BIRDS OF ETHIOPIA AND KENYA COLONY 425 



gave its range as the region from Kilimanjaro and Olgos northward 

 to Ukamba and Kikiiyn. 



This race is similar to xanthomelas but has longer rectrices, paler 

 under wing coverts, and the females and the males in off-season 

 plumage are gi-ayer, less tawny-brownish, especiall}^ on the under- 

 parts, than xanthomelas. Granvik " writes that young birds resemble 

 the old females but are considerably paler. "In the adults, for in- 

 stance", he says, "the fore-neck is dark yellowish brown, but in the 

 young birds it is pale greyish brown. The lesser wing-coverts, which 

 in the former are distinctly olive-yellow, are in the latter pale olive- 

 yelloAv, and so on." At first sight this would seem to indicate that 

 the character of the grayer coloration ascribed to the females of 

 kilime7isis is merely a matter of age, but a fine series from Ukamba 

 and Kikuyu districts shows that this is not the case. 



3. E. c. zamhesiensis : Mozambique and the lower Zambesi Valley 

 to Nyasaland, north through the coastal belt of Tanganyika Terri- 

 tory to Tanga, inland to Morogoro. I consider litoris Neunzig to be 

 a synonym of this form. I have examined three topotypes of litoris 

 and find they agree absolutely Avith material from Inhambane, Mozam- 

 bique. This race is like hilimensis but smaller, with noticeably shorter 

 wings (male, 65-70 mm (occasionally 72 mm) as against 70-78 mm 

 in kilimensis). Sclater considers litoris a synonym of hilimensis but 

 it is really identical with zambesiensis. 



Sclater " and Belcher ^* both consider Nyasaland birds as '■'■xantho- 

 melas'''' and not as zainbesiensis. I have seen 10 specimens from 

 Nyasaland (Zomba and Chilwa) and find them all to be zambesiensis. 



4. E. c. approximans: Zululand, Natal, and the adjacent parts of 

 the Transvaal and of the eastern Cape Province. This race differs 

 from all the above in that the adult breeding males lack the yellow- 

 ish margins on the primaries. 



A race that does not directly affect the present report but that may 

 be mentioned here is the heavy-billed Ruwenzori bird, crassirostris. 

 It ranges from the lower, northern slopes of Ruwenzori to the Ituri 

 and Uele districts of the Belgian Congo. 



I have seen no Ugandan material and do not know whether the 

 birds of that country are xanthomelas or kilimensis or intermediate. 



The graph (fig. 28) illustrates the real nature of the difference in 

 the tail length in xanthomelas and hilimensis, based on adult males 

 only. From this it may be seen that the great majority of speci- 

 mens of hilimensis have longer tails than do comparable specimens of 

 xanthomelas. The specimens of hilimensis that approach and over- 

 lap the caudal measurements of xanthomelas come from the Kikuyu 



"Journ. fiir Orn., 1923, Sonderheft, pp. 168-169. 

 " Systema avium .iEthiopicarum, pt. 2, p. 762, 1930. 

 '* Birds of Nyasaland, p. 319, 1930. 



