THE INTELLIGENCE OF MAMMALS 255 



of another with the end of securing some advantage which 

 the other animal enjoys. Imitation of this character 

 involves a species of inference, and its occurrence in animals 

 is less widespread than was commonly supposed some 

 years ago. Thorndike tested cats and dogs with the view 

 of ascertaining if they could learn to get out of a puzzle 

 box by seeing other cats and dogs get out. After witness- 

 ing a number of times the successful exits of the animals 

 which had learned the means of escape these cats and 

 dogs were unable to get out by themselves any more 

 quickly than other individuals which had not the ad- 

 vantage of seeing how the escape was effected. In all the 

 experiments performed the animals did not show the least 

 tendency to imitate the performances of their successful 

 comrades. The experiments of Cole and Davis on the rac- 

 coon yielded similar negative results. 



Small, in his studies on the white rat, finds evidence of a 

 very simple kind of imitation, but no clear cases of imitation 

 of the inferential type. When one rat sees another digging 

 it is apt to dig also, and when one runs over a box it is apt 

 to be followed by others. Berry, who has made a more 

 thorough study of imitation in the white rat, found better 

 evidence of intelligent imitation. "When two rats were 

 put into the box together, one rat being trained to get out 

 of the box and the other untrained, at first they were in- 

 different to each other's presence, but as the untrained rat 

 observed that the other one was able to get out while he 

 was not, a gradual change took place. The untrained rat 

 began to watch the other closely; he followed him all about 

 the cage, standing up on his hind legs beside him at the string, 

 and pulling it after he had pulled it, etc. We also saw that 

 when he was put back the immediate vicinity of the loop 

 was the point of greatest interest for him, and that he tried 



