PATTERN-DISCRIMINATION IN VERTEBRATES 221 



The same animal was later trained in a short time to discrim- 

 inate between two visual objects. During the training period 

 the stimuli differed in area, pattern and luminous intensity, but 

 only the latter difference was effective. When in a control test 

 it was eliminated, and a 1 to 4 difference in mean brightness was 

 presented with a 4 to 1 difference in area, the animal ceased 

 to discriminate and did not learn to discriminate on the new 

 basis. In this experiment and the one preceding reliance was 

 placed on the animal's (assumed) ability to accommodate, and 

 obtain a clear retinal image. The validity of the assumption 

 could not be established on empirical evidence. 



The same animal, continued in training by the same method, 

 was now presented with a field covered with an extremely coarse 

 pattern, to be discriminated from a uniform field of the same 

 area, form and mean brightness. The conditions of the experi- 

 ment insured the formation of a sharp image of the test-fields on 

 the animal's unaccommodated eye. He failed to learn the 

 problem or to make any improvement in 500 trials. 



The cumulative evidence yielded by the above experiments 

 indicates that the dog is very deficient in detail- vision, and that 

 the immediate cause is the relative insensitivity of his retina to 

 differences of distribution of brightnesses over it. 



Nela Park, Cleveland, Ohio. 



