SIGNIFICANCE OF KUPFFER’s VESICLE. 
5 
nally derived by karvokinetic division from the nuclei of the 
cells at the edge of the blastoderm. Thus another supposed 
case of autogenous origin of nuclei falls to the ground. 
I have given the exact relations of the vesicle at the time of 
its full development in the herring, as shown by a series of 
sections, in fig. 3. Figs. 2 and 4 show the relations of the 
layers behind and in front of the vesicle. It will be seen that 
at the period when the vesicle exists, the hypoblast is not dis- 
tinctly differentiated from the mesoblast, and is not columnar. 
There is no distinct separation between the notochord and 
hypoblast, but the mesoblast is very sharply marked off from the 
notochord on each side. The two layers of the epiblast are 
distinctly seen, and the lower one is continuous with the 
neurochord. In the latter there is no central cavity; this 
appears later as seen in the subsequent figures. Both in front 
of the vesicle and behind it the hypoblast and periblast are in 
contact. The mesoblast does not extend far to the sides of the 
embryo ; over the lateral and ventral parts of the yolk the 
periblast and epiblast are in contact. 
The ventral wall of the gut in Elasmobranchs is formed by 
the differentiation of cells round the yolk-nuclei ; in Amphi- 
bians by the differentiation of the superficial layer of the yolk- 
cells. There is little doubt that the floor of the gut in the 
herring is formed in an exactly similar way. As is shown by 
fig. 5, from an embryo sixtv-four hours old, no periblast, or 
scarcely any, exists below the floor of the intestine, which is in 
this region complete : it is certain that no periblast nuclei are 
present here though they are seen beneath the lateral meso- 
blast and up to the side of the intestine. The same thing is 
shown by fig. 6, which is a transverse section passing through 
the otocvsts. We may conclude, then, that this portion of the 
periblast has been used up to supply the cells of the floor of 
the gut. In the stage represented, figs. 5, 6, and 7, the intes- 
tine was not formed in the most anterior region of the body, 
and here, in front of the notochord, the neurochord comes into 
contact with the hypoblast of which there is a thin layer 
between the neural tissue and the periblast. This is shown in 
