398 HEMICHORDA OR ENTEROPNEUSTA. 



(2) " Gillslils." — Numerous gill slits (Fig. 124, ^j. ) open from the 



anterior region of the gut to the exterior, and are separated 

 from one another by skeletal bars, which in some ways 

 resemble the framework of the respiratory pharynx in 

 Amphioxus. There are, however, many differences in 

 detail, thus the slits open dorsally, not ventrally ; the 

 skeletal bars are differently disposed, the blood supply is 

 different. Nor is it certain that the gullet of Balano- 

 glossus is endodermic like that of Vertebrates. Still, the 

 possession of these respiratory slits is one of the most 

 satisfactory of the alleged Vertebrate-like characters of 

 Balaiwglossits. 



(3) '"'' Dorsal nerve cord.^^ — A dorsal median insinking {Fig. 125, 



d.n) of ectoderm, especially strong in the region of the 

 collar, may be compared with the medullary canal of Verte- 

 brates. But it must be noticed that there is also a ventral 

 nerve cord (Fig. 125, v.n] which cannot be ignored as 

 subsidiary in character. 



Mr. Bateson has also noted that the mesoblast arises, as inAmphioxtis^ 

 &c. , in the form of ccelome pouches, but this is true of many Inverte- 

 brates. He states that the history of the anterior ccelome pocket, which 

 grows forward into the proboscis of Balaiioglosstcs^ is closely like that in 

 Amphioxus, but this is denied by Spengel. He compares a slight fold, 

 which grows backwards from in front of the gill slits, with the epipleural 

 folds of Amphioxus (Fig. 124, op. ), but the fold appears to be restricted to 

 one species. It is still uncertain what weight should be attached to the 

 fact that Balanoglossus is unsegmented. 



Affinities with Annelids (after Prof Spengel). 



(i) The larva (Tornaria) (Fig. 127) may be regarded as a modi- 

 fied Trochosphere, but this points at most to a far-off 

 common stock. Moreover, the nephridia, usually present 

 in the Trochosphere, are unrepresented in the Tornaria. 



(2) The body cavity is fonned from segmentally arranged 



ccelome pouches ; but there is a pair of pre-oral pouches 

 absent in Annelids, and segmental arrangement in the 

 organs of the body in Balanoglossus, is, to say the least, 

 very vague. 



(3) The heart lies, as in some Annelids, dorsal to the gut, not 



ventral as in Vertebrates ; the dorsal vessel carries blood 

 forwards, the ventral backwards, as is usual in Annelids. 

 But the double nervous system is essentially different from 

 that of Annelids ; and the gill slits are also, so far as we 

 know, unrepresented. If there be a relationship between 

 Enteropneusta and Annelids, it must be a very distant one, 

 perhaps restricted to origin from some common stock. 



Besides these affinities, others have been ingeniously detected. Those 

 alleged to exist between Enteropneusta and Nemerteans, e.g. , the exter- 

 nal ciliation, the unsegmented musculature, the correspondence of the 

 " notochord " and the Nemertean proboscis, are even more unsatisfac- 



