120 L. W. COLE 



trials, and with 590 units only 16.66 trials. The same relation 

 holds time for medium discrimination and stimuli of 220, 350 

 and 480 units, but when a stimulus of 590 units was employed 

 the number of trials required for learning to make the dis- 

 crimination increased froin 40 to 50. In order to make certain 

 that this increase in the number of learning trials was due only 

 to the strength of the stimulus I repeated the test with a second 

 group of six chicks and the average was practically the same, 

 namely, 46.66 trials. With medium difference of brightness of 

 the two screens, therefore, the optimal stimulus lies nearer the 

 threshold than under the easy condition of discrimination. 



The responses of the chicks to the third, or difficult, condition 

 of discrimination are less easy to interpret. With the weakest 

 stimulus used for this condition, 350 units, none of the six chicks 

 failed, with the medium stimulus one failed, and with the strong 

 stimulus two out of five failed. Moreover, the utmost patience 

 was required of the experimenter in order that all should not 

 fail. Each trial also required much more time than in medium 

 and easy discrimination. If, however, we consider only the 

 chicks that learned to make the difficult discrimination the 

 relation stated for easy discrimination appears once more, i. e., 

 the stronger the stimulus the more rapid the learning. It seems 

 clear, therefore, that, with difficult discrimination, the strong 

 stimuli divided the chicks into two groups, (i) those which 

 after a few trials ceased to try to escape and would no longer 

 step on the electric wires, and (2) those which chose with greater 

 and greater caution and, therefore, learned to choose correctly 

 after a small number of trials, each of which consumed much 

 time. 



To what shall we ascribe this dual result under the third 

 condition of discrimination? It seemed possible that the chicks 

 were divided into the two groups according to their sensitive- 

 ness to the electric stimulus. That is, the more sensitive chicks 

 might learn most rapidly under the influence of a weak stimulus, 

 be slow to learn under the influence of a strong one, and fail 

 completely when under the influence of both a strong stimulus 

 and a difficult condition of discrimination. 



In order to answer this question twelve chicks were selected 

 of which number six had a threshold of stimulation of 90 units 

 and the remaining six of 150 units (relative values). The former 



