COLOR VISION IN FISHES 481 



By alternating the gray and yellow backgrounds he could thus keep the 

 fish at a constant brightness of skin pattern. Left on the yellow back- 

 ground for an hour or more, the fish turned yellowish; but it would 

 never do this on the matched gray. 



Similar results were obtained with Crenilabrus pavo and Trigla corax. 

 And, when yellow and blue fluids, both so concentrated as to appear 

 black, were used as backgrounds, Frisch found that gradual dilution 

 evoked graded brightnesses of skin coloration but an expansion of the 

 yellow chromatophores occurred only in the case of the yellow solution. 

 He chose two shades of yellow papers and found a gray which gave the 

 same brightness-stimulation as the lighter of the yellows. When the skin 

 of a fish had adapted to the darker yellow, substitution of the gray for 

 the dark yellow caused an immediate paling of the skin. This demon- 

 strated that the two yellows differed more in brightness than did the 

 lighter yellow and the gray. Prolonged exposure to either yellow now 

 caused a yellowing, which would not take place on the gray, or on green, 

 blue, or violet backgrounds — on these, the xanthophores contracted. 



Freytag shortly repeated some of Frisch's work with Phoxinus and got 

 negative results. His fishes responded to the shade of the background 

 but not to the color, even after twenty-four hours. Reeves later rather 

 lamely suggested that Freytag had not waited long enough for the color 

 change to take place. It is far more likely that Freytag's specimens simply 

 came from the wrong river! At about the same time, Haempel and 

 Kolmer were reporting their work on Phoxinus Icevis and Cottus gobio, 

 where the only color-mimicry they observed was a reddish response of 

 Phoxinus to red backgrounds, no reaction being given to gray, or to other 

 colors. Their specimens had come from a red-bottomed river, the Wiirm. 

 Years later, in 1920, Schnurmann found that while Munich specimens 

 behaved just as Frisch had described them, others from Ulmar gave no 

 color-response to yellow, orange, or red backgrounds. 



Another fish which is a favorite with physiologists is our own killifish, 

 Fundidus heteroclitus; but unfortunately its dermal color-repertoire is as 

 limited as that of Phoxinus. Connolly placed killifishes on backgrounds 

 illuminated with spectral lights carefully equated in intensity by ther- 

 mopile measurements. It took several days for the specimens on red and 

 yellow grounds to become distinct from those on blue. A more versatile 

 species was found by Frisch — Crenilabrus roissali, which adapts to red, 

 green, and blue grounds as well as to yellow ones. To the achromatic 

 (scotopic) human eye the brightnesses of the backgrounds offered by 



