 2 109 4.8 16 5 455 15 65 4 25 0 110 8 160 15 183 8.4 17 6 120 15 280 9 33 0 168 4 39 2 128 6.0 18 6 120 4 164 10 81 4 101 5 85 4 110 5.4 19 6 132 12 78 7 110 6 40 2 151 12 102 7.8 20 7 258 10 223 16 33 1 92 5 37 1 129 6.6 21 7 110 7 23 3 44 4 20 4 305 23 100 8.2 22 7 100 4 60 8 167 15 44 7 58 4 86 7.6 23 8 43 1 179 7 356 6 34 3 65 3 135 4.0 24 8 92 5 56 5 42 3 17 1 23 1 46 3.0 25 9 85 5 114 3 62 3 129 8 31 0 84 3.8 26 9 30 2 36 4 109 15 12 1 34 2 44 4.8 27 9 69 5 40 4 85 6 36 3 16 1 49 3.8 28 10 169 7 80 3 28 0 142 5 35 2 89 3.4 29 10 155 5 266 8 91 5 27 0 37 2 115 4.0 30 10 29 1 25 2 124 14 83 6 111 12 74 7.0 31 10 465 6 208 8 95 3 65 3 159 13 198 6.6

On the basis of two tests per day, two common mice, a white one and a gray one, quickly learned to escape from labyrinth _A_ by the shortest path.

The time of escape for the gray individual (Table 37) decreased from 180"

in the first test to 21" in the tenth, and the number of errors from 6 to 1. Similarly in the case of the white individual, the time decreased from 122" to 8", and the errors from 5 to 1. A fraction of the number of tests to which the dancer had been subjected sufficed to establish a habit of escape in the common mouse. It is evident, therefore, that the dancer differs radically from the common mouse in its behavior in a maze, and it is also clear that the labyrinth method, if it is to be used to advantage, must be adapted to the motor tendencies of the animal which is to be tested.

TABLE 37

RESULTS OF LABYRINTH A TESTS WITH COMMON MICE

GREY MOUSE WHITE MOUSE TEST T E T E

1 180" 6 122" 5 2 26 2 80 6 3 37 1 56 4 4 18 0 27 1 5 68 2 33 2 6 10 1 19 1 7 11 1 17 1 8 13 1 17 1 9 10 0 8 1 10 21 1 8 1

The behavior of the dancer made obvious two defects in labyrinth A. Its pa.s.sages are so large that the mouse is constantly tempted to dance, and it lacks the basis for a strong and constant motive of escape by the direct path. To obviate these shortcomings labyrinth B was constructed, as is shown in Figures 23 and 24, with very narrow pa.s.sages, and a floor which was covered with the wires of an interrupted electric circuit so that errors might be punished. The length of this labyrinth was 52 cm. and the pa.s.sages were 2.5 cm. wide and 10 cm. deep. Dancing in these narrow alleys was practically impossible, for the mice could barely turn around in them. In the case of all except the common mice and two dancers, a depth of 10 cm. was sufficient to keep the animals in the maze without the use of a cover.

As an account of the behavior of the dancer in labyrinth B has already been given in Chapter XI, I may now state the general results of the experiments. In all, thirty individuals were trained in this labyrinth.

Each individual was given tests at the rate of one per minute until it had succeeded in following the correct path five times in succession. The weak electric shock, which was given as a punishment for mistakes, provided an activity-impelling motive for escape to the nest-box.

An idea of the extreme individual difference in the rapidity with which the labyrinth-B path was learned by these dancers may be obtained by an examination of Table 38, from which it appears that the smallest number of training tests necessary for a successful or errorless trip through the maze was one and the largest number fourteen. It is to be remembered that each mouse was given an opportunity to pa.s.s through the labyrinth once without punishment for errors, and thus to discover, before the training tests were begun, that a way of escape existed. This first test we may designate as the preliminary trial. Table 38 further indicates that the females acquired the labyrinth habit more quickly than did the males.

TABLE 38

RESULTS OF LABYRINTH-B EXPERIMENTS, WITH TWENTY DANCERS

MALES FEMALES

NO. OF NO. OF FIRST NO. OF LAST OF NO. OF NO. OF FIRST NO. OF LAST OF MOUSE CORRECT FIVE CORRECT MOUSE CORRECT FIVE CORRECT TEST TESTS TEST TESTS

76 8 14 75 4 15 78 5 20 77 7 11 86 13 22 87 12 22 58 2 14 49 1 5 50 6 23 57 3 20 60 13 37 59 14 28 410 6 20 415 4 13 220 4 8 225 6 18 212 3 7 211 6 10 214 10 28 213 5 14

AV. 7.0 19.3 AV. 6.2 15.6

A graphic representation of certain of the important features of the process of formation of the labyrinth-B habit is furnished by Figure 26 in which the solid line is the curve of learning for the ten males of Table 38, and the broken line for the ten females. These two curves were plotted from the number of errors made in the preliminary trial (P in the figure) and in each of the subsequent tests up to the sixteenth. In the case of both the males and the females, for example, the average number of errors in the preliminary trial was 11.3, as is indicated by the fact that the curves start at a point whose value is given in the left margin as 11.3.

In the second training test the number of errors fell to 3.3 for the males and 2.7 for the females. The number of the test is to be found on the base line; the number of errors in the left margin. If these two curves of learning were carried to their completion, that for the males would end with the thirty-seventh test, and that for the females with the twenty- eighth.

[Ill.u.s.tration: FIGURE 26.--Curves of habit formation, plotted from the data of labyrinth-B tests with ten males and ten females. The figures in the left margin indicate the number of errors; those below the base line the number of the test. _P_ designates the preliminary test. Males ____[solid line]; Females ----[broken line].]

Time records are not reported for these and subsequent labyrinth tests because they proved to be almost valueless as measures of the rapidity of habit formation. At any point in its progress through a labyrinth, the dancer may suddenly stop to wash its face, look about or otherwise examine its surroundings; if a shock be given to hurry it along it may be surprised into an error. It is my experience, and this is true of other animals as well as of the dancing mouse, that a long trip, as measured in time units, does not necessarily indicate the lack of ability to follow the labyrinth path correctly and rapidly. Hence, whenever it is possible (and the experimenter can always plan his tests so that it shall be possible), the number of errors should be given first importance and the time of the test second place. I have presented in Table 38 the number of the first correct test, and the number of the last of five successive correct tests. s.p.a.ce cannot be spared for records of the errors made in the several tests by each individual.

In general, labyrinth B proved very satisfactory as a means of testing the ability of the dancer to learn a simple path. The narrow pa.s.sages effectively prevented dancing, and the introduction of the electric shock as a punishment for mistakes developed a motive for escape which was uniform, constant, and so strong that the animals clearly did their best to escape from the labyrinth quickly and without errors. This maze was so simple that it did not tend to discourage them as did the one which is next to be described. It must be admitted, however, that, though labyrinth B is perfectly satisfactory as a test of the dancer's ability to learn to follow a simple path, it is not an ideal means of measuring the rapidity of habit formation. This is due to the fact that the preliminary trial and the first training test play extremely different roles in the case of different individuals. A dancer which happens to follow the correct path from entrance to exit in the preliminary trial may continue to do so, with only an occasional error, during several of the early training tests, and it may therefore fail for a considerable time to discover that there are errors which should be avoided. The learning process is delayed by its accidental success. On the other hand, an individual which happens to make many mistakes to begin with immediately attempts to avoid the points in the maze at which it receives the electric shock. I was led to conclude, as a result of the labyrinth-B experiments, that the path was too easy, and that a more complex labyrinth would, in all probability, furnish a more satisfactory means of measuring the rapidity of habit formation.