ORIGIN OF METAMERIO SEGMENTATION. 73 



divided at the anterior opening (fig. 15) into two streams, one 

 of which passed through the mouth into the alimentary canal, 

 while the other passed through the neural canal. 



There was probably an olfactory sense organ developed from 

 the epiblast close to the front end of the neural canal over 

 Avhich this water rushed. 



The anterior convex wall of the neural canal now becomes 

 bulged out forwards, and gives rise to a large anterior lobe, 

 whose cavity opens behind into the neural canal, close to its 

 opening into the mouth (fig. 16). This anterior lobe carries 

 with it the olfactory epithelium, which, however, remains in 

 connection with the mouth by grooves or canals. It becomes 

 bi-lobed and transformed into the cerebral hemispheres of 

 living Vertebrates. 



The neural canal now closes both in front and behind, and 

 assumes some other function than that of respiration. Behind 

 the closing leaves no trace, while in front remains of the con- 

 nection are seen at the present day in the infundibulum, and 

 in the pituitary body. 



It will be evident from the above hypothetical account of the 

 origin of the Vertebrata, that I believe that the mouth and 

 anus of the Vertebrata are homologous with the corresponding 

 structures in the Invertebrate segmented animals. I have 

 stated above, that I suppose that the blastopore of the Verte- 

 brata is a specialised larval structure derived from the primitive 

 mouth of a two-layered ancestor. It will be obvious also, that, 

 according to my view, the position of this primitive mouth 

 coincided with the middle line of the dorsal surface of the 

 Vertebrate embryo, and that supposing it persisted in its 

 primitive form in the embryo until the adult mouth and anus 

 were formed, it would appear as a slit extending from the 

 mouth anteriorly and ventrally round the front end of the 

 head, along the whole surface of the medullary groove to the 

 primitive streak round the hind end to the ventrally placed 

 anus. 



In the first part of this essay which deals with the blasto- 

 pore, I have attempted to show that the mouth and anus of 



