64 C. J. CONNOLLY. 



that condition on a black background, which they would not do 

 if the eye were still functioning. 



Spaeth (1913) has shown that light, temperature, and also 

 various salts have a direct effect on the contraction of the chro- 

 matophores, both melanophores and xanthophores, when the scales 

 with the superficial chromatophores are separated from the fish. 

 But in our experiments with the normal fish, these factors did 

 not influence the results, as the conditions, except background 

 were uniform for all. Secerov (1909) maintains that the light 

 has a direct action, but this seems disproved by von Frisch. 



B. Effect of Spectral LigJits or Different Wave-lengths. 



The eye being the receptor of the light stimulus, the results of 

 the above experiments can only be explained on the assumption 

 that Fundulus objectively discriminates light of different wave- 

 lengths. Since, however, intensity apparently plays an important 

 role in the different shades assumed by the fish and to some extent 

 in color, by the relative degree of expansion of the melanophores 

 thus covering or exposing other color elements, a more rigid test 

 was made by using spectral lights of different wave-lengths but 

 of the same intensity. 



In much of the work done on color vision in lower animals 

 where the method of food association with definite colors was used, 

 the factor of intensity was either totally disregarded or arrived 

 at only approximately by comparing the relative brightness of the 

 differently colored objects presented. This method of comparison 

 being subjective, can give no exact quantitative data. Filters 

 whether liquid or glass, are likewise unsafe as they are not al- 

 ways monochromatic, nor do they always exclude the infra-red 

 and ultra-violet rays, and of course differ considerably in the 

 amount of light which they allow to pass through. It is therefore 

 necessary that either the brightness of the various spectral lights 

 should be equated by the use of a flicker photometer, as Reeves 

 (1919) has done in food association experiments, or the lights, 

 varying in quality, should be made equal in the quantity of radiant 

 energy. This latter method was first employed by Laurens (1911) 

 at the suggestion of Prof. G. H. Parker. 



More recently Laurens and Hooker (1917) have described an 



