academy or sc.hnces] CENTERS OF DISTRIBUTION. 19 



course, in many instances, but it is not absolutely necessary to accept it in all cases. As a 

 matter of fact it seems very difficult to appreciate the occurrence of Papaver pyrenaicum in 

 the Rocky Mountains and the Pyrenees in any other way than by accepting Schouw's hypothesis, 

 and so with a number of others, and especially of such as are Alpine rather than Arc tic- Alpine. 

 As for the latter, the Arctic-Alpine, the question of these being, to some extent, remnants of 

 a glacial vegetation must necessarily be considered. 



By examining the geographical distribution of our Alpine species it will be seen that many 

 points are in favor of the acceptance of Schouw's hypothesis. This may be demonstrated by 

 outlining the actual limits of the present distribution, and especially of the larger genera, besides 

 giving an account of association with allied species. 



With respect to the variation of the species, it may be mentioned in this connection that 

 several instances occur where variation according to northern latitude coincides with that of 

 altitude, but less so with that of longitude farther south. 



Anemone narcissiflora L. 



By De Candolle 10 this species is a member of the section Omalocarpus together with 

 A. umbdlata Willd., and A. Siberica L.; two other species — A. demissa Hook. f. and A. poly- 

 anthes Don. — from the Himalayas are referred to the same section by Hooker. 11 Of these, 

 A. narcissiflora is a very variable plant, and the following varieties are described by De Candolle: 

 fasciculata, monantha, and vilososissima, besides pedicellaris and frigida; the two last mentioned 

 may, however, according to this author, represent distinct species. The geographical distri- 

 bution of the typical plant has been given above, and concerning the varieties fasciculata occurs 

 in South Europe, in Cappadocia, and in Altai; monantha in dry, barren places in the European 

 mountains, furthermore in Davuria, eastern Asia, Alaska, and west America; vilososissima has 

 been reported from Altai, Kotzebue Sound, and Unalaska. In accordance with Turczaninow 12 

 A. narcissiflora is very frequent in the Baikal Mountains, and represented by several forms, 

 among which are fasciculata and monantha; the latter of these is said to occur in the Alpine 

 region, is of dwarfed habit, and with the flower only half as large as in the typical plant. C. A. 

 Meyer 13 states that the species is frequent throughout the Caucasus Mountains ; Kjellman " 

 observed only the variety monantha in dry as well as damp soil at St. Lawrence Bay. In Colo- 

 rado the typical plant is the most frequent, but one-flowered specimens are sometimes met 

 with; in these, however, the flower is by no means smaller than the typical plant. On the 

 other hand, there is a one-flowered form in Alaska which is very distinct from the others by the 

 much larger size of the flower, especially when compared with the Siberian plant (Baikal); 

 specimens of the Alaskan form were kindly presented to the writer by Mi - . Thomas H. Kearney, 

 who had collected it on Hall Island, Bering Sea, and near Orca, Prince Williams Sound. 



The fact that the section, Omalocarpus, is represented by several species in the Himalaya 

 Mountains, elsewhere by only one — A. narcissiflora — seems to indicate that the most important 

 center of geographical distribution and of development of the section is located in these moun- 

 tains; a characteristic of the Himalayan Anemone narcissiflora is that it is densely villous 

 as stated by Hooker, as is also the case with the plant from Baikal, thus representing the 

 forma villosissima of Alpine regions. The European and North American forms are pilose 

 rather than villous, and we have seen that the one-flowered variety differs as to the size of 

 flower. The species evidently developed in the Altai Mountains, and spread from there to the 

 Himalayas, westward to the Caucasus and the European Alps, where the plant is now widely 

 dispersed, but of a somewhat different habit — pilosa, fasciculata, monantha. In the Rocky 

 Mountains of Colorado the species has larger flowers than the European plant, as is also 

 the case of the one-flowered form (monantha), while in Alaska the latter is very distinct from 



"> Regni vegetahilis systems naturale. Vol. I, p. 212. Paris, 1818. 



11 Flora of British India. Vol. 1. London, 1S75. 



i! Flora Baiealensi-Dahurica. Moscow, 1842. P. 43. 



is Verzeichniss der Pfianzen Caucasus. St. Petersburg, 1831. P. 203. 



n Asiatiska Beringssunds-Kustens Fanerogamflora ( Vega-Expeditionens vetensk. arbet. P. 542). 



