DEVELOPMENT. 75 



In their earlier stages the sporosacs may be seen as minute hollow tubercles, projecting 

 from the sides of the blastostyle. They are composed of two layers, cndodcrm and ectoderm, 

 directly continuous with the corresponding layers of the blastostyle, with whose ca\'ity that of 

 the young bud is in free communication. At first we can detect no change beyond a simple 

 increase in size ; but wc soon find the ectoderm separated from the cndodcrm by the interposition 

 of a minutely granular mass between them. This mass constitutes the basis of the generative 

 elements, and is afterwards to become ova or spermatozoa. In the mean time the ectoderm has 

 itself become difierentiated into two layers ; and we have thus laid down the foundation of all 

 the parts which we meet with in the full-grown gonophore. The wall of endoderm which 

 surrounds the central cavity of the developing gonophore, and is itself immediately surrounded by 

 the generative elements, is the spadix ; the more internal of the two layers into which the 

 ectoderm has divided is the endotheca, the more external the ectotheca. 



The sporosac now becomes more and more distended by the increasing volume of the 

 generative mass, while the spadix at the same time continues to grow, and now constitutes a 

 club-shaped hollow organ, extending through the axis of the mass, while fioating particles from 

 the cavity of the blastostyle are freely admitted into its interior, where they may be seen per- 

 forming active rotatory movements. 



The sex of the gonophore becomes evident at an early period, by the appearance of ova with 

 their germinal vesicle and spot in the generative plasma of the female, while in the male the 

 interval between the spadix and endotheca continues still to be occupied by a uniform grumous 

 plasma, in which, at a somewhat later period, spherical cells and ultimately free-moving 

 spermatozoa may be detected. 



The gonophore of Ilijdractima ecMnata does not pass to any higher grade of development 

 than that here described ; but in some other forms of adelocodonic gonophore a further differ- 

 entiation takes place by the development of an additional membranous sac or mesotheca, with 

 gastrovascular canals, between the endotheca and ectotheca {Tubularia indivisa — PI. XX, fig. 3, 

 and PI. XXIII, figs. 8 and 11). I have never succeeded in following the development of the 

 mesotheca, and cannot say under what condition it begins, or how it proceeds, the membrane 

 appearing always fully formed from the moment it is recognisable. 



It will be seen that in the above account I differ in some important points from the inter- 

 pretation given by Agassiz to the appearances which present themselves in the development of 

 the adelocodonic gonophore. In his account of this process in his Clava leptodyla, Agassiz' 

 regards the perigonium or walls of the gonophore as simple, and as homologous with the umbrella 

 of a medusa. In Clava nwUicoriiis, however, the existence of two membranes may with care be 

 demonstrated in these walls, though I admit that I have frequently failed in detecting more than 

 a single one. In no case, however, can the walls of the gonophore in Clava be regarded as the 

 homologue of an umbrella. When two membranes can be demonstrated in them, these will be 

 an endotheca and ectotheca ; if only a single membrane be present, as Agassiz l)elieves to be the 

 case in his Clava leplostyla, this wOl be an endotheca, while the part which would really 

 represent an umbrella, namely, a mesotheca, is not developed." 



> Op. cit. vol. iv, p. 221. 



' In my earlier researches into the anatomy of the reproductive system iu the Hydroida {" On 

 the Anatomy aud Physiology of Cordylophora," Phil. Trans. 1853), I entertained the view here advo- 



