BIRDS OF ETHIOPIA AND KENYA COLONY 321 



in northeastern Tanganyika Territory and southern and central 

 Kenya Colony; and neumanni in southern Somaliland. The last- 

 named form is said to be like graculinus in color, but smaller in size. 

 I have seen no undoubted Somaliland birds, but one specimen in the 

 United States National Museum from the collection of A. Donaldson 

 Smith, without data, may have come from southern Somaliland, as 

 it has a wing length of 124 mm, as large as that of graculinus. If 

 it came from Somaliland, it is important in that it casts doubt on 

 the validity of neumanni^ but it may have been part of a collection 

 bought by Smith in East Africa to fill out his own series. It is note- 

 worthy that in his account of Smith's Somaliland collection^ 

 Sharpe "^ does not list Sigmodus retzii. 



These six races may be told as follows: graculinus and neumanni 

 differ from each other chiefly in size, the latter having a wing length 

 of from 114 to 120 mm as against 120 to 130 mm in the former, and 

 both differ from all the other races in that neither of them has a 

 broad white band on the inner webs of primaries, while retzii^ 

 nigricans, tricolor, and intermedius have this band well developed. 

 (It is noticeable only on the underside of the wings.) These four 

 subspecies differ in the color of the back, inner remiges, and upper 

 wing coverts. These feathers are dusky grayish brown in retzii; 

 more grayish, rather brownish ashy gray in nigricans; hair brown 

 in intermedius; and pale drab in tricolor. 



The two adults collected by the Frick expedition are very similar, 

 but one has a very narrow, short white line on the margin of the 

 inner web of each primary, while the other has no white. Both the 

 young birds have a narrow white band on the inner webs of the 

 primaries, the band being considerably more prominent in one bird 

 than in the other, but in neither is the band even a third as wide a« 

 in a corresponding specimen of tricolor. Van Someren °^ found that 

 in his series of 14 birds from Kenya Colony "some adults show traces 

 of white on the inner webs of the primaries, and all the young and 

 immature birds exhibit this character, indicating a very close rela- 

 tionship to 8. r. intermedius and tricolor P 



Roberts ^^ has proposed a genus Eressornis for Sigmodus retzii 

 on the basis of the longer rictal bristles. "In the typical Sigmodus'"', 

 he says, "the base of the bill is exposed ; but in S. retzii ( Wahlberg) 

 the base of the bill is hidden by the frontal bristles, the longer of 

 which extend over the nostrils, and * * * would therefore place 

 it in a new genus under the name EressornisP I have compared 

 this species with caniceps, the genotype of Sigmodus, and find that 



"Proc. Zool. Soc. London, 1895, pp. 457-520. 



^Nov. Zool., vol. 29, p. 109, 1922. 



»0Ann. Transvaal Mus., vol. 8, p. 248, 1922. 



