COLOR VISION IN FISHES 489 



water converts the extreme ultraviolet. The fixing of the short-wave limit 

 of the true visible spectrum for such an animal is very difficult. In Gaster- 

 osteus, in contrast to the frog, no retinal action-current can be detected 

 in ultraviolet illuminations; and yet a pure ?l366m[i beam, apparently 

 visible to the fish, causes the complete pattern of retinal photomechanical 

 changes. Quinine-fluoresced light will do likewise, however. Through 

 fluorescence, ultraviolet light may paradoxically lead to seeing without 

 actually itself being seen. The exact position of this lower spectral limit 

 for fishes in general is of little or no biological importance anyway, for 

 a meter or less of water eliminates all of the ultraviolet in sunlight. 



Something, then, is known of the spectral limits, hue-categories, and 

 hue-discrimination in different spectral regions for these two fishes, and 

 we can hope eventually to learn all about the system they employ — 

 whether it has three component central processes, or more. We can 

 already be certain that they are not dichromatic, for they distinguish too 

 many hues and, according to Hamburger, get no sensation of 'white' 

 from any monochromatic light, and therefore have no neutral point. 

 Hamburger made a beginning at an analysis of the laws of color mix- 

 ture as they apply to Phoxinus. He found that fishes trained to white 

 light not only discriminated it qualitatively from every spectral region 

 but recognized, as white, human complementary mixtures of yellow and 

 blue, red and blue-green, orange and blue-violet, and so on. A similar 

 demonstration of complementary colors for Betta splendens, by an en- 

 tirely different technique, was incidental also to the work of Beniuc 

 (see pp. 364-5). 



It is one thing to be able to say that all cone-rich teleosts assuredly 

 have color vision, and quite another thing to say how much color means 

 to fishes. The environment of the average species is rather drab. We 

 have seen that particular colors — red and blue — may provoke particular 

 species to vigorous responses. Miss Reeves' fishes tended, however, to 

 pay more attention to brightness than to hue. How successfully might 

 form also compete with color for the attention of a fish? Is the shape of 

 an artificial lure perhaps more important than its color, even though the 

 latter is perceived? Recently some experiments have been made along 

 this line by a Japanese, Horio : 



This investigator trained carp positive to a red disc and negative to a 

 blue one. They learned this discrimination readily — not so readily, how- 

 ever, the discrimination of a white triangle (positive) and a white square 

 (negative) . The better to compare the effectiveness of form and color, 



