ADJUSTMENT OF FLATFISHES 465 



fish wearing a black mask and placed in a dark -walled tank. In 

 this case, likewise, no change of pigmentation should occur, if a 

 visual comparison between the animal's skin and the surrounding 

 bottom is a necessary element in the reaction. Yet such changes 

 did occur in a considerable number of my experiments, and in 

 several cases they occurred under such circumstances as to go 

 far, I believe, toward refuting the hypothesis in question. 



In two such instances, where a white mask was employed, the 

 fish (at first dark) did become pale upon a white bottom, and this 

 reaction, in the case of each specimen, occurred again after a 

 second trial. 



Once more, a pale fish, which had been kept for ten days on 

 white, was stained (anterior parts only) with potassium permanga- 

 nate. This produced a continuous dark brown mask, covering as 

 much of the animal's skin as it was enabled to see without bending 

 the body. The fish, after return to the white box for a while, was 

 later transferred to a black one. It became plainly darker in five 

 hours and fairly dark in eight hours. The next day it died. Another 

 specimen gave similar results, though not so well marked. 



It must be allowed that in none of these cases was the shade 

 assumed as pale (or as dark), as in normal specimens, but this I 

 believe was due to the inhibitory effect of such severe treatment. 

 That the latter is the true explanation is rendered probable by the 

 fact that reactions were quite as likely to be inhibited which con- 

 formed to the requirements of the visual comparison hypothesis 

 as were those which were contradictory to it. Thus one dark fish, 

 which was covered with a dark mask and then transferred to white 

 showed little or no change in the course of three days. 



Fairness compels the mention of a case in which the reaction (in 

 this instance to a white bottom) was almost wholly inhibited for two 

 days by the presence of a white mask, but occurred within the next 

 few hours after removal of the latter. In this case, the stitches 

 and marginal portions of the cloth were left in situ, and it cannot 

 be said that the effects of injury had been lessened by removing 

 the other parts of the mask. This result, which was obtained 

 much less conclusively with two other specimens, may be held to 

 support the view that what the animal sees of its body is a deter- 



THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 10, NO. 4 



