72 HAROLD C. BINGHAM 



the source distances within a maximum of 40 cm., and by irregu- 

 larly changing the shifter from left to right. 



Chick 27 learned the problem so much more rapidly than the 

 other subjects that it was evidently ready for the one-two relation 

 before the others had acquired the one-three habit. In view of 

 the possibility of the inequality in the two stimuli of the dark- 

 light alternations, it was deemed desirable to reduce the difference 

 rates towards the threshold. This would not only indicate the 

 chick's threshold of difference, but also make less the possibility 

 of different light-dark ratios serving as the basis of discrimination. 

 Since early experience had emphasized the importance of concen- 

 trating the work as much as possible while the chicks were in 

 satisfactory condition, further work with the other chicks was 

 abandoned and the mechanism was changed for one-two dis- 

 crimination. The change seems to have been fortunate, for the 

 other chicks dropped off shortly, and chick 27 did not last long 

 enough to complete more than an elementary phase of this par- 

 ticular task. 



The details of the learning of this task by chick 27 are furnished 

 in Chapter VIII where tables 15 and 16 present quantitatively 

 several phases of behavior in addition to the criterion of correct 

 choice per series of ten tests. Table 10 presents the discrimination 

 conditions, reactions of the chick, and the controls employed. 



To determine the specific rates of the two stimuli, the method 

 was adopted of recording with a stopwatch the time elapsing 

 during fifty flashes of either stimulus. Thus the time relation is 

 given as 1-3 or 39" - 13" for the first twenty series. In series 

 21 the same relative rate was continued, 1-3, but the specific rates 

 were made slower, the time for 50 flashes being respectively 54 

 and 18 seconds. The next change, introduced in series 30, January 

 10th, results in both relative and specific variation. The flicker 

 relation becomes 1-2 and the respective periods for 50 flashes 

 are 36 and 18 seconds. Slight changes were made in series 31, 

 and in series 32 the effect of substituting without additional train- 

 ing, 80-40 and 20-10 periods was observed. 



At the close of series 20, the controls had become rigorous 

 enough to indicate that the flicker difference, or some aspect of it, 

 was really perceived. On January 8, at the close of the 26th series, 

 an unusually severe control was introduced. The upper lamp was 

 changed from a 4 c.p. to a 16 c.p. illumination, and in addition, 



