THE ASSOCIATION CENTERS 239 



ported are adapted to test sense physiology, stereo- 

 typed reflex and instinctive behavior, and learning by 

 trial-and-error better than inventiveness. Mazes and 

 problem boxes of the usual sort do not provide the 

 best setting for bringing out inventiveness, as Kohler 

 has pointed out. The animals claw or bite or dig or 

 run about aimlessly. Whenever a rat, as sometimes 

 happens, at the beginning of an experiment goes di- 

 rectly to the latch or other key-part of the apparatus 

 and begins to claw or bite it, it is suggested that pre- 

 vious experience in a somewhat similar situation is 

 being utilized, and this might be a first step in the 

 development of intentionally controlled behavior. 

 This much is common to all sorts of learning. Cortical 

 learning is apparently roughly proportional to the 

 complexity of associations that can be made without 

 the assistance of completed overt trials, but we have 

 no satisfactory quantitative studies which confirm 

 this supposition. So far as I know there is little un- 

 ambiguous evidence that the rat under experimental 

 conditions shows this type of behavior. 



Kohler's observations on chimpanzees do reveal 

 such behavior, apparently of a rather high order. 

 Some laboratory workers may discount his results be- 

 cause they are not supported by statistical tables, 

 graphs, and kymographic records. But the very na- 

 ture of inventiveness makes such standardization 

 difficult. Doubtless such observations can be made 

 under conditions of more satisfactory laboratory con- 



