476 FRANCIS B. SUMNER 



lutely small amount of light reflected from it. In the same way, 

 the interiors of the white vessels used in my experiments seemed to 

 the outside observer-" to be white, and those of the gray vessels to 

 be gray, despite the fact that the latter actually appeared far 

 lighter in a photograph. 



Are the perceptions of the fish similarly determined ? How can 

 such a thing be possible, in cases where the uniform walls of the 

 receptacle constitute practically the entire visual field of the ani- 

 mal ? Without any outside standard of comparison, how can a 

 heavily shaded surface of white appear paler than a brightly hghted 

 surface of gray ? 



We have seen that one simple solution of this difficulty would 

 be to assume that the fish makes a direct visual comparison be- 

 tween its own body surface and the bottom on which it lies. If 

 the former is adjusted to the latter, the absolute degree of illumi- 

 nation which is common to the two is a matter of no possible 

 consequence, for the object of this adjustment is the concealment 

 of the animal. This hypothesis was, however, rejected, in view 

 of pretty conclusive experimental Gv:':l3nce. Furthermore, it 

 does not accord well with the fact (p. 443) that the behavior of the 

 nsh does not seem to be influenced in other ways by the color- 

 phase in which the animal happens to be. 



What, then, is the standard of comparison by which the fish (or 

 its unconscious nervous mechanism, if the reader prefers) deter- 

 mines the shade to which the skin is to conform? In other 

 words, if, as has been demonstrated, the absolute amount of light 

 reflected from the background is not the only factor in the 

 effective stimulus, what other one is there? As just stated, the 

 human observer would decide the point by reference to other ele- 

 ments of the visual field. From the fish's point of view, the only 

 other element of the visual field, besides the bottom and walls of 

 its tank, is the illuminated area overhead, representing the source of 

 light — commonly sunlight reflected from objects outside the tank. 

 May not, then, the ratio between the light reflected from the near- 



*" At least they did so to me. It is quite possible that one who did not appreciate 

 the density of the shadow might have judged otherwise. 



