STIiUCTURE AND DEVELOPMENT OF EKISSxVER’s FIBRE. 91 
tlie sinus inesocoelicus (Text-fig. 7, s. m.). It continues 
antero-dorsally as a short wide tube which divides very soon 
into two canals, one passing anteriorly (figs. 21, 22, and 
Text-fig. 7, s. c. c.), and the other almost directly dorsally 
(figs. 21, 22, and Text-fig. 7, op.). 
Of these, that which extends anteriorly is quite constant 
in its relations. In all my specimens it runs forwards to a 
point slightly postero-venti-al to the habenular ganglia. In 
one case only was it seen actually to communicate in front 
with a remnant of the third ventricle. In the others it ended 
Text-fig. 7. 
A slightly diagrammatic median sagittal section through the 
brain of Bdellostoma (Pol istotrema) stouti. (The lines 
2d. 24, 25 indicate roughly the levels at which the sections 
represented in PI. 3. figs. 23, 24. and 25 are taken. The sections 
are not, however, nearly so obliquely cut as would appeal-, for 
the specimen which was cut transversely was one in which the 
brain cavities w-ere much more extensive, and in which, there- 
fore, it was possible for the various cavities to a^jpear in sections 
which were much more nearly transverse.) c.c. Can alls 
centralis, i.c. Isthmic canal. infd. Infundibular cavity. 
op. Optoccel. p.c. Posterior commissure, r.f. Reissner's fibre, 
s.c.c. Sub-commissural canal. s.c.o. Sub-commissural organ. 
s.ni. Sinus mesoccBl icus. I’.c. Ventricular canal. III. Third 
ventricle. IV. Fourth ventricle. 
blindly, the third venti'icle being in this region wholly obli- 
terated, and represented merely by the scattered nuclei of its 
vestigial ependymal epithelium. 
In transverse sections (fig. 23, s. c. c., which should be 
