1 88 BULLETIN OF THE BUREAU OF FISHERIES. 



the white areas except one near the pectoral fin disappear. In general, it may be 

 said that the larger the proportion of the background covered by the black areas the 

 larger the proportion of black in the skin and vice versa, and the smaller the black 

 and white areas in the background the smaller those in the skin. This is well shown 

 in figures 9-17. 



These statements hold, however, only within certain limits. If the black areas 

 are small it requires a larger proportion of black, under certain conditions, to produce 

 the same effect than it does if they are larger. This is shown in some of the records in 

 table III and in plates xxxi, xxxii. A background with dots 3 mm. in diameter, 20 mm. 

 apart, produces very nearly as much black in the skin as one with dots 2 mm. in diameter 

 and 10 mm. apart (fig. 52, 54), although in the former there is only 1.71 per cent black 

 while in the latter there is 3.14 per cent black. And if the dots are very small there 

 are no black areas produced in the skin, no matter how great the proportion of the 

 background covered with black (fig. 49). In the specimen used in obtaining the results 

 recorded in table iii, dots 0.5 mm. in diameter, 2.5 mm. apart, produced no black 

 spots in the skin, while dots i mm. in diameter, 10 mm. apart, did, although in the 

 former case 3.39 per cent of the background was black and in the latter only 0.78 per 

 cent. If the figures are very small the black acts just as it does if uniformly distributed. 

 That is, a background containing numerous very small spots has the same effect as a 

 uniform gray background. This matter was, however, tested only in a single indi- 

 vidual. 



If the light and dark areas are excessively large the skin also tends to assume a 

 uniform shade; or, at any rate, there is no simulation of the background. The size 

 of the figures required to produce this varies with the size of the fish. The 2 cm. squares 

 represented in figures 11 and 15 are very near if not slightly beyond the limit in size 

 for simulation in specimens of the size, 14 cm. long, shown in these figures; while in 

 smaller specimens, 10 cm. long, represented in figure 18, the limit in size, of the figures 

 simulated is about one-fourth as great. The effect of relatively coarse-grained bottoms 

 like these depends, however, upon the position of the fish. If the fish is held so that 

 the head is on a black area and the edge of the anterior end of the animal is on white, 

 as in figure 15, the skin assumes a much larger proportion of white than it does if the 

 head is held on a white area and the edge is on black. The production of a uniform 

 shade results if the figures are so large compared with the size of the fish, that one eye 

 is ordinarily affected mainly by black and the other mainly by white. We shall con- 

 sider the cause of this later. 



While large specimens come to resemble the background more closely than small 

 ones if it contains relatively large configurations, the opposite is true if the configura- 

 tions are small. In the smaller specimens, 5-15 cm. long, adaptation also proceeds 

 more rapidly than in the larger ones, 20 cm. and up. 



Relation between the pattern in the skin and that in the background. — Our evidence 

 shows conclusively that both on natural and on artificial backgrounds there is a con- 

 figuration produced in the skin which resembles that of the background. This is clearly 

 shown in the photographs already referred to. This resemblance between the fish and 

 the background was, however, even more remarkable in specimens which had been for 

 several weeks in a variegated dark blue and white granite pan. (PI. xxvii, xxviii.) 

 Is this resemblance due to an actual reproduction of the configurations in the back- 



