RELATIONS OF YOLK TO GASTRULA IN TELEOSTEANS. 9 
are from this stage onwards the nuclei of the periblast, and, as 
shown by the researches of Agassiz and Whitman, they 
multiply in two directions—outwards from the edge of the 
blastoderm over the uncovered yolk, and inwards beneath the 
blastoderm. Whether the continuity between the nucleated 
periblast and the cellular blastoderm is as completely broken 
as I have now stated, or at later stages cells are formed from 
the periblast which are added to the blastoderm, is a question 
which will be discussed further on. It is certain, at all events, 
that the nuclei in the periblast are at first confined to that 
part of it which is beneath the blastoderm or near its outer 
edge, there are none at the lower pole of the yolk. In sections 
which are in my possession of the herring ovum at the stage 
when the blastoderm has enveloped a little more than half the 
yolk, the nuclei of the periblast are still confined to the neigh- 
bourhood of the edge of the blastoderm, and do not extend to 
the pole of the uncovered part of the yolk. 
There can be little doubt that the origin of the so-called free 
nuclei in Elasmobranchs and Sauropsida is exactly similar to 
that which is now demonstrated for the nuclei of the periblast 
in Teleostei. It was at first supposed that the yolk nuclei 
in those Vertebrates in which the yolk is not cellular arose 
spontaneously, a supposition which Balfour in his ‘Comp. 
Embr.’ considered improbable. A new and surprising pro- 
position concerning the origin of the nuclei of the Teleostean 
periblast has recently been sustained by H. Hoffmann (11), 
namely, that the first nuclear division after fertilisation takes 
place in a plane parallel to the surface of the blastodisc, that 
the upper of the daughter nuclei is the parent of all the 
nuclei of the blastoderm, and the lower the parent of all the 
nuclei of the periblast. The results of Agassiz and Whitman 
are in direct opposition to those of Hoffmann, and the conclu- 
sions of the former are based on very strong evidence, 
while the existence of the state of things represented in fig. 12 
is altogether incompatible with Hoffmann’s views. 
If it be admitted that the nuclei in the Teleostean periblast, 
as well as the yolk nuclei in Elasmobranchs and Sauropsida, 
