68 ROBERT M. YERKES 



greatly to the surprise of the experimenter, for no indication had 

 previously appeared of this approaching solution of the problem. 

 It seemed possible, however, that the successes were accidental, 

 and it was anticipated that in a control series Julius would again 

 make mistakes. But on the following day. May 12, the presen- 

 tation of the original series of ten settings, which, of course, 

 differed radicahy from the settings used from May 4 to May 11 

 was responded to promptly, readily, and without a single mistake. 



Julius had solved his problem suddenly and, in all probability, 

 ideationally. 



Only three reactive tendencies or methods appeared during 

 Julius's work on this problem: (a) choice of the open door near- 

 est to the starting point (sometimes the adjacent boxes were 

 entered); (b) a tendency to avoid the " nearest " door and select 

 instead one further toward the left end of the group; (c) direct 

 choice of the first door on the left. 



The curve of learning plotted from the daily wrong choices 

 and presented in figure 18, had it been obtained with a human 

 subject, would undoubtedly be described as an ideational, and 

 possibly even as a rational curve ; for its sudden drop from near 

 the maximum to the base line strongly suggests, if it does not 

 actually prove, insight. 



Never before has a curve of learning like this been obtained 

 from an infrahuman animal. I feel wholly justified in concluding 

 from the evidences at hand, which have been presented as ade- 

 quately as is possible without going into minutely detailed 

 description, that the orang utan solved this simple problem idea- 

 tionally. As a matter of fact, for the solution he required about 

 four times the number of trials which Sobke required and twice 

 as many as were necessary for Skirrl. Were we to measure the 

 intelligence of these three animals by the number of trials needed 

 in problem 1, Sobke clearly would rank first, Skirrl second, and 

 Juhus last of all. But other facts clearly indicate that Julius 

 is far superior to the monkeys in intelligence. We therefore 

 must conclude that where very different metlwds of learning ap- 

 pear, the number of trials is not a safe criterion of intelligence. 

 The importance of this conclusion for comparative and genetic 

 psychology needs no emphasis. 



