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



mature responses was significantly increased, the positive series 

 was reacted to correctly, while the negative series gave more 

 positive than negative responses. Such results might naturally 

 be expected from the animal's excitement and impatience. In 

 the few cases where the levers worked smoothly and the animal 

 seemed in a normal mood, discrimination failed for the minimum 

 of spatial difference. However, we must confess that our test 

 was not decisive. 



1 6. A final series of tests covering three days of experimen- 

 tation was given (i) to determine the composition of the two 

 lever groups which renders them effective as specific stimuli, 

 and (2) to determine the relative effectiveness of the three 

 members constituting any lever group. Since the cards are 

 ineffective, they were not removed. Dr. Hunter operated in 

 the enforced absence of Mr. Gregg. The results are given in 

 the following tabular statement. 



Stimulus Positive 



W-B-R (1-2-3) 44 



R-R-R (3-3-3) 5 



W-W-W(l-l-l) 13 



R-R-R (3-3-3) 1 



W-R-R (1-3-3) 11 



R-R-R (3-3-3) 1 



B-B-B (2-2-2) 6 



R-R-R (3-3-3) 1 



"W-W-W(l-l-l) 25 



B-B-B (2-2-2) 6 



W-B-B (1-2-2) 23 



B-B-B (2-2-2) 4 



As to the character of a group constituting a specific stim- 

 ulus we may say: — (a), The group is highly general, never 

 definite and particular. Positive reactions of food getting may 

 be stimulated successfully by any of the following groups — 

 1-2-3, ^"^i 1-2-2, 2-2-2, or i-i-i. Likewise, inhibition, or 

 negative responses, may be stimulated by either group 3-3-3, 

 or 2-2-2. (b), The nature of either stimulus is relative to the 

 character of the other group with which it is alternated. The 

 group 2-2-2 may serve equally well either as a positive or nega- 

 tive stimulus. It excited the negative response with an average 



