x SOME EXPERIMENTAL RESULTS 195 



The pole ran out along and beyond the side of the cage, 

 so that by pressing it, the door, which otherwise shut with 

 a spring, could be opened. I held the cage above Jack's 

 head, in such a position that the door would open down- 

 wards, and a piece of meat placed inside would then fall 

 out. For six trials (after each of which he was shown) 

 Jack merely jumped up and " scrabbled " at the cage. He 

 then pulled the lever but did not get <c fixed " in the habit 

 immediately. 



There were three or four more failures out of eleven trials in 

 this series, and he had to be shown three times the next day. 

 He then became perfect in it, and when for the cage I substituted 

 a box which could only be opened by a projecting lever, he soon 

 went at the lever after a little preliminary scratching at the box, 

 and without being shown. 



It is clear that accident might easily enter into this experiment, 

 especially in the case of the box. Jack, however, did not seem 

 to learn it by any accidental clawing. He naturally aimed 

 straight at the food, and he had to learn to claw or bite (he used 

 both methods) at a bar some inches away. He seemed to do 

 this quite definitely all of a sudden, 1 though he sometimes forgot 

 it afterwards. 



Tim failed to learn this trick in three trials with the 

 bird-cage, but acquired it at the third trial with the box, 

 and did not afterwards fail. 2 



Of other animals, the otter learnt to use the lever by being 

 first accustomed to open the door, which for his benefit was left 

 with a crevice afterwards filled up. The elephant had no diffi- 

 culty with it. One collie, Rose, absolutely failed, or perhaps 

 refused 3 to open a box thus by lifting the lid. Her son opened 

 it after being shown three times. 4 



1 The notes show that he once before nearly pawed the lever in his 

 first trial. 



2 When a general statement of this kind is made it refers only to the 

 same experiment under the same conditions. Sometimes an acquired 

 trick enters as an element into a more complicated experiment, and then 

 failure in that experiment would put the animal off its simpler performance. 

 This occurred in the present case. 



3 After showing her several times, her master arranged so that she 

 began pushing her nose in as he lifted. Even this had no effect. 



4 He failed several times the same night, but I found afterwards that 

 the lever would rub uncomfortably against his nose. I altered this, and 

 next night he succeeded. 



O 2 



