244 Proceedings of the Royal Physical Society. 



With regard to the subdivisions of the Lumbrici, it is 

 clearly necessary to indicate in the arrangement their prob- 

 able phylogenetic relationships. This is not indicated by 

 Eosa in his scheme of classification, although he does do so 

 later in his paper in the " Stammbaum/' which I have copied 

 into the present communication. 



Kosa's classification will doubtless commend itself to many 

 for the reason that it is based upon the total of a large 

 number of characters. If we exclude those which are found 

 in more than one family, we get the following diagnoses of 

 Kosa's families : — 



Lurtibricidce — 



Male pores in front of clitellum. Gizzard behind sexual 

 organs. 



Geoscolecidce — 



Copulatory setae longer than the others, and of a different 

 form. 



Acanthodrilidce — 



Four groups of penial setae (connected with the four atria). 



Eudrilidce — ? 



Perichcetidce — 



Setae very numerous in each segment. 



All of these families cannot, as constituted by Eosa, be 

 diagnosed at all. Further research, particularly the discovery 

 of the genus Deinodrilus and the species Perichceta stiiarti, 

 has rendered it at least difficult to distinguish the Perichse- 

 tida3 and the Acanthodrilidce. 



On the other hand, the Lumbricidse and Geoscolecidse 

 appear, so far as we know at present, to be natural families. 

 It is, in fact, necessary, in order to arrive at a tabular expres- 

 sion of the real affinities, to combine some of the groups into 

 larger ones, and to split up others into smaller ones. This 

 is, to a certain extent, done by Eosa in his " Stammbaum." 



He places the Acanthodrilidse quite apart from the others, 

 and at the base of the series. 



How far is this justified by our present fuller knowledge 

 of this group and of others ? 



