228 'Journal of Comparative Neurology arid Psychology. 



trials per day for five days; No. 3 approximately 140 trials each 

 day tor 5 days. The number of trials per day varied slightly with 

 the degree of the animal's hunger. 



601-690 



75 or 83^% 



It is evident from these tests that many more trials are required 

 to learn to distinguish red from green, than to discriminate black 

 from white. As already stated, this may be evidence of a response 

 to difference in brightness alone. 



At this point I devised a "card displayer" by which the two 

 colors could be shown in succession instead of simultaneously; 

 it was also necessary to arrange the experiment so that it could 

 be carried on by one person (Fig. i). 



r:r^. 



Fig. I. 



The front of the card-displayer consisted of a board twelve 

 inches high. A round pin or pivot on which two levers could be 

 turned was inserted in a hole near the lower edge ot the board. 

 In the upper ends of these levers colored cards were fastened so 

 that raising one of the levers to a vertical position displayed red, for 

 example, raising the other displayed green. During one test red 

 would be on the forward lever one inch in front ot the other, dur- 

 ing the next test on the rear lever. The animal could not, there- 

 fore, react to the position of the cards. On account ot the diffi- 



