METHODS OF STUDYING VISION IN ANIMALS 45 



of human vision, but because of their low intensity, in com- 

 parison with prismatic spectra, they are less satisfactory than 

 the latter where a wide range of intensities for a given hue is 

 demanded. As soon, however, as it becomes possible to rule 

 a large concave grating with short focus, the diffraction grating, 

 in all probability, will supersede the prism, since with such 

 a grating collimating and objective lenses would be unnecessary. 

 The grating would have the additional advantage of rendering 

 it possible to get a good yellow. This is extremely difficult 

 with the prism. 



Prismatic spectra seem to us, at present, the most satisfactory 

 chromatic stimuli. We have therefore decided to use them in 

 connection with our standardized apparatus for the investigation 

 of color vision in animals. 



The advantages which spectral stimuli obtained by "diffrac- 

 tion" or "refraction" have over chromatic stimuli obtained 

 by "reflection" or by "transmission" (as used in this report) 

 may be thus stated: 



(i) They are perfectly under the control of the experimenter 

 with respect to (a) color, (b) saturation, and (c) intensity. 



(2) They are measurable, and therefore describable, with a 

 ■degree of facility and accuracy which is not attainable in con- 

 nection with other methods. 



Inasmuch as the remainder of this report is to be devoted 

 to the description of a method of using spectral light, it is need- 

 less to discuss the matter further at present. 



3. Measurements of stimulus 



Whatever the means employed for obtaining chromatic stimuli, 

 they should' be accurately describable in terms of wave-length 

 (quality or composition) and intensity (psycho-physiological 

 and physical). 



As has been pointed out, it is practically impossible to describe 

 stimuli obtained from surfaces which exhibit selective absorp- 

 tion and reflection. It is considerably less difficult, however, 

 to describe stimuH obtained by the use of ray filters. Finally, 

 it is possible to obtain highly satisfactory descriptions of the 

 stimuH used in the " color " apparatus now to be described. 



