1 2 6 WHEELER. [Vol. V 1 1 1 . 



had paired openings on the surface of the body. This view, 

 which I fully endorse, has a good basis of anatomical data ; 

 but Nusbaum has not shown that there is a double opening 

 to the sexual organs or even a distinctly paired Anlage to the 

 ectodermal portion of the sexual apparatus. According to 

 his own figures the vagina and ejaculatory duct are unpaired 

 from the first ; the structures on which, he laid stress as being 

 paired ectodermal portions were nothing more nor less than the 

 unmodified terminations of the mesodermal ducts — the ter- 

 minal ampullae. 



4. Gejicral Considerations. 



The foregoing account of the development of the sexual 

 organs differs sufficiently from the accounts of other authors 

 to justify a brief consideration of some general questions. 



First in regard to the germ-cells. These arise as six meta- 

 meric pairs of clusters in the splanchnic walls of the meso- 

 dermal somites. Since the single layer of cells forming the 

 walls of each somite corresponds to the peritoneal epithelium 

 of Annelids, Heymons' conclusion that the germ-cells of insects 

 arise in essentially the same manner as the germ-cells of 

 Annelids, is certainly well-founded. In both groups the germ- 

 cells are local proliferations of the epithelium lining the body 

 cavity. In Annelids the germ-cells lose their connection with 

 the peritoneum and drop into the body cavity where they 

 undergo maturation. I have called attention to a similar 

 process in the XipJddiiim embryo. Whether these germ-cells 

 disintegrate or again attach themselves to the wall and become 

 invested with epithelial cells, I must leave undecided. I am 

 inclined to adopt the latter alternative, since I have found no 

 traces of germ-cells in the coelomic cavities in stages but little 

 older than the one figured. (Fig. 53.) Heymons has observed 

 in Blatta a similar migration of the germ-cells into the 

 coelomic cavity. 



I have alluded to the fact that Xiphidium exhibits more 

 pronounced metamerism in the early arrangement of the 

 germ-cells than Blatta. Strictly speaking the germ-cells in 

 the form studied by Heymons are not at all metameric since 



