^?6 yournal of Comparative Neurology and Psychology. 



:>3 



Obviously, factor "i" does not account for all the 266 first 

 choices given in this table. From this total we must subtract {a ) the 

 88 first choices that started the 88 series; (b) the 49 correct first 

 choices that occurred in the course of the series (although it is a 

 fair inference that factor "i" accounts for many of this latter 

 group, such an inference cannot be drawn with certainty, since 

 the animal often terminated a series by choosing the attached 

 pedal, only to return to the favorite pedal of the series thus inter- 

 rupted at the next trial); {c) the 42 first choices that occurred as 

 the second member of a series when the initial first choice of that 

 series was successful (such reactions are tabulated under factor 

 "2"); and (d) the 25 times when the animal might have been fol- 

 lowing up a successful color or odor in making his second succes- 

 sive first choice of a given pedal. 



To summarize: (i) there are left 266-88-49-42-25 = 62 

 first choices that cannot be accounted for on the assumption of a 

 transiently formed association between the desired result and an 

 immediately preceding successful pedal, color, or odor, or to the 

 influence of the directing sign board. These 62 first choices — all 

 of them incorrect — may be ascribed to factor " i." 



(2) In loi trials the animal made (incorrect) first choice of the 

 pedal that had just let him out. 



(3) In 102 trials he made (incorrect) first choice of the pedal 

 bearing the color or odor of the immediately preceding attached 

 pedal. This occurred in 83 (25.9 per cent) of the 320 experi- 

 ments with colors alone; in 15 (15 per cent) of the 100 experiments 

 with colors reinforced by odors; and in 4 (20 per cent) of the 20 

 experiments with odors alone. 



(4) Out of the total 440 experiments in which colors, colors 

 and odors, and odors alone were used, he made correct first choice 

 of the attached pedal 122 times (27.7 per cent). It is here that 

 we must look for any light that the experiments may have thrown 

 upon our problem. The question at issue is, "Did the presence 

 of the directing sign board increase the adequacy ot the animal's 

 reactions to the situations in which it entered ?" Inferences 

 drawn from a mere numerical tabulation of results must, of course, 

 be qualified by the character of these 122 objectively correct first 

 choices. A great deal depends on whether his choice of the 

 attached pedal appeared to be haphazard, "hit-or-miss," or to 

 be attended by such preliminary activities as would have made for 



