468 FRANCIS B. SUMNER 



for the other changes which are here described. In other words, 

 the skin pigments which were displayed seemed to be restricted 

 to components of the more habitual backgrounds encountered 

 by such fishes. 



3. Upon a homogeneous ground, the pigment of the skin was 

 commonly much more uniformly distributed than upon a back- 

 ground having a diversified appearance. 



4. Upon a mixed background, such as was afforded by one of 

 the ordinary sands or gravels of its customary habitat, the fish 

 took on a definite color pattern, which varied with the texture of 

 the material, and was oftentimes in striking harmony with the 

 latter. 



5. Artificial backgrounds, containing variously distributed 

 areas of pure black and white, called forth far more contrast in 

 the skin patterns than did the less contrasted tones of the sand 

 and gravel. 



6. The principal markings constituting these various skin 

 patterns were found to be permanent, in the sense that they always 

 reappeared in the same positions, and even when the animal 

 adapted itself to a homogeneous background, the outlines of most 

 of these spots were still distinguishable. In the case of Rhom- 

 boidichthys podas, the arrangement of these spots was, in its 

 essentials, constant for all members of the species. Regarding 

 the other fishes used, I cannot speak with the same certainty. 



7. The patterns assumed were consequently limited, in great 

 degree, by fixed morphological conditions. Thus squares, cross- 

 bands, circles, etc., were never copied in any true sense, by the 

 fishes. 



8. Within the limits thus imposed, the capacity of one of these 

 species (Rhomboidichthys podas) to adapt itself in respect to the 

 distribution of its skin pigments was often remarkable. For ex- 

 ample, experiments with painted squares and circles of black and^ 

 white showed that the resulting skin patterns depended not only 

 upon the relative amounts of black and white in the background, 

 but upon the degree of subdivision of the areas of the latter. As 

 an example of this last point, when the background was divided 

 into areas 2 millimeters square, a finer grained appearance was 

 produced in the fish than when 1-centimeter squares were used. 



