TRAINING IN WHITE RATS UPON VARIOUS SERIES OF MAZES 25 



alley 1, 60 per cent of those made by group 1 and 74 

 per cent of those made by group 2 were made in the same 

 blind alley. All the conditions under which the three 

 groups learned the maze were the same, and the only 

 difference among the groups was in the number of mazes 

 they had learned before being placed in D. Group 3 

 had learned one maze, group 2 had learned two mazes, 

 and group 1 had learned three. It would seem therefore 

 that the difference in the records of group 3 on the one 

 hand and of groups 1 and 2 on the other must have been 

 due to some factor or factors acquired prior to the learning 

 of the C maze. Reference to the patterns of the mazes 

 concerned will furnish a clue as to the nature of this factor. 

 In maze C the true pathway proceeds directly from the 

 entrance without necessitating any change in the orientation 

 of an animal that was introduced straight through the en- 

 trance; but in mazes A and B the true pathway turns 

 to the left, demanding a turn to the left in an animal so 

 introduced. The marked tendency of the rats in groups 

 1 and 2 to turn into blind alley 1 in maze D would seem 

 accordingly to come from a habit brought to maze C, 

 leading them to seek the true pathway in the same position 

 which the blind alley occupies in maze D. 



It would be reasonable to affirm that, on the basis of 

 the principles adduced to explain the cases of negative 

 transfer, we should look for negative transfer in the group 

 which learned mazes E, A, B and C before learning the 

 E maze, for in the E maze as well as in mazes A and B 

 the true pathway lies to the left of the entrance. It would 

 seem equally reasonable to expect that the tendency to 

 take the first runway to the left of the entrance in maze 

 D would be more pronounced in group 1 than in group 2, 

 and more pronounced in group 5 than in either 1 or 2, 

 since group 5 learned three mazes in which the true path- 

 way takes this course, group 1 learned two, and group 2 

 learned one. Contrary to these assumptions, we find that 

 the transfer of group 5 on maze D is positive. But it is the 

 least of all the positive transfers, being only 4.33 per cent; 

 and when we examine the distribution of errors according 

 to the blind alleys and the trials, we see the same tendencies 



