EXPERIMENTS ON THE RABBIT 167 



that the rabbit would then make its red and black choices on 

 the basis of the experience gained with red and grey, rather 

 than on the basis of experience gained with the red and black 

 combination itself. Dark Nose had given some indication of 

 seeing the velvet black as darker than the red, and the Stoelting 

 black had apparently been for Light Xose and Dark Nose of 

 equal brightness with the red. If Abednego could be trained 

 to choose the darker impression as such, he should choose the 

 velvet black oftener than the red in these interpolated tests, 

 while no difference should appear in the numbers of choices of 

 red and Stoelting black. The 150 red and grey tests were made 

 without a single error. In all, there were 26 interpolated red 

 and veil ct black tests. In these he chose the black 46 per cent. 

 of the time and the red 54 per cent. There were 31 interpolated 

 red and Stoelting black tests: the red was chosen 51.5 per cent. 

 of the time and the black 48.5 per cent, of the time. The series 

 offered no evidence that the velvet black was seen as darker 

 than the red It is possible, though in view of the other results 

 with red and black and of the unlikelihood that the rabbits could 

 accumulate experience from a few experiments of one kind 

 scattered among experiments of a different kind, not at all 

 probable, that the absence of discrimination between red and 

 black was due to the fact that both doors were open in these 

 red and black tests. 



. A special treatment of the results of the red and black inter- 

 polated tests was made to investigate the possibility that the 

 perfect record which the rabbit was making with the red and grey 

 was due to the formation of a habit of going to either side alter- 

 nately, instead of to a visual discrimination. If such an alter- 

 nating habit was being followed in the red and grey tests, then 

 when a red and black test was introduced into the series the 

 rabbit should have had a tendency to choose the door on the 

 opposite side to that chosen in the experiment immediately 

 preceding. That such was not the fact is shown by the follow- 

 ing figures. Out of a total of 57 interpolated red and black 

 experiments, in 47 per cent, the rabbit went to the same door 

 that he had opened in the preceding test, and in 53 per cent he 

 went to the opposite door. Evidently his perfect series of choices 

 had nothing to do with a tendency to alternate from side to 

 side. The series lasted from April 2$ to May 8. 



