PHYLOGENETIC ARRANGEMENT 163 



be tenable ; but these, and other equally or nearly equally different types, are 

 connected by so many intermediate forms, that the difficulty is not in uniting them 

 into the same large group, but in finding adequate characters by which most of the 

 families and genera can be differentiated. There is in fact no doubt, in my opinion, 

 that the group Oligochaeta is a perfectly natural one. This being the case it is 

 desirable to attempt a general sketch of the interrelationships of the several families, 

 which shall indicate the most probable evolution of the class. In attempting a 

 solution of this problem, no help is to be obtained from Palaeontology, and very 

 little from Ontogeny; the development of only four types is known, viz. Lumbricus 

 (various species), Acanthodrilua, Rhynchelmis, and Enchytraeus (s. I.), Many different 

 opinions have been held as to the relative positions of the different families; ROSA. 

 has argued in favour of the archaic position of the Acanthodrilidae ; I have urged 

 the claims of the Perichaetidae to occupy the lowest place in the series ; BENHAM 

 has recently put forward the view that the terrestrial Oligochaeta are the most 

 modern group, and that some family (not particularized) of the ' Lumbricomorpha 

 minora' are to be looked upon as their ancestors. 



ROSA'S opinion was mainly based upon the occasional doubling of the dorsal 

 vessel in the Acanthodrilidae, which is known to be an embryonic trait, and upon 

 the presence of eight hephridia per segment in Octochaetus multiporus. Since then 

 so many worms have been found to possess a ' plectonephric ' condition, and the 

 double dorsal vessel has been met with so frequently, that more is wanted before 

 the Acanthodrilidae can be accepted as the primitive family. My own arguments 

 chiefly rested upon the structure of the nephridia in Perickaeta ; embryology has 

 apparently shown that the plectonephric condition was preceded by paired tubes, 

 though it is still not proved that the paired nephridia of existing genera are the 

 precise equivalents of the embryonic nephridia of Octochaetus and Megascolex. 



As regards the earthworms, at any rate, it seems to me to be not so difficult 

 a task to arrange them, if only the starting point could be agreed on ; the aquatic 

 families, on the other hand, are more puzzling ; this is a good deal due, I should 

 imagine, to the fact that our knowledge of them has by no means kept pace with 

 our knowledge of the terrestrial forms ; this is particularly to be regretted, in view 

 of the fact, that all probability points to the origin of the terrestrial forms from 

 aquatic and not vice versa ; this cannot, I think, necessarily mean that we must 

 look for the oldest type of Oligochaeta among the existing aquatic genera already 

 known. Continuing our general resume' of the conditions which would on a priori 

 grounds suggest an archaic position, we may refer to wide and discontinuous 

 distribution ; many undoubtedly ancient forms show this ; a familiar example is 



y 2 



