STARVATION OF VEGETABLE AND ANIMAL TISSUES, 
65 
be directly affixed to the epidermal cells themselves. On the 
formation of a blank space or hiatus in the epidermis^ one or 
more connective-tissue corpuscles, according to the size of the 
area to be filled up, are drawn into it. Tiie connective-tissue 
elements at first form an open reticular tissue, which, as the 
nuclei enlarge and the processes diminish, is gradually converted 
into a coherent patch of new cells intercalated among the older 
epidermal elements, and occupying a space originally formed by 
the degeneration and removal of one cell. If this interpretation 
of the phenomena be correct, one main function of the granule 
cells under ordinary circumstances is mechanically to enlarge the 
area of the inner layer of the epidermis. 
Observations did not decide how the outer epidermal layer 
accommodates itself to the increased area of the inner one. The 
occurrence of any general desquamation was never observed, and 
there were no appearances of the addition of new elements from 
beneath. Trom the extreme variations in the form and size of 
the cells of the outer layer it seems not improbable that a good 
deal of the accommodation is accomplished by means of mere 
extension and flattening of the constituent cells. 
It is evident that the granule cells may also play an impor 
tant part during the resorption and disappearance of the fin, for 
if developed in large numbers without equivalent addition of new 
elements from the deeper tissues, they must tend to the ultimate 
removal of the deep layer of epidermis. This, however, is a 
subject not calling for discussion on the present occasion. The 
important points to recollect here are, that when developed in 
due proportion, the granule cells provide for the growth of the 
tissue in which they occur, and that in such proportion their 
development is strictly physiological for the organism, although 
pathological for the individual ceils. 
The remaining structures in the tail of the larvae call for little 
detailed notice, as their characters are familiar to all physio- 
logical observers. The branched amoeboid cells beneath the 
epidermis are very inconspicuous during the life of the animal, 
and may readily escape notice unless carefully looked for {vide 
fig- 5 , a). 
The connective-tissue elements present no special peculiarities. 
The vascular loops are full, and the current of blood is strong 
and rapid. Many of the blood-corpuscles are characterised by 
containing two or three minute shining oil granules. 
This brief account of the normal anatomy of the tissues may 
serve to render an account of the ])henomena associated with 
dq)rivation of nutritive material intelligible. The following 
description is taken from the notes recorded during the examina- 
tion of a tadpole which had been for a fortnight in distilled 
VOL. XX. NEW SER. E 
