348 



JOSEPH PETERSON 



second trials. The only escape from this difficulty to the theory 

 would seem to be an assumption that frequency factors are wholly 

 nil up to a certain point, after a given number of trials and suc- 

 cessful arrivals at the food box, and that they then begin sudden Jy, 

 or gradually, to operate, — an incredible hypothesis, It is evident 

 from a careful study of individual records that frequency effects 

 show themselves from the beginning of the experience of the 

 animal in the maze, and that other directive factors are opera- 

 tive in bringing about learning in even as simple a case as this. 



II JIB 



2 



1 



■E 



n 



2 



?'. 



— — ■* 



10 



fOOD 

 BOX 



CAST 



-~ -, 2 



i 



"5 





.--, s 



/a 



FOOD 

 BOX 



J 



EAST 



Fig. 2. — Mazes IB, IIB, and IA as used for present experiment. Dotted cross- 

 lines mark the end of blind alleys in B-mazes. These mazes differ only in 

 the relative lengths of blind alleys. 



Our second objection is a question of fact. Do rats learn the 

 maze in accordance with expectations based on frequency alone, 

 or on frequency and recency? The question can be answered 

 by a detailed examination of the actual records of several animals. 



The method of tabulating individual records for detailed ex- 

 amination requires more space than can be expected for this 

 article. A couple of samples of such tabulations up to points 

 of nearly complete learning will be given, and results calculated 

 from all the records at hand will be shown in more condensed 

 form. The results are taken from experiments described else- 

 where 10 on twenty-four white rats which were made to learn in 

 "Peterson, Jos. Op. cit. 





