64 
D. D. CUNNINGHAM. 
During this accumulation the cell generally increases con- 
siderably in size^ so as to fill a space twice or three times as great as 
that which it originally occupied, and the nucleus is at the same 
time pressed to one or other side of the cell. After the highest 
stage of development has been attained, a certain amount of 
condensation usually occurs, and the granular mass shrinks 
and lies more or less free in the space which it formerly occupied. 
The nucleus and cell-wall finally disappear, and a mere mass of 
fat granules remains behind, lying in a large space in the epidermal 
tissue. Such appears to be the normal course of development 
of these structures, but in some cases more or less pigment is 
deposited along with the fat, and persists after the solution of 
the latter. 
The only other bodies which these cells in some degree resemble 
are the granule cells discovered by Kolliker in the epidermis of 
the lamprey. They differ from them, however, in their simple 
structure, having no processes connected with them which can be 
regarded either as gland ducts (Kolliker) or sensory filaments 
(T. Schultze). 
In the normal state of the healthy larvae, the granule cells are 
developed in varying, but never in excessive numbers. In so 
far as a careful study of their development and subsequent 
history affords any information on their functions they appear to 
subserve important ends in the processes of growth of the tissue 
in which they occur, as well as in the involution of those 
portions of the latter which are peculiar to the larval condition, 
and disappear on the assumption of the adult form. As pre- 
viously mentioned, the granule cells during the course of their 
development increase considerably in size, forcing apart the 
surrounding cells of the epidermis, and, on their dissolution, 
leaving spaces much larger than those occupied by the unaltered 
elements of the tissue. Under normal conditions these spaces 
are now filled up from below by the intercalation of new struc- 
tures. These may in some cases be derived from the system of 
sub-epidermal free amoeboid cells; but, as a rule, they are, I 
believe, derived from the connective-tissue system. I have 
frequently been able to trace a direct connection between fusi- 
form nuclear bodies lying in the blank spaces and the common 
connective-tissue system of the interior of the fin; and in 
cases where this cannot be done, the appearances presented 
by the new elements in spaces which are about to be filled 
up cannot fail to render such an explanation of their origin 
probable. The peripheral ramifications of the connective 
tissue appear to be everywhere closely connected with the inner 
surface of the epidermis, a system of connective-tissue corpuscles 
lying scattered over the latter, while many processes appear to 
