238 BULLETIN 15 3, UNITED STATES NATIONAL MUSEUM 



and the British Museum contains Abyssinian material. It may be 

 that this refers to Lord Lovat's specimen, or it may be that there are 

 other unpublished records. Salvadori never listed this dove in his 

 papers on the avifauna of Shoa, and likewise it is not mentioned by 

 Von Heuglin in his great work or by Riippell. 



The most detailed statement that can be made regarding the dis- 

 tribution of this species is that it occurs north to Aletta and An- 

 deratscha in Ethiopia, while to the westward it is not known north 

 of the forested areas in Uganda, being entirel}'^ absent in the Anglo- 

 Egyptian Sudan (Nile Valley, Kordofan, and Darfur) apparently 

 not living north of the outlying patches of the Congo forest, but in 

 the western Sudan once more ranging northward around Lake Chad, 

 whence its northern limit extends to Sierra Leone. The southern 

 limit of its range is marked by the valley of the Zambesi River, 

 south of which it is replaced by the typical race. However, within 

 this vast area it is by no means ubiquitous, being entirely confined to 

 forested areas, wooded stream banks, and here and there in the 

 denser thickets in the parklike savannas. The absence of the bird 

 from the Anglo-Egjq^tian Sudan is probably wholly ecological in its 

 significance, and it may yet be found in some of the few small 

 wooded areas in the Bahr-el-Ghazal district. 



The present subspecies is characterized by the color of the upper 

 parts which are darker than in typical tympanistrla. In studying 

 this species I have examined 37 specimens, of which only 2, a single 

 male and 1 female, are tymfanistria^ the rest being fraseri. The 

 metallic wing spots are bluish or purplish blue in most of the birds 

 but in a few they are gre6n. In one case, a male with blue spots 

 and one with green ones were collected together (at Kome, Mw^anza, 

 Tanganyika Territory) so that geography, season, and wear have 

 nothing to do with the variation. Occasionally the spots are absent 

 in one wing, or at least fewer in numbers on one side than the 

 other, recalling the aberrant, spotless individual on which Hart- 

 laub "^ based his Tympamstrla vlrgo. Some of these cases seem 

 better explained by the loss of those feathers during the process 

 of preparing the specimens than by their actual absence in life. 



The size of the birds varies considerably but not geographically. 

 According to Hartert'^'' southern birds average larger than more 

 northern ones, but this difference is not constant. The Aving length 

 varies from 108 to 119 millimeters in the males, and from 108 to 

 113 millimeters in the females. 



Van Someren ^^ has described the plumages of this species in 

 detail. My observations confirm his as to the fact that the birds 



8«Nov. ZooL, vol. 25, 1918, p. 435. 



OTIbis, 1886, p. 2. 



««Journ. E. Afr. and Uganda Nat. Hist. Soc, No. 32, January, 1928, p. 162. 



