216 THE CONDOR Vol.XXY 



of the series into harmony with the division indicated between the birds from 

 these two island localities, a task of some difficulty, for it means attempting 

 to indicate by definite names and definite geographic boundaries average 

 differences that do not lend themselves to hard and fast diagnoses. It is hoped, 

 however, that the division here indicated, in accord with the facts illustrated 

 in an extensive series of specimens, will be satisfactory as a rather gross separa- 

 tion of races. For any minute and detailed study of geographic variation in 

 Melospiza melodia in this general region, I do not believe that our system of 

 nomenclature can be used to advantage. That is, I do not think that it is 

 practicable to give names to any divisions smaller than are here indicated. 



Comparing the extremes mentioned, from the Queen Charlotte 

 Islands and from Vancouver Island, the outstanding characters of the birds 

 from the Queen Charlotte Islands are large size, long bill, and dark coloration; 

 from Vancouver Island, smaller size, rather stubby bill, and bright reddish 

 color. Other birds with characters the same as those of the specimens from 

 the Queen Charlotte Islands were found from the westernmost islands of the 

 Alexander Archipelago, Alaska, including two adults from Sitka, the type 

 locality of Emheriza rufina Brandt. Also, specimens closely similar to Van- 

 couver Island skins were found from the eastern islands of the Alexander 

 Archipelago and from the adjoining mainland. So. miaking due allowance for 

 variation in each series, a fairly satisfactory division of races could be made, 

 as is indicated in the accompanying map (fig. 61.) 



Birds from Chichagof, Baranof, and the Queen Charlotte Islands are 

 alilce and are here taken to represent rufina in the extreme manifestation of 

 that race. Specimens from Kuiu, Prince of AVales, "Warren, and Forrester 

 islands are intermediate in character, between rufina and morphna. Two 

 specimlens from, Duke Island are distinctly like rufina. The apparent anomaly 

 in dist'-ibution in this last-mentioned case (see map) assumes added interest 

 in view of the fact that a white-footed mouse (Peromyscus) is similarly isolated 

 on Dnke Island, the species found there (P. sitkensis) being the same as the 

 one that occurs on the distant Barauof and Chichagof islands, 



As regards names for the two subspecies of song sparrow here recog- 

 nized, the following appear to be applicable. To the song sparrow of the 

 western islands of the Alexander Archipelago, Alaska (Chichagof, Baranof, 

 Kuiu, Prince of Wales, Dall, Duke, and ad.jacent small islands), and the 

 Queen Charlotte Islands, British Columbia, the name Melospiza melodia 

 rufina (Bonaparte) will apply. This name dates from: [Passerella] rufina 

 Bonaparte, Consp. Av., t, July 15, 1850, 477 (Sitka, Alaska; ex "Emheriza 

 rufina Brandt, Deeic. Av. Eossic, 1836, tab. ii, 5"). (Citation copied from Ridg- 

 way, IT. S. Nat. Mus., BiiU. 50, part 1, 1901, p. 374.) 



Melospiza melodia morphna Oberholser (Auk, xvi, April, 1899, p. 

 183), substitute subspecific name for guttata, preoccupied, type locality near 

 moutli of Columbia River, will apply to the song sparrow of the eastern is- 

 lands of the Alexander Archipelago (Admiralty, "Wrangell, Revillagigedo, etc.) 

 and the ad.iacent mainland coast from'; Glacier Bay southward, over most of the 

 mainland of British Columbia (except the extreme northern and northeastern 

 sections), and Vancouver Island, and south at least to central Oregon. 



This arrangement of the song sparrows of the section of the northwest 

 const under consideration is in a sense a recession from my attitude in previous 



