27 

 to determine when the maze is learned was shown by 



types of movements in this experiment which were not 

 noticable in learning the latch box problem. These must 

 be eliminated before it can be said that the maze problem 

 is solved. Running movements are chiefly in demand and 

 such movements are far simpler than those involved in 

 raising the latch of the latch box. Probably the most 

 complex organized movements required in running the maze 

 are those made in turning into another path quickly. As 

 soon as these are integrated the time is materially les- 

 sened, for in nearly every case in learning the maze, speed 

 was the first thing attained, the elimination of errors 

 coming later. Errors were made as often when running in 

 the paths as when turning from one path into another one. 



They seem to have been made because of interfering 

 stimuli of the sort which cause distractions. There 

 must take place, in order that a run be error-free, a more 

 consistant or fixed interaction of stimuli which bring 



