THE COLORATION OF REEF FISHES 553 



unsuited to its modified condition seemed worth investigating. 

 The turtle grass (Thalassia) provided suitable material for mak- 

 ing the test. The rootstock of the plant is usually deeply im- 

 bedded in the sand. The leaves spring from short offshoots^ 

 and for half their length, on the average, are not exposed to the 

 light. Their buried portions are white, therefore by cutting off 

 the green parts and making shallow plantings of the basal seg- 

 ments it was possible to manipulate the materials in a small 

 tank, so that the fishes were exposed to two sets of surroundings, 

 whose essential difference was in color. 



In the white grass Iridio bivittatus, Monacanthus hispidus, 

 Siphostoma mackayi and Sparisoma hoplomystax all reacted 

 as they did over bare sand, except that one young Siphostoma 

 w^as even lighter in color than in that case. To the green they 

 responded as indicated in table 1. Hence color changes in 

 fishes appear to be induced directly by the color of surrounding 

 objects,^ and not indirectly through suggestion derived from 

 familiar forms. There is, therefore, nothing here to lend col- 

 lateral support to such hypotheses as that of Steinach ('01), re- 

 jected indeed by Cowdry ('11), that the chromatophores in the 

 octopus are controlled by reflexes from the suckers, and that, as a 

 consequence, the texture of the bottom, rather than its color^ 

 determines the color changes that occur. 



At this point one should refer for a moment to an idea one 

 frequently encounters, and which seems in fair way to become 

 an article of faith in the matter of animal coloration. The 

 reasoning upon which it rests is wholly illogical, as the reader will 

 observe. Sometimes it is simply affirmd ex cathedra: "Absence 

 of movement is absolutely essential to protectively colored 

 animals." (Beddard, '92, p. 90.) . Sometimes it is stated with 

 some attempt at justification: ''No color whatever could make a 

 flying butterfly invisible to its enemies, because the background 

 against which its body shows is continually changing during its 



^ In view of what is known of the path of nerve-impulses, the fact that the 

 fishes are able to match different colors proves, of course, that they possess 

 color vision. The point has already been made by Mast ('14), but the demon- 

 stration of adaptive color changes in additional unrelated species emphasizes 

 a conclusion that has been commonly mistrusted. 



