262 Proceedings of the Royal Physical Society. 



Eosa regards his family Geoscolecidse (=my group Geo- 

 scolecini) as being more nearly related to the Lumbricidse than 

 to any of the other groups. Criodrihcs, according to him, is 

 the connecting link. The Acanthodrilidee he thinks bring 

 them into relations with other forms. Some Acanthodrilidse 

 have a complete clitellum ; these lead to the Perichcetidce. 

 In others, as in Trigaster Lankesteri, the clitellum is ventrally 

 incomplete ; this leads to the Geoscolecidse and Lumbricidse. 

 It seems to me that Eosa lays too much stress upon the form 

 of the clitellum/ as of classificatory value ; a strict adherence 

 to the principle laid down by him would necessitate the 

 removal of Diachceta from the Geoscolecidse; for in this 

 genus, as Benham informs us, the clitellum " completely 

 surrounds the body as in Perichceta., Bigaster, etc." 



The entire group Geoscolecini is, in fact, intermediate in its 

 characters between the Acanthodrilini and the Lumbricini, but 

 its relations with the Acanthodrilini are not, I believe, with 

 the family Acanthodrilidse, but rather with the Cryptodrilidse. 

 The satisfactory definition of this group and of the Lum- 

 bricini is rendered difficult by the fact of its intermediate 

 character ; it shades off at one end into the Cryptodrilidse, and 

 at the other into the Lumbricidse. 



One of the characteristic features of the group (which it 

 shares with the Lumbricini) is the modification of the 

 clitellar setse, and also the fact that these and sometimes the 

 setse elsewhere are ornamented. Among the Acanthodrilidse 

 nothing of the kind has as yet been described ; but among 

 other families of the Acanthodrilini such variations in the 



^ The purely saddle-shaped clitellum of the Lumbricidse {cf. Rosa, 

 I lumbricidi del Piemonte, Torino, 1884, figs. 1, 4, and 5) is so far 

 modified in such Geoscolecidse as Rhinodrilus {cf. Beddard, On the 

 Structure of a new Genus of Lumbricidse, Thamnodrilus Gulielmi — P. Z. S., 

 1887, fig. 1, p. 155, fig, 2, p. 157), that the anterior part has a much 

 narrower ventral gland-free area than the posterior part. The next stage, 

 which is exemplified not only in Acanthodrilus, but in such " Eudrilidse" as 

 Deodrilus, shows an entire disappearance of the ventral non-glandular area in 

 front, but a broad non-glandular tract is still left behind. Finally, we have 

 the "complete" clitellium of Perionyx, etc. Apart altogether from clas- 

 sificatory difficulties which are involved if the modifications of the clitellum, 

 as used by Rosa, are retained, it is impossible to say where the line is to be 

 drawn. The clitellum of UrocJiceta and Rhinodrilus appears to be exactly 

 intermediate between those of Lumbricus and Acanthodrilus. 



