No. 2.] EPIPHVSTS OF TELEOSTS AND AMIA. 251 
order that each fibre may execute a curve which shall bring it 
at once and independently in contact with the pia-mater. At 
least, in every case in which (as in Fig. 16) one can trace the 
fibres individually for some little distance after they leave the 
bundle, they are seen to pass toward the brain. 
In the distal ends of many of the groups of cells seen in the 
ventral wall of the stalk are found spherical cavities, one in 
each group, filled with a slightly granular colloid-like mass 
(Fig. 15a, mt. col.). Such groups of cells are near the central 
cavity, and their cells have not been seen to give off nerve 
fibres. A few of them are found also in the distal part of the 
epiphysis on the ventral side. It is possible that we have here 
to do with a process of degeneration. 
The fourth region: The groups of nerve-cells gradually 
disappear in the proximal part of the stalk, and are replaced 
by cells similar to those in the distal end of the epiphysis. A 
few of the same cells are to be found scattered between the 
groups of nerve-cells throughout the organ. 
Along the dorsal wall of the whole stalk there is a bundle of 
fine nerve fibres which passes to the posterior commissure. 
These fibres at first sight appear to have the double contour 
which is common to medulated fibres, but on careful examina- 
tion it is found that this appearance is caused by two parallel 
fibres running close together, and that what seems to be the 
axis cylinder is merely the line of contact between these fibres. 
Cross-sections fail to show any myelin sheath, so that these 
fibres probably never pass beyond the condition of embryonic 
nerve-fibres. As these fibres are traced into the epiphysis, 
many are seen to be continuous with the nerve-cells in the 
dorsal wall of the stalk. Although they are collected into a 
compact bundle in the proximal end of the stalk, they separate 
from one another distally, and are difficult to follow. 
CATOSTOMUS. 
In Catostomus teres, Mitch. 8 mm. long (25 days old), 
Fig. 5, the epiphysial vesicles have the position and form 
described for Salmo, 7 mm. long, and it is therefore unneces- 
