266 
Annals of the Transvaal Museum 
rhynchus {Ann. Transvaal Mus. vi. 117); it has the culmen length over 
20 mm., as against not more than 19 mm. in the typical capensis, and the 
whole bill is proportionately very heavy. In specimens from Knysna the 
culmen measures only 16 to 17 mm. in length, as against 18-19 mm. in 
capensis, and for this subspecies, which is intermediate between capensis 
and approximans (the latter with the length of culmen only 15-15-5), I 
propose the name of X. capensis Knysnae subsp. nov., type an adult male 
in the Transvaal Museum collection. No. 5114. It is a mistake to regard 
X. xanthomelas (Riippell) as a subspecies of capensis, as it is quite readily 
distinguished by having the primaries wholly dark brown, not margined 
with yellowish externally as in the subspecies of capensis and the mandible 
is white instead of normally black. X. capensis approximans occurs along 
the Drakensberg as far as Lydenburg District, beyond which its place is 
taken by xanthomelas. The Drakensberg and Woodbush specimens have 
the culmen of the same length, but in the Lower Zambesi valley the culmen 
measures only 13-5-14-5 mm., and for the latter I therefore propose the 
name of X. xanthomelas zambesiensis subsp. nov. ; in the typical xantho- 
melas from Abyssinia the length of the culmen is given by Riippell as 
15-5 mm., the wing as 76, tail 40 (? 50-60), tarsus 20 mm.; the three last 
dimensions in zambesiensis are 67-70, 53-55 and 21-22, respectively, 
showing that it is in all respects a smaller bird. The type is from Villa 
Pereira, Boror, taken by Kirby and myself, 25 May 1908, No. 4524 in the 
Transvaal Museum collection. 
The estrildine Finches require careful rearrangement, the genera as at 
present recognised in text-books being mixed up in a confusing manner. 
Lagonosticta of authors contains at least two genera and several subgenera, 
which should be recognised to avoid confusion. The typical Lagonosticta 
ruhricata Licht. has the second primary usually, though not invariably, 
thinning towards the tip, jamesoni having this feather normally not 
thinning and haematocephala having it as in ruhricata] again in L. hrunnei- 
ceps the second primary is normally deeply emarginate ; but this character 
is not sufficiently constant to warrant its being taken as the subgeneric 
character. In the colour of the bill, however, we find the typical L. ruhri- 
cata has it dark blue-black except at the base of the mandible, which is 
pale; in jamesoni and haematocephala the bill is wholly dark, but in hrunnei- 
ceps the bill is pale and with dark stripes on the culmen, tomia and under 
the mandible. In colour L. ruhricata, jamesoni and haematocephala have 
the under tail coverts black, but in hrunneiceps pale like the abdomen. In 
the proportions of the tail feathers, in the typical Lagonosticta and L. rhodo- 
pareia the difference between the longest and shortest is greater than the 
length of the culmen, in jamesoni not less than three-fourths of the length 
of the culmen and in hrunneiceps less than half. We see, therefore, from 
these comparisons that L. ruhricata, haematocephala and jamesoni are 
allied, but that L. hrunneiceps is well marked off and it may therefore be 
referred to the genus Rhodopyga Heuglin, type R. rhodopsis (Heuglin). 
Lagonosticta may be separated into several subgroups, but their distri- 
bution does not appear to overlap so consistently as does that of Rhodopyga, 
Lagonosticta being a temperate climate species, whose place is taken in the 
hotter low-country of the east by haematocephala, and in the drier west 
(and even overlapping on the east) its place is taken by jamesoni, I have 
