THE METAMORPHOSIS OF ECHINODERMS. 99 
primary tentacles, it would be very unsafe to attach any im- 
portance to such a speculation. That an atrial cavity was 
present in the common ancestor at some stage or other is 
rendered almost certain by its presence in such widely separated 
groups as Crinoids, Echinids, and Holothurians ; and whether 
it arose at the stage now under consideration or at a later one 
is a question of no importance. 
The only other point requiring justification is the limit 
placed on the forward extension of the posterior body-cavities. 
It has apparently escaped the notice of Semon and Bitschli 
that these cavities rarely extend beyond the anterior margin of 
the stomach ; even in Ophiurid Plutei (5, fig. 4) where they 
extend further than elsewhere, they still lie posterior to the 
water-pore. Asterid larve must, of course, be judged by those 
early stages in which the anterior and posterior cavities are 
separated on the left side, and they will then be found to agree 
with other larvee in this respect. 
It will be noticed that I have given this ancestor neither 
preoral lobe nor point of fixation; a rudiment of the former 
(which probably existed at an earlier epoch) may have been 
present, and there may have been on it a sucker for temporary 
fixation (though I refuse to admit that complete fixation can 
have occurred without leaving stronger evidence behind) ; but 
such assumptions, though leaving the main outline of my 
hypothesis untouched, are wholly unnecessary, and to my mind 
not justified by the evidence at present before us. 
Transition to Radial Symmetry. 
In all recent adult Echinoderms (with the exception of 
Holothurians, to which we shall return later) the plane of the 
water-vascular ring is at right angles to the axis of the ceso- 
phagus and parallel to the longitudinal mesentery of the larva. 
To arrive at this condition from that of our hypothetical 
ancestor, it is only necessary to assume a movement of the 
cesophagus, with the water-vascular ring, into the left side of 
the animal. Of course I do not mean to assert dogmatically 
that an actual migration, and that alone, of the cesophagus 
