PIGMENT. 79 
tween the cells of the epithelium, and are therefore frequently 
curved : they are in general thickest in the neighbourhood of 
the cell, and diminish as they proceed from it ; but they some- 
times also swell out slightly at some distance from the cell. 
These fibres give off others at different points. The presence 
of the cell-nucleus, and the fact that all the stages of transition 
from indubitable pigment-cells to these bodies may be demon- 
strated, are sufficient evidence that these black spots, with the 
fibres proceeding from them, are actually cells, and that the 
fibres are hollow prolongations of them filled with pigment. 
These transitions are delineated in plate II, fig. 8, just as they 
existed close together in another part of the tail of a tadpole: 
@ is an indubitable pigment-cell, scarcely differing from an 
ordinary one; it has also its nucleus. The only circumstance 
which distinguishes the majority of the primitive cells of these 
stellated pigment-cells from common pigment-cells is that they 
are generally smaller, and more closely filled with pigment. 
6 is a smaller cell, which has commenced to taper; and ¢ is 
distinctly elongated into a fibre. A slightly clearer spot is the 
only indication of the nucleus in both imstances. d and e 
elongate at both ends into fibres, one of which (the upper end 
of d) terminates in a knob witha defined outline. At the spot 
where this knob unites with the body of the cell, a shading, 
indicating a cavity, may be clearly perceived, the pigment 
being more closely deposited in the neighbourhood of the cell- 
wall than in the centre; and lastly, fis a cell which elongates 
into fibres on three sides. When a small piece of the 
skin of the tadpole is torn in water, separate portions of 
these pigment-fibres, or prolongations of the cells filled with 
pigment, may be observed to float about isolated. Instances 
sometimes occur in which one of these pigment-fibres passes 
uninterruptedly from the body of one cell to that of another ; 
for example, fig. 9, a. We may imagine this to be effected 
by the prolongations of two cells meeting at one point. 
As the pigment does not move from one cell to another, we 
cannot accurately determine whether the partition-walls be- 
come absorbed at such a point or not. Such, however, 
may be supposed to be the case, otherwise an interruption of 
the pigment corresponding to the double thickness of the cell- 
wall must be seen at the spot where the prolongations are in 
