PATTERN-DISCRIMINATION IN VERTEBRATES 321 



one feeding-place and reject another. Taken at their face value, 

 some of the results in audition of the dog reported by members 

 of Pavloff's school seem to indicate that this is actually true. 



In the present study it seemed advisable at the outset to set 

 some problems of a different type from those heretofore attacked, 

 and to develop methods of studying those problems. This arose 

 from consideration of certain limitations of apparatus used in the 

 standard methods of studying vision in animals, and of certain 

 limitations of the current problems themselves. 



STANDARD METHODS OF STUDYING VISION 



For some years vision in animals was investigated by more 

 or less impromptu methods of great diversity. The data thus 

 obtained were at best incomparable. The conclusions drawn 

 from them were often vitiated by the lack of control of physical 

 factors. 



In response to a widespread demand for improvement and 

 standardization, Yerkes and Watson 2 published in 1911 the results 

 of an extensive investigation of apparatus and procedure. For 

 studying color-vision they recommended an apparatus consisting 

 essentially of a high-intensity spectrometer system, equipped 

 with devices for selecting monochromatic bands and presenting 

 them to the animal. The stimuli thus obtained are accurately 

 measurable and highly controllable as to wave-length, intensity 

 and saturation. The apparatus meets all physical requirements. 

 Since it has yielded quite definite behavior results, both positive 

 and negative, it seems well established as the standard. No 

 other adequate method has been described up to the present time. 



To prepare stimuli for testing the sensitivity of animals to 

 differences in brightness, size and form of visual objects, the 

 authors recommended the use of standard forms presenting a 

 constant difference only in the physical characteristic whose 

 effect is under study. They report that ' differences of the 

 compared stimuli as to . . . color . . . texture, posi- 

 tion, odor," etc., can be eliminated in their method. The animal 

 to be tested is stimulated by light transmitted through two opal 

 glass screens whose respective illumination is independently 

 controllable. They are covered with mutually interchangeable 



2 Yerkes, R. M. and Watson, John B.: Methods of Studying Vision in Animals. 

 No. 2, Behavior Monographs, Cambridge, Holt, 1911. 



