572 r. W. GAMJJLK. 



Hippolyte ^ive the first indication that the purplisli colour 

 is actually deveh^ped in a few weeks when the animal is 

 exposed to <^reen light. 



The significance of the scai'let colouring, so characteristic 

 of abyssal Crustacea and of certain more shallow-water forms, 

 e. g. Hemimysis lamornee, is still obscure, but the obser- 

 vations made above as to the development of red (vermilion) 

 pigment in young specimens kept in darkness may throw 

 some light upon the subject. With i-egard to Hippolyte 

 varians, the facts so far ascertained are these : 



The red pigment is the first to appear. It arises in the 

 larva, even if tiiis is reared in darkness, and the amount at 

 the time of hatching is roughly proportional to that in the 

 mother. In adolescent specimens subjected to darkness the 

 scarlet pigment increases in auiomit. 



VI. Slmmaky ov Results. 



Crenilabrus melops. 



(1) The colouring of young specimens is due in part to the 

 blue endo-skeleton and in part to chromatophores. 



(2) On backgrounds of weeds these fish assume varied 

 coloration. On brown weed they become brown, on green 

 weeds green, on red weed green. 



(3) In light transmitted through weeds, Crenilabrus 

 assumes a colour, the complement of that which is most 

 strongly represented in the incident light. Thus, in light 

 mainly green, a brownish red colour (due largely to red 

 pigment) develops. In light mainly red, a green colour (due 

 largely to yellow pigment) develops. 



Hippolyte varians. 



(1) In any brood the amount of larval pigment (which is 

 always red) is constant, and is correlated with the amount of 

 red pigment present in the female parent in all colour- 

 varieties except green. 



(2) A given green Hippolyte throws one ot" three kinds 



