BIRDS OF ETHIOPIA AND KENYA COLONY 97 



The birds taken on the Arussi Plateau were shot in a patch of 

 juniper woods at an elevation of 9,000 feet (2,700 meters) above the 

 sea. The male is not fully adult and has the top of the head and 

 nape more or less washed with pale tawny, especially the head, 

 some of the feathers of which part are bluish gray only medially and 

 have broad tawny edges, while others are almost entirely brownish. 

 The tail feathers are tawny chestnut, suffused with bluish gray for 

 their basal three-fifths ; the upper tail coverts and rump are chiefly 

 bluish gray but still retain some barred rufous tawny feathers. 



A long series of specimens of the races of this species in the United 

 States National Museum and the Museum of Comparative Zoology 

 indicates that the intensity of the coloration of both the upper and 

 the underparts varies not only with sex, age, season, and wear, but 

 individually as well to an extent that leads me to feel that the 

 Egyptian race rupicolaeformis Brehm is probably not distinguish- 

 able from typical tmnuncidus. Swann's Arabian series of ruficolae- 

 formis (collected in the Aden Protectorate by Meinertzhagen) are 

 certainly not different from European birds, which latter no one 

 would say were other than tjq^ical tinmmculus. 



For that matter, all the races of this kestrel are based on average 

 differences, and are therefore not alwaj^s identifiable when only a 

 few specimens are available. However, the series at hand is ex- 

 tensive, and I can recognize the following subspecies: tinnuncnlus, 

 canariensis, dacofiae, japonicus^ interstinctus, neglectus, carlo, and 

 rupicolus. The South African rupicolus seems almost worthy of 

 specific rank. A specimen of carlo from Ethiopia is small enough 

 to fit the characters of neglectus to which form it certainly does not 

 belong. F. t. 7ieglectus is small, dull, dark, and barred in both sexes, 

 and is therefore obviously different from carlo. The form do^^riesi 

 was rejected by Meinertzhagen ^^ after a study of Siberian specimens. 

 Two specimens (from the Swann coll.) from Wad Medina, Blue 

 Nile, attributed to dorriesi are typical tinnuncidus. It is true that 

 they have tails of more than average length, but the difference is 

 too slight to be worthy of nomenclatural recognition. These two 

 specimens are no paler than many European and African examples 

 of tin7iunculus. 



Although the European kestrel was known to migrate to Africa 

 as far south as central Tanganyika Territory, it had not previously 

 been found high up in the Ethiopian highlands where carlo is the 

 resident form. The specimens from the Arussi Plateau indicate 

 that the species migrates not only along the Nile Valley and the 

 Red Sea coast, but over the lofty inland plateau of Ethiopia as well. 



" Ibis, 1922, pp. 60-61. 



