214 EDMUND B. WILSON. 
with it a loop of the intestine, would destroy the balance of 
the larval body, render the przeanal belt of cilia useless, and 
increase the size of the body, so that the swimming appa- 
ratus would be insufficient, and the pelagic life would have 
to be abandoned for a period of transition. On the other 
hand, if the ventral wall, as it grew, should become so folded 
on itself as to preserve the outlines of the body unaltered, 
an extensive growth would be possible, without the need for 
any transition period. Abundant room for such an infold- 
ing is afforded by the spacious perivisceral cavity of Actino- 
trocha, and the pouch may be regarded as having been 
derived in this way. Strictly parallel cases among other 
animals are not numerous. Perhaps an imperfect illustra- 
tration may be drawn from the foldings in the embryonic 
wings of insects which render them capable of being stowed 
away in a very small space in the pupa, and yet permit 
their rapid expansion at the time when the imago state is 
assumed. 
A much more striking instance is afforded by the meia- 
morphosis of certain Polyzoa. We owe to Barrois (see 
‘Annales des Sciences Naturelles,’ August, 1880) the very 
interesting observations that the so-called “internal sac,” 
which has been considered as the rudiment of some portion 
of the alimentary canal, becomes evaginated at the time of 
metamorphosis and then forms a large peduncle (plague 
guadrangulatre), by means of which the larva first attaches 
itself. The similarity of this process to the metamorphosis 
of Actimotrocha is obvious ; and it seems equally clear that 
it has arisen for the same reason. The larval Polyzoon thus 
makes ready for the act of attachment without detriment to 
its own adaptation to free-swimming life. 
The infolding in the larvee of the ancestors of Phoronis 
must have been at first slight and more or less indefinite, 
becoming more extensive and definite as the dorsal 
region of the adult diminished in consequence of the 
gradual obliteration of the flexure. From such a primitive 
infolding the pouch of Actinotrocha could by natural selec- 
tion be readily derived. The external opening would tend 
to become restricted for the exclusion of parasites or foreign 
bodies, and the gradual deepening of the infolding would at 
length reduce it to the form of a long pouch. The fibrous 
or muscular connection between the pouch and the walls of 
the alimentary canal are perhaps homologous with some of 
the similar fibres connecting the stomach wall and body wall 
in such Gephyrea as Thalassema or Bonellia, which are in 
turn comparable to the septa of the Annelides. 
