On July 27, 63 days later, the same test was repeated with the same animal. Time required, .33 min. Conditions were the same as before. The apparatus was freshly washed, and smoked paper used to show the movements. The only difference was that the guinea pig was taken at the usual feeding time, and it had not been handled for two months.

At this time this and several other of the experiments were repeated on other individuals, and the memory in every case was almost perfect.

I.

Conclusions from the labyrinth experiment.

The guinea pig can learn a complex path to food. II. The time curve for learning is very abrupt for the adult, and for any one individual is also irregular. It tends, however, to reach a minimum at which point it is, after a few trials, nearly constant. In the labyrinth used this minimum will be observed to be . 166 min.

III. The curve for elimination of random movements follows very closely the time curve, as random movements necessarily increase the time required.

IV. There are two kinds of random movements: (1) Those made in attempt to reach the food, as biting the wire, running into blind alleys; (2) those of superfluous activity or curiosity, as exploring the cage, running about, and jumping. When the guinea pig seemed in too playful a mood to attend to business, it was a sign that it was not hungry, and therefore conditions were not uniform.

Test IX. Learning without the aid of vision.

Granted that the guinea pig can learn a complex path, the problem arose, What sense elements contribute to this result? What does the guinea pig remember?

The guinea pig may orient itself by means of vision, by

means of smell, or by kinesthetic impressions. It will be remembered that when the simple wire box was placed before the guinea pig, it ran in exactly as it was accustomed to do, although no food was inside. The box itself was then the stimulus, as being associated with food.

To determine how quickly the odor of food alone could set up the association, a series of experiments was designed, in which the visual factor was eliminated.

Text-figure 8.

The apparatus used in this test was a simple wooden box, 6% by 61⁄2 by 61⁄2 inches, with numerous small holes bored in it (Text-fig. 8). A door was swung by hinges at the top A copper spring inside the door made contact with a plate in the top of the box when the door was pushed open. Connection was made with an electric light in a dark box entirely outside the experimental cage. The experimental cage was covered with a black cloth frame, and all experiments were performed at night, so that no light should be present. Noises and other accidental

disturbances were thus diminished.

In the first series of experiments it was possible, after the reaction had been made and the light turned on, for a faint glow to penetrate the cloth covering of the experimental cage. At a later time these results were verified in the dark room where no light could enter, and the arc connected with the apparatus was so arranged that no illumination of the room was possible.

The typical series given was taken from an adult guinea pig about nine months old, of the smooth English variety. It had not been used before for any experimental work, and therefore the first thing that had to be done was to tame it, and ac

custom it to being brought upstairs. As usual, the quickest way seemed to be to associate the experience with food. On Feb. 12 it was taken to the dark room at the usual feeding time and put in the experimental cage. After having been left there about half an hour it was returned to the guinea pig room and fed lightly.

Feb. 13. It was taken to the experimental cage, food having previously been placed there. In 5 min. the food was found and was pulled to one side of the cage to be eaten.

Feb. 15. The food was put in a wire box, but was not found. The guinea pig was now becoming tame, and behaved naturally when removed to the experimental room. Fright had disappeared.

Feb. 16. The electric food-box was used with the door open. The food was not visible in the dark, and the door was left open in order that the slight grating of the hinges, the noise of the contact of the spring with the plate, and the touch of the door itself might not frighten the timid animal. The time was recorded from the moment of placing the guinea pig in the experimental cage until the sound of pulling out the food was heard. (When the door to the box was closed the appearance of the electric light gave a more accurate time limit.) The food was found, seemingly accidentally, in 2.264 min.

Feb. 27.

Rattling of windows in the wind frightened the guinea pig, and

This increase in time was probably due to two things; (a) not so hungry

The results of this series of experiments when compared with a similar series taken in the light are these: (1) The range of variation in reaction-time is greater in the dark than in the light; (2) A longer time is required to form a definite habit of entering the cage for food; (3) The average time required, even omitting the excessively long periods, is longer than that required for the analogous experiment in the light. This is true in spite of the greater activity of the animal in the dark and the greater freedom with which exploration is made; (4) It follows, therefore, that the number of random movements is much greater in the dark than in the light. This the smoked paper shows to be almost invariably the case.

Conclusions from the four tests.

In those tests in which only a visual stimulus of food was permitted in a situation not previously associated with food, there was no attempt to obtain the food; it apparently did not attract attention. Other experiments, particularly the choice experiment of test VII, gave negative results as to the efficiency of a visual stimulus when not reinforced by other stimuli.

From the sixth and seventh tests it was concluded that, after a situation is once connected with food, it is reacted to as a whole with the appropriate movements. An odor stimulus of food is then not a part of the situation essential to the reac tion.

From the eighth test we found that the situation might be considerably complicated without diminishing the appropriateness of the reaction. A situation which presents difficulties of the kind which the animal would meet in its natural environment, is rapidly learned and reacted to almost automatically.

The ninth test has shown that vision is an important element in learning the problem, but cannot be the only element, since the problem was learned without it, though more slowly. after eating the bite or two allowed at each entrance, (b) a little fright and discouragement from being repeatedly removed from food. The lengthened time of reaction was often noticed if the experiments were repeated several times in succession, and therefore too frequent repetition was hereafter avoided.