290 THE PROrOCHORDATA. 



while the usefulness of the pair of gill-slits, in allowing the 

 surplus water to pass out of the pharynx, is evident. 



The notochord is more difificult to explain, and the fact 

 of its occurrence in the proboscis of Balanoglossus and 

 in the tail of the Ascidian tadpole is very puzzling. The 

 mode of its occurrence in Balanoglossus is undoubtedly 

 divergent, and not in the direct line of Vertebrate descent. 

 It is possible that the notochord has not arisen through a 

 process of elaborate change of function from a pre-existing 

 structure, but simply as a solidification of the endoderm 

 which was continued into the caudal or post-anal extension 

 of the body to form the axial support for a locomotor tail ; 

 while the subsequent extension of the notochord into the 

 pras-anal region of the body is not difficult to understand. 

 The general capacity of the endoderm for producing 

 skeletal tissue is already present in some of the Medusas 

 and Hydroid polyps whose tentacles are stiffened by a 

 solid endodermal axis. 



From a purely morphological point of view it now 

 seems as though the praeoral lobe and in a lesser degree, 

 perhaps, the hypophysis, would materially assist in furnish- 

 ing the key to a correct appreciation of the relationship 

 between the craniate Vertebrates, the Protochordates, 

 and the Invertebrates. 



As we have indicated above, in the formulation of the 

 Annelid-theory* no allowance has been made for the prin- 

 ciple of parallelism in evolution ; but it is impossible to 

 doubt that this is a very potent factor which should always 

 be borne in mind in estimating the genetic affinity between 

 widely different groups of animals. The closer the super- 

 ficial resemblance between an Annelid and a Vertebrate 

 (in the possession of somites, segmental organs, etc.) is 

 shown to be, the more perfect appears the parallelism 



