90 
GKOHGE E. NICHOLES. 
hi'iiin, pi'acticiilly all have failed to refer to the extraordinary 
development of the ependymal epithelium (of the snb-com- 
missnral organ), nnd to its remarkable distribution upon the 
ventricular walls. 
Bdellostoma (Pol i stotr ema) stonti. 
In the brain of this animal the reduction of the ventricular 
spaces has not, in my specimens at least, proceeded so far as 
is the rule in Myxine. Nevertheless, practically the whole 
of the third ventricle may be obliterated — the infnndibnlar 
cavity alone persisting — for the so-called “trigouum 
cinereum,” though frequently present, is apparently not 
always to be made out. Where the reduction has progressed 
least there may be found a canal, the “ canal is con uec ten s” 
of Holm (’01), which runs upward from the infundibular 
cavity to join the mesocoelic cavity. Into the “canalis 
connectens” there may open a canal which runs back- 
wards from the trigonum cinereum. In others, this 
canalis connectens may be represented merely by a 
broken chain of isolated spaces, or even marked only by a 
band, of varying width, of scattei-ed nuclei, the remains of 
the cells of the ependymal epithelium which, earlier in life, 
lined such spaces. Even these traces of a one-time con- 
nection may be altogether lost. 
The fourth ventricle, too, is enormously reduced, owing, 
as I believe, to the fusion of its walls, mesially, at the level 
of about half the height of the ventricle, followed by the 
more or less complete obliteration of the upper chamber so 
cut off from the rest of the ventricle. Here, again, there is 
considerable variation, the vestiges of the upper portion 
of the ventricle being much more consideiable in some 
specimens than in others. 
’Ihe reduction of the ventricles has been least marked in 
the mid-brain, and the mesocoel forms the most considerable 
of the brain cavities. Postero-ventrally it is dilated into a 
large subs{)lierical chamber, which I shall distinguish as 
