THE METAMORPHOSIS OF ECHINODERMS. 87 
fore it seems to me that the presence of stem-joints in the 
larval Antedon indicates that this region is not a pure preoral 
lobe, but a mixture of this with a structure (the stalk) belong- 
ing to a much later (phylogenetic) date. The forward pro- 
longation of the right body-cavity is due to the same preco- 
cious development of the stalk; for it is impossible to compare 
this, which arises so late, with the right anterior enterocel of 
other Echinoderm larve, which is one of the first parts of the 
ccelom to appear. 
(3) The left body-cavity grows round the csophagus at the 
time of metamorphosis, forming with the right body-cavity a 
mesentery parallel to the plane of the hydrocel ring. The 
two horns of the left body-cavity certainly both reach the 
interradius of the water-tube, and appear to me to form a 
mesentery supporting the latter; but this,in theface of Seeliger’s 
researches (30, p. 292) must be admitted to be doubtful. 
(4) In the free-swimming larva the right and left body- 
cavities are approximately equal in size (Seeliger rightly points 
out that in my diagrams I represented the right cavity too 
small). In the Cystid, however, the left cavity is much smaller 
than the right, and though, when the arms begin to grow out, 
its relative size increases for a time, yet the subsequent exten- 
sion of the right cavity into the arms apparently neutralises 
this. The important point for us, however, is the fact that the 
actual metamorphosis cannot in any way be said to be due to, 
or even accompanied by, “ predominance of the left posterior 
body-cavity ” (19, p. 434). 
(5) The order of development of the first-formed pairs 
of tube-feet in other Echinoderms is a little doubtful, 
but apparently they are from the first (as the later pairs 
certainly are) in centrifugal succession. In Antedon, how- 
ever, the second pair is undoubtedly formed later than the 
first pair of each ray; while the development of the later 
tentacles in triplets (26, p. 177) is utterly unlike anything 
observable elsewhere, and goes far to establish an absence of 
homology between the arms of Crinoids and those of other 
brachiate Echinoderms. 
