[115] EMBRYOGKAPIIY OF OSSEOUS FISHES. 569 



ing tlie early stages of develoi^meut I believe, in some cases, at any 

 rate, to have been purely the products of reagents. Tbe separation of 

 the cells during cleavage is a very probable occurrence, and originates 

 by the cells pushing and displacing each other somewhat during this 

 process, as suggested by Whitman. The evidence wliich I liave been 

 able to gather, both from the living eggs and sections, leads me to the 

 conclusion that the KeimhUhle of Strieker and CElhicher is the true blas- 

 toccel of the Teleostean ovum, as Ziegler has more recently urged. The 

 yelk hypoblast is its floor, and the at first two-layered epiblast its roof. 

 Its development is constantly the same in all of the forms studied by 

 me, and I have not yet found any evidence of the existence of species 

 without the epiblastic or outer covering of the yelk-sack, as asserted by 

 Yon Baer. 



I see no valid reason for not regarding the yelk as an active part of 

 the ovum, through the intermediation of the yelk hypoblast, and it 

 seems evident that the segmentation cavity is simply a space filled 

 with fluid which facilitates the gliding of the blastoderm over the yelk 

 during growth, and that it is placed between the blastoderm and tlie 

 yelk, with its free nuplei peripherally displaced to a remarkable degree. 

 In other words, I would regard the yelk as an integral x)art of the Q-^^, 

 taking a share in segmentation only at a very late period. In conse- 

 quence of the almost entirely i>assive condition of the yelk during the 

 earlier stages, the blastoderm is obliged to spread to an extreme degree, 

 and in parts becomes remarkably attenuatecl. On this account I would 

 still hold to the view first expressed in my paper on the development 

 of Tylosurus, that the germ-disk alone is practically the homologue of 

 the whole Batrachian or Marsipobranch ovum, since we actually do not 

 find any intimate connection of the yelk with the embryo, except by 

 way of the vascular system, which develops late in most forms. In 

 Alosa the yelk miglit be removed at any stage without taking away any 

 essential x)art of the embryo except the floor of the segnjentation cavity. 

 The mode of development of the gastrula and cceloma, is, we find, greatly 

 modified by the ])rcsence of the yelk, but it is not an active factor in 

 the development of either by means of any process of segmentation. 



The free nuclei of the yelk hypoblast apparently proliferate as the 

 blastoderm spreads. They are, at any rate, at first confined to the 

 germinal pole of the ovum, and are only found at the opposite pole 

 after the yelk-globe has been included by the blastoderm. The infer- 

 ence, therefore, is that they spread and multii)ly with the lateral growth 

 of the blastoderm. It is these nuclei possibly which are the centers 

 of certain free cells around the margin of the germinal disk when the 

 latter has attained the morula stage, as in Gyhium and Tylomrns, as 

 shown in Fig. 3, PI. XIX, of my essay on the latter form. If such is 

 the case, it is possible that the germinal wall [KcimicaU) at the edge of 

 the blastoderm of the chick is homologous with the yelk hypoblast of 

 the fish ovum. In fact, it is highly probable that there is a yelk hypo- 



