672 'Journal of Comparative Neurology and Psychology. 



It seems evident from a comparison of Table 15, with the preced- 

 ing tables and with the following one, that light colors were most 

 difficult for the animals to discriminate. If such is the case, the 

 rather long time required to learn the series which matched Gray 5, 

 is partly accounted for. Group 5 was the first one we tried, and 

 some time was required for the animals to learn to operate the mech- 

 anism. As this was soon learned, however, our tables show that dis- 

 criminations were more readily made among dark than among light 

 colors. 



TABLE 16. 

 VS 1 in Group 20. 



TABLE 17. 

 GBS 2 in Group 30. 



It seemed to us that No. 3 made most of his errors when he seemed 

 to be looking at the colors with his left eye. Since his twenty-nine 

 correct choices in the first thirty trials of the above series showed 

 that he could discriminate the food-color, we compelled him to walk 

 along the row of glasses from left to right in the second series of 

 thirty trials. His correct choices were at once reduced to fourteen. 

 Our suspicion that his left eye was defective arose from the fact that 

 the animal, unless prevented from doing so, invaribly went to the 

 right end of the row of glasses, then along it to the left with his right 

 eye thus nearest to the row. The above reduction in the number of 

 his right choices, and our record of his earlier errors, indicate that 

 his vision with the left eye was poor as compared with the right eye, 

 and as compared with the vision of Raccoon No. 2. 



The records above do not give all the training in discrimination of 

 colored papers which was demanded of the raccoons, for some of the 

 control tests required many trials. The tables do show, however, that 



