Disputed Points in Teleostean Embryology. 209 



both the subblastodermic protoplasmic layer and the proto- 

 plasmic envelope of the yolk outside the blastoderm. Thus 

 the segmentation-cavity is potentially established at this early 

 stage, though there is no actual space between the central cells 

 of the blastoderm and the subblastodermic protoplasm. Both 

 the four central cells and the twelve marginal cells now go on 

 dividing, the former principally dividing in the horizontal 

 plane so as to form several layers of cells. When a marginal 

 nucleus divides it divides vertically, its two halves lying 

 horizontally side by side ; the internal half becomes com- 

 pletely separate and joins the blastoderm, while the external 

 half remains continuous with the cortical protoplasm of the 

 ovum. At a later stage the nuclei divide faster than the 

 protoplasm, and while some cells are separated off from the 

 marginal cell to join the blastoderm, the multiplying nuclei 

 extend into the cortical protoplasm both internally beneath 

 the blastoderm and externally outside the blastoderm. Thus 

 the nucleated syncytium called the periblast comes to be 

 established. Finally the centre of the blastoderm becomes 

 thinner and lifts itself up from the subblastodermic periblast, 

 and thus the actual segmentation- cavity is established. At 

 the same time the cells which are constantly being separated 

 off from the marginal periblast pass inwards and form a defi- 

 nite layer beneath the germinal ring. 



Now let us turn to the account of these processes given by 

 M'lntosh and Prince in their bulky memoir in the Edinburgh 

 Transactions (9). They say that the blastodisc is formed by 

 the segregation at one pole of protoplasm, which, moreover, 

 constitutes a superficial and tenacious layer around the 

 vitcllus ; and that this layer is itself derived by centrifugal 

 transference from the scattered protoplasm mingled with the 

 general matrix of the yolk. These authors entirely ignore 

 the distinction which exists with regard to the relations of the 

 protoplasm between pelagic eggs with a continuous yolk and 

 other eggs with a discontinuous yolk made up of yolk-spheres. 

 In the former case, as I have shown in my memoir " On the 

 Eggs and Larva? of Teleosteans " (5), all the protoplasm of 

 the ovum at the time of deposition is cortical, and there is no 

 " scattered protoplasm mingled with the general matrix of the. 

 yolk." In the herring's egg and all eggs with discontinuous 

 yolk the yolk-spherules are contained in a network of proto- 

 plasm. Jn the former case the protoplasm collects to form the 

 blastodisc only from the cortex ; in the latter case it collects 

 from the internal network as well. M'Intosh and Prince cite 

 various authors in support of the idea that the protoplasm 

 duiing development is nourished and grows at the expense of 



