THE EMBRYOLOGY OF THE PORIFERA. 213 



of small cells being largely developed and three or four 

 cells deep all round. 



The embryo at the later stages of segmentation has 

 a striking resemblance to the cases of unequal segmenta- 

 tion, followed by epibolic gastrulation, which are of such 

 common occurrence in other groups of animals. In fact, 

 were it not for the subsequent fate of the two kinds of 

 cells, we should have no hesitation in identifying, by 

 analogy with other Metazoa, the micromeres as ectoderm 

 cells, and the large, granular yolk containing cells as 

 endoderm. But, as a matter of fact, the micromeres give 

 rise to the ciliated cells of the larva and the collar cells of 

 the adult, while the macromeres become the inner mass of 

 the larva and the flattened epithelium and mesoderm so 

 called of the adult. 



The next step is the histological differentiation of the 

 cells composing the two layers which hitherto have had 

 the appearance of blastomeres. At the same time that 

 the first spicules appear in the interior of the embryo, the 

 cells of the peripheral layer begin to arrange themselves 

 to form an epithelium, composed of cells which are first 

 rounded, then cubical, then columnar, and finally exces- 

 sively elongated and attenuated, each bearing a single 

 flagellum. The cells of the inner mass do not remain 

 uniform in character but differentiate into the various cells 

 composing the inner mass of the larva, such as spicule cells, 

 contractile cells, cells of the future epidermis, and finally 

 undifferentiated amoeboid cells. With respect, however, to 

 the exact composition of the inner cell mass of the larva, the 

 statements of Maas and Delate are at variance. The larva 

 is now fully developed and ready to leave the maternal body, 

 which it does in all cases, probably, by the osculum. 



The free swimming larva is oval in shape and covered 

 with cilia, or rather flagella, over the whole surface in 

 Spongilla and the horny sponges, but in the Cornacu- 

 spongicE proper, i.e., the suborder HalicJwndrina of the 

 Monaxonida of Ridley and Dendy, the flagella are absent 

 at the hinder pole. Two distinct types can further be dis- 

 tinguished in the partially flagellated larvae of Cornacu- 



