THE EMBRYOLOGY OF THE PORIFERA . 209 



of the adult, as well as probably to the lining of the 

 gastral cavity ; in short, that the ciliated cells of the 

 amphiblastula become the future "endoderm,'' and not, as 

 had been assumed before, the ectoderm. On the other 

 hand, the cells in the amphiblastula larva, which become the 

 ectoderm of the adult sponge, are those cells which are 

 laden with the food yolk of the ovum, the reverse of what 

 usually occurs in the development of other types of animals; 

 for, as a general rule, food yolk, if present, is contained in 

 cells of the endoderm. These anomalies in development 

 did not receive an explanation, as might have been hoped, 

 by comparison with the lowest calcareous sponges (Ascetta) ; 

 for in the latter a mode of development occurred which 

 apparently differed widely from the amphiblastula type, 

 though, as we shall see later, the difference is probably not 

 so great as it seems. 



While the course of development in calcareous sponges, 

 and especially in the amphiblastula type, was thus made out 

 more or less clearly, even if it appeared somewhat strange 

 when compared with the development of Metazoa other than 

 sponges, the embryology of siliceous sponges, on the other 

 hand, remained very imperfectly understood until quite 

 recently. The typical larva of Silicispongise was shown by 

 all the investigators to be a solid organism, consisting of a 

 superficial layer of ciliated cells enclosing an internal mass 

 of granular cells. The ciliated layer sometimes completely 

 enclosed the granular cells, and sometimes left them un- 

 covered at one of the poles. In the latter case a comparison 

 with the amphiblastula of Calcarea seemed obvious, 1 but was 

 negatived by the fact, in which all investigators were 

 agreed, that the ciliated cells of the embryo gave rise to the 

 ectoderm of the adult, while the collar-cell layer arose from 

 the granular inner mass of the larva, the exact reverse of 

 what occurs in the amphiblastula type. It is true that in 

 the larva of Spongilla Gotte described the ciliated cell layer 



1 Such a comparison was made by Metschnikoff {Zcitschr. f. zviss. 

 Z00L, xxiv., 1874, p. 12) and by Balfour {Co7nparative Embryology, vol. 

 i., p. 147). 



15 



