208 CARNEGIE INSTITUTION OF WASHINGTON. 



Coloration of Tropical Reef Fishes, by W. H. Longley. 



Between the middle of May and the first of August, my investigation of the 

 colors of tropical reef fishes was continued in Porto Rico and Tortugas. 

 Distinct progress was made in the attempt to determine what law, if any, maj' 

 be apparent in the character and distribution of the pigments externally visible 

 in this group of animals. Present and prospective advance is associated with 

 the recognition of three facts: 



First, in these fishes the dependence of the various color-phases upon the 

 character of the environment may be demonstrated most clearly by observa- 

 tion of individuals over a series of simple bottoms, covered, for example, by 

 bare white sand, brown algae, or the green blades of turtle grass {Thalassia 

 testudinum) ; for over mixed bottoms, where many colors may appear in 

 small patches, animals which seem to be exposed to the same conditions have 

 been recently or are being acted upon by different stimuli whose separate 

 effects it is very difficult to dissociate. 



Second, those species which may be observed at a given station may be 

 arranged according to habit in subclasses in which, upon the whole, there is 

 much narrower range of color than in the larger group. 



Thu'd, conspicuousness is largely dependent upon the relative position of 

 the observer and the observed object. 



Dr. Charles H. Townsend, Director of the New York Aquarium, has 

 described the color-phases of a number of Bermudian fishes confined in aquaria 

 in New York, and is inclined to correlate them with specific psychic states, 

 such as anger or fright, or specific activities, such as playing or feeding. Some 

 of these fishes, together with others whose color-phases are undescribed, have 

 been under my observation in their natural habitat, where their reactions lend 

 slight support to Dr. Townsend's hypothesis. But I find that in some species 

 I am able to evoke different color-phases at will by tempting the fishes to swim 

 from one locahty to another of different and definite character whose influence 

 one may forecast with precision. 



As a result of such experiments or of observations many times repeated upon 

 fishes whose movements were uncontrolled, I am able to state that color change 

 is commonly induced by change in the environment and is dominated by the 

 color character of the environment which actuates it. This conclusion applies 

 to the following species: Iridio hivittatus and I. maculipinna, Lachnolaimus 

 maximiis, Monocanthus hispidus (young), Sparisoma abildgaardi, and S.flaves- 

 cens, Sphyrcma barracuda, and the young of Thalassoma bifasciatus, which 

 are described as T. nitidus and T. nitidissima. Since some of these display 

 those bright hues or have patterns involving those strong contrasts in color 

 whose occurrence has led to the introduction into biological theory of the 

 hypotheses of warning and immunity color, we must believe that fishes are com- 

 mon in which conspicuousness is concomitant with obliterative counter-shading 

 and adaptive color change, or accept the conclusion that the alleged conspicu- 

 ousness of these colors and patterns is overrated through some misconception 

 of the mode in which they cooperate with counter-shading and color change to 

 secure the maximum degree of inconspicuousness in a very difficult situation. 

 Happily, in the choice of alternatives, one is not left to weigh these proba- 

 bilities, for the trend of the evidence is decidedly in favor of the second. 



The coincidence of a given color with a specific habit, e. g., the practical 

 restriction of red among shallow-water fishes to those which lie concealed by 

 day and feed at night, is in itself sufficient to raise the gravest doubts regarding 

 the truth of those hypotheses which postulate conspicuousness. For it is 

 remarkable that out of more than 100 species which I have observed, 5 bearing 

 this color, usually rated as conspicuous, fall in a single subclass defined by 



