TRAINING IN WHITE RATS UPON VARIOUS SERIES OF MAZES 27 



any one of the criteria was cumulative throughout the series, 

 but there were some instances where there was cumulation 

 for three maze-learnings and one where there was cumula- 

 tion for four. These are given in table 21. 



TABLE 21 



CuMtJLATivE Savings in Average Trials, in Average Errors per Trial and 

 IN Average Time per Trial 



Savings in Trials 



Group Mazes Percentages of Transfer 



in Mazes Named 



3 C to D, D to E, E to A 27, 68, 75 

 5 E to A, A to B, B to C 20, 34, 57 



Savings in Errors 

 2 B to C, C to D, D to E 33, 58, 95 



5 E to A, A to B, B to C 53, 60, 71 



Savings in Time 

 2 CtoD.DtoE, EtoA,AtoB 45, 73, 86, 92 



4 D to E, E to A, A to B 65, 85, 92 



Over against these five instances of cumulation for three 

 mazes and one of cumulation for four, there were eight 

 instances in which the transfer decreased in amount through 

 three maze-learnings. Three of these were in trials, four 

 in errors and one in time. They are presented in table 22. 



TABLE 22 

 Successive Decreases in Transfer of Training 



According to Average Trials 



Group Mazes Percentages of Transfer 



in Mazes Named 



1 A to B, B to C, C to D 59, 39, —3.58 



2 D to E, E to A, A to B 81, 76, 74 

 4 E to A, A to B, B to C 74, 60, 44 



According to Average Errors per Trial 



1 A to B, B to C, C to D 70, 59, 45 



2 D to E, E to A, A to B 95, 90, 78 



3 D to E, E to A, A to B 87, 80, 79 



4 E to A, A to B, B to C 84, 65, 54 



According to Average Time per Trial 

 1 B to C, C to D. D to E 88, 76, 71 



