TRAINING IN WHITE RATS UPON VARIOUS SERIES OF MAZES 65 



it consisted of four or sixteen. As shown in the average 

 total time per rat, it was positive throughout the entire 

 series in which the preliminary training took place in the 

 harder maze, and negative throughout the series in which 

 it took place in the easier maze, except that when these 

 sixteen preliminary trials took place in the easier maze 

 there was a positive transfer of .75 per cent. 



When two mazes were learned together, the first being 

 partially learned, then the second completely learned, and 

 finally the first completely learned, it was found that 

 no one of these methods enjoyed a superiority according 

 to all three criteria. The mazes chosen for this purpose 

 were the ones in which the greatest degree and the least 

 degree respectively of positive transfer had appeared in 

 the preceding experiments. When the one in which the 

 least had appeared and the one in which the greatest had 

 appeared were learned in this order for purposes of control, 

 it resulted in the two being learned with fewer trials and 

 errors than when the order was reversed; the time was 

 slightly greater, being an average of 1.22 as against 1.19. 

 The advantage, from the point of view of trials and from 

 that of errors, in learning the maze first in which the least 

 transfer-effect had appeared was manifest throughout the 

 entire series of learnings where this may be viewed as the 

 fundamental order. The advantage from the point of view 

 of time resides with the other order, in which the maze 

 manifesting the greatest transfer was learned first. 



When a maze was relearned after four intervening maze- 

 learnings, only one group showed, according to all three 

 criteria, signs of the retention of beneficial effects from 

 the original learning of the maze. Two groups showed 

 signs of the retention of a beneficial effect in time, but 

 with no unmistakable evidence of such benefit in either 

 trials or errors. 



