THE GENUS CHOEDEILES SWAINSON OBERHOLSER. 39 



ently represent the breeding form at these localities. Specimens 

 from Ashcroft, Lac La Hache, Hope, and Chilliwack, in southern 

 British Columbia, have considerable indication of the more brownish, 

 more heavily light-mottled upper surface of Chordeiles virginianus 

 hesperis, but are, on the whole, nearer Chordeiles v. virginianus. An 

 adult from Carberry, Manitoba (Xo. 57875, Am. Mus. Nat. Hist., 

 July 29, 1892), and a half-grown young from the same locality (No. 

 57876, Am. Mus. Nat. Hist., Aug. 1, 1892) are somewhat vergent 

 toward Chordeiles virginianus se7inetti, but are not pale enough 

 above for the latter and apparently belong under the typical form. 

 Another adult from Winnebago County, Iowa (No. 26724, A. N. S. 

 Phila., Sept. 1, 1879), which seems to represent the breeding bird 

 of this locality, is likewise intermediate, but nearer Chordeiles v. 

 virginianus. A typical example of Chordeiles v. virginianus from 

 Trail, extreme southeastern British Columbia (No. 2811, Victoria 

 Mem. Mus., June 14, 1902), is doubtless a migrant, since the breed- 

 ing subspecies at this place is Chordeiles virginianus hesperis. Simi- 

 larly migrant also doubtless are two adult male^ (No. 15638 and 

 No.^ 15639, L. B. Bishop, June 5, 1906) from Maple Creek, south- 

 western Saskatchewan, where the breeding subspecies is Chordeiles v. 

 hesperis. Two other adult males (No. 13449 and 3578, L. B. Bishop), 

 respectively from Stump Lake, N. Dak., August 2, 1905, and Towner 

 County, N. Dak., June 8, 1897, are likewise best considered migrant 

 Chordeiles v. virginianus, though verging a little toward C. v. hes- 

 peris, since the breeding bird of these localities is Chordeiles vir- 

 ginianus sennetti, and these two specimens are too dark for that 

 race. Another example (No. 63939, U.S.N.M.), from Pembina, 

 N. Dak., taken June 16, 1873, is unquestionably perfectly typical of 

 G. V. virginianus, and hence a belated migrant, for we have an actu- 

 ally breeding female of Chordeiles virginianus sennetti from this same 

 locality, taken three days earlier. A single individual from Cushion 

 Lake, southeastern Missouri (No. 205012, U.S.N.M.), is typical Chor- 

 deiles V. virginianus, but does not, in all probability, represent the 

 breeding form of this locality, which form is reasonably sure to be 

 Chordeiles virginianus chapmani. A single male, perhaps a migrant, 

 from Richland County, southeastern Illinois (No. 90781, L"^.S.N.M., 

 May 25, 1883), is apparently nearer the typical subspecies, though the 

 breeding bird of the adjoining county of Wabash is Chordeiles vir- 

 ginianus chapmani. 



Birds taken on Smiths Island, southeastern Virginia, prove to be 

 smaller than northern specimens, and are, in fact, about half way be- 

 tween Chordeiles virginianus virginianus and Chordeiles virginianus 

 chapmani; but, both size and coloration taken into consideration, 

 seem to be nearer the former. A single adult female, from Fort 

 Macon, North Carolina (No. 57903, U.S.N.M.), taken June 10, 1869, 



