50 PKOFESSOR VAN BENEDEN. 



thickness in the broadest part of the marginal welt, so that in 

 optical section it presents^ taken altogether, the form of a comma, 

 the head of the latter corresponding to the widened part of the 

 peripheral thickening. 



3rd. A layer formed by round cells, but little adherent to one 

 another, finely granular, and provided with nuclei having punc- 

 tiform nucleoli. This layer delimits the blastodisc interiorly, 

 but it does not exist throughout the area of the blastodisc ; it is 

 absent from the vault of the roof of the germinal cavity, or, at 

 any rate, is only represented in this region by a few isolated cells. 

 It is, on the other hand, fairly thick within the limits of the 

 marginal welt of the blastodisc, which rests by its intermediation 

 on the intermediate layer. 



The intermediate layer has also undergone some modifications 

 of importance. The median lenticular thickening no longer 

 exists, but in its place are found on the floor of the germinal 

 cavity certain cells which resemble in every respect the cells 

 which line the inferior face of the blastodisc, and which form 

 what we have called its third layer. It is manifest that the cells 

 which rest on the floor of the germinal cavity are derived from 

 the "median lens" of the intermediate layer; for we find by the 

 side of completely isolated cells other cells, which, whilst project- 

 ing into the germinal cavity, are still partly implicated in the 

 intermediate layer. 



The peripheral welt has preserved very nearly the same condi- 

 tion as in the preceding phases, and its form is scarcely at all 

 modified. Rounded cells, very like those which rest on the floor 

 of the germinal cavity, are found embedded in the welt. 

 Throughout the extent of the intermediate layer can be dis- 

 tinguished in contact with the deutoplasm a very regular range 

 of flattened oval nuclei. The limits of the cells are too indistinct 

 to justify one in saying that they form a simple pavement epithe- 

 lium; but it appears as though an epithelial layer, formed by a 

 single range of flat cells, were in the course of detaching itself 

 from the deeper part of the intermediate layer. This epithelium 

 rests directly on the surface of the deutoplasm. 



Such is the series of the phases of this development, which I 

 am able to describe with sulficient completeness to render it pos- 

 sible to draw positive conclusions from them The observations 

 which I have made on the ulterior stages are too imperfect lor 

 publication. It now remains for me to compare the facts which 

 I have above set forth with what is actually known relative to 

 the formation of the germ-layers in Teleostcan Fishes. 



I. What comes out most strikingly in the first place from my 



