ORIGIN OP MBTAMBBIO SEGMENTATION. 109 



genetic history been traversed by this slit. Consequently the 

 only course open is that the mouth should be formed as a 

 secondary perforation entirely independent of the blastopore. 



From the nature of the case it is exceedingly difficult to bring 

 forward any direct proofs derived from embryology in favour of 

 this view. But I think it can be shown that there is reason to 

 believe that the mouth and anus of the Vertebrata are placed in 

 the line of the original blastopore. Amphioxus, so far as I under- 

 stand its development, offers no support to my view, but the 

 case is different with the Ascidians and the higher Vertebrata. 



Weldon ^ has shown conclusively that the anus is formed 

 within the area of the primitive streak, though after the dis- 

 appearance of the latter structure. It is on all hands admitted 

 that the primitive streak is a part of the original blastopore. I 

 need, therefore, say nothing with regard to the anus. 



The mouth, however, is a great difficulty. Dr. Dohrn has 

 attempted to show that it is derived from a pair of gill slits. 

 Now, without considering the embryological facts opposed to 

 his view, which have been so ably pointed out by Balfour, I 

 venture to suggest that it is exceedingly improbable that an 

 animal should lose its mouth and develope a new one. It is 

 surely, on a priori grounds, far more likely that it would 

 change gradually the position of its mouth than that it should 

 lose it and go through the labour of acquiring a new one, 

 though that new one is supposed to be derived from pre- 

 existing structures. 



Turning to the actual development, I may mention here two 

 facts which appear to me of importance. 



(1) In Ascidians, Kowalewsky ^ has shown that the mouth 

 at a certain stage is dorsal (neural), and that the neural canal 

 opens into it (woodcut, fig. 2, V). The neural canal, also, at a 

 slightly earlier, if not contemporaneous stage, opened behind 

 into the gut. We thus find the hypothetical siphon stage 

 of the evolution of the neural canal actually repeated in the 



1 ' Quart. Journ. of Mic. Sci.,' 1883. 



2 Kowalewsky, 'Arch, f. Mic. Anatomie,' vol. vii, 1871. 



8 



