yo Jenkinson, Gervdnal Layeis of Vertebrates. 



We have now to consider whether the embryology of 

 other groups promises any better prospect of success. 



The two-layered — Gastrula or Planula — stage is, of 

 course, of widespread occurrence in all the Invertebrate 

 groups. Without pausing at present to enquire into the 

 origin of these two primary layers, let us merely follow 

 out their fate in a few cases, and in particular the fate of 

 the endoderm. Frequently, it is true, the endoderm gives 

 rise quite simply to some part of the alimentary canal ; 

 but it has always been a source of great difficulty that 

 the region of this canal which arises from the endoderm 

 is so exceedingly variable in extent. While it may and 

 generally does happen that the major part is endodermal 

 in origin, there are several cases amongst the Arthropods 

 in which only a very small midgut is derived from this 

 source, the remainder of the digestive tract being stomo- 

 daeal and proctodaeal. Apparently then, unless develop- 

 ment misleads us, ' there are,' to quote Balfour, ' instances 

 in which a very large portion of the alimentary canal is 

 phylogenetically an epiblastic structure.' ^^ Balfour felt 

 this so strongly that he suggested that there was a border 

 land between the epiblast and the hypoblast which might 

 be developed indifferently from either. Worse, however, 

 is to follow. The endodermal midgut, already reduced 

 in these cases almost to a rudiment, may in others 

 entirely disappear. 



In the Cephalopoda, the alimentary canal, according 

 to Watase,^" (though this is denied by Korschelt,") is 

 formed entirely of the stomodaeal and proctodaeal in- 

 vaginations, the yolk-sac epithelium — which is unani- 



'^■' "Comparative Embryology," vol. 2, ch. 13. 



'^ Stud. Biol. Lab. Johns Hopkins Univ., vol.4, P- 163 — 183, 1887 — 90. 

 . ^^ " Festschrift fiir Leuckart," Leipzig, 1892, p. 347 — 373. 



