WHITE-EUMPED SWIFT 309 



Measurements (in millimeters) : Exposed culmen, 6-7; tarsus, 10; 

 wing, 166-173 (Shaw), 176-184.5 (Hartert). Tail deeply forked, 

 77-79 (Shaw), 72-84 (Hartert). 



Food. — Entirely insects, taken on the wing, but little definitely 

 recorded. J. C. Kershaw (1904), however, says that on the Kwang- 

 tung coast it feeds largely on a species of beetle that infests the "paddy" 

 (rice). 



Behavio7\ — All writers call attention to its extraordinary powers 

 of flight, often at great heights, coming down to lower levels in 

 thundery and stormy weather. Przewalski describes it as spending 

 whole days on the wing, shooting through the air and among the 

 rocks, but in morning and evening coursing low over the steppes 

 and feeding. 



Voice. — On the whole it is not a particularly noisy bird. Cochrane 

 (1914) speaks of "a moderate amount of subdued screaming" while 

 hunting. C. Ingram (1908) also, comparing it with M. apus., says it 

 has very similar habits, but is a much more silent bird and more 

 sparing of its screamlike cry; and La Touche (1931) also speaks of 

 it as extremely silent, only a faint scream being now and then 

 audible. 



Field 7)iarks. — At close quarters the white rump, taken in connec- 

 tion with the large size, renders identification easy, as the white- 

 rumped swifts of the aifinis group are much smaller. It is fre- 

 quently seen on the wing in company with the large spine-tailed swift 

 {Chaetura caudacuta)., from which it is readily distinguishable by 

 the difference in the shape of the tail, square in ChaetuTa and deeply 

 forked in M. pacifiGus. 



Fall. — Toward the end of August the flocks disappear from Mon- 

 golia, but in Japan they seem to linger till October, and an extraor- 

 dinary incident occurred in 1897, when Mr. Owston saw more than a 

 dozen on December 26, when they should have been in Australia, as 

 related in A. J. Campbell's work (1901). It may be a coincidence, 

 but this was the year of the great bush fires in Tasmania and Aus- 

 tralia, the smoke from which covered thousands of miles at land and 

 also at sea. Swinhoe obtained a specimen at Amoy in November, 

 now in the British Museum, and at Hongkong departure takes place 

 early in September. It arrives in India about September, but the 

 bulk of the migrant horde comes in October and November. 



'Winter. — From the Asiatic mainland and the islands of the North 

 Pacific this species migrates across the Pacific Ocean to Australia and 

 also in smaller numbers to Tasmania. On the way it has been re- 

 corded from Borneo by Salvadori; also from New Guinea by the 

 same author, E. B. Sharpe, and O. Finsch; it winters in Australia, 

 rarely visiting Tasmania. During its stay in Australia it is usually 

 seen on the wing, coming down only in stormy weather. 



