THE POSITION OF SPONGES, ETC. 447 



is produced by a fusion of retromorphosed collar cells. 

 Goette (1886) also believed the ovum of Spongilla to be 

 multicellular in nature, but his descriptions have been shown 

 very clearly by Fiedler to be erroneous. More recently, 

 H. V. Wilson (1894, pp. 285-294 and 327-329) regards 

 the reproductive body from which the larva arises as in 

 some cases a gemmule produced by cell fusion, in other 

 cases as a true ovum [Tedajiione, pp. 341-343). That an 

 essentially similar free-swimming embryo should be in one 

 case a true larva produced from a fertilised ovum, and in 

 another case "a bud embryo exhibiting ancestral traits," 

 derived from an asexual gemmule, is a state of things which 

 would require very convincing proof before it could be 

 accepted. It seems, however, far more probable that 

 Wilson's statements, and also, we may add, those of Kent, 

 rest as Maas suggests (1896, p. 233) upon a mistaken 

 interpretation of a process of oogenesis. 



The last feature of sponge organisation which we need 

 discuss, namely, their embryonic development, is perhaps 

 the most important of all when considering the question of 

 their affinities. The ontogeny alone enables us to decide 

 how far the resemblances between the adult forms of 

 sponges and other animals are due to a similar origin, or 

 are secondarily and independently acquired ; in other words, 

 to distinguish between what is homologous and what is 

 analogous. And just as the reproduction seems to be of a 

 distinctly Metazoan type, so also may be said of the earlier 

 developmental history. The segmentation of the ovum 

 presents no special features and may be described as total 

 and regular. It results typically in a blastula of the usual 

 type, from which the diblastula, or two-layered embryonic 

 condition, arises in different forms in a variety of ways, all 

 of them, however, paralleled in other groups of Metazoa. In 

 the lower Ascons, in Halisarca, and in Plakina, it is formed 

 just as in many Hydroids, by an immigration of cells either 

 at any point or from the posterior pole. A simple modifica- 

 tion of the latter type leads to the amphiblastula larva, the 

 only larval form at all specially characteristic of sponges 

 where the cavity is too small to contain the immigrated 



