not enter here into any discussion about their origin , nor give 

 definite names belonging to the system of germinal layers to 

 these cells. I think it for the moment by far the best not to 

 apply terms as hypoblast or entoderm to the cells of larvae like 

 the one I describe here. There is no proof whatever to designate 

 the cells as mesoblast or hypoblast etc. There is an inner mass 

 of various cells, and an outer epithelium. Voila tout. Close un- 

 der the epithelium , i. e. inward the larva , there are some layers 

 of very minute cells, the nuclei of which have almost the same 

 size. These minute roundish cells occur abundantly throughout 

 the whole larva in the stage before the flagellated epithelium is 

 visible. Afterwards they occur especially in the region of the fla- 

 gellated epithelium. Characteristic for the layer of cubic cells is , 

 that between each cell a processus of gland-cells is distinctly 

 visible. The latter lay next to the cubic cells but inwardly, nea- 

 rer to the centrum of the larva. 



The central mass of the larva shows various cells ; amongst 

 them numerons silicoblasts , especially in the portion towards the 

 cubic cells. 



In this state a normal larva swims round in irregular curves, 

 but soon (1 — 2 days) becomes fixed. This begins at the pole of 

 cubic cells, and it is clear thafc the gland-cells are here active. 

 Already before the attachment becomes the larva less elongated, 

 more spherical. Afterwards it flatteus more and more and the 

 base of attachment becomes larger. At the same time become the 

 long flagellated cells more cubic and lose their flagella. Gland- 

 cells appear and push their processes between the epithelial cells. 

 In the mean time the inner mass grows and a large quantity of 

 semi-fluid or intercellular substance is secreted. One of the re- 

 sults is , that the cubic cells become fiatter and flatter. In some 

 parts there are quite flat pavement-cells , but in others there 

 are cubic cells even in the adult sponge. I have cultivated many 

 larvae and prepared them for microscopical research in various 

 periods after the attachment. And in this way I could observe 

 in complete series of sections how the flagellated epithelium step 



