Annals of the Transvaal Museum 
228 
may be utilised for the African species. Chamaetylas gurneyi has the 
following characters : first primary much longer than the primary coverts ; 
tarsi and feet comparatively slender; bill rather short and fairly stout, 
higher than broad at the base ; under tail-coverts about half the length of 
the tail ; upper wing-coverts with white spots and the inner surface of the 
wing showing a white pattern, as in Psophocichla and Pseudoturdus, but 
underparts of the body not spotted, the middle of the belly white and the 
throat, chest and flanks rich reddish fulvous. Its structural and colour 
characters exhibit the same relationship to the following genus as Pseudo- 
turdus does to Psophocichla. Peliocichla Cabanis, genotype Turdus pelios L., 
represented in South Africa by lihonyanus, differs from Geokichla in having 
the flrst primary shorter, about equal to the primary coverts, the cheeks 
and throat striped, and in lacking the white marks on the wing-coverts 
and the white pattern under the wing. Turdus olivaceus and T. cahanisi 
cannot be associated with the preceding genus, side by side with which 
they may occur, and may therefore be placed in a new genus under the 
name of Afrocichla, genotype T. olivaceus L. ; this genus differs from 
Peliocichla in having no white pattern on the abdomen, the flrst primary 
longer than the primary coverts, the second primary equal to the seventh, 
the third and fourth equal and longest. 
Turdus (= Afrocichla) milanjensis, recorded by Swynnerton from 
Chirinda, has been given the name of swynnertoni by Bannerman (Bull. 
B.O.C. XXXI. 56, 1913). 
Sharpe (Handlist, iv. 145, 1903) has placed all our species of Monticola 
under Petrophila ; but this arrangement hardly meets modern requirements. 
The typical Monticola is at once separable on the shape of the wing, notably 
in the shortness of the outermost primary, which is only about half an 
inch in length, the second and third primaries longest and the tail hardly 
more than half the length of the wing. Cyanocichla, genotype Monticola 
cyanus (Linne), has the third and fourth primaries nearly equal and longest, 
the second about equal to the fifth. Petrophila, genotype M. cinclorhyncha 
Vigors, has the third, fourth and fifth primaries subequal and longest, the 
second about equal to the sixth. M. rupestris (Vieill.) has the third to the 
sixth primaries subequal and longest, second about equal to seventh. If 
Petrophila is to be recognised, it becomes necessary to separate M. rupestris, 
and I do so, giving it the subgeneric name of Petrornis, of which it will 
be the monotype. Further subdivision also becomes necessary, as the other 
southern species cannot be assigned to the preceding subgenera. M. hrevipes 
Strickland and Sclater is a much smaller bird than rupestris, though having 
much the same wing formula, and may be separated under the name of 
CoLONOCiNCLA subgen. nov. and would include M. angolensis, which occurs 
within our northern limits. Finally, M. explorator (Vieill.) differs so com- 
pletely from all the preceding in its long and slender legs and feet, that it 
might well be allocated to a distinct genus ; but for the present I propose 
to give it the status of a subgenus only, under the new name of Notio- 
CICHLA, of which it will be the monotype. 
With regard to species of the above-mentioned genera, Monticola 
pretoriae Gunning and Roberts (Ann. Transvaal Mus. iii. 118, 1911) should 
be added to our list, as a subspecies of hrevipes. 
Sharpe, as mentioned above, admitted Petrophila as a genus, but over- 
