1920] Swarth: Revision of Avian Genus Passerella 77 



means of identifying specimens of Passerella readily as captured. 

 This, from the nature of the case, I fear can not be done, can never 

 be done. I believe it to be possible, at the present state of our knowl- 

 edge, to arrive at a solution of the fox sparrow problem that will 

 enable anyone to identify accurately perhaps ninety per cent of the 

 specimens taken, but it is to be doubted if diagnoses can be prepared 

 enabling anyone positivelj^ to identify all specimens of Passerella 

 secured. Even with the exceptional facilities I have enjoyed, I would 

 not guarantee the correct allocation of all the birds examined in the 

 present study. Intergrades between two forms may resemble a third 

 and, taken in their winter home, may have their characters wrongly 

 interpreted. Also immatures of one form may bear some resemblance 

 to adults of another, as is sometimes seen in the thick-billed subspecies 

 of the Schistacea group. Many specimens may be identified as of 

 certain subspecies on the basis of large comparative series, whereas 

 from one skin or a few it would be difficult to understand the 

 allocation. 



Some of the difficulties encountered in identifying specimens may 

 be understood from the following facts. Passerella is variable to an 

 extreme degree, and intergradation of characters apparently occurs 

 wherever two races come together. Thus, in the chain of subspecies 

 extending along the coast of southern Alaska and British Columbia 

 there are intergrades to be found between any two contiguous forms. 

 There is also, judging from winter birds, intergradation between each 

 of these coastal races and some inland form — iliaoa or altivagans, most 

 likely. The nature and extent of this latter type of intergradation, 

 as regards breeding birds, is still almost unknown, though winter 

 specim.ens apparently indicative of its existence are not uncommon. 

 It will be seen from this that in the Unalaschcensis group of sub- 

 species as a whole there is an extremely wide range of variation 

 possible. 



Of breeding birds of these six subspecies there are available fairly 

 extensive series of each race from several localities, but the map (fig. 

 N) will show what vast stretches of country there are as yet not rep- 

 resented in collections. Series of summer birds from any one place 

 are signally uniform in appearance, so it is evident that an explana- 

 tion of the great range of variation seen in series of migrants and 

 winter birds is to be sought in the study of specimens from many 

 sumjner stations — the variation in a series of stations rather than 

 among many specimens from one place. 



