SURFACE FAUNA OF THE GULF STREAM. a 
sion of the base of the coenosare may perhaps be considered as the first indi- 
eation of the formation of a float. In this case this rudimentary pneumat- 
ophore is attached to the ground or to a movable body-like Fusus, ete., and 
forms papilla, between which arise the two different kinds of polypites, the 
sterile tentacular polypite and the reproducive polypite, giving rise to either 
male or female colonies. This combination is exactly similar to that of Por- 
pita, in which we have a chitinous float and tentacular and proliferous poly- 
pites arranged in addition round a central sterile polypite. 
The homologies we have attempted to trace between Porpita and the 
Tubularian Hydroids might perhaps be still further extended. It is well 
known that in nearly all Tubularians the base of the coenosare, by which 
they are attached to the ground, extends either as filaments or rootlets over 
a considerable space; these filaments, or expansions of the chitinous tubes, 
forming either a connected series of canals, more or less complicated, as in 
Clava, Cordylophora, Coryne, ete., or a net-work of canals as in Dicoryne, Bou- 
gainvillia, Tubularia, Hydractinia, Podocoryne, ete. ; or else such filamentary 
processes as those of Corymorpha, in all of which there is a more or less active 
circulation connected with that of the cavity of the Tubularian. In Cory- 
morpha we find a series of longitudinal canals, more or less branching and 
anastomosing, extending along the ccenosare. Let us now imagine this long 
Corymorpha coenosare reduced in length and at the same time flattened so 
as to form a disk somewhat below the base of the tentacles, retaining its 
peculiar pointed terminal basal extremity. We could thus have a free Hy- 
droid differing but little from Velella and Porpita; that is, our Corymorpha 
would be transformed to a Hydroid, with a crown of marginal tentacles below 
the chitinous disk, in which there are canals of the vascular system. Between 
this row of marginal tentacles and the large central opening of the polypite 
we find clusters of reproductive Meduse. Imagine the same transformation 
in a colony of Hydractinia, or of Podocoryne, in which we find the chitinous 
disk already formed and traversed by a network of canals, and add to it a 
central sterile polypite, and we have all the structural features of a modified 
Porpita, namely, a disk, rows of tentacular polypites, next rows of repro- 
ductive and feeding polypites, while if we make the same comparison with 
Hydractinia we add a third kind of appendage, the so-called spiral zoiids. 
An examination of very young stages of Tubularians, such as, for instance, 
the very early stages of Endendrium figured by Allman (PI. XIII, Figs. 
14-16, Tubularian Hydroids), show such a chitinous disk to be compared in 
