2 journal of Cojnparative Neurology and Psychology. 



and raw meat, with the occasional addition of some vegetables. 

 They were fed only once a day, and the feeding time was imme- 

 diately after the experiment. Enough bread and milk were given 

 to keep the animals in good condition, and the raw beef was used 

 principally as a reward for the performance of the required act. 

 In all the experiments the trained cat was fed each time it performed 

 the act which the situation required. All of the cats were per- 

 fectly tame and very active. 



The general method of testing the imitative tendency of cats 

 was as follows : Either separately or together the cats were given 

 opportunity to learn to perform a certain act or series of acts. 

 In case all learned it of their own initiative no tests of imitation 

 could be made, butif one, or more, of the individuals failed, after 

 abundant opportunity, to discover the appropriate mode of 

 reaction, it was given a chance to learn by watching another cat 

 perform the act. 



II. EXPERIMENTS. 

 Experiment J. Jumptng from Box to Table. 



Method. — A box, open on one side, was placed upon a table 

 8i cm. high. A second table lo cm. lower was placed 56 cm. 

 from the first. The act to be performed was to jump from the 

 box to the meat, which was placed in plain sight on the lower table. 



Results. — ^X, who was put into the box alone, was afraid to 

 jump. M was put in with her, and she saw M jump from the 

 box to the table five times, but still she was afraid to follow. Y 

 and Z were tested in like manner with the same result. I now 

 drew the lower table 8 cm, nearer the first, and put the three kittens 

 into the box together, but still they were afraid to jump. M was 

 now put in with them. The second time she jumped to the meat, 

 X followed her. When X was put back with Y and Z she jumped 

 down at once, and in less than a minute was followed by Z. After 

 Y saw Z jump to the meat five times he jumped to the floor. No 

 further trials were made. 



Experiment 2. Opening Door by Pulling Knot. 



Method. — A wooden box (Box I) 73 cm. long, 50 cm. wide, and 

 56 cm. deep was closed on one side by a door of wire netting of half 

 inch mesh. Near one end of the opposite side was a door (20 x 15 

 cm.) which was held shut by a wooden crossbar. From the end of 



