124 



theory are practically homologous with the medullary folds of the 

 Vertebrata, which , it will at once be recognised, have precisely similar 

 relations to the blastopore and the central nervous system. The prae- 

 oral pigmented pits, which were shut off from the exterior by the 

 fusion of the ciliated folds, enlarged, and represent the optic vesicles 

 of Vertebrate embryos. The elongated nerve-ring, underlying the 

 ciliated band, represents the lateral components of the brain and spinal 

 cord of the Vertebrata. The adorai ciliated band I regard as the 

 homologue and predecessor of the peripharyngeal bands and of the 

 marginal ciliated bands of the endostyle in Tunicata, for I show in 

 my larger paper that these two bands — the one transverse , and the 

 other longitudinal — are directly continuous with one another , and 

 jointly take a course which is identical with that of the adorai baud 

 in Echinoderm larvae. The primitively V-shaped condition of the 

 endostjle in Amp hioxus , as described by Will ey, is probably tobe 

 explained in a similar way, for its limbs are continuous anteriorly with 

 the ventral ends of the peripharyngeal bands. The number of entero- 

 coeles became greatly increased with the evolution of the Chordate 

 type ; the anterior pair of the ancestor is represented in Amphioxus by 

 the lateral halves into which the anterior enterocoeles , according to 

 Hatschek, becomes divided in ontogeny, the left component acqui- 

 ring a communication with the exterior (ciliated pit), the right compo- 

 nent undergoing degeneration, as in the case of the Echinoderma. 



The Enteropneusta are to be derived from the hypothetical Auri- 

 cularian ancestor by a rather more complicated series of changes. The 

 lateral halves of the circumoral ciliated band did not approximate 

 along their whole extent, but, as is clearly indicated by Tornarla, 

 became divided at the apical pole into two portions, a prae-oral and a 

 post-oral band , each of which assumed a highly sinuous course over 

 the sides of the body. Nevertheless portions of the post-oral band 

 approximated dorsally in the middle region of the body , which even- 

 tually becomes the collar-region of the adult. To this region the 

 neural canal of the adult is also confined, from which fact I infer that 

 the ciliated bands of the ancestor fused in this region , and in this 

 region only, — a conclusion corroborated by the course of the ciliated 

 bands in Tornarla. This limitation of the area of fusion also explains 

 the curious fact that the prae-oral sense-organs of the metamorphosing 

 Tornarla are not enclosed in a medullary tube , as they are in the 

 Vertebrata, but remain on the external surface of the body (at the 

 apex of the proboscis); and also that there is no neurenteric canal in 

 Balanoglossus ^ the blastopore persisting as the anus. The origin of 

 the nervous system proper is discussed in my larger paper, but is too 



