560 PROCEEDINGS: BOSTON SOCIETY NATURAL HISTORY. 
ately posterior to the brain, but without any posterior opening 
whatever. 
At a later period (pi. 24, fig. 13) a narrow canal grows out from 
the posterior end of the stomach and forces its way backward 
among the endoderm cells. This narrow tube then bends dorsally 
and continues its growth posteriorly immediately beneath the pro¬ 
boscis sheath. It remains closed at its posterior end until it reaches 
a point well back in the body, when it opens directly through the 
dorsal wall into the lumen formed in the midst of the endoderm 
cells, that is, into the intestinal canal. It thus forms the pylorus of 
the adult, and as it opens far behind the anterior end of the intestinal 
canal, a considerable portion of the latter is left as a broad chamber, 
which passes forward from the opening of the pylorus and ends 
blindly immediately behind the brain. The chamber thus becomes 
the intestinal caecum of the adult, while the whole remaining por¬ 
tion of the canal formed within the endodermic mass and lying 
behind the pyloric aperture, is the intestine proper. 
As in Tetrastemma and Drepanophorus, the rhynchodaeum, 
esophagus, stomach, and pylorus are thus formed from the ecto¬ 
derm, the intestine proper and the intestinal caecum alone being 
derived from the endoderm. It is to be observed that both intestine 
proper and intestinal caecum are formed from the same kind of 
cells, and therefore it is not surprising to find both these parts of 
the alimentary canal exhibiting practically the same peculiarities in 
the adult. 
The cells which compose the pylorus are much smaller than those 
lining the stomach, but they are likewise covered with cilia, which 
are much finer and shorter than those of the latter organ. 
As the length attained by the pylorus before opening into the 
dorsal wall of the intestinal canal determines the length of the 
intestinal caecum in this species, it is reasonable to assume that the 
great differences in the length of the caecum in various species of 
Hoplonemertea are largely due to similar differences in the extent 
of growth of the pylorus before it opens into the intestinal canal in 
the embryo. 
The delamination by which the intestinal canal is formed in the 
midst of the mass of endodermic cells takes place in such a way 
that it is surrounded on all sides by at least two layers of these cells 
(pi. 24, fig. 13 ; pi. 25, fig. 18). The canal itself is commonly 
