BEHAVIOR OF RACCOONS 255 



groups 3-2-1 and 1-1-1, we would expect a high degree of dis- 

 crimination, with the group 1-1-1 being regarded as the positive 

 stimulus. As a matter of fact the animal was confused and 

 unable to give consistent results. Both stimuli were sometimes 

 regarded as positive and sometimes as negative, though the 

 tendency to a positive selection was the stronger in both cases. 

 The group 3-2-1, however, was selected as positive much more 

 frequently than was the group 1-1-1. These results offer three 

 possibilities of interpretation: — (a), An animal may adapt itself 

 to the alterations of the discriminative stimuli which are intro- 

 duced in the control tests. In this way the nature of the effec- 

 tive stimulus may be gradually altered. By interpolating but 

 a few controls in a longer series of the normal, one can prevent 

 any radical alteration except as to some minor details. For 

 example, much of the relativity of the two group stimuli in 

 our experiment was probably acquired during the controls. The 

 anomalous results of the first test can be explained upon the 

 hypothesis that the discrimination was originally based upon 

 both color and lever order, the breakdown was due to the func- 

 tional opposition of the two factors, and the element of color 

 order was subsequently neglected. There are several objections 

 to such a hypothesis. The visual element must have been 

 stronger than the factor of position to account for the strong 

 positive tendency in the responses, and it is improbable that 

 the stronger component should be subsequently neglected. The 

 change or adaptation should be gradual and there is no evi- 

 dence of this. The animal which stood in such a position that 

 lever contact was impossible failed to learn the discrimination 

 on a visual basis, (b), The first alterations of any accompany- 

 ing conditions upon which the discriminative responses do not 

 directly depend may cause confusion and disturbance and the 

 animal may soon learn to adapt itself to these novel disturbing 

 conditions. There was much in the behavior of the raccoon 

 which suggested such a hypothesis, (c), The results may also 

 be explained by additional assumptions as to the nature of 

 the group stimuli, which in no way conflict with our previous 

 conclusions. We may assume that both the first and second 

 members of the group are sometimes effective, that the animal 

 tends to react positively to any spatially distinct series of stim- 

 uli, and negatively to any series of identical members. While, 



