THE METAMORPHOSIS OF ECHINODERMS. 101 
into the water-vascular ring in the interradius through which 
this plane passes, would give (as the construction of a simple 
diagram will easily show) a very oblique course to the water- 
tube between the water-vascular ring and the pore; and in 
order to obtain the straight course for this tube, which it invari- 
ably possesses, we must assume either that the tube has shifted 
the position of its opening into the water-vascular ring, or that 
the latter has been rotated. That the latter is the more probable 
solution we shall see when we have examined the justification 
of the main features of my hypothesis afforded by ontogeny. 
If we may judge by external appearances, Ophiurid Plutei 
repeat with most completeness the features of the supposed 
ancestral migration of the cesophagus and water-vascular ring. 
On the other hand, in Bipinnaria asterigera the hydrocel 
lies from the first nearly in the middle of the left side of the 
stomach, and a new esophagus is formed in the centre of it, so 
that there is no migration of either organ. Between these two 
extremes we have several very instructive intermediate steps. 
In Crinoids there is no csophagus, but the atrial cavity 
and water-vascular ring both migrate, so as to place the plane 
of the latter more and more parallel to the ccelomic mesentery ; 
but the nature of the process is much obscured by the early 
change of position of the body-cavities and by the migration 
being directed towards what appears to be the posterior pole. 
In Echinids there is no migration of the cesophagus, but the 
atrial cavity and hydroccel move over the left side of the 
stomach, arriving fairly early at their final position. The forma- 
tion of a new cesophagus is obviously secondary, but otherwise 
the ancestral process seems to be fairly closely followed. 
Brachiolaria (including Asterina) is still more interesting. 
Here, as in Bipinnaria, the hydroccel is from the first nearly in 
the centre of the left side (its position there being in my opinion 
secondary ),anditconsequently undergoesnomigration. Butthe 
larval cesophagus is retained intheadult, andat the time of meta- 
morphosis is bent sharply into the middle of the hydroccel ring ; 
at the same time it undergoes a complete change of position 
