218 PROCEEDINGS OF UNITED STATES NATIONAL MUSEUM. 



Nordpol-Ealirt, II, p. 221 seqv). It is therefore absolutely necessary 

 that the comparison should be made between fully mature birds, in the 

 hlack plmnagej and n-ithout dusly tips to the wingeoverts. If an investijia- 

 tion based upon such material shows trenchant and constant characters, 

 then we have all that is needed to establish good and undoubted species. 



As all winter specimens in the light and mottled plumage and all im- 

 mature birds with mottled wing-speculum are to be rejected, a large 

 material, of course, is needed. I have had unusual facilities in that 

 respect, and I doubt that any ornithologist has ever had 78 good speci- 

 mens, besides downy young of these three forms, as I have now before 

 me. In this vast series are birds from almost all quarters where these 

 species occur: Atlantic and Baltic coasts of Scandinavia, Spitzbergen, 

 Orkneys, Iceland, Greenland, Cumberland Sound, Hudson's Bay, north- 

 eastern coast of North America, Point Barrow, Herald Island, north- 

 eastern corner of Asia, Alaslja, Kamtschatka, Aleutian Islands, and 

 "west coast of North America as far down as San Francisco and San 

 Miguel in California. Of these 78 specimens some 30 are adults in 

 the plumage indicated above. All of these have been examined, but 

 only the measurements of 25 have been given below, as the mounted 

 specimens have not been measured in order to secure perfect uniformity 

 of the measurements. I trust thai all necessary jirecautious to obtain 

 conclusive results have thus been taken. In thefollowing, consequently, 

 is only meant specimens in totally hlaclc (not even partially' mottled) 

 plumage icith no dusky tips on the ivhite upper unng-coverts, unless other- 

 wise stated. 



There is one character which in all ages and plumages is sufficient at 

 the lirst glance to distinguish C. colurnba from the two other species, 

 viz, the color of the under wing-coverts, these being always more or 

 less brownish gray or smoky in C. colurnba and pure white in C. grylle 

 and mandiii. This character is "unfailing," and not only distinguishes 

 the adult birds, for I have young before me still partially in the down, 

 in which it is as fnWy diagnostic as in adults in full breeding plumage or 

 in the light winter garb, and in ail the 78 birds no one exception or in- 

 tergradatiou. To this mark may be added several others, as will be 

 seen from the tables of dimensions, as given below; colurnba is alto- 

 gether the larger bird, the toes besides being disproportionately longer 

 than in the other species, the bill stouter, etc. As a rule C. colurnba 

 has 14 tail-feathers, while the other two have only 12, a very remarka- 

 ble feature, though one which is not always to be relied upon, as 

 individuals of grylle* occasionally are found with 14 and of coluniba 

 with 12 rectrices. The unconditional reliance upon this character 

 caused v. Heuglin to identify a bird with 14 tail-feathers from Spits- 

 bergen as G. colurnba, a mistake he never Avould have made had he 

 looked at the color of the under wing coverts. A further difference is 



*Brehm seems to have had specimen of mandtii with 14 rectrices. Cf. Naumaonia 

 1855, p. 300. 



