200 Memoirs of the Indian Museum. [Vol,. 1, 



Notiodrilus (Eodrilus), to argue as BENHAM does. I think it best to recapitulate 

 the considerations and arguments which led me to constructing the systematic 

 scheme of the family M egascolecidce . As I have pointed out, and as all my colleagues 

 in this special study, among them BENHAM, have acknowledged, the family 

 MegascolecidcB is a much-branched tree, which took its origin from the acanthodriline 

 primordial form, that is to say, from Notiodrilus {Eodrihis). The greater branches of 

 this tree, the different sub-families, are well defined in their distal parts. There is, 

 for instance, no doubt ohout Poly tor eut^is belonging to a special sub-family Eudrilince , 

 or Pheretima and Megascolex belonging to a sub-family M egascolecinœ , etc. The 

 difficulty lies in defining the proximal parts of these branches or sub-families. To 

 construct a systematic scheme I am obliged to cut off these branches from the 

 main stem. In any case I am bound to separate two nearly allied forms, for the tree 

 MegascolecidcB represents an almost completely continuous structure. There may be 

 a dispute at what points to make these separations. I resolved to make them as 

 near as possible to the main stem Notiodrilus (Eodrilus). The proximal ends of 

 these cut-off branches must necessarily be very nearly allied to the main stem-genus ; 

 e.g., Octochœtits, the proximal end of the generally well-defined sub-family Octochcetince, 

 is ver3^ nearly allied to Notiodrilus {Eodrihis) , and so also are Kerria, Chilota and 

 Diplocardia, these genera being the proximal ends of other sub-families or groups. 

 But I cannot see any reason in this near affinity for not making the separation at 

 this point, for an intersection at any other point would separate affinities just as 

 close. Eutyphoeus and Dinodrilus are allied to Octochcetus ^ quite as closely as the 

 latter is to Notiodrilus {Eodrilus) , and Hoplochœtella is as near to Dinodrilus as the 

 latter is to Octochcetus. I may as well separate Notiodrilus {Eodrilus) and Octochœtus 

 as any two others of these genera , which represent a continuous row of closely allied 

 forms. Furthermore, the difference between Notiodrilus {Eodrihis) and Octochcetus, 

 the first being meganephric, and the latter, like all the other Octochœtinœ micronephric, 

 seems to be of sufficient importance for choosing this point as the limit between the 

 OctochœtincB and the Acanthodrilince {Eodrihis). As little as I could formerly follow 

 BENHAM when he made the condition of the nephridia the chief distinguishing 

 character between the greatest groups of earthworms, can I follow him now when 

 he, falling into the opposite extreme, denies these characters of the nephridial 

 system any S5^stematic importance. It may be granted that there are often difficulties 

 in determining the character of the nephridial system. As the micronephric condi- 

 tion is derived from the meganephric condition, perhaps in various ways, there 

 are certain intermediate and doubtful states not easy to be ranged under these two 

 main groups of mega- and micronephric condition (for instance the genus Lampito). 

 Furthermore, the micronephric condition may secondarily produce a state that may 

 easily be mistaken for a meganephric condition (if the originally scattered microne- 

 phric villi approximate to one another and unite at each side of one segment to form 



^ The very near affinity between Eutyphoeus and Octochcetus is still more confirmed by the recent 

 studies of BED DARD and of myself. See the discussions about the genera Octochcetus and Eutyphoeus 

 below. 



