140 THE VOYAGE OF H.M.S. CHALLENGER. 



From the main stem of the Simple Ascidians (see table, p. 150) close to Ciona 

 several lines are shown diverging to the various groups of the Ascidiidse. Rhodosoma 

 was derived, I consider, from Ciona. It is a remarkable form,' in which the test on the 

 anterior part of the Ijody has been modified to form a hood which by muscular con- 

 traction can be shut down so as to cover 1 loth the apertures. One species of Ciona, 

 C. savignii, from Japan, shows a tendency towards the modification of the test found 

 in its extreme condition in Rhodosoma. 



Two important lines from near Ciona have diverged to Ascidia and its allies on the 

 Due hand, and to Corella and its allies on the other (see table, p. 150). In Ciona 

 the stomach and intestine form a simple loop extending slightly beyond the l>ranchial 

 sac, and the heart is a fusiform tube running antero-posteriorly alongside the stomach. 

 In Corella'- and its allies, however, the branchial sac has extended posteriorly so 

 as to cause the alimentary canal as a whole to lie upon the right-hand side of the 

 branchial sac, the stomach being anterior to the intestine, and the heart placed in front 

 oi the stomach. On the other hand, in the branch leading to Ascidia and allied 

 genera the l^ranchial sac must have extended downwards upon the right-hand side of 

 the liody so as to throw the alimentary canal to the left, and carrying the stomach 

 with it, that organ has come to be placed behind the intestine, and the heart behind 

 the stomach. 



Pachychlwna and Ascidiclla ami other subdivisions of Ascidia may all be 

 represented by short twigs springing from the line leading from the Ciona-like 

 ancestor to the typical Ascidia (see table, p. 150). Corynascidia, on the other 

 hand, belongs to the Corella branch; while Ahyssascidia"' is intermediate in 

 structure between Ascidia and Corella, and is l)est represented by a short liranch 

 springing from near the base of the line leading to Corella, and not far from the 

 point occupied by the common ancestors of all the groups of Ascidiidse. The very 

 remarkal)le Chchjosomct* is an extreme modification of a side Ijranch from near Corella. 

 As in the other Corellinas, the stigmata in the branchial sac have become curved, while 

 in tlie Ascidiinse they are straight ; Init unlike any other Ascidiidse, the test has become 

 modified into a set of regularly-shaped horny plates, of which eight surround tlie 

 Ijranchial aperture and six the atrial. An approach to this condition of the test is seen 

 in Stijcla tessellata and some other species of the Cynthiidfe. 



The main stem of the Simple Ascidians beyond the ancestral Ascidiidse leads to a 

 great series of forms in which the branchial sac is highly developed, and has its surface 

 largely increased by being thrown into a series of longitudinal folds. Before this 



' See Lacaze-Duthiers, Ann. d. Sci ISai., ser. o (Zool.), torn. iv. p. 203, 1865. 

 ^ See Herdman, Notes on British Ascidians, Jonrn. Linn. Soc. Lorn!., Zool., vol. xv. p. "274, 1880. 

 " For further details as to the relationships of the genera of the Asoidiidre, see this Report, Part I. p. 285. 

 -•Eschricht, Dan.Kl: Viil. Sehk. Afh., \x. i\ 1, 1842; and Traustedt, 17(7. MedJ. Nat. For. Kj^benhani, 1879-80, 

 p. 429. 



