89 



The syncA^tial influoiicc. towards the end of the second day 

 that is towards the end of the segmentation period, extends to the 

 sheet of protoplasm next the margin on the ventral side of the 

 blastodisc. This appears always to occnr by a number," not all, of 

 the cells of the blastodisc which are attached to this layer dividing 

 vertically (fig. Im) the ventral of the resulting cells sinking into and 

 merging with the protoplasmic sheet which thus becomes thickened, 

 the other remaining connected with the cells around it. Early in 

 the third day the same process takes place over the whole of the 

 ventral protoplasmic sheet of the blastodisc, converting the whole 

 of this into a s\Ticytial la^^er continuous with the marginal s\ai- 

 cytium. I have looked carefully over a great many preparations 

 of the stages concerned of the cod, whiting and plaice, and I cannot 

 find the least trace of an inwandering of the marginal nuclei into the 

 sub-germinal periblast as described by Agassiz and Whitman. In 

 the case of the plaice the marginal cells retain their individuality 

 longer than in gadoids, and the change described takes place almost 

 simultaneously over the whole of the sub-germinal periblast. In- 

 ferentially then an invasion of this area by marginal nuclei is 

 impossible. 



This is followed by a change of the greatest importance. The 

 prelude to it is the increase in number of, and the fusion of the 

 vacuoles of the marginal syncytium, dividing definitely an ectoplasm 

 continuous with that to w^hich the outer layer of blastomeres is 

 firmly connected (fig. 2a) from an entoplasm in which the marginal 

 nuclei are embedded, and which is continuous with the ventral sheet 

 of protoplasm of the blastodisc. Then the whole blastodisc becomes 

 gradually flattened, the margin at the same time advancing over 

 the yolk. The blastomeres near the margin press into the excavated 

 region external to them, and the lower s;yTicytial layer next the yolk 

 becomes defined by the withdrawal of the cells above it. A cavity 

 is thus formed, which becomes more and more apparent as the pro- 

 toplasmic strands are absorbed. 



By these processes the blastodisc has become resolved into a 

 many-layered ectoderm, and a sjaicj-tial entoderm united at their 

 margins. The outer protoplasmic sheet is still present, and the 

 outer layer of ectodermal cells now flattened is closety applied to 

 it. The entoderm is based on the primitive protoplasmic sheet 

 next the yolk, and now consists of a thickened sheet of protoplasm 



