425 
are almost entirely ependymal, the meningeal portion being dorsally 
situated and occupying a very restricted area. 
In the case of specimen C, where the fibre was severed before 
hardening, there has evidently been an extensive recoil, the fibre having 
sprung into a large tangled mass (Pl. I, Fig. 3) which has blocked the 
lumen of the central canal in the filum terminale at a point a little 
forward of the sinus terminalis. The recoil seems to have been a 
very violent one, for the knotted mass which resulted has ruptured the 
terminal filament, which is, at this point, simply an epithelial tube. 
At the same time the posterior end of the fibre seems to have been 
jerked free from its attachment and to have sprung forward, for I 
can find no trace of it behind the knotted mass. 
In the two specimens that were thoroughly fixed before the tail 
was cut off, the fibre may be traced in a direct and continuous line, 
from the point where the cord was severed, into the sinus terminalis, 
showing clearly that there has been no post-mortem retraction of the 
fibre. Here, again, owing to the end of the sinus terminalis having 
been, in one case, cut obliquely, the actual end of the fibre could 
be determined in only one specimen, where it is seen to pass into 
the apex of a transparent and very lightly staining conical membrane 
which lies at the end of the terminal sinus. In the other specimen, 
notwithstanding the obliqueness of the sections, the fibre could be 
traced as a perfectly taut thread into a terminal mass of ependymal 
cells at the end, but in addition to this a huge coiled mass of fibre 
could be made out at a point some way in front. This coil seemed 
unrelated to the present fibre and must represent the result of an 
accidental snapping of the fibre some time previous to the death of 
the animal. (It closely resembles the coiled mass which I have 
described as having been produced artificially in specimen C after 
death.) The normally stretched fibre must clearly have been produced 
by regeneration between the time of the snapping of the earlier fibre 
and the killing of the specimen. 
The only adult dogfish tail of which I have satisfactory sections 
(apart from the material used for the experiments) was quite completely 
hardened before the spinal cord was cut. In this, REISsSnEr’s fibre 
passes back in a straight line to end in a typical terminal plug forming 
a membranous cone, which is, however, slightly collapsed and passes 
quite gradually into its basal part, which belongs certainly to the 
posterior wall of the sinus terminalis. This chamber, in the dogfish, 
is bounded for the greater part of its extent by the ependyma, which 
loses its columnar character in the posterior part and passes into a 
