678 Jourtial of Comparative Neurology and Psychology. 



the food-glass was reached. In cases of turning back before food was 

 obtained they pulled at some of the glasses twice in each trial. Except 

 in the first fifteen trials with Raccoon No. 2 there was food in only 

 one glass, GIBS 2. 



It seems fair to conclude, from these experiments, that the animals 

 were not making their selections of the food vessel by means of the 

 odor of the food, of the pigment, nor of their own paws. Evidently 

 the discrimination was a matter of vision or of some sense unknown 

 to us. 



Visual Discrimination. Since the possibility of discrimination by 

 means of the sense of smell has been eliminated, we must inquire by 

 what visual criteria the different colored papers were distinguished. 

 Watson has pointed out that "the surfaces of the papers differ greatly 

 owing to accidents in manufacture, dyeing, ironing, etc.," and that 

 there is difficulty in "j^asting them upon surfaces so that slight differ- 

 ences do not appear."'^ Discrimination by means of these criteria 

 must be guarded against, and in the case of the raccoons it was done 

 as follows. (1) Not one food-glass alone was used but a half dozen 

 different ones of the same color. (2) As already stated (p. 674), at 

 the end of a test, we filled the glass holder with six glasses of the 

 same color to see whether the animals could pick out the single one 

 in which food w^as placed. They simply pulled at every glass. 

 Besides testing the animals, in this way, on several colored papers, 

 they M-ere also tried on gi'ay and white. It seems apparent, there- 

 fore, that they were not being guided in their choices by any secondary 

 criteria which the papers may have presented. The possibilities, 

 then, seemed limited to two. Either the animals discriminated 

 between the several glasses (a) by means of their brightness differ- 

 ences, or (6) by means of their differences in color. 



{a) As we have stated, the colors were selected so as to be of 

 equal and very nearly equal brightness for the human eye, and as 

 wide a range of brightnesses was used as the ninety colored papers 

 would furnish. The value of our tests rests on the assumption that 

 colors of equal brightness for the human eye may be somewhere near 



^'Watson, .1. B. Some experiments bearing; upon color vision in monlveys. 

 Jour. Comn. Neur. and Psych., vol. 19, pp. 3-4. 1909. 



