208 EDMUND B. WILSON. 
the circumoral ring of the adult. They never float freely 
in the perivisceral cavity, and cannot for a moment be con- 
founded with the true blood-corpuscles of this cavity. During 
the metamorphosis these masses suddenly break up, and the 
corpuscles are almost immediately carried along within the 
vessels by the peristaltic contractions of the latter. During 
the later larval stages they are sometimes elongated towards 
each other, and connected by a narrow band containing a 
few corpuscles (see fig. 10) ; and repeated observations of the 
metamorphosis has convinced me that, for our species at 
least, Metschnikoff’s account is incorrect. Owing to the 
failure of attempts to make satisfactory sections of the larvee 
I have been unable absolutely to demonstrate this point, 
which is of considerable importance from its bearings on 
the relation between the pseud-hemal system and the body 
cavity. 
Part II. 
Before advancing speculations as to the origin and sig- 
nificance of this most remarkable course of development, it 
is necessary to dwell for a mcment on the systematic rela- 
tions of Phoronis, and on certain structural features of the 
group to which it belongs. It is pretty well agreed that 
Phoronis is a Gephyrean, although a greatly modified and 
specialised representative of this peculiar group. Although 
the peculiarities of its development are so great as to lead 
so high an authority as Mr. Balfour to question the cor- 
rectness of this identification, I believe, nevertheless, that 
these doubts are not well-founded. In all essential ana- 
tomical characters, so far as they are known, Phoronis agrees 
closely with such forms as Sipunculus or Phascolosoma. Its 
most striking characters, such, for example, as the close prox- 
imity of the mouth and anus and the high development of 
the oral tentacles, are simply exaggerations of characters 
possessed by the last-named genera. Almost all of the adult 
characters are readily explicable as the result of extreme 
adaptation to a strictly tubicolous life. In regard to the 
development, I shall endeavour to show that its peculiarities 
are almost certainly due to secondary adaptations, correlated 
with the highly specialised structure of the adult. Many 
facts show that it is not worth while to attach much import- 
ance to soinconstant and variablea character as the arrange- 
ment of ciliated belts in pelagic larve. But if this point 
were of importance, their arrangement in <Actinotrocha 
would not present any great difficulties in the way of re- 
