1898-1902. No. 2.] VASCULAR PLANTS OF ELLESMERELAND. 51 



The description of Rydberg (1. c, p. 101-102) i.s, on the whole, 

 well in accord with this, the small differences are probably due to the 

 greater material which he has had at his disposition. He says, for in- 

 stance: — "Leaves pinnate of 2—3 approximate pairs and a sessile 

 terminal leaflet, silky above, more or less white-tomentose beneath", and 

 further "Petals obcordate, a third exeding the sepals". 



My specimens accord very well both with these descriptions and 

 with the figures quoted. The principal difference lies in the prevalence 

 of 3-digitate leaves, and in the reduction of the lower leaflets, if such 

 are developed. There is, however, a continuous transition between the 

 forms on the basis of which the original description is made (Lehmann's 

 material was collected at Great Bear Lake by Richardson, and that of Ryd- 

 berg still further to the south), and the most reduced arctic forms with 

 only 3-digitate leaves, such as are found on open gravel soil. Where P. 

 rubricaulis grew in rookeries, or other places with a richer soil, it al- 

 ways showed rudiments of at least one lower pair of leaflets, even if 

 these basal leaflets were very small. I have never seen really 5-digitate 

 leaves in it. In the Nat. Hist. Mus. there were several specimens col- 

 lected by Richardson, perhaps from the original locality, and at Kew 

 also from the arctic coast and from Mercy Bay, Banks Land, leg. 

 MiERTSCHiNG. The RiCHARDSON-specimens especially, were larger and 

 had larger and more deeply incised leaves than my Ellesmereland- 

 specimens, but they only had two pairs of leaflets (or even one only) 

 and the basal ones are at least somewhat smaller than the upper ones. 

 The difference between this form and the arctic one is consequently 

 not greater than might be expected. There may, however, be reason 

 for distinguishing the northern form as a separate variety: var. arciica, 

 n. var., characterised by low growth; mostly 3-digitate leaves, or by 

 very small basal leaflets, furnished only with one or a few feeble teeth. 



P. rubricaulis, and especially var. ardica, very much resemble P. 

 nivea in habit; there can also be said to exist certain points of like- 

 ness to P. pulchella. It might therefore perhaps, be suspected of being 

 a hybrid between the two last-named species, or perhaps that it had 

 been overlooked in Greenland and Spitsbergen : it might thus have been 

 the cause of the opinion quoted from Nathorst and Wulff (cf. under 

 P. pulchella). P. rubricaulis, however, is clearly distinct from P. pul- 

 chella in its much larger flowers, broader and darker petals, many- 

 flowered inflorescence, assurgent or erect stems, and in its broader, less 

 deeply incised leaflets. One fact among others which decidedly proclaims 

 against its being a hybrid is, that it is much more common within the 



