Nov. 29, 1883] 



NA 'JURE 



117 



lies in tlie zone of gemmation of the lateral hydranth at the spot 

 where the gonophore bud is formed. 



Ill the genus Eudendrium most remarkably there is a difference 

 in the formation of the gonad elements in the case of different 

 species. In Eudendrium racemosum the gmophores are not 

 b )rne by the hydranths but on blastostyles, which bud out only 

 from the lateral hydranths. Both male and female germinal 

 celU have their place of origin not in th? gonophires or blasto- 

 styles, but in the ccenosarc ; the gonophores are only the ripening 

 places of the cells. The blastostyles are not regarded by the 

 author as liydranths which in an ontogenetical sense become 

 atrophied in the history of each colony, in consequence of the 

 exhaustive effect of the development of gonophores on them, but 

 as special structures probably derived originally from hydranths, 

 but which have undergone a permanent phylogenetic modification 

 (at all events in E. racemosum and E. capillare) to adapt them for 

 their peculiar function. The developing buds from which blasto- 

 styles are formed are very early to be distinguished from those 

 forming hydranths, and do not vary in colonies of the same sex, 

 though they show a constant difference in form in the two sexes. 

 The male blastostyles have no hypostome, mouth, or trace of 

 tentacles. The female have also no hypostome but have a double 

 crown of tentacles, and appear at the time when tlie gonophores 

 are ripe to have a small temporary mouth, which it is suggested 

 may possibly swallow the spermatozoa to effect fertilisation. 



In the female stocks of Eudendrium racemosum when in full 

 sexual maturity the ccenosarcal tubes at all the free ends of 

 the branches contain large quantities of ovicells. The fine twigs 

 are often full of hundreds of them. They occur in both ectoderm 

 and endoderm, but far more abundantly in the former, where 

 they are found in all stages of development, whereas in the endo- 

 derm scarcely any but large egg-cells are found. The primitive 

 germinal cells are derived from ordinary young ectoderm cells, 

 with which in rapid process of multiplication the whole germinal 

 zone is filled. This zone lies ouly in the principal hydranths, 

 coaimencing a little below their necks and extending a shorter 

 or further distance down the stem, but as a rule not further than 

 the second lateral hydranth (K%o, Fig. 3). Within this zone the 

 production of new ovicells is almost entirely restricted to its 

 uppermost region. As the principal hydranth grows, the germinal 

 zone, which maintains a constant length, rises with it, and as 

 soon as it rises above the point of junction of any lateral 

 hydranth, this hydranth is cut off from any further supply 

 of ovicells. The ovicells never occur in the endoderm within 

 the germinal zone, but are only found in that layer within 

 the hydranth and gonocope. This is because of the remark- 

 able migrations which the cells perform, which take place in 

 perfectly definite directions at definite times. The cells remain 

 in their place of origin, the ectoderm of the germinal zone, 

 until a new lateral hydranth bud begins to be formed, and 

 into this they migrate through the ectoderm, not at once, but as 

 soon as the hydranth has attained a well defined stem. They 

 wait here in the ectoderm, growing considerably, until they have 

 attained a certain size, and then bore their way into the endoderm, 

 nearly all the cells in each lateral hydrocope effecting the pene- 

 tration of the basement membrane simult.ineousIy, just at the time 

 when a blastostyle bud commences to form. The cells hold on to 

 the basement membrane on its inner face by one end, and stretch 

 forwards the other in the direction of the position of the future 

 blastostyle, and become remarkably elongate, their free ends being 

 drawn out into long slender filaments amongst the endoderm cells. 

 As soon as a hollow is formed in the blastostyle bud they creep 

 in, still clinging to the basement membrane and always to its 

 endodermal face. As the hollow enlarges, more and more creep 

 in, and the bud takes on a pear shape. As the gonophores are 

 budded out from the blastostyle the cells pass into the endoderm 

 of these, then almost simultaneously bore through the basement 

 membrane again, and reach the ectoderm layer of the sporophores, 

 their final ripening place. The ovicells never reach maturity on 

 the hydranths in which they originate, but always in the blasto- 

 style of a lateral hydranth. 



In the male stocks of Eudendrium racemosum the place of 

 origin of the germinal cells is the ectoderm of the region of 

 gemmation of the lateral hydranth>. Thence they migrate by 

 the endoderm into the sporophores, and then like the ovicells 

 bore their way out into their ripening place, the ectoderm of the 

 sporophores. 



In the other species of Eudendrium examined, E. capillare, 

 the place of first appearance of both male and female germinal 

 cells is in the endoderm. 



The results obtained as to the history of the generative elements 

 in the various species examined are given in a concise tabular 

 form under a series of headings, the importance and distinctness 

 of which will now be recognised. The case of Podocoryne is 

 taken as an example. The German terms are not easy to find 

 English equivalents for. 



Podocoryne carnea 



Keimstdtte. Germinal place. \ Male gejrminal cells : the ecto- 

 (Layer in which the earliest ( derm. 



appearance of the germinal I Female germinal cells ; theendo- 

 cells can be detected.) ) derm. 



I-Ceimzone Germinal zone. 

 (Region of the colony where ( 

 these cells are earliest de- | 

 tected.) 



Abkunft. Actual origin of the 

 mo t primitivegerminal cells, 

 (in very many cases a matter 

 of inference only). 



Ripening place. 



Migrations. 



In male stocks : the manubrium 



of the Medusa buds. 

 In female stocks : tht^ endoderm 



sac of the gonophore bud. 



Male germinal cells : young ecto- 

 I derm cells. 



, Female germinal cells : proba- 

 bly ectoderm cells which have 

 ' migrated into the endoderm. 



The ectoderm of the manubrium 

 I of free-swimming Medusae. 



' The male cells none. 

 The female cells out of the pri- 

 mary endoderm sac of the 

 I gonophore bud into the spadix 

 and thence into the ectoderm 

 [ of the manubrium. 



The facts with regard to all tht investigated species, when 

 thus placed in a tabuUr form, appear at first sight so varied 

 and complicated as to defy all reduction to uniform law. The 

 germinal cells appear to be developed sometimes here, some- 

 times there, without rule of any kind and without definite 

 relation to the germ layers. A most remarkable fact lies in 

 the circumstance that the greatest differences in these matters 

 occur in closely allied genera and even species. But, since this 

 can occur without affecting the general evidences of these re- 

 lationships, "the variations must depend on such differ- 

 ences as can occur amongst nearly related forms." And in 

 this circumstance really lies in Prof. Weismann's opinion the key 

 to the w hole matter. By careful use of the comparative method, 

 he has arrived at the conclusion that the differences in the posi- 

 tion of the place of first appearance of the germs depend on a " phy- 

 logenetic shifting "of this position, and have ensued /irW />a^i« 

 with the degeneration of the primitive free medusx unto sessile 

 brood sac-. The advantage gained by the animal in the shifting 

 which has brought this about, has lain in the earlier ripening of 

 the gonad elements. 



In accordance with a widely accepted view, the sessile gono- 

 phores of all the attached hydromedusa; except hydra, are 

 probably to be regarded as degenerated medusa;. In the an- 

 cestral medusae the gonad elements of both kinds originated 

 in the ectoderm of the manubrium, and ripened there as they do 

 now in six out of seven Tubularine genera bearing medusae exa- 

 mined by the author, viz. Dendroclava, Bougainvillia, Perigoni- 

 mus, Cladonema, Corymorpha, Syncoryne. Both the origination 

 and ripening of the germinal cells occurred during the free life 

 of the medusje. Certain causes rendered the free medusa stage 

 dis.advantageous, and in many instances the gonophores in con- 

 sequence became sessile, whilst the sexual elements originated 

 and ripened in them at an earlier sta'^e. At first the elements 

 retained the same place of origin as in the free medusse, a con- 

 dition which survives in the medusoid gonophores of the existing 

 Cladocoryne. But it became advantageous that the elements 

 should not wait for their formation by cell division and for their 

 gradual maturation until the process of construction of the gono- 

 phores by budding had been completed, and thus the formation 

 of the ovicells became shifted, and appeared in an earlier stage. 

 What may be regarded as a first stage in this process is repre- 

 sented in Pennaria and Tubularia, in which the germinal cells of 

 both sexes first appear in the endocodon (see Fig. i) of the gono- 

 phore bud, being carried afterwards, as development proceeds, 

 to the original ripening place, the manubrium. As a further stage 



