EARLY DEVELOPMENT OF GfiANTtA COJIFRESSA. 267 



much darker than the flagellated cells. The nuclei of the flagellated cells 

 are basophile and the cytoplasm only faintly oxyphile. The mitochondria of 

 the amphiblastula, larva are not pigmented in the fresh state. In PI. 23. 

 fig. 31 is drawn at a higher magnification a posterior granular cell. In this 

 there are both yolk and mitochondrial spheres (Y and M) and also probably 

 Golgi elements (GX) ; the nuclei of these cells are most characteristic, being 

 large, pale, and with an open reticulum. By the chrome-osmium method of 

 fixation, the ground-cytoplasm of these cells is quite smooth and fine. 



The minute structure,- of the flagellated cell is interesting, for the yolk- 

 granules are aggregated into a special group beneath the nucleus and at the 

 inner pole of the cell, as in PL 23. fig. 33, Y ; the mitochondrial granules 

 also lie nearly in the same region. On the periphery of the flagellated cells 

 are found certain enigmatic granules whose histochemical reactions are 

 unlike those of either yolk or mitochondria ; these fine granules are marked 

 OG in PL 23. fig. 33. 



The nucleus is formed of a dense basophil ' matrix in which a nucleolus is 

 visible. On the outer side of the nucleus is a coarse granule from which 

 originates the flagellum. This coarse granule is apparently the Golgi 

 apparatus, which surrounds the centrosome (see GA in PL 23. fig. 33). 



Lying inside the small blastocoel are found a few small amoeboid cells, which 

 have a nucleus like that of the flagellated cell ; the cytoplasm of these cells is 

 crammed with mitochondria; in PL 23. figs. 28 and 32 such cells are marked 

 GAM, and in PL 23. tig. 30 one of these cells is drawn at a higher power. 

 I could not find any yolk-granules in these cells — only mitochondria. 



Apart from these three cell-constituents, the sponge larva, while still in 

 situ within the maternal tissues, is found to be invested by a cell-capsule, as 

 in PL 23. fig. 32 at LGC ; these cells are seen to contain a large number of 

 mitochondria of somewhat irregular shape ; in imperfect preparations this 

 layer early peels off from the posterior granular cells. In PL 23. tig. 29 one 

 of these capsule-cells is drawn at a higher power. They contain mainly 

 mitochondria, but no certain yolk-spheres. 



So far as I can understand, the capsule around the Grantia larva is formed 

 just after the period of differentiation of the flagellated hemisphere, from 

 amoeboid elements of the mesoglea. The matter, however, is far from being 

 settled. There is one point to which I must draw attention : the maternal 

 nutrient capsule often simply surrounds the cells of the posterior pole of the 

 larva as in PL 23. fig. 32, but in other cases the maternal cells partly tend 

 to penetrate between and well into the granular posterior cells as in 

 PL 23. fig. 28 at LGC. This matter will be further investigated by me. 



