134 B Canadian Arctic Expedition, 1913-18 



And even if such dual origin may actually be a rare occurrence, I cannot 

 explain the distribution of Anemone trifolia, Papaver pyrenaicum, Lloydia, and 

 some other plants in any other way than that they have originated from severail 

 centres. 



In the preceding pages I have discussed the remarkable development of 

 the genus Sieversia at several points remote from each other, and although the 

 genus is both arctic and alpine, the origin of the genu? may be interpreted in 

 the same light as Pyrola, Moneses, and Chimaphila. 



We have seen that S. glacialis, S. Rossii, S. calthifolia, S. radiata, and S. 

 triflora are natives of this continent, and that S. glacialis and S. Rossii occur 

 also in eastern Siberia, S. calthifolia in Kamtchatka; furthermore, that S. 

 reptans and S. montana are endemic to the Alps and Pyrenees, S. elata to the 

 Himalayas; that *S. anemonoides is a native of Kamtchatka and Japan, S. 

 haratavica Reg. et Schmalh. of Turkestan. 



In other words, there must have been more than one centre from where the 

 genus developed, two on this continent, viz.: the arctic coast; a second one in 

 the Appalachian mountains; a third one in the Alps and Pyrenees; and a fourth 

 one in the Himalayas. The arctic element may well be connected with the 

 species in Kamtchatka and Japan, but the other species are so isolated that no 

 connection seems possible. With reference to the species themselves, the struc- 

 ture of the style, it being not jointed, wholly persistent and straight, is a constant 

 character in the genus. But while the style is glabrous in S. Rossii, it is pilose 

 in all the other species; furthermore, the petals are erect, not spreading, and of 

 a reddish colour in ^S. triflora, but in all the others the petals are yellow and 

 spreading. The members of Sieversia thus show the same sectional structures as 

 Geum with regard to the flowers, viz. : Caryophyllastrum : petals white, greenish- 

 yellow to golden-yellow, spreading; Caryophyllata: petals purplish, erect. 



The nearest ally of Sieversia is Geum, represented on this continent by eight 

 species, four of which are endemic. Among these G. macrophyllum extends to 

 Kamtchatka. G. strictum is, on the other hand, widely distributed through 

 Siberia to Central Russia, including Caucasus and the Altai mountains. G. 

 rivale, though not infrequent especially throughout the northern parts of this 

 continent, is more generally distributed in Europe and Siberia; G. urbanum has 

 become established at Cambridge, Mass. 



In the Alps and Pyrenees Geum is represented by G. rivale, G. urbanum, G. 

 pyrenaicum Willd.,and G. sylvaticum Pourr.; of these G. urbanum is the only 

 species recorded from the Himalayas. 



Another near ally of Sieversia is Dryas. which, as we have seen, occupies 

 much the same territory as the northern species of Sieversia, besides occurring in 

 the Alps and the Pyrenees. 



Sieversia is thus an excellent example of a genus having developed at 

 immensely remote stations, giving rise to species very closely related to each 

 other, yet absolutely distinct. The rule that the most natural genera, or those 

 genera in which the species are most closely related to each other, are generally 

 confined to the same country or, if they have a wide range, that their range is 

 continuous,^ can not be applied to Sieversia. And several corresponding instances 

 may be cited from other families. We remember the genus Pleuropogon of which 

 there is a high-northern type, P. Sdbinii, and two ones endemic to California, 

 viz.: P. refractum Benth., and P. californicum Benth.; furthermore, Claytonia^ 

 which is one of the most characteristic genera of North America is in Australia 

 represented by C. australasica Hook, and in New Zealand by C. calycina 

 Colenso. 



' Darwin, C. Origin of Species, l.o. p. 422. 



* Holm, Th. Types of Claytonia. Mindeskr. for Japetus Steenstrup. Kjpbenhavn, 1913. 



