Cole and Long, Visual Discrimination in Raccoons. 673 



nine different colors and one gray have been selected by the raccoons 

 from colors wbicb were equally bright for the human eye. In all, 

 thirty-nine different colors were used, and from these the ten food- 

 colors had to be chosen. Up to this point, however, we cannot be sure 

 that the selections were not made by the sense of smell. 



The Possibility of Selection by Odor Differences. We have al- 

 ready stated that the raccoons never swerved toward the food-glass 

 from a distance of one or two feet. In fact, they seemed to be unable 

 to distinguish the food-glass from the otbers at a distance of more 

 than four or five inches. Thus the animal's nose was always very 

 close to the glasses and we were always confronted with the possibility 

 that a keen sensitiveness to odors might account for the behavior we 

 observed. The discrimination by smell might have come about in 

 any one of three ways, (1) The animal might have detected the 

 odor of food when near the glass which contained food. (2) He 

 might, as it were, follow his o^vn trail, and this might be done in 

 either of two ways: (a) since the animal ate the food from his paws 

 and also laid hold of the food-glass more than the others he might soon 

 get an odor of food from the outside of the glass; (6) he might in a 

 similar way detect the native odor of his paws as stronger on the 

 food-glass than on the others, due to repeated pulling at the glass even 

 though he did not communicate the odor of food to it. The last 

 supposition while hardly probable is, nevertheless, possible. (3) The 

 animal might be able to discriminate between the odors of the pig- 

 ments of the colored papers. 



With closed glasses the first possibility is easy to deal with. We 

 used bread as food and often allowed it to become so dry that it had 

 hardly any odor for us. In addition to this we put food in all the 

 glasses, usually the same amount that we put in the food-glass, some- 

 times more than that glass contained. From the first day's work 

 with closed vessels our rule was to put food in all of them during at 

 least half of the trials. In a large part of the experiments, we put 

 food in all at the beginning of the day's work. During a smaller 

 portion of the time all the glasses contained food during the last 

 thirty or sixty trials. In no case did this seem to modify either the 

 animal's behavior or his record of discriminations. 



