216 MIND IN EVOLUTION CHAP. 



running, which suggests that, if I had continued the experiment, 

 there would have been an improvement. 



The older cats did better. Teufel, in the first series, 

 was right four times out of six ; in the second series was 

 wrong once only, and right three or four times running. 1 

 The two year old Persian was right three times out of five. 



The little Rhesus monkey, Jimmy, who was kindly lent 

 to me by Messrs. Jennison for some days, also failed 

 signally in this experiment. 



He was generally right if I let him sit in front of the drawers 

 while 1 put the food in ; but if I banished him to a chair, he got 

 quite wrong, and by degrees I noticed that he was elaborating a 

 method of his own to save him the trouble of attention. To 

 make quite sure, he would open the drawers systematically, from 

 the bottom upwards. 2 



Success or failure, in an experiment of this kind, seems 

 to be a matter principally of attention, partly of reten- 

 tiveness. 3 



1 In this series bread and butter were used instead of meat, to avoid 

 guidance by smell, if possible. This, however, is hardly a danger in this 

 experiment, as the cat opens with its claw, and at that distance the smell 

 could hardly guide it as between one drawer and another. 



2 He did not, indeed, get this system perfect, as he often forgot the 

 bottom drawer ; but it seemed to be growing as he went on. The point 

 of it was that when an upper drawer was pulled out, the one below was 

 difficult to reach, so Jimmy made sure of avoiding this difficulty. In 

 quite a similar way the chimpanzee, who had to identify two boxes out of 

 seven (marked B and containing banana), frankly gave up the attempt, 

 and took them quite systematically in order, sometimes finishing the row 

 after he had secured two pieces of banana, and sometimes leaving it. 

 In discrimination experiments of this kind the penalty of failure is not 

 sufficient to stir the animal's attention. Somebody unkindly suggested 

 that if we put a live wasp in each of the blank boxes, we should soon 

 teach the chimpanzee his alphabet. 



3 Some further experiments confirming this view will be mentioned 

 lower down. 



Mr. G. V. Hamilton (J. A. B., 1911, pp. 33 ff.) has described a series 

 of experiments with human beings, monkeys, dogs, cats, and a horse, 

 where the problem was to open one of four closed doors. The rule was 

 simply that the same door was never locked twice. Only the human 

 beings showed by their behaviour that they grasped the relation. But 

 some of the animals, particularly the monkeys, took to opening all the 

 doors in succession either in regular or irregular order. This is precisely 

 the behaviour to be expected of a semi-intelligent animal. Mr. Hamilton's 

 trick would be more baffling to such an animal than to one that was quite 

 stupid, for it would always tend to be guided by the result of the last 

 experiment, which would always be wrong. It gets out of this by the simple 

 process of trying all four doors in succession, the best thing it can do in 

 the circumstances. 



