Annals of the Transvaal Museum. 
R. mascareniensis D.B. — The Rev. R. Godfrey has taken this species 
at Pirie, the most southern record known; Victoria Falls (Honourable 
P. Methuen). 
R. grayi Smith.— The Woodbush specimen (Honourable P. Methuen) 
came from an altitude of about 5500 feet. Probably humidity is an 
important factor in determining the distribution of this species. Mr. 
Methuen took the allied species fasciata Tschudi in the same neighbourhood 
at 5800 feet. 
R. ( Pyxicephalus ) ruddi Boul. — This will probably prove to be a 
synonym of R. ornata Pet. A careful comparison between the excellent 
figures published with the original descriptions of these two forms shows 
a striking resemblance, even to details of colour pattern. Mr. Boulenger 
stated however that his ruddi is ‘ 4 very closely related to P. ornata Pet . . . 
but £ ' differs in the shorter web between the toes ” a difference which is 
hardly appreciated in the figures and represents only a slight range of 
variation. 
R. (Pyxicephalus) adspersus Bib. — Dr. Werner’s figure of the tadpole 
is a very poor one, and quite incorrect so far as the tail is concerned, the 
swimming web being much better developed than represented in that 
figure. 
Phrynobatrachus. — It is hard to believe that the three forms natalensis 
Smith, ranoides Boul., and capensis Boul. really represent so many distinct 
species. The two latter are almost certainly very young forms — natalensis 
exceeds 30 mm. , whereas the type of ranoides was 22 mm. and capensis 
15 mm. In my Grahamstown series of natalensis there is a juvenile 
specimen of 17 mm. which agrees well with the description of ranoides, 
and I do not doubt but that ranoides at any rate is invalid. The distinguish- 
ing characters of the key (Boulenger) are not trustworthy, as the tympanum 
varies according to the method of preservation, and the length of the 
hind-limb also varies, though not to a great extent in adults. As regards 
capensis, there are two characters cited in the description which do not 
occur in our series, viz. absence of median papilla on the tongue and 
absence of the tarsal tubercle. 
Rappia. — This whole genus requires critical revision. Judging from 
the very extensive and overlapping distribution accorded to several closely 
allied forms, it may be doubted if they are all distinct species. For instance, 
undulata Boul. was first described from the Congo, afterwards recorded 
by the author of the species from Pietermaritzburg and now from Cape 
Division and from Port Elizabeth ; cinctiventris Cope has a similar distri- 
bution, Senegal, Congo, Angola, Zambezi, Natal, and Cape Colony ; the 
original description of undulat stated that it is to be distinguished from 
-cinctiventris by the very small head, and now the key (Boulenger) gives 
“ fingers one- third webbed — undulata ”, “ fingers not one- third webbed — 
cinctiventris The Cape Peninsula record (Boulenger) for marmorata 
is the first record of this species from the Western Province. R. argus 
Pet. was recorded by Peters from Boror, Upper Zambezi and Zanzibar 
Coast, and Mr. Boulenger ’s general distribution includes Zululand, but I 
have seen no published records from that district , though quite likely it 
will prove to occur there. 
