THE METAMORPHOSIS OF ECHINODERMS. 15) 
side, is certainly not easy to determine with relation to the 
radial vessels, which are on the outer side; but I have fully 
satisfied myself by means of sections of this and later stages 
that it was correctly given in my previous paper (5, p. 22), 
and I cannot admit the plea of variability with which Semon 
meets my criticism of his figures (33, p. 9). 
The whole hydroccel is pushed more and more towards the 
ventral surface of the larva by the elongation of the water- 
tube; but the obliquity which brings its posterior end even 
nearer this surface than its anterior becomes more marked, 
while the anterior end bends over at the same time more 
towards the middle line than the posterior (fig. 3). 
The formation of the water-vascular ring round the ceso- 
phagus follows rapidly on the stage shown in figs. 2 and 8. 
The posterior end of the hydroccel bends round on the ventral 
side of the oesophagus (close to its junction with the stomach) 
about as far as the middle line. The anterior end, on the 
other hand, crosses over on the dorsal side of the cesophagus 
to the right side of the larva, and then, bending posteriorly 
and ventrally, passes round to the ventral surface, where it 
eventually joins the other end of the hydroccel, and completes 
the water-vascular ring. 
It will be remembered that according to Metschnikoff the 
most posterior pouch of the hydroccel Auricularia is the rudi- 
ment of the polian vesicle, while the most anterior becomes 
one of the radial canals. Semon expresses some doubt on this 
point ; but a series of sections through the larva during stage 
A (before the closure of the water-vascular ring) sets this 
question at rest. The posterior pouch (which at the close of 
this stage may even be to the right of the middle line, though 
it seems to vary somewhat) turns slightly inwards, and pushes 
before it the wall of the left body-cavity (see figs. 10 and 18, 
in which, however, but little more than the peritoneal covering 
of the vesicle is visible). None of the other pouches ever 
“ project into the body-cavity in this way; and from this and 
later stages it is abundantly clear that Metschnikoff was right 
in identifying this posterior pouch with the polian vesicle. It 
