BIRDS OF THE ANAMBA ISLANDS. 29 



The comparison of the above specimen withfexamples of this and 

 allied forms has led incidentally to an examination of Hirundo 

 rustica rustica, Hirundo rustica transitiva, Hirundo savignii, Hirundo 

 gutturalis, Hirundo tytleri, and Hirundo eryihrogastris, with par- 

 ticular reference to their relationships with each other. After careful 

 study of the specimens in the United States National Museum and of 

 the literature on the subject, I am led to consider them all geo- 

 graphical races of one species, and thus to agree completely with the 

 view expressed by Dr. R. B. Sharpe x and more recently by Dr. E. 

 Hartert. 2 This is also the same opinion as that held by Mr. Henry 

 Seebohm, 3 except that he unites Hirundo tytleri with Hirundo ery- 

 ihrogastris. 



That Hirundo rustica rustica, with its usually unbroken dark 

 jugular band, intergrades with the broken-banded Hirundo rustica 

 gutturalis, both geographically and individually, there seems to be 

 not the slightest reason to doubt, for there are altogether too many 

 intermediate specimens. Furthermore, some of the darker exam- 

 ples of Hirundo rustica gutturalis are exceedingly close to the lighter 

 specimens of Hirundo rustica eryihrogastris, and are difficult to dis- 

 tinguish without actual comparison; in fact the two forms overlap 

 individually in all characters. Similarly, the dark extreme of 

 Hirundo rustica eryihrogastris overreaches individually the light ex- 

 treme of Hirundo rustica tytleri, and thus connects these two forms. 



After careful examination and comparison of a large series of 

 Hirundo rustica eryihrogastris from various parts of its range, in- 

 cluding a considerable series from western Alaska, I am entirely 

 unable to see any even average differences which entitle the Alaska 

 bird to subspecific separation as Hirundo erythrogastra palrneri;* since 

 all the supposed distinctions of both size and color seem to be merely 

 individual variations. 



Reverting again to Hirundo rustica tytleri, it is evident that inter- 

 mediates between this and Hirundo rustica gutturalis would be practi- 

 cally indistinguishable from Hirundo rustica eryihrogastris; and that 

 such intergradation actually takes place in northeastern Siberia 

 where the breeding range of Hirundo rustica tytleri approaches that 

 of Hirundo rustica gutturalis is evident from the fact that birds from 

 Lake Baikal are practically like specimens of Hirundo rustica ery- 

 ihrogastris, 5 although for geographical reasons referable of course to 

 Hirundo rustica gutturalis; and from the various winter specimens 

 from Burma and Cochin China that so closely resemble Hirundo 

 rustica eryihrogastris that Doctor Sharpe and others have identified 

 them as such. 



i Cat. Birds Brit. Mus., vol. 10, 1885, pp. 126-140. 



2 Vogel palaarctischen Fauna, vol. 1, 1910, pp. 800-804. 



3 Hist. Brit. Birds, vol. 2, 1884, pp. 171-172. 



4 Grinnell, Condor, vol. 4, May 15, 1902, p. 71 (Amaknak Island, Unalaska Harbor, Alaska), 

 s See Sharpe, Cat. Birds Brit. Mus., vol. 10, 1885, p. 127. 



