ORIGIN OF METAMBEIC SEGMENTATION. 101 



partly through the other end, and partly through the excretory 

 pores of the alimentary pouches. Some of these alimentary 

 pouches became, on the development of a vascular system, spe- 

 cialised as respiratory organs and retained their communica- 

 tion both with the exterior and with the alimentary cavity. 



Thus gill slits are serially homologous with nephridia. 

 This view of their origin is entirely supported by their devel- 

 opment from pouches of the hypoblast of Vertebrate embryos, 

 and by the fact that the kidney system in Vertebrata does not 

 overlap them, but begins immediately behind them. A diffi- 

 culty to this view lies in the fact that the coelom does overlap 

 the gill slits ; but I think this difl&culty is not a serious one 

 when we remember that the coelom being originally a vascular 

 space had to extend in the region of gill slits as elsewhere, and 

 that this extension might easily have proceeded either from 

 the mesoblastic somite next behind the last gill pouch, or from 

 a compression of the body in this region so that many somites 

 (probably after separation from the archenteron) extended into 

 the region of the gill slits. 



Summary. 

 The hypothesis suggested in the preceding pages are all based 

 upon the gastrsea theory, developed by Lankester and Hseckel. 

 I take the gastrsea as my starting point and do not inquire how 

 the gastrsea itself arose. I first (p. 82 to p. 91) by following 

 the gastraea theory to its logical conclusion — and there seems 

 recently to have been a disinclination on the part of some 

 morphologists to do this — attempt to show that the gastrrea 

 mouth is not only homologous with the Coeleuterate mouth, 

 but that the blastopore of the embryos of the Triploblastica is 

 homologous with the gastrsea mouth, and therefore homologous 

 with the Coeleuterate mouth ; and, finally, that if these neces- 

 sary deductions from the gastrsea theory are correct — and it 

 should be noticed that the gastrsea theory itself stands or falls 

 with them — it necessarily follows (from the consideration of 

 the Peripatus embryo) that the mouth and anus of the Triplo- 

 blastica are deriyed from the gastrsea mouth, i. e. Coeleuterate 



