COLOR CHANGES IN TELEOSTS 531 



coal may be reflecting more photic energy than the snow had done. In 

 some way, our perceptual machinery (not our thinking processes — it 

 works too fast for them to be involved) makes allowances for the intens- 

 ity of the general illumination. We can easily be led to see white paper 

 as gray, or black paper as nearly white, if our clues to the overhead illum- 

 ination are eliminated in an experimental situation. Similarly, the even 

 more fully automatic 'allowance-making' mechanism of the fish can be 

 deceived. If, by such devices as the use of translucent material lighted 

 from below, the substrate is made lighter or darker than the overhead 

 illumination would call for, the skin of the fish changes accordingly. 



Sumner early suspected that this ability of the fish to adapt the chrom- 

 atophores to background albedo was due to a vertical polarization of the 

 retina. The retina was thought to control the pigment cells in sympathy 

 with the relative illuminations of its upper and lower halves, correspond- 

 ing respectively to the lower part of the visual field (the substrate) and 

 the upper part (the source of natural light) . Von Frisch soon produced 

 experimental evidence for this view, to which Sumner and others have 

 since added a great deal. 



By means of vaseline-lampblack paint, and by fitting celloidin caps, 

 blackened in various patterns, over the corneas of fishes, Frisch and Sum- 

 ner have shown that when the upper half of the cornea is left clear and 

 the lower half blacked out, the fishes will darken greatly regardless of 

 the tone or albedo of the substrate. If only the lower cornea is clear, pale- 

 adapted fishes remain pale on either white backgrounds or dark gray 

 ones. All-black covers did not always prevent all shade-changing ability, 

 probably because light could still reach the retina through the translucent 

 tissues of the head. Ordinarily, however, fishes so provided darkened up 

 as though they were eyeless. In the entire situation, then, we can see 

 certain tendencies: 



A. When no light is striking the fish (with or without its eyes), the 

 melanophores 'contract'. 



B. When light strikes only the skin (whether the eyes are present or 

 not), the melanophores 'expand'. 



C. If more of any light entering the eye strikes the upper part of the 

 retina, the melanophores 'contract' despite Tendency B. 



D. If more of the light entering the eye strikes the lower part of the 

 retina, the inhibitory effect of Tendency C upon Tendency B is ineffec- 

 tive, and the melanophores 'expand'. 



