76 H. G. SIMMONS. [sec. arct. exp. fram 



to Altai and Tibet, Arctic Russia, European mountains down to Spain 

 and Portugal, Great Britain, Faeroes, Iceland, Novaja Semlja, Spits- 

 bergen, Franz Joseph Land, Jan Mayen. 



Saxifraga rivularis, L. 



S. rivularis, Linnaeus, Sp. plant., 1753 ; Sternberg, Revis. Saxifr. ; Engler, Mon. 

 Saxifr. ; Lange, Consp. Fl. Groenl.; Krubse, List E. Greenl. ; Nathorst, N. W. 

 Grijnl. ; Hart, Bot. Br. Pol. Exp.; Geeely, Rep.; Hooker, Fi. Bor. Amer.; 

 Britton & Brown, 111. Fl.; Kjellman, in Vegaexp. ; Ledebour, Fl. Ross.; An- 

 dersson & Hesselman, Spetsb. karlv. ; Kruuse, Jan May. ; S. vaginata, Stern- 

 berg, 1. c, Suppl. II; S- Laurentiana, Seringe, in Decandolle, Prodr. ; S. 

 petiolaris, R. Brown, List of pi., et Chlor. Melv. 



Fig. Linnaeus, FI. Lapp., T. 2, fig. 7; Sternberg, 1. c, T. 12, fig. 4; Sv. 

 Bot., T. 729; FI. Dan., T. 118. 



Rather variable in size, shape of the leaves, hairiness, colour of the 

 flower, &c. I have not, however, been able clearly to distinguish any 

 of the forms, which Engler, 1. c, p. 105, has described, but I think 

 that some of my specimens from rather dry, mossy localities may be refer- 

 red to the var. purpurascens of Lange, 1. c, p. 62. They have red flowers, 

 but that seems also to be the case with specimens of Rob. RR0V\rN's 

 S. hyperborea, which is also a variety of the present species. Lange 

 has the var. hyperborea also as different from his own form, but surely 

 it is very difficult to draw the line between them. My specimens with 

 red flowers have generally a. single stem only, but Lange says about 

 var. purpurascens "dense pulvinato-caespitosa", which again holds true 

 also for some of the hyperborea specimens, at least, that I have seen 

 in the Nat. Hist. Mus. Of S. hyperborea there exists no figure, as it 

 is a mistake when Engler, 1. c, quotes Sternberg, Suppl. H, T. 15. 

 Some plants belonging to S. rivularis are indeed figured, but of S. 

 hyperborea no figure is given. Another pecuHar form, which, however, 

 is not sharply defined, was found in several localities (464, 640, 1103); 

 it is almost stalkless, very small, and generally densely tufted. The 

 plants designed with the names S. vaginata, Sternberg, Suppl. 1, p. 

 39 and T. 15, S. petiolaris, R. Brown, List of pi., and Chlor. Melv., 

 S. Laurentiana, Seringe in Decandolle, Prodr. IV, p. 35, are not to 

 be distinguished even as varieties. The name S. Chamissonis is also 

 used for S. rivularis by Sternberg, but only in labels (I have seen a 

 specimen in the Stockholm herbarium "ex insula Chamissonis, misit 

 Chamisso"), but later, he has transferred the name, altered to S. 

 Chamissoi, to another plant (cf. S. tricuspidata). 



S. rivularis is spread over the territories visited, but is not plenti- 

 ful. It grew generally in wet places, among moss or in clay soil, the 



