Yerkes instituted experiments with Cambarus affinis, a relative of the fresh-water crayfish. He placed these crustaceans, in a specially shut-off portion of a box, from which they could get out into the main portion of the box through an opening in the middle. From the main portion two outlets, separated from one another by a partition, led into a vessel of water (fig. 3). The animals as often made use of the right as of the left exit. Next, the right hand outlet was closed by a glass plate. During a period of 3o days each crayfish underwent 6o experiments, in which the boxes were always carefully cleansed to exclude the presence of possible smell stimuli. The aver age results of the performances showed in the first ten trials so% of correct solutions, in the second ten 6o%, in the third 75.8% and in the sixth 90%. By a correct solution is meant that the experimental animal imme diately chose the correct exit.
One of the crayfishes, after 450 experiments, made only one mis take in 5o trials. It is plain that such a method permits the investigator to test how long the lesson is retained by the experimental animal when no practice takes place in the interim. After a Cambarus had progressed so far in its lessons that it made one error only in ten trials, the experiments were interrupted for a fortnight. When, on the f our teenth day, the experiments were recommenced, it was evident that the animal still remembered what it had learnt. But now three errors were made per ten trials.
Yerkes carried out experiments on frogs similar to those already described with Cambarus. He used a box like that mentioned above, except that the two exits were contrived in a rather more complicated manner. The outlets led to a tank of water but the direct path was closed by a transparent glass plate, so that the animals were obliged to make a detour (fig. 4). Daily series of ten experiments were made. After 100-120 trials the frogs no longer made mistakes. They made for the water by the shortest path. The effect of the training was measured by the number of correct and incorrect attempts at the two critical spots, the entrance and exit of the labyrinth (fig. 5). After the experiments had been discontinued for one month, it was found that the frogs had not yet forgotten their lesson.
It is generally known that birds and mammals can learn or be trained. But in addition we see everywhere in nature how the most varied types of animals turn their acquired experiences to account. It is the faculties just described, the plasticity of in stincts, the capacity for memory, and the ability to form associa tions, which permit of such con duct. But the foundation for the
behaviour resulting from these faculties, which is frequently complicated, is always provided by inborn reflexes and instincts. With the object of analysing animal conduct more closely, numerous experiments have been made, particularly by American workers, utilizing apes, dogs, cats, hedgehogs, rabbits, rats, mice, various birds and even reptiles, such as tortoises. The object of these experiments is to discover the capacity for learning, the ability to acquire experience. The outlet from more or less complicated laby rinths or mazes has to be found and learnt, or problem boxes are used, in which the animal must open a door to get out by latches which are more or less complicated according to the nature of the animal studied. These fastenings consist of levers, hooks, bolts, wires, that must be pulled, levers to be pressed down with the foot, and so on (fig. 6). The animal is obliged to discover the mechanism of opening the door for itself by chance trial in the course of its planless efforts. The duration of the learning, or here of the formation of a habit (the dropping of useless movements), is judged by registering the failures (figs. 7 and 8). Companions are sometimes placed in the company of animals which have already learnt the solution of such a task, to see whether they will themselves learn more rapidly by imitating the movements of their fellows. It has been proved that this rarely occurs. Even when the animals are shown the necessary manipulations by the experimenter putting the limbs of the animals through the movements, this does not help. Each animal is obliged to learn everything through its own experience. A mass of results has been accumulated by these American "be haviourists," the value of which will only become apparent when they have been co-ordinated.
In conclusion of this section it must be pointed out that in experiments with animals it can easily happen that the subject trains itself to small occurrences unnoticed by the experimenter and not taken into account by him. Slight noises, movements, for instance the reversal of an electric switch, and so forth may unwittingly be the cause of the establishment of a habit. Then the results obtained from the experiments do not give a picture of the capabilities of the animals which it was attempted to test, but of their capacity of turning experiences to account.