40 



Appendices to Fourth Annual Report 



tion cavity is to a great extent obliterated. In front of the marginal thicken- 

 ing the new layer is not, however, in contact with the archiblast, but a 

 slight fissure separates the two layers. As cells are added from the floor 

 of the segmentation cavity (parablast), the cavity itself is not completely 

 filled up, so that the fissure between the archiblast and primitive hypoblast 

 marks the original extension of the segmentation cavity. It is this fact 

 which I think has led to the conclusion (which is still generally accepted) 

 that the primitive hypoblast is derived by an infolding (invagination) of 

 the archiblast. It has been argued that if the new layer is not derived 

 from an ingrowth from the old one, why should not the separation 

 between the two be complete at the periphery as well as nearer the 

 centre. The reason for the close union of the two layers at the margin 

 appears to me to be clear from an examination of fig. 21. The archiblast 

 in the region of the peripheral rim rests directly on the parablast, so that 

 the cells derived from the latter layer become closely united with those 

 of the archiblast. More central wards the two layers are separated from 

 one another by the segmentation cavity, and I take it that it is for this 

 reason the cells derived from the floor of the cavity do not unite with 

 those forming its roof. In sections stained with carmine the two layers 

 are quite distinct throughout their whole lengtL The primitive hypo- 

 blast does not stain so deeply, for the reason that the nuclei of very few 

 of its cells are in the restingr stage. 



While the primitive hypoblast is in process of formation the extension 

 of the whole blastoderm over the yolk becomes more and more marked. 

 At the same time the archiblast, in what may now be distinguished as 

 the caudal region of the embryo, thins out more than that in the anterior 

 region of the body axis (see figure 23). The thickened rim is formed 

 all around the blastoderm, but afterwards there is no increase in the 

 anterior portion. In the posterior portion the new layer grows inwards 

 as a shield-like structure (embryonal shield), thicker in the centre, the 

 cells of which are ultimately collected into the body axis. In the anterior 

 region there is a considerable thickening of the archiblast, which soon 

 becomes distinguishable as the head of the embryo and the primitive 

 hypoblast gradually thins off as it reaches this part. The anterior portion 

 of the rim, which in the herring is more elongated than in many other 

 forms, grows downwards over the yolk, leaving behind it a row of 

 archiblast cells, and a row derived from the parablast, the two being 

 separated by a fissure. The cells in the thickening itself are derived 

 from a fusion of the two layers. The method by which the yolk becomes 

 completely enclosed by cells of the blastoderm is practically the same in 

 all teleostean fishes with which I am acquainted, and has been frequently 

 described. In figure 14 the thickened rim is seen to have reached the 

 yolk pole, and a considerable extension of the caudal extremity has also 

 taken place. In figure 15 the yolk is almost completely enclosed, while 

 in figure 17 the thickening which has been pushed down from the 

 anterior extremity has united with the caudal fold. Figure 16 represents 

 a transverse section of the axis of the embryo, shortly prior to the closure 

 of the blastopore. The relation of the layers at about the stage of figure 

 17 will be understood from a comparison of figure 24, which represents 

 a longitudinal section of the same stage. 



I conclude from the examination of a large number of sections that the 

 animal pole of the ovum gives rise to the ectoderm. In many forms the 

 animal pole at the time of the formation of the segmentation cavity consists 

 only of true archiblast cells. In the herring, and probably some other 

 forms (about which I hope to write later), the animal pole receives an addi- 

 tion of cells from the parablast prior to the formation of the segmentation 



