184 H. M. JOHNSON 



limits of the bird's discriminative ability which I regard them 

 as being. It will be seen that the threshold- values for the chick 

 vary irregularly from 33% to 42% of the width of the striae 

 on the standard field. The variations may be explained by 

 assuming that discrimination was difficult throughout this region 

 of stimulus-differences. If the bird received punishment several 

 times in close succession shortly after the difference approached 

 this region, he "gave up" earlier, and yielded a larger "thresh- 

 old" than if his errors were more widely distributed. 



Table 5 also shows a great disparity between the "upper" 

 and the "lower" threshold values given by the monkey when the 

 width of the striae on the standard field was 0.520 mm. This 

 fact pointed to a large effect of practice. The magnitudes of 

 the thresholds obtained at the smaller values of the standard 

 stimulus are so much lower than those obtained in the earlier 

 stages of the work at the larger stimulus- values, that it was 

 necessary to make a control test to discover if these differences 

 were not due to the effect of training, instead of being a function 

 of the absolute w T idth of the striae. This test was made between 

 March 4, 1915 and March 27, 1915. It shows quite clearly that 

 the differences were due to the effect of training. The results 

 suggest strongly that if training had been continued sufficiently 

 long after the full effect of practice had been obtained, the 

 values for all the difference-thresholds where the striae on the 

 standard field were over 0.3 mm. wide would have borne a rela- 

 tion to the absolute width of the members of the standard sys- 

 tem analogous to Weber's law for brightness. This relation 

 probably does not hold for absolute widths below 0.2 mm., 

 under these experimental conditions, since such fine systems 

 become increasingly hard for the monkey to distinguish as 

 striate. Due to pressure of other work and the small likelihood 

 of Monkey 2 living through many more months, I did not feel 

 justified in carrying this exploratory study farther at the time. 

 Should a similar study ever be made, it would seem advisable 

 to select fewer points at which to determine the animal's thres- 

 hold, and to give a large number of presentations — several 

 hundred, at least — of a number of differences in either direction 

 from each point. In work on another problem with this animal 

 I found such procedure quite fruitful. 



