222 



BULLETIN OF THE BUREAU OF FISHERIES. 



SELECTION OF BACKGROUND. 



No one has heretofore obtained any_ evidence indicating that fish show any ten- 

 dency to select a background which harmonizes with the shade, color, or pattern of 

 the skin. Sumner (191 1, p. 443) made a number of tests with Rhomboidichthys and 

 obtained negative results. He is very conservative, however, in his conclusion based 

 upon these results, stating merely that they "render it improbable that the fish exer- 

 cises much selection in respect to the shade of its background." The results here 

 noted of experiments on both Paralichthys and Ancylopsetta show very clearly that 

 such selection occurs. They throw but little light, however, on the question as to 

 the extent and the significance of the selection. 



In these experiments two methods were pursued. One may be called the group 

 method and the other the individual method. 



GROUP METHOD. 



In the group method a given number of specimens, usually 10, which had been 

 either on a white or a black background continuously for several days or longer, were 

 released in an aquarium one half of which was black and the other white. Their 

 movements were then studied and their positions at given intervals were recorded. 

 The aquarium used for this purpose was 55 by 70 cm. The bottom of this aquarium 

 was divided lengthwise into four equal areas. Two of these, with the adjoining side 

 walls, were painted white and the other two, alternating \vith the first, were painted 

 black. In these tests, specimens of Paralichthys were exclusively used. They varied 

 in length from 12 to 20 cm. At the close of the first test two new individuals were 

 added. With this exception the same individuals were used throughout the tests. 

 In making these tests the individuals were put into the aquarium and left undisturbed. 

 At intervals of about five minutes their positions were noted and the number on the 

 white and on the black, respectively, was recorded. At the close of each test these 

 numbers were added. The resulting sums appear in table v. 



Table V. — The Relation Between the Shade of the Skin andThat of the Background Selected. 



