Development of the Food-Fishes. 449 



and possess a nucleolus, appear at first close to the marG;in of 

 the disc and are crowded together, but soon are distributed 

 over a wide though variable area called the " nuclear zone." 

 The origin of these nuclei is still undecided, and as they 

 appear primarily quite at the periphery of the blastoderm, and 

 increase row by row over the periblast-stratum, they have been 

 derived by some authors from the nuclei of the disc. Not 

 only do they extend outwards, but, as Agassiz and Whitman 

 noted, they extend inward beneath the disc, and are promi- 

 nently seen studding the floor of the segmentation cavity. They 

 are often more numerous in some parts of the periblast, and 

 less numerous or wholly absent in others. 



Invagination of the Rim. 



Towards the close of the first or on the second day the 

 blastodermic rim ap}iears. Its mode of origin is uncertain, 

 though appearances in the living ovum strongly suggest its 

 growth as a true invaginated layer, separated from the cells of 

 the disc above by a distinct fissure which cannot be traced to 

 the periphery. 



Henneguy"^ holds that the rim is really inflected, but 

 that the outermost or " corneous epiblast " layer takes no 

 part in the process, an opinion which is directly opposed 

 by Kingsley and Conn. Certainly Ql^llacher's view (with 

 which Ryder agrees), that the hypoblast arises in situ by a 

 simple differentiation of cells, presents this formidable diffi- 

 culty, that a great part of the floor of the segmentation cavity 

 is permanently periblastic, and that the rim merely ijiter- 

 poses between the disc and the periblast beneath the embryonic 

 radius and in proximity to the margin. Further, the rim 

 clearly proceeds from the periphery towards the centre, be- 

 neath the disc, and this is inexplicable if the process be one 

 of delamination. Nor do appearances strongly favour the 

 theory that the rim is solely derived from the periblast ; but 

 as periblast cells are undoubtedly added to the periphery of the 

 disc, the rim is probably a derivative from both. 



In Petromyzon the epiblast layer extends by marginal 

 addition, by the conversion, in fact, of the non-embryonic yolk- 

 cells into epiblast cells f ; and by a like process, doubtless, 

 epiblast and lower-layer cells in Teleosteans increase at the 

 margin, the converted periblast cells being immediately re- 

 flected, along with archiblast cells, to contribute to the growing 

 and extending blastoderm. With the invagination of the 



* Bull. Soc. Philom. de Paris, Apr. 1880. 



t A. E. Shipley, " Mesoblast of Lamprey," &c., Proc. Ro}-. Soc, Nov. 



1885. 



