No. 3.] GROWTH AND METAMORPHOSIS OF TORNARIA. 445 



tive functions, and it is certainly significant that the cavity of 

 the notochord of Balanoglossus is continuous with the strongly 

 marked dorsal groove above the oesophagus. It has been 

 objected to Bateson's comparison that the dorsal aorta of the 

 Vertebrates is beneath the notochord, and the dorsal blood-ves- 

 sel in Balanoglossus is above it, and that in the latter the blood 

 flow is forward, while in the former it runs backwards. The 

 objection is trivial, inasmuch as it assumes the necessity of the 

 identity of the two blood-vessels. Indeed, it seems to me that 

 this very difference of the vessels in the two groups is of 

 great assistance in a comparison between them. As the noto- 

 chord arose in Vertebrates from before backzvards by the union 

 of the two sides of the digestive wall, the lateral blood-vessels 

 of the folds, at first double, would tend to unite into a single 

 trunk, and we find in the ontogeny of Vertebrates this pro- 

 cess of union of the blood-vessels to form the dorsal aorta is 

 actually brought about. Bateson's reasons, then, for including 

 Balanoglossus with the Vertebrates in the large group Chordata 

 seem to me to be valid. However, it does not follow that the 

 whole length of the Balanoglossus of to-day represents the 

 length of the ancestral form from which the Chordata arose. I 

 should rather regard the greater part of the length of Balano- 

 glossus as a secondary acquirement ; as an adaptation to its 

 peculiar habitat in a tube in the sand. The Nermertines and 

 many Annelids furnish us analogies. 1 



If we grant this much concerning the relationships of Balano- 

 glossus on the one hand to the Echinoderms, and on the other to 

 the Vertebrates, we have gone perhaps as far as the facts per- 

 mit. To follow out detailed comparisons between groups that 

 must have separated so long ago, and to give each an exact 

 place in a newly constructed phylogenetic tree is likely — and 

 we have obvious examples — to be disastrous. 



1 No one knows how long the ancestral Vertebrate may have been. We may have 

 been unconsciously prejudiced by too close comparisons with Annelids. For instance, 

 it does not seem to be plausible that the anal end of the Vertebrate was ever at the 

 distal end of the " tail"; that is, to have opened out from the last tail metamere, as in 

 Annelids. On the contrary, it seems more plausible that the vertebrate tail may rep- 

 resent a dorso-posterior outgrowth of the body beyond the anus. If the trunk was 

 already metameric it is conceivable that the tail may have secondarily become so; 

 or, the metamerisation of the trunk and tail may have been acquired at the same 

 time and each due to the same laws. 



