ANATOMY AND PHYSIOLOGY. 57 



therefore not to be confounded with the true vascular portion 

 of the gill. To turn . Pyrosoma into a Kalpa, little more 

 seems necessary than to arrest entirely the growth of the 

 primary branchial vessels, and to give to each individual a 

 separate test. 



An arrest of development of these vessels is carried to a 

 much greater extent in Duliolum. In this form the secondary 

 vessels have not only disappeared, but the longitudinal bars 

 are also absent, and the primary vessels themselves only very 

 imperfectly developed. The two bands named by Huxley 

 respectively the " epipharyngeal " and " hypopharyngeal " in 

 this curious form, undoubtedly indicate the line of the great 

 ventral channel and oral lamina, bent up in accordance with 

 the peculiar development of the creature. In the Ascidl& 

 that have the branchial sac prolonged behind the mouth, 

 the ventral channel extends likewise behind the mouth, as 

 well as in front of it; and if we suppose the endostyle to 

 be shortened in these species, and the posterior portion of 

 the sac to be consequently drawn backward and upward, the 

 corresponding extremity of the ventral channel would pass 

 up the dorsal side of the pallia! or branchial chamber; and 

 thus this axis of the gill would at once take up the position 

 it occupies in Doliohtm : that is, part would be above or in 

 front, and part below or behind the mouth; part would form a 

 "hypopharyngeal" band, and part an "epipharyngeal" band. 



Now the primary vessels or "tubular bars" originate in 

 the sides of these bands, and are, as already stated, very im- 

 perfectly developed, extending, as they do, only for a short 

 distance, and then terminating by opening through the lining 

 membrane of the respiratory cavity into the pallial sinuses, 

 just as we have supposed the similar vessels to do in Pi/ro- 

 fioma. The vessels or "bars/"' however, of the middle portion 

 of the gill, according to Prof. Huxley, do not so terminate, 

 but end in free loops. The branchial sac is, indeed, in such 

 a rudimentary condition that one step more in its degrada- 

 tion and it would entirely disappear, and Doliohtm would be 

 scarcely distinguishable from Salpa. 



In Appe iidini lari a the gill is wholly absent; but the oral 

 lamina is represented by the " ciliated band/' which adheres 

 to the ventral surface of the respiratory cavity; and it is 

 interesting to find that the anterior extremity of this band 

 divides into two branches, which, passing towards the dorsal 

 region., encircle the cavity a little below the ganglion, just as 

 the anterior ciliated band does in Salpa, and as the anterior 

 band or collar does in Ascidia. 



