Vol. XIX"| Grinnell, Geographical Abrasion in Birds. I 29 



bearing upon the study of subspecies should not be underrated, 

 since differences due to such a factor may be found correlated 

 with different areas in the habitat of a species. The disposition 

 now is to grasp at any perceptible character common to a series of 

 specimens from one locality and to use it to distinguish a ' new ' 

 subspecies. I believe the discrimination of even the slightest 

 differences to be of importance. But I would urge that a character 

 which is purely adventitious and due to external causes cannot 

 serve to characterize a subspecies ; for I believe that a subspecies 

 is an incipient species, and that only what we can judge to be in- 

 cipient species should be called subspecies. Direct mutilations 

 from external sources must not be confused with innate manifesta- 

 tions, developed from individual variations by natural selection 

 and perpetuated through inheritance. The latter constitute sub- 

 species and species. 



I have suggested that variation in abrasion may exist, and that 

 such variation should be discriminated against by those who seek 

 minute color characters. But I do not believe there has so far 

 been much error on that score. Fortunately, color characters are 

 usually accompanied by differences in extent of markings, propor- 

 tions of measurements, etc. It might be advisable, however, here- 

 after to use as types of detailed color descriptions, especially in 

 the case of geographical races, specimens having newly-acquired 

 plumages. Colorations at other stages of feather wear might then 

 be intelligibly explained in comparison. 



I wish to call attention to one case to which the above remarks 

 seem to apply. A subspecies of the Russet-backed Thrush has 

 been distinguished (^Hylocichla nstulata flf^Z/Vrt;), the habitat of which 

 is given as " California, excepting the northern coast ; north in the 

 interior to southern Oregon " ; etc. The habitat of Hylocichla 

 nstulata ustulata is thus restricted to the " Northwest Coast region." 

 I have before me 32 specimens of the Russet-backed Thrush from 

 the Pacific Coast, all collected by myself, as follows : Pasadena 

 (10), Pacific Grove (2), Palo Alto (r i) , Seattle (i), Sitka (8). 

 These represent habitats of the two alleged subspecies, as 

 defined, by 23 specimens, and 9 specimens, respectively. 

 Turning to the original description of «^/m (Auk, XVI, Jan. 1899, 

 pp. 23-25), we find it characterized as being similar to nstulata, 



