No. 3-] NOTES ON DEVELOPMENT OF SOME SPONGES. 5 1 5 



rounding cells becoming flattened to form the epithelioid wall. 

 The various canals and cavities, originally independent, open 

 one into the other, and to the exterior, by simple perforation of 

 the intermediate tissue. The oscula and pores are at first 

 indistinguishable, and are scattered over the surface of the 

 sponge, with no attempt at arrangement. Even in the adult I 

 can see no morphological distinction between the pores and 

 oscula. The difference in size is the only difference, and that 

 loses its significance because of the occurrence of apertures, 

 which hold several intermediate positions in this respect be- 

 tween pores and oscula. 



The flagellated chambers arise as independent structures, 

 which subsequently acquire connection with the canal system. 



There are in the larva, when it first attaches, a large number 

 of mesoderm cells, distinguishable from the rest by their size 

 and bulky shape. Such cells I may call formative cells. They 

 diminish greatly in number, and grow smaller in size during the 

 metamorphosis. The formative cells contain as a rule several 

 nuclei, and are destined for the most part to split up into much 

 smaller cells. The particular way in which the flagellated 

 chambers are formed in any larva depends on the behavior of 

 the formative cells. (All the cells of the larva, I may add, are 

 connected together by fine processes.) In some larvae the for- 

 mative cells arrange themselves round a central cavity (inter- 

 cellular space) so as to form a hollow sphere. Numbers of such 

 spheres, consisting of comparatively large cells, are found in 

 some larvae. Division of the cells then ensues, and the hollow 

 sphere gradually assumes the nature of a flagellated chamber. 

 In some larvae, on the other hand, all the formative cells may 

 break up into fine cells before the marking out of any flagellated 

 chambers. The mesenchyme of such larvae consists of a solid 

 mass of fine cells, with here and there a formative cell. The 

 flagellated chambers of such a larva must be formed directly 

 from a group of fine cells, probably by some rearrangement of 

 the cells round a central cavity. In other larvae, both processes 

 go on at the same time. Some of the formative cells arrange 

 themselves in hollow spheres and form chambers directly ; 

 others break up into solid masses of small cells, which subse- 

 quently acquire a cavity. That a single formative cell itself 

 ever forms a chamber, I do not believe. 



