TRAINING IN WHITE RATS UPON VARIOUS SERIES OF MAZES 55 



the data for the solution of the problem, and any attempt 

 at an explanation here would be but the advancement of 

 conjecture. 



The first three trials of the groups returning to the D 

 maze after they have learned the E maze do not all show- 

 positive transfer, as is the case in the E maze after having 

 learned the D. The group that had had two preliminary 

 trials in D showed a saving both in errors and in time; 

 but the group with four preliminary trials showed the 

 same average number of errors as did the control-group, 

 while those with eight and sixteen preliminary trials re- 

 spectively made more errors than the control-group. With 

 all four groups there was a constant saving in time. 



