578 W. H. LONGLEY 



to some differences in habit upon the part of the three species 

 last mentioned. 



The coloration of Tylosurus raphidoma is simple and appar- 

 ently unchangeable. Its surroundings are equally simple and 

 almost as invariable, for it has not been observed at any time 

 more than a few inches beneath the surface, where it may lie 

 for considerable periods, scarcely covered and almost motion- 

 less. Sphyraena has much in common with Tylosurus, but 

 sometimes sinks to the bottom. As has been noted, under such 

 conditions it changes in shade if necessary, and assimilates itself 

 with the substratum. The vertical range of Abudefduf exceeds 

 that of Sphyraena, and its relations with the bottom are more 

 intimate, yet, in spite of the fact that its color is inconstant, 

 definite adaptive color changes have not been observed in the 

 species The fish bears instead a relatively stable color com- 

 bination, which is compounded of characteristic tones from its 

 environment. 



There is, therefore, no inconsistency in Abudefduf's retention 

 of its dark bands. Its coloration, like that of most other fishes 

 as yet studied carefully in their relation to a complex environ- 

 ment, is a patent compromise. Its range is extreme, and as a 

 result the vicissitudes to which it is exposed are many. But the 

 risks it runs are well distributed; for its pigmentation is such 

 that it is at once both more and less distinctly visible than it 

 should be, if it were all of any one of the several colors perfectly 

 suited to a limited portion of its range. 



To conclude, there is undoubtedly correlation between at least 

 the lighter blues and the probability of being seen from below 

 against an aqueous background. Moreover, the interpretation 

 which it seems necessary to place upon this fact is in no wise 

 opposed by the appearance of others in combination with 'water 

 colors' in one pattern. 



There is httle novelty in the facts concerning the distribution 

 of green among the fishes studied. They are added in the inter- 

 est of completeness and are thoroughly consistent with what has 

 gone before. This color appears upon surface-swimmers, or 

 upon bottom fishes in correlation with the occurrence of green 



