A SIMPLE MAZE 



359 



TABLE 2 



The average number of trials required for learning the simple maze by animals trained 



under diverse conditions 



is 36. 3 4.8, and as this is nearly eight times its probable error 

 indicates a real effect of the distribution of practice. 



But among the groups trained by different methods the effects 

 of distribution of practice vary enormously, from a minimum of 

 53 per cent to a maximum of 338 per cent increase in practice 

 required for learning following equal lengthening of the practice 

 periods. The relatively slight individual variation within the 

 groups, indicated by the small probable errors, makes the dif- 

 ferences significant. 



The methods of training most effective in producing variations 

 from the rate of learning determined in the customary way were 

 preliminary exploration of the maze without food and presence 

 of screened food in the cul de sac. The significant differences 

 found are : 



(C and A are practically identical.) 



The influence of these different methods of training was in the 

 same direction for both concentrated and distributed practice, 

 but only in concentrated practice were marked effects produced. 

 The simplicity of the movements involved in traversing the 

 maze made it possible to record the behavior of the animals in 

 detail and to distinguish characteristic differences in the behavior 

 of the different groups which seem to be correlated with the rate 



