x SOME EXPERIMENTAL RESULTS 219 



" No." He then came to me, knocked the iron off with his 

 paw, and got the biscuit. Next time he did this almost at once. 

 Next time he began by biting at the string. The iron fell off, 

 but I prevented the lid from falling, and replaced the weight. 

 He then began jumping on the box, but, being restrained, looked 

 carefully at the iron and pawed it off. Next time I had to 

 restrain him from pulling the table cloth, and he then pawed off 

 the iron. Next time he knocked it off without hesitation. The 

 next day he first bit at the string, and the iron tumbled off", and 

 he got the biscuit. In the next trial he repeated this, but I 

 cheated him by keeping my fingers on the lid, and replaced the 

 iron. He then pawed it off the hook, and repeated this four times 

 in succession. 



This experiment seemed to me strongly suggestive of 

 learning by perception of results. I cannot think that any 

 accident would have led Jack to pay the slightest attention 

 to the weight. 



Failures of Jack. 



(a) 1 tried without success to arrange a converse experi- 

 ment to that of the weight. 



A basket was hung by a string from the door of the box, and 

 the plan was to balance it so that when a weight was put into 

 the basket, it would weigh the lid down. Unfortunately my 

 arrangement was very defective ; and furthermore, Jack soon 

 found out that to put his own paw in the basket was quite as 

 effective as putting the weight in it. I tried to defeat him by 

 frankly holding the lid in place when he used his foot, and I 

 gradually got him to the point of knocking down and pawing the 

 piece of wood which I used as a weight, and which was placed 

 near the basket. At other times he would take the wood in his 

 teeth, and three or four times he knocked the wood on to the 

 basket. The failure may be ascribed partly to the complexity of 

 the task ; but equally, I think, to the knowledge of another and 

 simpler method. 



(b} Hook. 



The same explanation applies to an experiment in which I tied 

 the latch of a box by a string to a hook, which was then fastened 

 to a nail, first, on the box itself ; later, at a distance of about a 

 yard from the box. 1 The string held the latch in place, and to 

 free it, it was necessary either to break the string or remove the 

 hook. I began by testing whether Jack would learn this without 



1 See diagram, p. 280. 



