248 F. M. GREGG AND C. A. McPHEETERS 



based primarily, if not wholly, upon the initial members of the 



two series. 



CONTROL SERIES 



Jack failed to discriminate between the two series, but he 

 did learn to inhibit his responses until the final card had ap- 

 peared. The control series of tests was instituted to determine 

 the stimulus of these responses. (i), The color cards were 

 removed and the levers alone were presented. No disturbance 

 of normal conduct resulted. (2), The levers were so operated 

 that they made their usual sound while they did not appear 

 above the screen board. The visual-auditory stimulus was thus 

 replaced by an auditory one alone. No change in behavior 

 resulted. (3), A series of but two sounds invariably stimulated 

 the normal response. When a series of four or five sounds were 

 given, Jack always inhibited his reaction until they ceased. 

 Since all possibility of contact stimuli is eliminated by the 

 distant position of the animal from the levers, these controls 

 indicate that any completed series of sounds constitutes the 

 sensory stimulus for this raccoon. 



To determine the basis of Jill's discriminative reactions, it 

 was necessary to proceed cautiously for fear of destroying or 

 radically altering the coordination. Because of the crucial im- 

 portance of these tests, they will be described with some detail 

 in the order in which they were given. 



1. To test the relative efficiency of color versus lever order, 

 the red and white cards were interchanged. The normal order 

 of colors was given and as a consequence the order of lever 

 positions was reversed from that of the normal. Hereafter in 

 describing the order of lever positions, they will be numbered 

 according to their distance from the animal. The results are 

 given in the table. 



Stimulus Positive responses Inhibitions 



W-B-R, 3-2-1 31 9 



W-B, 3-2 (premature) 8 



R-R-R-, 1-1-1 »... 17 14 



Counting the eight premature responses as correct, we have a 

 percentage of 67 in 79 trials on the assumption that the reac- 

 tions were made to color order. These results are very poor as 

 compared with previous records. On the preceding day the 

 percentage of correct behavior was 91, and the average per- 



