80 K. KISHTXOUYE. 



opening whi^li was at first pre- appendicular gradually shifts its position 

 backward and becomes surrounded by tlie cephalothoracic appendages. 



Tlie procf-cxlaeum is not well marked out. In the sagittal section 

 represented in fig. 50, we see an invagination at the posterior end of 

 the body. The invaginatioii is crescent-shaped, as represented by tlie 

 horizontal section, fig. 53, and is continued to the grooves produced 

 by the ventral reflexion of the pleurae (pi.). Therefore the invagina- 

 tion can not he considered as the proctodneum. 



I could not find anv trace of the mesenteron at all, even in the 

 embryo of the trilobite stage (fig. 17). According to l^rooks and 

 Bruce, the proctodaeum and the mesenteron do not appear in the 

 embryonic stages. 



Coelom. 



The early development of the cœlom may easily be understood 

 by examining figs. 12-17. The cœlomic cavity is produced by the 

 splitting u]) of the mesoblast into the somatopleur and the splancli- 

 nopleur. The somatopleur is thicker than the splanclmopleur, the 

 former being many cells thick, while the latter consists of a single 

 layer of cells (figs. 29-32). AVhen the cœlomic cavity is ]rro- 

 duced, the mesoblast is already divided into many transverse seg- 

 ments, in every one of which the mesoblast is separated into 

 lateral halves, so that the cœlomic cavities are always paired in a 

 seo-ment. Thouo-h the lateral halves of the mesoblast afterwards 

 fuse toofcther secondarily at the ventral median line, the cœlomic 

 cavities of both sides remain always separate. The walls of the first 

 cœlomic cavity fuse together along the whole line in whicli the two 

 lateral members meet — from ihe ventrum up to the dorsum. Fron; 



