592 W. H. LONGLEY 



diversity of coloration to be observed at particular points is 

 due simply to the fact that most extensive overlapping or strati- 

 fication of specific ranges occurs there; they are foci at which 

 different habitats converge or interdigitate in greatest numbers. 

 About the coral stacks, for example, the diurnal reef-rangers 

 meet and mingle mth fishes, which all day long s\\T.m in and out 

 through the interstices between the heads. An occcasional red 

 recluse is seen dimly or emerges momentarily from the shadow, 

 and conveys the impression that its kind associates closely ^^^th 

 those which lie at ease or move sluggishly about, as digestion of 

 last night's gains goes on apace. Others rising or falling, break- 

 ing the water's surface or lost in the gaping darkness of the reefs, 

 like animated shuttles bear the weft that binds them in an al- 

 most indissoluble whole. 



When one becomes steeped in such facts as these, much of the 

 current literature of animal coloration is hopelessly unconvincing. 

 One reads without enthusiasm such sentiments as the follomng:^ 



On every Indian lake three species of kingfisher pursue their pro- 

 fession cheek by jowl. . . . It is obvious that all .... of 

 these diversely plumaged birds cannot be protective^ coloured. It 

 may perhaps be objected that their piscatorial methods differ in de- 

 tail. We admit that this is the case, but would maintain, at the same 

 time, that these comparatively slight differences in habit do not ac- 

 count for the very striking differences in plumage. 



In view of established fact one feels that the significant 

 idea expressed above is precisely that which its authors dismiss 

 as immaterial; but, hereafter, although the correlation between 

 color and habit is conclusively demonstrated in only one class of 

 animals, the burden of proof of such statements as the second 

 in the preceding paragraph lies \AT.th those who make them, 

 whether they refer to distinct species or to the two sexes in dimor- 

 phic forms. 



It may not be settled conclusively at present whether the cor- 

 relation between color and habit demonstrated in this paper is 

 due to natural selection or the action of the environnient, since 

 intensive researches upon the degree of protection afforded their 



9 Dewar and Finn ('09), p. 202. 



