EXPERIMENTS ON THE RABBIT 177 



for the red. This was especially striking in Series 4 and 10, 

 where, when the red and black tests were begun, the interval 

 between successive tests on the same animal was greatly reduced, 

 a circumstance which should have strengthened a motor habit 

 of alternating from side to side, whereas in fact the alternation 

 which had occurred while red and grey were being used now 

 gave way to a habit of choosing the door on one side only. 



b) When in the midst of red and grey series in which red 

 was being regularly chosen as it occurred on alternate sides, 

 single tests were interpolated where black was substituted for 

 grey, the tendency to choose the door on the same side as the 

 one chosen in the preceding test was about as strong as the ten- 

 dency to choose the door on the opposite side (Series 22). 



c) In series where a high percentage of correct choices was 

 being made the rabbits would often move first in the direction 

 of the wrong door, checking themselves while a foot or more 

 away and turning to the correct door. This would hardly have 

 been the case had a motor habit been formed. 



8. The results were not based merely on the formation of 

 an inhibition of the impulse to push the grey door Breed, in 

 his work on chicks, found evidence that in color tests they were 

 not so much learning to choose one door as to avoid the other. 

 Thus, after they had learned to choose black instead of blue, 

 they chose white instead of blue, indicating that the discrimi- 

 nation was not based on a brightness difference, but rather on 

 an acquired avoidance of blue. 7 Is it possible to explain our 

 results with red by supposing that the rabbits learned in the 

 red-grey experiments to avoid grey, rather than to choose red 

 as the darker impression, and that they failed in the red-black 

 tests because the grey, to which they had learned to check 

 reaction, was no longer before them, so that their impulses were 

 equally divided between the two doors which they saw? Un- 

 doubtedly the learning did involve the acquired inhibition of an 

 impulse to push the wrong door: not infrequently, as we have 

 said, the rabbit would run to within a short distance of the 

 wrong door, stop abruptly, and turn to the other one. At other 

 times, for long periods, the animal would dash at once to the 

 right door as if the sight of it set off the impulse instantly. A 

 careful examination of our results shows that the difference 



7 Op. cit. 



