IX ENTEBOPNEUSTA—PHYLOGENY 591 



Gonads. — The development of the gonads has already been sufficiently described 

 above, p. 5S6. 



B. The almost Direct Development of the Balanoglossus Kowalevskii. 



AVe can only select a few of the principal jioints in tliis develoi)nient for 

 description. Segmentation is total and equal, and leads to the formation of a 

 coeloblastula, out of which, by invagination, a coelogastrula is produced. This 

 latter becomes covered witli cilia, and a ciliated ring forms round the blastopore, 

 which diminishes in size and finally closes ; this ring corresponds with the principal 

 ciliated ring of the Tornaria larva. At this stage the larva leaves the egg to live 

 at the bottom of the sea, without showing any trace of the form, or of the charac- 

 teristic ciliated rings, of Tornaria. 



The differentiations which take place in the archenteron are important. Its 

 anterior part becomes constricted oft' as a semilunar vesicle lying transversely. This 

 takes up the whole of the most anterior j^a-rt of the blastoccel and becomes the 

 proboscidal coelom, which thus, according to these observations, is an enteroccel. 

 Two pairs of lateral outgrowths become constricted off from the rest of the archen- 

 teron, the anterior being the rudiment of the collar coelom, and the posterior that 

 of the trunk coelom. 



The blastoccel is small from the first. 



The mouth is said to arise by the simple breaking tlirough of the intestine out- 

 wards, and tlie anus in a similar way, in the place of the original blastopore. Thus 

 the whole of the intestinal wall is of an endodermal origin. 



The developmental processes in B. Kowalevskii cannot here be further described : 

 we refer the reader to the account of the formation of the organs in the Tornaria 

 larva, given above. 



XII. Phylogeny. 



The systematic position of the Enteropneustan must still, or rather again, be 

 considered as altogether uncertain. In any case, the Enteropneusta are not closely 

 related to any single large division of the animal kingdom. Special affinities with 

 the Chordata, the Echinodermata, and the Nemertines have been long suggested, 

 and in quite recent times also with Cephalodiscus and Rliabdopleura. 



A. The relation of the Enteropneusta to the Chordata lias been maintained 

 on tlie following grounds : — 



1. The Chordata and the Enteropneusta show a ver}' far-reaching and extra- 

 ordinary agreement in their gills. This agreement holds good even in details 

 (branchial tongues, branchial skeleton, synapticulse) if the gills of Amiihioxus are 

 taken for comparison. 



2. Tlie i)roboscidal diverticulum of the Enteropneusta is, in structure and origin, 

 comparable with the chorda of the Chordata. 



■3. Tlie proboscidal skeleton of the Enteropneusta corresponds with the sheatli 

 of the chorda. 



4. The body cavities in the two groups are of enteroccelomic origin ; the pro- 

 boscidal ccelom corresponds with the anterior unpaired mesoderm vesicle of 

 AmpMoxus. 



5. The collar cord of Balnmxjlossus corresponds with tlie dorsal cord of tlie 

 Cliordata, and arises in the same way as the neural tube of Vertebrates, by sinking 

 in and covering over. 



The most recent researches liave, however, yielded results unfavourable to tliis 

 assumed homology. 



