ON THE REPRODUCTIVE SYSTEM IN THE HYDROIDA. 363 
not be supposed that these two forms are constructed upon plans widely 
different from one another. We find, on the contrary, that the most exact 
parallelism admits of being demonstrated between them ; for though they 
may at first sight appear very different, it can nevertheless be shown that the 
closed generative sac of a Clava or a Hydractinia is an easily understood 
modification of a medusa*. 
In comparing the two classes of gonophores with theview of determining 
their homological relations, their composition out of the two membranes ecto- 
derm and endoderm must be carefully kept in mind. 
Commencing with the central parts of a phanerocodonic gonophore (fig. 5 B), 
and comparing these with the central parts of an adelocodonic gonophore 
(fig. 5 A), we shall find that in the former we have a manubrium in the form 
of a more or less elongated tubular body occupying the axis of the gono- 
phore. The walls of the manubrium are composed of two layers, an internal 
or endodermal layer and an external or ectodermal; and in all phanero- 
codonic gonophores of the sexual type (gonochemes) these two layers become 
ultimately more or less separated from one another by the development of the 
generative elements between them. 
In the adelocodonic gonophore also we have a double-walled tubular body 
occupying the axis ; but while in the phanerocodonic gonophore this body is 
in almost every case perforated by a terminal mouth, in the adelocodonic 
forms it is completely closed. The generative eloments are here also de- 
veloped between the two layers exactly as in the gonocheme ; but, in conse- 
quence of the absence of a mouth, the central organ assumes, by the increasing 
volume of these elements, the appearance of a single-walled sac, filled with 
ova or spermatozoa, and having a cecal diverticulum (spadix) plunged into 
the middle of the mass. This cecal diverticulum is plainly the equivalent of 
the endodermal portion of the manubrium in the phanerocodonie gonophore, 
while the wall of the sac (endotheca), which thus immediately confines the 
generative elements, represents the ectoderm of the same organ. 
The umbrella and gastrovascular canals of the phanerocodonie gonophore 
have their equivalents in the mesotheca and canals of the adelocodonic gono- 
phore, when these happen to be present, though in many cases they are never 
developed ; while the ectotheca holds exactly the same position and relations 
in the two forms. 
It would seem that in no case is a velum or its homologue developed in the 
adelocodonic gonophore, while the marginal tentacles of the phanerocodonic 
forms are, except in the ‘‘ meconidium”’ (see below, p. 376, fig. 12), also without 
their representative in the sporosac; for the tentacula-like tubercles which 
crown the summit of the adelocodonic gonophore of some of the Tubularidce 
'(Tubularia coronata for instance) are of an entirely different significance, 
being merely processes of the ectotheca. 
We are thus enabled to trace a close parallelism between the two kinds of 
gonophore ; but another comparison of great interest in this inquiry here sug- 
gests itself, that, namely, between the gonophore and the polypite. Now there 
Oceania octona, Fleming, and O. turrita, Forbes, meduse belonging to the type of the 
gonocheme, are, on the other hand, described by Forbes as having a lithocyst imbedded in 
the tentacular bulb just below the ocellus. 
_ * Jt is now many years since I endeavoured to demonstrate that the so-called “ ovarian 
vesicles” of the Tubularida, and the fixed sacs contained within the gonangium of the 
Sertularida and Campanularida, were in all cases strictly homologous with the free me- 
dusee—that they possess a true medusal structure in a more or less degraded or disguised 
condition. (‘On the Anatomy and Physiology of Cordylophora,” Phil. Trans., June 1853.) 
