Disputed Points in Teleostean Embryology. 213 



rated from the ring of periblast at the same time that the 

 nuclei in that ring continually become more numerous and 

 extend outwards and inwards from the marginal region of the 

 blastoderm. Agassiz and Whitman do not admit that cells 

 are separated from the periblastic layer after it is once formed. 

 They admit apparently that the marginal cells of the blasto- 

 derm, which are continuous inferiorly with the sub-blasto- 

 dermic and extra-blastodermic pellicle of protoplasm, continue 

 to divide almost up to the appearance of the actual segmen- 

 tation-cavity, and that the inner cells join the blastoderm 

 while the outer remain continuous with the protoplasmic 

 envelope of the yolk. But after this time, when the nucleated 

 syncytial periblast is definitely formed, they say it remains 

 at every stage so perfectly distinct from every other portion 

 of the embryo that they see no ground for suspecting that it 

 enters into any of the permanent embryonic layers. The 

 hypoblast they believe to arise from the division and centri- 

 petal ingrowth of cells from the margin of the disk. 



I believe myself, from a comparison of the various descrip- 



ions published and from my own observations, that some 



portion or other of the periblast is always, throughout its 



existence, budding off cells, which go to join the blastoderm 



or tissues of the embryo. 



In this belief I agree to some extent with Brook. This 

 observer, in his paper " On the Germinal Layers in Teleostei " 

 (1G), leaves the question of the first origin of the nuclei in the 

 periblast open, his preparations of the herring's ovum not 

 having afforded conclusive evidence on the subject. But he 

 maintains that the hypoblast of the germinal ring is formed 

 by cells segmented off from the nucleated periblast beneath 

 the ring. I think he is most likely right, and therefore I 

 have to abandon the view I formerly took, that the hypo- 

 blast layer was formed solely by a centripetal ingrowth from 

 the edge of the blastoderm. Of course my general view of 

 the morphological significance of the " invagination " in the 

 Teh ostean ovum remains unaltered. 



Next we have my own conclusions concerning the segmenting 

 off of cells from the periblast at a much later stage to form 

 the floor of the intestine in the region of Kupffer's vesicle (1) ; 

 my views on this point I see as yet no reason to change. 



Again, in the late embryonic history of the mackerel the 

 large oil-globule becomes entirely surrounded by periblast. 

 No ingrowth of cells from any part of the embryo ever occurs 

 bttween the oil-globule and the periblast, yet nevertheless 

 stellate chromatophores exactly similar to those of the meso- 

 blast of the skin make their appearance on the deep internal 



