158 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 



appearance, and is not sufficient taken by itself to deter-' 

 mine even the family, it is impossible to say what the 

 species is. The figure (Lesueur, pi. i, fig. 3) is rather 

 more like E. thurstoni than any other known form, but 

 the superficial similarity is not sufficient to prove identity. 



The primary classification of the Clavelinidse is into 

 those genera with internal longitudinal bars in the 

 branchial sac, and those with none* ; while it is a character 

 of the family as a whole that the bars, if present, are not 

 provided with proper papillae. The forms with internal 

 longitudinal bars are those which have been discussed in 

 the preceding pages, and the only genera remaining 

 without such bars are Glavelma, Peroplwropsis and (partly) 

 Perophora. But it has been stated in works on the 

 Tunicata that Perophora has "papillae" on the inner 

 surface of its branchial sac ; these structures, however, 

 are not p.ipillae comparable with those found on the 

 internal longitudinal bars of the Ascidiidae, but are 

 merely papilliform connecting ducts exactly like those on 

 each side of the endo style and dorsal lamina in some 

 species of Ecteinascidia (see PI. VI, fig. 7). 



I have recently examined the two species Perophora 

 .listeri (from Liverpool Bay) a M d P. viridis (from the Coast 

 of North America, the specimen being one of those named 

 and sent out by the United States Fish Commission) with 

 .the object of determining afresh their specific characters, 



* Laliille dismembers tbe family chiefly on this account, but I do not 

 think he is justified in doing so. I believe it forms a natural group, the old 

 " Social Ascidians " of H, Milne-Edwards, 



