236 Mr. G. A. Boulenger on 



(Miers), for example, shows a similarity in the disposition of 

 the grooves of the carapace, particularly in the marked trans- 

 verse branchial groove (or posterior branch of the cervical), 

 which suggests that the nearest relatives of P. niloticum need 

 not be sought for in distant parts even of the same continent. 

 If this be so, a similar argument applies with greater force 

 to attempts that have been made to trace a connexion between 

 the African Potamonidse and those of the New World. 

 Ortmann suggested that the South-American Pseudothelphu- 

 sinte were linked to the Old-World river-crabs by this same 

 Acantliothelphusa nilotka. This conclusion was disproved by 

 Alcock, who showed that the affinities of the Pseudothelphu- 

 sinse were with the Gecarcinucina?, a group which probably 

 does not occur in Africa at all. Bouvier now suggests (C. li. 

 Acad. Sci. clxv. 1917, pp. 617 & 753) that the African 

 Acantliothelphusa (with Erimetopus) forms a transition to the 

 other American subfamily, the Trichodactylinse. Here, again, 

 the argument is greatly weakened if it can be shown that the 

 essential characters of Acantliothelphusa have been acquired 

 independently in different parts of Africa by various groups 

 of Polamonautes ; nor am I yet convinced, any more than 

 were Ortmann or Alcock, that the Trichodactylinse are 

 Potamonidse at all. 



XXVI. — On the Papuan, Melanesian, and North- Australian 

 Species of the Genus liana. By G. A. BoULENGEU, 

 F.R.S. 



(Published by permission of the Trustees of the British Museum.) 



HAVING recently undertaken a much-needed revision of the 

 south-easternmost representatives of the large genus Rana, I 

 feel able, thanks to the extensive material accumulated in the 

 British Museum since the publication of the ' Catalogue of 

 Batrachians ' in 1882, to offer for consideration new views on 

 the delimitation of the species and on their synonymy, as well 

 as an attempt at a better classification of them. 



The species fall under three groups, which may be regarded 

 as natural subgenera j — 



1. Rana, s, str., represented by one species only, R. grun- 

 niens, Daud., a close ally of the widely distributed R. niacro- 

 </<>/i, D. & B„ from which it differs in the absence of tooth- 

 like processes in the lower jaw and in the shorter tibia. 



