TRANSFER FACTORS IN MAZE LEARNING 343 



an animal could be sensitive to direction-in-general regardless of 

 what his auditory, visual, olfactory, kinesthetic, and other stim- 

 uli are, is a proposition almost defying interpretation in psy- 

 chological terms. The matter may be given point by repeating 

 an assertion made by a former colleague of the writer: "If I 

 were to find an animal orienting himself in a certain absolute 

 direction I should at once want to discover what the particular 

 cue is that is guiding him." That numerous cues may so func- 

 tion is conceivable and an unravelling of the efficient cues and 

 of their respective roles would make material sufficient for a 

 complete series of experimental studies by itself. Any slight 

 sloping of the maze floor, any shadows cast, the presence of the 

 experimenter in a constant position, the presence of a post or 

 piece of furniture, a constant difference of temperature on dif- 

 ferent sides of the room, a constant direction from the nests where 

 other rats are feeding and making slight noises these and other 

 factors are all conceivable as effective here. For the present 

 study, however, the writer must be content with having shown 

 that a general orientation in what was the food box direction, 

 whatever the cues used, became gradually established through a 

 series of changing mazes having this direction constant. 



II. SECOND SERIES 



The foregoing experiments were carried on in the winter of 

 1918-19. The following spring and summer a complete repeti- 

 tion of them was made for two reasons. In the first place, it 

 was felt that on account of the limited number of subjects used, 

 the experiments could be called only preliminary; whereas a 

 checking of results by a new set of animals used at another time 

 would give them more definite value. In the second place, the 

 ending of the curves for the experiments with the test maze Q 

 at the 8th trial (fig. 7) is unsatisfactory because it leaves partially 

 undetermined whatever the relation is between the previous prac- 

 tice on the other mazes and the later runs on this. Does previ- 

 ous maze practice affect only the initial attack upon a new maze 

 or does it accelerate or retard the whole learning process? The 



