262 Papers from the Marine Biological Laboratory at Tortugas. 



and Hepatus may reach the sliore and mingle there temporarily with the 

 colonies of predaceous gray snappers. They invariably return in a short 

 time to the reefs. 



The colors of the reefs themselves are subdued. The coral rock is gray, 

 in places nearly white. The mantle of coral polyps seen in mass varies from 

 cream to light-brown or delicate pea-green. It is almost a monochrome. 

 Only the dark shadows in the fissures or an occasional purple sea-fan lends 

 contrast to the picture. 



The coral-reef fish are nearly all conspicuous (plates i to 5, except fig. 9) 

 either because of bright colors, patterns of contrasting colors, bizarre form, 

 erratic movements, or because of some combination of these qualities. They 

 often combine with conspicuousness disagreeable qualities in the form of 

 defensive spines. The characters of some of these fish are given in table 13, 

 page 299, and are discussed in section VIII of this paper. The conspicuous- 

 ness is that commonly found in insects said to be warningly colored and the 

 combination of this with defensive spines suggests the combination of con- 

 spicuousness and disagreeable qualities of many warningly colored insects. 



In contrast to the fishes of the reefs are those commonly seen over the 

 bottom of coral sand near shore. These are inconspicuous. The gray 

 snapper, Lntianus griseus, as it appears over the coral sand (plate 4, fig. 9), 

 is an example. It is piscivorous and its close agreement in color with its 

 environment appears to be a case of general aggressive resemblance. Its 

 common prey during my stay at the islands was the so-called sardine or 

 hard-head, Atherina laticeps. This also is inconspicuous. It is almost invis- 

 ible when seen from below against the surface film, as it ordinarily appears 

 to the gray snapper. The fish is a plankton feeder. It does not find its food 

 on the reefs nor seek their shelter. Its color is an instance, apparently, of 

 general protective resemblance. Photographs were obtained of this species 

 in its natural environment. Although the photographs were excellent the 

 fish were so inconspicuous that the loss of contrast in the pictures, inevitable 

 in the process of reproduction, made it inadvisable to make plates from 

 them, as they would have shown practically nothing. 



II. THE PROBLEM STATED. 



The conspicuousness of coral-reef fish appears to me to exclude any 

 explanation based on protective or aggressive resemblance {cf. Wallace. 1891, 

 p. 266). Semper (1879, p. 386) says, speaking apparently of coral reefs 

 in general : 



Tlie surface of a reef lying just under water lias often been compared to a gay 

 garden of flowers and the splendour of such a " bed " of animals is in fact quite 

 astonishing. It is as tliough mother nature had here given free play to the fancy she 

 is elsewhere compelled to restrain in some degree, by indulging her delight ni lavish- 

 ing all the colors of the rainbow and by inviting a motley company of cre.itures to 

 disport themselves among the flowers and fruits of her submarine garden blue and 

 red star-fish, Holothuriae, of every hue, and gaudily painted fishes. 



