76 MORPHOLOGY. 



Again, in the gonopliores of Ihjdractimn jmJ^cHna, Agass., TuhiJaria (Paryplia) crocea, 

 Agass., and Tnhularia {Thamnocnidia) sjiedabUis, Agass., Agassiz correctly figures the two 

 membranes which enter mto their walls ; but he assuredly assigns an incorrect origin to the more 

 internal of these membranes when he describes it as rising, subsequently to the formation of the 

 generative mass, from the proximal end of the gonophore in the manner of a cup closely pressed 

 against the outer wall, and, at least in Hydractinia and Tubularia spectabiJis, ultimately closing 

 over the contained structures so as to form a continuous internal wall. 



Now, the internal wall in the gonophore of Hydractinia is undoubtedly formed, not after, 

 but simultaneously with the appearance of the generative mass, and is nothing more than the 

 internal of the two layers into which the ectoderm of the primary bud has become divided 

 simultaneously with its separation from the endoderm by the interposition of the generative 

 elements ; it is thus the endotheca of the sporosac, while the more external layer is the 

 ectotheca. 



Having had no opportunity of examining the development of the gonophores in either of the 

 two Tubularice cited above, I am unable to bring any direct observation into opposition with the 

 views of Agassiz as to the gonophores of these hydroids ; but the analogy of Hydractinia and of 

 other hydroids, whose adult gonophores correspond in all essential points with those of the 

 American forms, leads me to believe that the process is in all the same as in Hydractinia. 



It is only in those cases where a mesotheca becomes developed, as in Tubularia indivisa, 

 that the adelocodonic gonophore presents any true representative of the umbrella of a medusa, 

 the mesotheca being properly the homologue of this part. Agassiz, in his account of Tubularia 

 Couthouyi, Agass., ignores the existence of any membrane between the well-developed mesotheca 

 of this species and the generative mass which surrounds the spadix. In Tubularia indivisa, 

 however, this membrane cannot be overlooked, especially in the male, though in the female it 

 would seem to disappear at an early period, and may thus escape detection. 



d. Development of the Medusa. 



The medusa, whether gonocheme or blastocheme, shows itself at first in every case as a 

 minute hernia (woodcut, fig. 34 A), consisting of endoderm and ectoderm, and having its 

 cavity in free communication with that of the blastostyle, or of the trophosome from which it 

 springs ; thus in no respect differing at this period from the corresponding stage in the develop- 

 ment of the adelocodonic gonophore, or, indeed, in that of a hydranth branch. 



It is very difficult to follow satisfactorily the several steps by which this primordial tubercle 

 becomes ultimately converted into a medusa. I have bestowed great attention on it in different 

 species of Hydroida, and have more recently subjected the development of the medusa bud 



cated by Agassiz, as to the homology of the parts in question. Subsequent more extended observa- 

 tions, however, have induced me to modify in some respects the views then expressed, and to adopt 

 those which are advocated in the present INIonograph. (See my paper " On the Reproductive Organs of 

 Serlularia tamarisca,'' in the Report of the British Association foji- the Advancement of Science, 

 1858.) 



