332 Annals of the South African Museum. 



mostly from rotten willow logs and stumps behind Cogill's Hotel, 

 at Wynberg, in the Cape Peninsula, and the results of his im- 

 portant investigations were published in 1874 (Phil. Trans., clxiv., 

 pp. 757-782). 



In 1879 Wood-Mason described a species with 21-22 pairs of 

 legs from South Africa, which he named P. moseleyi. The locality of 

 these specimens is probably King William's Town. 



In 1883 Mr. Adam Sedgwick paid a visit to the Cape, and collected 

 a large amount of material of P. capensis and P. balfonri for 

 embryological investigation. His material was apparently obtained 

 from the Newlands slope of Table Mountain, and in the first part 

 of his 'Development of P. capensis' (Q. J. M. S. (2), xxv., p. 450, 

 1885), the smaller Cape species with 18 pairs of legs was first 

 separated as a distinct form (balfouri) from capensis. All the Cape 

 species known to him were finally elaborately described in his 

 ' Monograph on the species and distribution of Peripatus ' in 

 1888. 



Since that time no new South African species have been described. 

 I have, however, published a notice of several of the new forms 

 described in the present paper, but without attaching names to 

 them (Trans. S. A. Phil. Soc., ix., pp. xviii and xix, 1897). 



Classification. In 1894 Pocock proposed to break up the old 

 genus Peripatus into three genera, and introduced the terms 

 Peripatopsis and Peripatoidcs as new generic names for P. capensis 

 and P. novcB-zealandice respectively, while Peripatus was restricted 

 to the American forms (Journ. Linn. Soc. Lond., xxiv., pp. 518- 

 526, 1894 ; also in Lydekker's Eoyal Nat. Hist., vi., p. 288, 

 1896). Although Pocock's suggestion does not appear to have 

 commended itself to subsequent authors, I believe it to be perfectly 

 justified, and have adopted it in the following pages. The species 

 of Peripatidce are neither very few in number nor very similar in the 

 structure of all their organs, and as these structural differences are 

 associated with differences in the geographical distribution the 

 establishment of several genera in the place of one becomes a matter 

 of considerable convenience. 



For instance, I am acquainted with no less than six perfectly 

 distinct and five doubtful South African species of Peripatopsis, 

 which differ from one another only in such slight characters as the 

 number of the legs, the condition of the foot in the terminal rudi- 

 mentary pair of legs, the shape of the papillae, and the colour, 

 while all agree closely in every other character, as far as at present 

 known. Moreover, these eleven forms constitute a compact group, 



