08 Proceedings of the Royal Society of Victoria. 



some of the latter nuclei — as shown in Figures 1, 2 and 4 — 

 are found out of the row, and apparently approaching the 

 segmentation cavity, seems to indicate that it may possibly 

 be derived from the layer of columnar cells. On the other 

 hand, the granular cells, as I have already said, ai'e already 

 showing signs of proliferation, and may possibly have given 

 rise to the tissue in question. 



As development goes on the granular cells proliferate 

 aapidly, especially towards the middle of the layer, where 

 they become smaller and more numerous than at the 

 periphery. As they go on increasing they occupy more 

 and more space, and hence, as they cannot project outwards, 

 on account of the I'igid layer of spicules beneath them, they 

 become invaginated, and give rise to the Fseudogastrukc 

 (Figs. 2, 4). The Pseudogastrula, then, is due to a mechanical 

 invagination of the layer of granular cells, caused by their 

 active growth and the peculiar situation of the embryo. 

 These cells do not now, however, form a single layer, as is 

 usually supposed, but a layer several or many cells thick. 

 Such is certainly the case in Grantia lahyrinthica, and, 

 judging from the observations of MetschnikofF, to which I 

 shall refer more in detail later on, I think it will probably 

 be found to be the case also in Sycandra rajjlianus. The 

 advantage of thin serial sections in the determination of such 

 a point is obvious, and probably this method of investigation 

 will lead to similar results in the case of the latter species. 



At about this period of its life history the embryo leaves 

 the maternal tissues, and escapes into a flagellated chamber, 

 by rupture of the outer wall of the capsule (Fig. 3). This 

 rupture of the capsule takes place in a ring, ai'ound the base 

 of the blister which the embryo causes on the wall of the 

 flagellated chamber. It involves, of course, not only the 

 wall of the capsule, but also the wall of the flagellated 

 chamber, which by this time has become tightly stretched 

 and the collared cells composing it apparently flattened out. 

 The outer part of the endothelial lining of the capsule and 

 the portion of the wall of the flagellated chamber immediately 

 overlying it appear to come away with the embryo when 

 the latter breaks loose, forming a more or less structureless 

 membrane, closely adherent to the layer of columnar cells 

 (Fig. 4 s.m.) After the escape of the embryo the remains 

 of the capsule appear on the wall of the flagellated chamber 

 as a shallow recess lined by flattened endothelial cells 

 (Fig. 3 e.c.) 



