1890.] WORMS OF THE GENUS PKIUCH.ETA. 59 



the feeble development of the gizzard. The subnervian vessel, com- 

 monly believed to be absent from I'erichcetu, is found, at least in some 

 species. The reprodurti%e organs, although not presenting any 

 specially archaic characters, are not at any rate more modified than 

 those of other Earthworms. In short it cannot be urjied that the 

 organization of the Perichsetidse, as a whole, is opposed to the view 

 that these are the most primitive Oligochseta ; while the structure of 

 the excretory system in my opinion favours the supposition. 



(4) The most striking evidence, however, in favour of the deri- 

 vation of tlie paired arrangement by a gradual reduction of a 

 continuous circle of setse, is afforded by the struclxxxQ oi De'modrilus. 

 This genus is a native of New Zealand, and is in many respects 

 intermediate between Ferichata and Acanthodrilus. It is at present 

 the only Oligochaete known which possesses more than 8 setce in each 

 segment ^ and yet has not the continuous circle of setae of Perichceta. 

 Deinodrilus has 12 setae in each segment, disposed at approximately 

 equidistant intervals ; it tlierefore furnishes a connecting link 

 between the continuous circle of setae and the paired seta). Deino- 

 drilus has diffuse nephridia, more like those of certain species of 

 Acanthodrilus than those of Perichceta ; the nephridia of a few of 

 the anterior segments are more concentrated, as also are the corre- 

 sponding nephridia of Trigaster hnikesteri (Benham),*a species 

 which, in the opinion of Horst, should be referred to the genus 

 Acanthodrilus, and which is at any rate closely allied to that genus ; 

 this concentration reaches its extreme in A. mvlliporus, where the 

 nephridia of these segments are metamorphosed into a gland opening 

 into the buccal cavity. The reproductive organs are exactly like 

 those of Acanthodrilus, but the clitellum, which occupies segments 

 1-4-16, and is developed continuously round the body, is like that 

 of Perichceta. There is, therefore, a strong case for believing that 

 Deinodrilus represents a stage in the eTolution of Acanthodrilus 

 from Perichceta, or of Perichceta from Acanthodrilus. The question 

 is, which of these two alternatives is the more probable 1 The 

 species of Acanthodrilus which come nearest to Deinodrilus are 

 evidently those which have a diffuse nephridial system, i. e. A. multi- 

 porus, beddardi, schlegelii, b'dttikoferi, and antarcticus ; all these 

 species furthermore agree with Deinodrilus in having an incom- 

 plete prostomium (not dividing buccal lobe) and dorsal pores, 

 while the first and last have the persistent double dorsal vessel of 

 Deinodrilus. The species which are furthest away from Deinodrilus 

 are such forms as A. dissimilis, where the prostomium completely 

 divides the buccal segment, the nephridia are paired, and the dorsal 

 pores have commenced to disappear. These extremes are connected 

 by A. annectens, which has the incomplete prostomium and paired 

 nephridia, but the anterior pair of nephridia are much specialized 

 and open into the buccal cavity, as in A. multiporus. There are, 

 moreover, other intermediate forms. The question is real!}' inti- 

 mately connected with the development of the nephridia ; if the 



' The statement that HypogcBon has 9 setie in each segment requires verifi- 

 cation. 



