412 
a more or less winding course among the epithelial elements) 
somewhat thicker. 
Finally, fibrils spring from this superficial intra-epithelial 
network which, after attaining the free surface, divide into two 
branches. ‘These at first run horizontally in opposite direc- 
tions, and soon take a course towards the depth of the cornea, 
and lose themselves in the superficial intra-epithelial net- 
work above described. 
No one will be disposed to dispute that it is a matter of 
great difficulty to determine in a preparation in which the 
epithelium-elements are not distinctly seen, whether a nerve 
filament attaims the under or the upper surface of the most 
superficial layer. I cannot, however, regard this difficulty 
as insurmountable, on the following grounds :—It is in the 
first place to be borne in mind that by comparison of surface- 
preparations with vertical sections of the same part of the 
cornea, it is always possible in any given preparation to de- 
termine whether or not the whole of the epithelium is pre- 
sent. Further, we have a certain guide for the recognition 
of the true surface in the precipitate which adheres to 
it. Secondly, I have observed. when I have exercised the 
greatest caution in cutting the cornea and in preparing it, 
that the more perfect the preparation was, 7. e. the more 
completely the intra-epithelial network could be distinguished, 
the more rarely did it happen that filaments could be seen 
either on the surface or between the superficial cells which 
could not be shown to be connected on either side with others. 
It follows, then, from this that the terminal knobs of Cohn- 
heim are only intercalated swellings occurring in the course 
of fibres, and that the appearances he has described depend 
on imperfections in the preparations. 
To sum up what has been stated as to the intra-epithelial 
network of the cornea of the rabbit, we have to distinguish 
two nervous networks, the one at a level corresponding to 
the ends of the pallisade-epithelium, the other separated 
from the surface only by one or two layers of cells. The 
former we propose to designate, the deep intra-epithelial 
network; the latter zhe superficial intra-epithelial network. 
The latter is distinguished from the former in the greater 
density of the network, in the greater thickness of the 
fibres, and in the existence (in addition to the smaller vari- 
cosities which also exist in the deep network) of yet larger 
swellings above described, which are merely found at the 
junctions of two filaments. From this network filaments 
spring similar in character to those of which it is formed, 
