106 B Canadian Arctic Expedition, 1913-18 



S. radiata grows in the Alleghany mountains, North Carolina, and a variety 

 Peckii Gray has been recorded from alpine tops of the White mountains, New 

 Hampshire; S. Rossii is common in the alpine region of the Rockies in Colorado, 

 extending northward to the arctic shore and the archipelago, and westward to 

 Alaska; S. gladalis is known only from the arctic seashore west of the Mackenzie 

 river; S. calthifolia is known from the northwest coast, Unalaska and Sitka. 



According to Ledebour, S. gladalis and S. Rossii occur in eastern Siberia, 

 S. calthifolia in Kamtchatka. Beside these a fourth species is enumerated by 

 Ledebour (Flora Ross. I.e.), S. anemonoides R. Br., also from Kamtchatka. 

 Finally, there are two species in the Alps and Pyrenees, iS. montana (L.) Spreng., 

 and S. reptans (L.) Spreng.j and one in the Himalayas, S. elata Royle (alt. 9,000- 

 15,000 ft.). 



We have thus before us a genus of quite an extensive geographical distribu- 

 tion, but disconnected and seemingly unexplainable. 



None of the species are circumpolar; none have been reported from arctic 

 Europe, and nevertheless, two species, one of which is very characteristic by its 

 stoloniferous habit {S. reptans), inhabit the Alps and Pyrenees. On the Siberian 

 north coast from 173° W. Long, to 68° E. Long. S. gladalis is the only species 

 recorded so far (Kjellman, I.e.); the genus is absent from the Altai and Baikal 

 mountains which otherwise harbour so many arctic species; in the northeastern 

 corner of Asia, however, S. Rossii, S. calthifolia, and S. anemonoides are indigen- 

 ous. But between these regions and the Himalayas no species has been recorded 

 and, as mentioned above, only one occurs in these mountains. 



On this continent, however, the distribution is fairly well connected. For 

 even if S. gladalis and S. calthifolia are very rare on the northwest coast, and 

 none of them crossing the Mackenzie river, we have in S. Rossii a species quite 

 extensively distributed in the arctic region, including the islands of the archi- 

 pelago, and extending south to the alpine summits of the Rockies; furthermore, 

 S. triflora which is not arctic but widely distributed fropi the Atlantic to the 

 Pacific slope and southwards to Colorado, following the Rocky mountains. 

 Finally, the Alleghenies and the White mountains are the home of S. radiata. 

 But in spite of this fair representation of the genus on this continent, none has 

 been discovered in Greenland. 



If now the question be asked where these species, or let us say the genus, 

 originated, the answer cannot possibly be in the south, nor in the north alone. 

 The South European elements did certainly not come from the north, since 

 they are endemic to these mountains; they must have originated there, where, 

 furthermore, they are associated with two near allies: Geum and Dryas. 



With reference to the arctic species, S. gladalis and S. Rossii, these must 

 have developed in the polar regions, and of these the latter, S. Rossii, did not 

 altogether leave the south when the arctic flora returned for, as mentioned 

 above, it is still in existence on the alpine summits of the Rockies, in Colorado 

 for instance. S. triflora and S. calthifolia are -evidently of younger origin, both, 

 however, from a northern centre evidently located in Canada. Another centre 

 of development must have been in the Appalachian mountains, so far as con- 

 cerns S. radiata. Finally, with regard to S. elata, so widely secluded from the 

 other species, this must have originated in the Himalayas. 



Considering these data it seems to be characteristic of the southern element 

 of the genus that so very few species have become developed, and that these 

 are endemic to the southern mountains: S. elata to the Himalayas, S. reptans 

 and S. montana to the Alps and Pyrenees and, finally, S. radiata to the Appal- 

 achian mountains. The reason may be that the centres are of a more recent 

 date, and therefore entirely independent of each other. And it is a point of 

 great importance, I believe, that these independently developed species never- 

 theless show the typical habit of true Sieversia; the only distinction appears in 



