x SOME EXPERIMENTAL RESULTS 199 



seems to get angry with her tools, and, if she can, gets hold of 

 them and smashes them. 



I first showed Lily the box unbolted, and she quickly learnt to 

 open it by the knob. 1 I then put in the bolt, and she pushed it 

 out by a random movement. That it was random was proved 

 by the two following trials in which she failed and was shown. 

 At the fourth trial she succeeded, and here she seems to have 

 learnt the trick, for three more successes followed, and no fail- 

 ure. 2 The next day, however, she was awkward about it, and I 

 showed her twice. 3 Some days later she pulled out the bolt at 

 the first trial, though she failed to get the lid open until I turned 

 it conveniently for her. 



I do not think there is any doubt that the elephant 

 learnt the trick substantially at the fourth trial. How 

 she learnt it is another question which I cannot decide. 

 The first purely random success apparently gave her 

 no help ; and, indeed, my observations suggest that, as a 

 general rule, the more a success appears to be purely 

 accidental, the less it is likely to be repeated. The same 

 tendency to random movement caused her partial failures 

 (trying the knob first, and pulling the bolt the wrong 

 way) later on. It is to be observed that this was a double 

 movement learnt regressively i.e., the act to be done last 

 was learnt first. Having learnt to open the door by the 

 knob, she had then to learn that it was useless to pull the 

 knob till the bolt was out. To learn this without getting 

 lisgusted implies a certain measure of articulateness in 

 experience. Two things must be, in some sense, known, 

 and their order kept in mind. 



Billy, the otter, a most mercurial little animal, was now 

 taught to bolt. 



1 I kept the box sideways to avoid the complication of raising the lid. 

 Even so, the opening was not easy, as the bars of the cage would get in 

 the way. 



2 After the seventh trial I turned the box up. This baffled her at first, 

 and she kept fumbling about the side where the bolt had been. (This 

 suggests that she goes more by feeling than by sight.) However, she 

 succeeded in the end. 



3 In her first trial she tried the handle before the bolt, and also tried 

 to pull the bolt the wrong way. She then took to knocking the bolt out. 

 I replaced it, and showed her. After this she got vexed, and tried to 

 confiscate the box. I rescued it, and showed her. Two or three times 

 this day she was awkward and uncertain in her actions. 



