Cole, Intelligence of Raccoons. 223 



of a man who makes the same mistake each time in adding a long 

 column of figures. Not only was no routine order followed, but 

 often the raccoons worked one or more fastenings more than once, 

 never, however, was a latch operated after the door was open. 

 The temptation is strong to say that the raccoon has no memory 

 of having already released a latch, since he operates it a second 

 time. I think this interpretation of the animal's conduct incor- 

 rect, however, for in Boxes 11 to 13 inclusive, in which a thumb- 

 latch had to be released last, it was operated almost twice as many 

 times as any other fastening. The animals would work one or 

 two latches, then try the thumb-latch, and so on. This indicates 

 that the animal had a distinct association of the opening of the 

 door immediately after the depression of that latch, that is, the per- 

 ceptual factor of the opening door ivas a part of the association. 

 No such perception was possible after the release of any other of 

 the latches so they were worked at random because in the past 

 each of them had resulted in the opening of, the door. If the door 

 did not open they were worked again. Surely it is asking too 

 much of the animals to expect them to know that each latch when 

 released partly unfastens the door although the door does not 

 move until the last latch is worked. Such a view would demand 

 of the animals either reasoning or a human being's knowledge of 

 bolts and pulleys. His association, his idea, if he has one, is that 

 the last act opens the door. It is noticeable, too, that the two 

 buttons and the hooks which could be seen to be out of the way 

 when unfastened were not operated a second time nearly so often 

 as the loops and platform which presented no perceptible change 

 of place after having been depressed. Consequently it seems 

 quite as fair to argue that the raccoon pulls a loop a second time 

 because no desired result perceptible to him followed the first 

 pull, as to urge that he pulls it a second time because he has no 

 memory of having pulled it the first time. 



Since, therefore, so far as the animal can see, only the final act 

 opens the door and gains the reward of food, the conditions of 

 the experiments were probably quite unfavorable to the acqui- 

 sition of a fixed order in performing the acts. Combinationlocks, 

 the second element of which could not be unfastened until the 

 first had been operated did not entirely obviate this difficulty in 

 experiments with monkeys (see Kinnaman, p. 124), so it remains 

 to add to this device some means of making the effect of releasing 



