FACTORS IN LEARNING BY WHITE RATS 349 



different groupings two mazes of ten cul de sacs each and two 

 of six each. For a description of the mazes, the procedure in 

 the experiments, and of the animals the reader is referred to the 

 other report. The results there given will be assumed to be 

 familiar to the reader. In this connection we shall examine the 

 records of seventeen of the rats in the mazes there described 

 as IB and IIB (each with ten blind alleys), and IA (six blind 

 alleys). Only the records of wwtrained animals, or animals 

 without previous maze experience of any kind, will be consid- 

 ered. For convenience of reference the cuts of the mazes are 

 here reproduced (Fig. II). Our present data are taken from 

 animals running twice daily up to a point in the learning at 

 which but few errors were made. The records of two animals 

 in Maze IB and of one in Maze IIB were made useless for our 

 present purpose by slight omissions in the record of the first day's 

 trials. 



Tables II and III show the several reactions of two rats 

 tabulated in the manner explained in connection with Table I; 

 the first is of an animal experiencing considerable difficulty 

 (Rat 11), while the other is of one which much more quickly 

 acquired the habit of keeping the correct path with but few 

 errors (Rat 13). The records of the seventeen animals were 

 all tabulated as those shown in the tables, except that the 

 various " tallies " were first made w : th single lines. All records 

 were then gone through carefully and the critical choices were 

 marked in such ways as to indicate agreement or disagreement 

 with expectations based on frequency and recency factors. 

 Whenever an animal comes to a bifurcation in the maze that it 

 has passed previously while going in the same direction, it is 

 at what is called here a. critical position, and its going into 

 one of the rival alleys is a " critical choice." The second time 

 that it arrives at cul de sac 1 on a forward run is an example. 

 If it passed 1 the first time but now enters it, it goes contrary 

 both to frequency and to recency expectations. The first return 

 run in any section of the maze has critical positions only when 

 the animal arrives at blind alleys which have been entered in 

 the forward run. At such positions both frequency and recency 

 factors favor entrance again, rather than a continuance of the 

 return past this position in the true path. Entrance to such 

 a cul de sac is in accordance with the expectations based on fre- 



