1884.] 79 [Hyatt. 



asserts, that " the course of development invariably leads to one 

 constant result, i. e., the formation of a folded sac, the walls of 

 which are invariably composed of two layers of similar cells, with 

 nothing except an almost imperceptible residuum of structureless 

 blastema interposed between them." This layer of " structureless 

 blastema " is evidently the layer which has been described as the 

 cellular and partly granular mesoderm, its aspect being possibly 

 due to the modes of preparation, or some abnormal conditions in 

 the specimens examined by Dr. Sollas. If Dr. Sollas is not in error 

 we shall have to admit that the single wall of the blastula may 

 originate in two ways in the same group, and even in the same 

 species, in one case from free cells, as claimed by him, and in the 

 other from segmentation, as claimed by all other observers. His 

 descriptions and drawings certainly seem to show that a gas- 

 trula forms the three-layered embryo in the stage having cylin- 

 drical cells in the endoblast as well as the ectoblast, but there are 

 serious objections to the folded gastrula (pi. 37) as a normal young 

 stage of Halisarca lobularis. It seems contrary to all other obser- 

 vations, that ampullae should arise in this way as folds in the 

 walls of a gastrula, or that there should not be some representa- 

 tion of the Cinctoplanula and Ascula in this species. It would, 

 if true, be an extraordinary case of abbreviated development and 

 of great importance to the morphogenesis of Porifera. That the 

 endoblastic hemisphere may become collared and ciliated is shown 

 not only by these stages, which cannot be otherwise explained, 

 but also by the direct observations of Saville Kent (Ann. Nat. 

 Hist., vol. ii, ser. 5, p. 142, pi. 6.). He describes the so-called 

 swarm gemmule in Grantia compressa as passing from an amphi- 

 blastula (f. 11) with one hemisphere ciliated and collared, to a 

 stage in which (f. 12) all the external cells are furnished with 

 these organs. This author testifies to having seen the last stage 

 in other sponges, and considers it a natural sequence to the first. 

 That the collared stage occurs in Halisarca lobularis, however, is 

 distinctly shown by Barrois (Ann. Sci. Nat. Sec. 6, vol. 3, pi. 15, 

 f. 31-32.) 



A stage similar in all respects to the one figured by Schultze in 

 Halisarca and referred to above was observed in Hym. caruncula, 

 and Chal. limbata. A section of the former showed the interior 

 to be filled with small cells of uniform size and the external mem- 



