PHYLOGENETIC ARRANGEMENT 165 



Most of these characters will be dealt with in the succeeding pages ; it is only necessary 

 here to briefly recapitulate. The characters upon which I lay stress are these — 



(i) The continuous circle of setae. 



(a) The (generally) undifferentiated character of the calciferous glands, which 

 are often incompletely separated from the oesophagus. 



(3) The frequent presence of two pairs of egg-sacs. 



(4) The general existence of copulatory glands which are possibly the fore- 



runners of the spermiducal glands. 



(5) The existence of a series of small sacs along the septa, which may be the 



remains of a continuous series, of which some have been differentiated 

 into the sperm- and egg-sacs. 



Some of these characters are shared with members of the family Acanthodrilidae 

 and Cryptodrilidae ; but it is not an easy matter to distinguish these three families 

 as has already been pointed out. Arguments of nearly — but in my opinion not 

 quite — eqoial value might be put forward for placing either of the two remaining 

 families of the Megascolides in the position of most primitive earthworm. The 

 distribution of the Cryptodrilidae is obviously much like that of the Perichaetidae ; 

 that of the Acanthodrilidae seems to be not sufiiciently world-wide to use it as an 

 argument for the archaic nature of these worms. If all the Perichaetidae had the 

 racemose form of the spermiducal glands, which I believe to be secondarily derivable 

 from the tubular form characteristic of the Acanthodrilidae and many Cryptodrilidae, 

 it would militate against their being regarded as the most primitive group. The 

 Acanthodrilidae it will be remembered have occasionally (Plagiochaeta) continuous 

 circles of setae ; the fact that the majority have lost this arrangement is, in my 

 opinion, a further argument against regarding the Acanthodrilidae as at the base of 

 the series of earthworms. It is no doubt an exceedingly difficult matter to adjust 

 the conflicting claims of the Perichaetidae, Cryptodrilidae, and Acanthodi'ilidae, to 

 the position of the most primitive family in the group. 



With regard to other earthworms, there would seem to be no doubt that the 

 Eudrilidae and the Geoscolicidae are closely connected, and that the latter are 

 intimately related to the Lumbricidae. . The arguments for these statements are put 

 forward in the resumes which I give below of the structural characters of the several 

 groups. The question now to be considered is the relation which they bear to the 

 group Megascolides, and their relative position with regard to each other in the 

 phylogenetic scheme. 



Rosa and others have placed the Eudrilidae, with some or all of the Cryptodrilidae, 

 in one family ; this position seems to me to be quite untenable. The Eudrilidae diflfer 



