32G H. M. JOHNSON 



CASTEEL'S EXPERIMENTS IN PATTERN-DISCRIMINATION 



Casteel 7 worked on the painted turtle, using a modified form 

 of the Yerkes box. The animals were trained to choose between 

 two food-boxes, equally illuminated. The food-boxes were 

 ' identical in every respect," except that each carried a different 

 pattern, painted directly on the box or on cardboard covering the 

 box. The patterns consisted of alternate " black and white " 

 stripes respectively equal in width. The experimenter set two 

 problems: discrimination between a system of horizontal striae 

 and one of vertical striae; and between two systems of striae of 

 different widths but lying in the same direction. With respect 

 to the former, it may be remarked that Casteel did not ascertain 

 how small a system could be distinguished as a striate field: i.e., 

 what band-width was just distinguishable at a given distance from 

 a field of equal mean brightness uniformly disposed. This would 

 have added greatly to the value of an already interesting set of 

 results. In both experiments Casteel obtained discrimination from 

 several animals in less than 600 trials. It is unnecessary to point 

 out the difficulty of adapting Casteel's method to quantitative 

 work. 



ELEMENTARY PROBLEMS IN PATTERN-DISCRIMINATION 



Broadly the term pattern-discrimination may be used to desig- 

 nate discrimination between visual fields equal in outline, area 

 and average brightness, differing only in the respective disposition 

 of their brightnesses. The present work is an attempt to study 

 four elementary problems included in such a study. 



1. The stimulus-threshold for striation. Given two striate 

 fields, the striae on one of which are too small to be resolved by 

 the eye at the given distance, the striae on the other field being 

 large enough to be distinguishable: How great must be the 

 width of the individual striae on the second field to occasion 

 discrimination between it and the sensibly uniform field ? The 

 visual angle subtended by one of the striae may be taken as a 

 convenient measure of the animal's visual acuity under the 

 experimental conditions. 



2. The difference-threshold for size (and conversely for 

 number) of visible striae. Given two systems of striae, the width 

 of the members of one system being greater and their number 



'Casteel, D. B., "Discriminative Ability of the Painted Turtle." Journal of 

 Animal Behavior, I, 1911, pp. 1 ff. 



