TRANSFER FACTORS IN MAZE LEARNING 



335 



entered three times by one of the four animals of the group, two 

 times by another; alley B was entered four times by one rat, 

 twice by each of two rats; alley C was taken three times by one 

 rat; alley E was taken twice each by two rats. (Of course, the 

 alley names, A, B, C, D, E, do not refer to the same alleys in 

 successive mazes the method of tabulation is used for conven- 

 ience.) Apparently, the animals were gradually learning not to 

 re-enter a cul-de-sac previously entered on the same run. 



Comparing the different methods of training for the four 

 groups, we see that group I was given no acquaintance with any 

 feature of the maze to be used later for testing. Group II was 

 given acquaintance with the sorts of materials of which the maze 

 was to be constructed, but was given no culs-de-sac to learn to 



TABLE 2 



eliminate .and hence no true maze situation. Group III was 

 given a complete maze problem of the usual type. Group IV 

 was given mazes but was not allowed to mechanize the true path 

 in any case, having always new blind alleys to chance into. If 

 this group were to show after effects of the twenty-five trials, 

 they would be due obviously not to the learning of specific turns 

 and twists as in the case of group III, but to the development of 

 more general traits. / 



The testing for transfer effects was given in the form of the 

 maze problem Q shown in figure 6. Unfortunately, the results 

 were invalidated after the eighth trial by an accidental factor. 

 The records after this trial were very irregular, the behavior of 

 the subjects on a following day being almost unpredictable. This 

 was not only quite unlike the results of other investigators but 



