EMBRYOLOGY OF THALASSEMA MELLITA 205 



By further divisions the cavity is increased in size and its wall 

 becomes composed of columnar cells filled with yolk granules 

 (Text-Fig. 6, B). According to Conn {^^6), "the central 

 cavity arises by an absorption of the digestive mass." This is 

 certainly not the case, for sections show plainly that it is en- 

 tirely due to the nature of the cell divisions. The macromeres 

 give rise to the dorsal and dorso-lateral part of the stomach 

 and intestine wall, while the progeny of the fifth and the a, b 

 and c quadrants of the fourth quartet produce the ventral and 

 ventro-lateral regions. The enteric cavity is secondarily divided 

 into stomach and intestine by a shelf which grows out from the 

 dorsal wall (Text-Fig. 6, B, C and U). Its origin is first in- 

 dicated by the projection of a single row of cells, but this is 

 immediately followed by a double row. In cross-section it 

 is accordingly, wedge-shaped. This shelf, finally, completely 

 divides the stomach from the intestine except on the ventral 

 side where a circular opening is left. A similar description is 

 given by Conn for this form and also by Treadwell for Podarke. 

 The partition becomes attached much nearer the oesophagus 

 on the right side than on the left, so that in the completely 

 differentiated trochophore the greater part of the stomach lies 

 on, the left side of the body and the intestine on the right (Text- 

 Fig. 7, B'). There is no proctodaeum, such as we find in Ain- 

 phitrite and Podarke. The anus,^ itself breaks through very 

 late, even after the trochophore has begun to feed. This seems 

 to be the case also in Eupomatus (Hatschek, '85). 



5. Formation of the CEsopJiagiis. 



The oesophagus is formed after gastrulation by the invagina- 

 tion of the greater part of the progeny of the second and third 



^ Conn has made several statements in this connection which seem to be founded 

 on errors in observations. He implies that the anus is formed at a point which 

 corresponds to one end of an elongated blastopore. This is true only in the most 

 general sense. Again, he says that the alimentary tract is complete in 12 hours. In 

 all the embryos that I have examined this is never the case until at least the eighteenth 

 hour. But the most radical error lies in the assertion that the cavity in the entoderm 

 opens directly to the exterior by the blastopore. The enteron during its formation 

 is, in reality, a closed sac and does not open to the exterior until it cornes second- 

 arily into connection with the ectodermal oesophagus. This is also the case in 

 Cyclas (Ziegler, '85), Pisidhim (Lankester, '75) and the Unionidoe (Lillie, '95). 



