ON THE DRVRLOPMENT OF SAGITTA. 361 



opening lying longitudinally on the ventral side of the head, 

 and it therefore is visible in a number of consecutive trans- 

 verse sections. When traced backward to the neck the 

 alimentary canal suddenly becomes laterally compressed, and 

 lies more toward the dorsal surface of the body, marking the 

 point where the ectoderm of the stomodfeuni meets the true 

 endodermic gut. The latter is from the first very narrow, 

 and appears simply as a sort of partition between the masses 

 of mesoblast lying on each side (PL 19, fig. 8). Later it 

 becomes still narrower, and forms a thin lamina, slightly 

 thicker dorsally and venti'ally than in the middle, and this 

 condition persists until a considerable time after hatching 

 (PL 19, figs. 13, 14, end. sep.). Just at the posterior end, 

 however, this extreme lateral compression does not take 

 place, and there, up till near the end of embryonic life, the 

 cellular nature of the septum remains visible (fig. 14, 

 e7id. Sep.). 



The mesoblast of the trunk and tail region is from the 

 beginning sharply distinguished into splanchnic and somatic 

 layers, which have different origins, the somatic being 

 derived directly from the primary hypoblast, the splanchnic 

 from the outer walls of the folds. In early stages a distinct 

 coelomic cavity is seen enclosed by the mesoblast ; it is tri- 

 angular in transverse section, and placed so that the 

 splanchnic layer forms the base, and the somatic layer the 

 other two sides of the triangle (fig. 8). In these coelomic 

 spaces lie the genital cells, two on each side, and in contact 

 with the splanchnic layer ; and behind them, at a stage when 

 the folds have not yet finished growing backwards, the two 

 cavities open into one another, and a transverse section shows 

 a single archenteric cavity. 



This condition with open coelomic spaces does not, how- 

 ever, persist very long. As the embryo increases in length it 

 becomes correspondingly narrower, and since it is enclosed in 

 the egg-shell its growth is restricted, and apparently in con- 

 sequence of this its internal cavities become obliterated. In 

 sections rather later than those described above (PL 19, figs. 



