194 BULLETIN OP THE BUREAU OP FISHERIES. 



The photographs show that the light reflected from the fish had very nearly the 

 same chemical effect on the plates as that reflected by the background, no matter what 

 color it was, indicating that the fish were well adapted with reference to brightness 

 or shade. By examining the autochromes it will be seen at once that the color of 

 the fishes differs greatly. The specimens on the blue appear blue, those on the 

 green appear green, those on the yellow appear yellow, and those on the red appear 

 brownish. This becomes strikingly evident if the plate is covered with the gray sheet 

 fastened to it so as to eliminate the background from view. There can be no question 

 but that the autochromes show a correlation between the color in the background and 

 that in the skin of the fishes, but the correlation seen in these autochromes is not as 

 marked in many respects as it actually was in the living specimens. The blue, the 

 pink, and the green are fairly accurately reproduced; the red and the yellow not so 

 accurately. This may have been due, at least in part, to the action of the water on 

 the light; for in order to prevent the formation of waves by the water forced through 

 the gills in the process of respiration it was necessary to have the water over the fish 

 1-2 cm. deep. 



Autochromes were taken of specimens on all of the nine different colors or shades, 

 a total of nineteen, but those of the two shades of brown could scarcely be distinguished 

 and there was very little difference between them and those of the two shades of yellow. 

 The difference between the vermilion and the maroon was also much greater than the 

 autochromes indicate. In general, there was actually much more difference in the color 

 of the fishes than the reproductions indicate. 



The original plates of nine of the autochromes reproduced on plates xxxm-xxxv 

 were exhibited at the meeting of the American Society of Zoologists held in Philadelphia 

 in January, 19 14. There was much surprise expressed at the remarkable contrast in the 

 color of the fishes and the correlation between it and that of the different backgrounds, 

 but in spite of the fact that the autochromes show less difference in color than there 

 actually was in the living specimens and minimize the correlation between their color 

 and that of the background, it seemed inconceivable to many that there could have 

 been even as much as shown ; and not a few suggested that the animals must have been 

 translucent so that the background showed through them, or that the color must have 

 been due to the reflection of light from the sides and the bottoms of the boxes. 



I shall present evidence showing conclusively . that these suggestions are not 

 valid. An autochrome of a specimen of a given color in a box of a different color 

 would have settled this whole matter, but there was so much difficulty in keeping the 

 specimens quiet long enough for the purpose required, without removing them from 

 the box in which they were, that, owing largely to lack of time, this was not attempted." 

 Fortunately, however, in one of the autochromes on the green color taken at the close 



a During my second season at Beaufort (1914) after this section of tlie paper was completed. W. P. Hay very generously took 

 a number of autochromes of flounders of a given color on backgrounds of a different color. Nearly all of these autochromes were 

 taken in direct sunlight. Thus the time of exposure was greatly reduced, and consequently much of the trouble experienced 

 during the preceding season in keeping the animals quiet was eliminated. 



Several of these autochromes are reproduced on plates xxxiii-xxxv (fig. 60, 64. 65. 66. 67). Figure 64 represents a specimen 

 autochromed in a green bos inmiediately after it had been taken from a blue box in which it was fully adapted. Figure 65 

 represents four specimens autochromed in a green box immediately after they had been taken from blue, green, yellow, and 

 pink boxes, respectively. Figure 67 represents the same individuals autochromed on a white backgroimd. These three figures 

 show conclusively that the color seen in the autochromes of the fishes is not due to reflection of colored light from the background, 

 and that it therefore must be due to the structure of the skin. I shall not, however, eliminate the other evidence presented in 

 favor of this conclusion, since it is of value in elucidating other characteristics associated with the process of adaptation. 



