24 APtT. 11. — >S. TTATTA. 



the definitive epithelium, or the indifferent cells wedge in be- 

 tween other cells so as to produce an epithelial layer one 

 cell tliick. If the former mode were the case, the cells of the 

 inner rows must either fall into the segmentation cavity, or shift 

 downwards along the inner surface of the outermost row, l)ecoming 

 heaped up, as it were, at the foot of the animal layer. In reality, 

 there are never and nowhere detected such indications. On the 

 contrary, in the part where differentiation is going on, the cells 

 of the outer row and those of the inner rows are found pushing 

 in between one another, and the layer of such condition passes 

 over gradually into the part which has already become a true 

 epithelium {Figs. 14-19 and Fig. 23). The latter supposition 

 seems, therefore, to be in accordance with facts. And this is 

 reasonable, since the animal laj^er has to extend exceedingly in 

 order to invest completely the macromeric hemisphere, i. e., to 

 accomplish gastrulation, as will be shown in future pages. 



b) As regards the invagination, a part of the flattened 

 macromeric field which is still seen projecting from the surface 

 below the blastopore (Fig. IG, f) comes first under notice. In 

 a little further advanced ovum {Fig. 17), this part is lost from 

 sight ; the invagination, on the other hand, goes deeper. The 

 invaginating pocket is not turned towards the center of the 

 segmentation cavity as in younger ova {Fig. IG), but pushes its 

 way dorsad and forwards nnder withdrawal of the protruded 

 part which at first is in the floor of the pocket bnt must 

 sooner or later lie in the roof {Fig. 17). The dorsal wall of 

 the blastoporic pocket is conseqnently brought to involve the 

 macromeric field, and here the archenteron {ar. en.) is formed 

 for the first time. It is therefore ,. plain that this additional 

 invagination {Fig. 17) is made enth-ely by the cells which formed 



