The animals did not come in contact with the second partition until they had almost learned to pass the opening of the blind alley (see stage 6). This was probably because the turn to the left which was made on approaching x was associated with visual data derived from points farther along the course (y), and when the animals reached z, compartment 2, these same data were received and were sufficient to occasion the turn to the left there also, thus bringing the birds against the partition. The adjustment was made principally on the basis of new sense-data arising from running against the wire, looking at it more closely, etc. For a few trials the birds made the turn at x too quickly, and thus failed to reach the third compartment. One of my most intelligent subjects made this mistake in the third test, and again in the sixth and seventh, and retained the act almost unchanged through the tenth, eleventh, twelfth, thirteenth, fourteenth, sixteenth, seventeenth, twentieth, and twenty-first tests, so strong was the tendency to continue a movement once begun, though it was really disadvantageous.[200]

FIG 3. Labyrinth M. E, entrance; F, food-box; height of large box and width of passages the same as in Labyrinth L.

After reaching the food and satisfying their hunger, the animals often returned to the maze passages, seeking an exit; but they never "explored" passages or showed other evidence of "free curiosity" and "desire to know all their new surroundings," as Small reports concerning rats.[201]

B. Habits in Labyrinth M

Five of the animals previously used were next tested twice daily, forenoon and afternoon, for five days, in a larger, more complicated maze. It had two blind alleys, and the food-box was near the centre (see Fig. 3). The animals' general behavior was similar to that before observed. The periods are given in Table II.

TABLE II. TIME REQUIRED TO REACH FOOD IN LABYRINTH M

Animals
Trials,ABCEFAverage
2 daily.' "' "' "' "' "' "
(1)16:252:556:333:264:116:42
(2):554:102:243:36:232:18
(3)1:12:558:279:062:074:21
(4):48:442:31:341:041:08
(5):27:16:14:16:14:17
(6):18:32:25:15:27:23
(7):14:11:31:44:30:26
(8):12:16:57:16:48:30
(9):12:18:23:16:16:17
(10):10:19:16:15:28:18

Although this maze was much more difficult, it will be seen that the animals learned the route to the food far more readily than before. The first period in this series was only about one fourth as long as the first period in the other, and the course was mastered sooner (by the fifth trial instead of by the ninth). There was less pressing against the wire than before, and unsuccessful movements were sooner discontinued. This improvement was probably due entirely to experience gained in dealing with the first maze. Thorndike speaks of the gradually increasing ability of animals to deal with successive contrivances.[202] The average results given in Tables I and II are plotted in Fig. 4, next page.

C. Habits in Labyrinth H