8 BULLETIN OF THE BUREAU OF FISHERIES. 



differ somewhat in size and form from those nearer the proximal ends of the scales. 

 Those lying superficial to the proximal areas of the scales and, consequently, beneath 

 the overlapping areas of other scales, are characterized by relatively few irregular long 

 and slender radial processes. Over the distal areas of the scales the majority of the 

 melanophores lie immediately beneath the epidermis. In these areas, as well as over 

 the proximal areas of the scales, the xanthophores lie somewhat deeper than the 

 melanophores. Many of the guanophores Ue at the level of the melanophores, while 

 others lie somewhat deeper. 



Relatively few chromatophores occur in the deeper layers of the skin. In many 

 instances they are closely aggregated around blood vessels in the corium and in the 

 subcutaneous tissue. However, these chromatic organs in the deeper layers of the 

 skin are probably too few and too far removed from the surface to play any important 

 part in the shade or color assumed by the fish. 



DEVELOPMENT AND DEGENERATION OF CHROMATOPHORES. 



In surface preparations, as well as in sections of the skin, distorted and fragmented 

 melanophores are not infrequently observed between the cells of the epidermis (text 

 fig. 2), sometimes actually protruding from the surface. Doubtless these melanophores 



are being eliminated through the epidermis. 



' *■ - This phenomenon was observed in the 



rs:/^ 



t majority of the specimens examined, regard- 

 less of their shade or color. It was quite 

 as common in specimens which were newly 



Fig. 2. — Camera lucida drawing showing fragmented melano , 1. * i_ 1 j . 



phores between the cells of the epidermis, (i, epidermis; taken aS m thOSe WhlCh had been kept m 

 2. layer of chromatophores and guanophores just beneath aquaria f Or shor'ter Or longer periodS. Obvi- 



ously the elimination of melanin pigment, 

 and also its more or less continuous production, must be looked upon as normal physi- 

 ological processes. A similar elimination of xanthine pigment through the epidermis 

 was not observed. However, this negative observation affords no evidence that xan- 

 thine pigment is not eliminated in this manner. Xanthine pigment is highly unstable 

 and would soon disintegrate following the fragmentation of the xanthophores. 



When a fish undergoes a change from a darker to a lighter shade and maintains 

 that shade more or less constantly, a material reduction in the number of melanophores 

 in the layer just beneath the epidermis takes place. Averages of a large number of 

 actual counts of the melanophores in this layer in a given unit area of the skin of speci- 

 mens newly taken or adapted to dark backgrounds and of specimens adapted to a 

 white background indicate a reduction in the latter of approximately 30 per cent in 

 the number of melanophores in the superficial layer of the skin. The rate at which 

 this reduction took place in a single specimen which was placed on a white background 

 as soon as it was brought into the laboratory is indicated by the curve ab in the accom- 

 panying figure (text fig. 3). This curve is based on the averages of a large number of 

 actual counts of the melanophores in a given unit area on the surface of detached scales 

 taken from various parts of the body. The figures in the vertical line indicate the 

 number of melanophores in the given unit area, those in the horizontal line indicate the 

 number of days following the initial counts. The initial counts were made when the 

 fish was placed on the white background. This curve shows, furthermore, that a 

 maximum reduction of approximately 28 per cent was reached in 11 days and that 



