ON THE GARTRÜLATION TN PETRO^fYZON. o3 



the archeiiteric cavity in tliis j^art, thougli visible for some time 

 here and there as irregular spaces, is finally lost from sight 

 {F'lcf. 22). 



c) Passing now to the translocation of the boundary groove 

 and obliteration of the segmentation cavity, I have left the 

 history of the boundary groove at a stage when the visible 

 blastopore appears for the first time {Fig. IG). Thenceforth the 

 groove shifts farther and farther forwards, as the archcnteric 

 pocket is added in length, to some extent by inward migration 

 of the micromeric epithelium, but largely by invagination of 

 the macromeric hemisphere, which process reduces the segmen- 

 tation cavity (Figs. 17-19). It is thus plain that the forward shift- 

 ing of the groove indicates the extent into which the segmentation 

 cavity is obliterated. It must however be borne in mind that 

 the ovum soon arrives at a stage, in which side by side with 

 the active invagination of the macromeric hemisphere there is 

 going on another process, viz., the backward shifting of the outer 

 micromeric epithelium. Henceforth the translocation of the 

 groove is, as stated on p. 13, only apparent. In other words, 

 the obliteration of the segraenatation cavity by the invagination 

 of the macromeric hemisphere, is very little when compared with 

 the reduction of it by the shifting backwards of the micromeric 

 epithelium. In this way the segmentation cavity is at length 

 entirely obliterated long before accomplishment of the gastrula- 

 tion [Fig. 21) ; and consequently the boundary groove is also 

 lost from sight. 



The invagination of the macromeric hemisphere and the 

 backward shifting of the outer micromeric layer represent, however, 

 by no means the sole cause for the obliteration of segmentation 

 cavity. The macromeres which form a part of the hemisphere 



