THE COLORATION OF REEF FISHES 543 



Labridae of Marseilles and assigns them to different stations. 

 Noe and Dissard ('94) find that Labrids are brown among rocks 

 and green in algae. Jordan and Evermann ('98) note that 

 Iridio bivittatus is variable in color, and Reighard ('08) records 

 the same of this species and of Thalassoma bifasciatus; but in 

 every instance one would infer from the context that the observ- 

 ers were impressed by what they beheved comparatively fixed 

 color types, rather than by any marked power of color change 

 in individuals. 



That notable color changes do occur rapidly in many species 

 of tropical fishes, may fairly be described as a matter of common 

 knowledge among zoologists. Those acquainted with the fact, 

 however, place very different interpretations upon it. Some 

 idea of the diversity of opinion which prevails will appear from 

 the following considerations. 



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

 has studied fishes confined there in exhibition tanks, and has 

 published detailed descriptions of a large number of them. While 

 he finds that coloration is determined to some extent by the 

 construction of the tank and by the position of the fishes in it, 

 Dr. Townsend states ('09, '10) that frequent changes are de- 

 pendent upon activity, rest, play, fright, temperature, food and 

 so forth, and thus suggests the association of the various color 

 phases with specific psychic states. Dr. F. B. Sumner ('11) 

 thinks that these changes are probably of no more utility than 

 blushing and various other evidences of emotional disturbance 

 in ourselves. Prof. S. J. Hickson ('10) believes, upon the con- 

 trary, that there can be little doubt that some are phases of 

 protective coloration, of which the excitement phase is not one, 

 and that much work must yet be done before it can be deter- 

 mined whether any phase has definitely a warning significance. 

 Mr. Thayer ('09), convinced of the truth of his theory of animal 

 coloration, but, as he informs me, without other ground for his 

 opinion, has hazarded the suggestion that the chief use of such 

 changeableness is adaptability to various backgrounds. Finally, 

 Professor Poulton ('08) states categorically that rapid adjustable 

 protective resemblance is suited to wandering forms which must 



