BEHAVIOR OF RACCOONS 253 



correctness of 73% in two of the above series; it likewise func- 

 tioned on one occasion in inducing positive responses with a 

 correctness of 100%. (c), When two groups are alternated, 

 that nearest to the animal serves as a positive stimulus, while 

 the more distant one becomes the negative stimulus. The 

 group 2-2-2 was positive when alternated with 3-3-3, but it 

 functioned negatively when given in conjunction with 1-1-1. 

 (d), The effectiveness of any two groups in inducing differential 

 responses is proportionate to their distance apart. The groups 

 1 -1 -1 and 3-3-3 were discriminated with an accuracy of 96% 

 for 27 trials, while the above two groups were distinguished 

 from 2-2-2 with an accuracy of but 87% for 55 trials. 



The above principles may be stated in terms of positions as 

 follows: — When two groups of positional stimuli are alternated, 

 the one nearest to the animal becomes positive while the more 

 distant one is negative. The accuracy with which the two 

 groups may be discriminated is proportional to the degree of 

 depth distinction between them. The animal thus responds to 

 objects according to their relative position in depth. 



As to the relative effectiveness of the three members com- 

 prising any group stimulus, our results indicate with a high 

 degree of probability the following conclusions: — (1), The dis- 

 crimination can be based exclusively upon the difference between 

 the first members of the two groups. Only these initial factors 

 were effective in a majority of the tests. This conclusion is 

 founded upon several groups of facts. In no. 16 of the control 

 tests, 66 trials were given in which the last two members were 

 identical in the two discriminated groups, yet these trials gave a 

 discriminative accuracy of 85%. Moreover, this accuracy was 

 proportionate to the degree of space difference between the 

 two first members; for example, the groups 1-3-3 an d 3-3-3 

 gave an accuracy of 95% as compared with a percentage of 

 80 for such groups as 1-2-2 and 2-2-2. When two groups are 

 identical as to the first member but different in respect to the 

 last two positions, discrimination is impossible. In no. 5, the 

 groups 1-2-3 an( i 1-1-1 were distinguished in but 12% of the 

 trials, a record which can be accounted for on the basis of 

 chance. As before noted, the discrimination was mediated 

 through contact and the first member of each group was usually 

 nosed. The inhibitory response to the negative series was 



