HEREDITY OF PIGMENTATION 163 
indistinct pale cells showing white with a closed diaphragm which 
fora time I took to represent the redchromatophores. Later, how- 
ever, it was found that ‘these cells were probably the processes 
of the black chromatophores from which the pigment had with- 
drawn itself. For, all over fishes that were slightly older similar 
cells were found, with beautiful branched processes showing 
white by reflected light, and at the center of almost every one 
a small mass of contracted black pigment. We have then at 
the time of hatching what appears to be a clear case of the domi- 
nance of the F. heteroclitus character ‘presence of red chromato- 
phores’ over the F. majalis character ‘absence of red chromato- 
phores.’ 
Immediately after hatching, however, this state of affairs 
began to change, for the red chromatophores began to fade, and 
when the fish were fed well had entirely disappeared in three or 
four days. When the fish were starved these chromatophores 
were visible in some cases for several days longer. These pig- 
ment cells did not contract or die but usually remained well 
branched and expanded as long as they were visible. The pig- 
ment, however, faded until it could only be seen in a few of the 
cell processes, usually lasting longest at the tips of these processes. 
Then it became practically invisible by transmitted, though still 
visible with reflected light; and finally could not be made out at 
all. 
The possibility that in order to be visible this pigment needs 
something that it had been obtaining from the yolk but which is 
absent in the ordinary food naturally suggested itself, and per- 
haps receives some support from the fact that the pigment faded 
sooner when the fish was fed. But on the other hand, this rapid 
fading might equally well have been due to a general accelera- 
tion of development due to the feeding. An attempt was made 
to test the matter by feeding yolk but the close of the breeding 
season prevented conclusive results. 
