172 H. M. JOHNSON 



under mydriasis for Chick 1 was 0.75 D. hyperopia in the right 

 eye and 1.25 D. hyperopia in the left eye. For Chick 2 the 

 error was between 0.25 D. and 0.50 D. hyperopia in both eyes. 

 No astigmatism was discoverable in either bird. After the ani- 

 mals had recovered from mydriasis, Dr. Cobb repeated the 

 tests, but, since the birds were continually changing their ac- 

 commodation, the results were quite indefinite. In the emme- 

 tropic human eye an hyperopia as great as that found for Chick 

 2 usually exists during mydriasis, owing to lack of tone in the 

 ciliary muscle. It seems fair, therefore, to regard this bird's 

 eyes as practically emmetropic, since no error was discoverable 

 without mydriasis. 



Chick 1 failed to establish a perfect or highly accurate habit 

 in 300 trials, although the width of the members of the positive 

 system of striae was three times as great as that of the members 

 of the negative system. At this point I abandoned the attempt 

 to train him. Later on I gave him the problem of discriminat- 

 ing between two regular systems of striae the members of which 

 were equal in width, but differed in direction by 90°. He failed 

 to learn this problem also. When I introduced a 2 to 1 differ- 

 ence in width in addition to the difference in direction, he estab- 

 lished a perfect habit in 150 additional trials, and maintained 

 discrimination until the relative width of the members of the 

 two systems was reduced to a 5 to 4 relation. The bird's record 

 over 1100 trials indicated that a difference in either width or 

 direction is ineffective if it is presented alone, but that either 

 difference is effective if it is made sufficiently large and is pre- 

 sented with a maximal difference in the other characteristic. 

 This fact raised some questions of great interest, which I was 

 prevented from studying by an injury to the bird's left eye. 

 I shall, therefore, omit a detailed report on the work done on 

 Chick 1. 



Chick 2 and Monkey 2 learned the problem in less than 100 



trials each. 



BASIS OF DISCRIMINATION 



In attempting to train an animal to respond to a difference 

 between two stimuli either of which may be made the variable, 

 it is important to know whether the animal is reacting on the 

 basis of relative difference, or by choosing or rejecting a familiar 

 stimulus. If in work on this problem, the animal can be trained 



