64 BULLETIN OF THE 
processes, but it is extremely doubtful if they should be so interpreted. 
They are situated too far from the ora serrata and too near the free 
edge of the iris. They are probably in some way merely incidental to 
the extensive pigmentation of the iris. 
In no instance are the layers of the retina in Typhlogobius as fully 
differentiated as in the normal eyes of fishes. I will first describe them 
as they are found in the majority of specimens, beginning with the inner- 
most layer, and will afterwards speak of the cases that show deviations 
from the common condition. 
An internal limiting membrane, distinct from the layer of nerve fibres, 
I have been unable to find. Corresponding with the exceedingly rudi- 
mentary condition of the optic nerve, the layer of nerve fibres is very 
thin, even in the immediate vicinity of its entrance, where, in the normal 
retina, it reaches its greatest thickness. In many sections only frag- 
ments of it are to be seen; and for considerable areas no traces at all 
are found. It is possible that its absence is due to its having been 
broken away during the preparation of the sections; but, however that 
may be, it is certain that, wherever present, the layer is very thin (Plate 
II. Fig. 5, Plate III. Figs. 18 and 21, st. for. opt.). The next layer, viz. 
that of the ganglion cells (Figs. 18 and 21, cl. gn.), is always distinct, 
and is from one to three or four cells in thickness. As a rule only the 
nuclei are distinguishable ; but occasionally the cell bodies can be made 
out. Examined under high powers and with careful focusing, some of 
these are found to possess one or more processes (Fig. 18, cl. gn). A 
nuclear membrane can usually be seen, as can also one to several darkly 
stained particles within the nuclei. The nuclei are in general very 
nearly spherical, though there is a tendency for them to become ellip- 
tical with the longer axis radially directed. 
The inner reticular layer (st. rtl. 4.) is well developed in all cases, and 
is essentially the same in structure as in the normal eye, though I have 
not detected any of the fibres running parallel with the surface of the 
retina that are found in normal fish eyes. Within this layer the radial 
fibres of Miiller (fbr. Mú.) can usually be made out, though they ap- 
pear to be few and indistinct. 
The remaining portion of the retina, as far as to the external limiting 
membrane, is never fully differentiated into the layers that are found in 
the normal eye between this membrane and the inner reticular layer ; 
and in many specimens scarcely any indication of a differentiation can 
be seen. About the average condition is shown in Figure 18, st. rtn./, 
where a layer of nuclei (st. nl. i) about two or three deep may be dis- 
