PATTERN-DISCRIMINATION IN VERTEBRATES 183 



seemed room for doubt whether the animal's errors were due to 

 the magnitude of the stimulus-difference or to disturbance from 

 some other cause, I presented a larger difference at a number 

 of trials in the same daily series with the smaller difference. If 

 all or most of the incorrect choices were made at the small 

 stimulus-difference, I concluded that discrimination at that stim- 

 ulus-difference was becoming difficult; if the animals' percentage 

 of correct choices was low for the large stimulus-difference as 

 well, I assumed that the source of disturbance was extraneous. 

 The values obtained on the two animals are not strictly in- 

 tercomparable. In work on the monkey I reduced the stimulus- 

 difference by smaller gradations than I could use in the work 

 on the chick, owing to a limitation of the optical instrument 

 by which the field was formed. As the angle of rotation of the 

 gratings over each other becomes very small, as is the case 

 where the width of the visible stripes thus formed is large, a 

 very slight change in the angle of rotation makes a large differ- 

 ence in the width of the visible stripes. In the work on the 

 monkey a larger angle and smaller gradations could be em- 

 ployed, since he was sensitive to much smaller widths than was 

 the chick under similar conditions. I took for "threshold-differ- 

 ences" for the monkey the differences at which his average per- 

 centage of correct choices most nearly approximated 75. For 

 the chick I took the stimulus-difference at which the first break- 

 down of discrimination not apparently due to disturbance from 

 other causes occurred. This procedure is open to criticism in 

 that I did not ascertain to what extent the bird could be made 

 to overcome his uncertainty by continued training. At the 

 time, however, this bird did not react well under punishment. 

 As soon as the stimulus-difference became relatively small he 

 usually refused to inspect both fields and adopted a position- 

 habit. I recognized this defect in the procedure, and in a later 

 piece of work attempted to settle the question. 7 I found that 

 this bird could be made to improve to a limited extent after 

 very long continued training. The degree of improvement which 

 I was able to elicit is not sufficiently large to invalidate the 

 results herein presented as the rough approximations of the 



7 Johnson, H. M. Visual pattern-discrimination in the vertebrates. IV. Effec- 

 tive differences in direction of visible striae for the monkey and the chick. To 

 appear in this journal. 



