1893-1902. No. 2.] VASCULAR PLANTS OF ELLESMERELAND. 163 



florescentiae subattingeute, ligula elongata; panincula virginea dense 

 coarctata, spiculis subbifloris, glumis longe acuminatis, atroviolaceis, 

 paleis basi viridibus alboinarginatis, apice purpureo marginatis". 



It must also seem peculiar that the characters are always found 

 together which I deem most important, viz., nari'ow spikelets with long 

 pointed glumes and a long pointed or erose ligule. The culm is rather 

 short and stout, and the proliferous Ellesmereland specimens have 

 generally only one transformed flower. As no proliferous form of P. 

 glauca is previously described, however common this phenomenon may 

 otherwise be within the genus, I was rather doubtful at first whether 

 the plant could belong to it, and I was for a time very much inclined 

 to refer it to P. stenantha, Trinius, Gram. gen. et spec, the description 

 of which, such as it is rendered by Hooker, Flor. Bor. Amer., and 

 Gelert, 1. c, seemed to apply rather well to it. Moreover Kjellman, 

 As. Beringss. Fan., p. 558, presumes that the P. stenantha from Senja- 

 vin Bay mentioned by Ledebour, 1. c, might in fact be P. glauca. 

 Indeed I think it very probable that the proliferous form of the latter 

 species is found in the Bering Strait region, but the real P. stenantha, 

 which is also very often proliferous, is rather a different plant, as I 

 immediately found when, in London I had access to the original descrip- 

 tion of Trintos and saw a great many specimens. 



The original specimens of Lange's var. atroviolacea, to which I 

 think it best to refer my plant, are from Umanak in Danish Greenland, 

 collected by Rink, and also specimens brought home by J. Vahl from 

 the same tract fully agree with mine, except for the proliferous spikelets 

 of the latter. But even if there are specimens collected in other parts 

 of Greenland, by far the greater part of the specimens that are referred 

 to it in the Copenhagen herbarium do not really belong to it. Even 

 Lange himself, seems later on to have laid stress only upon the charac- 

 ter expressed in the name "atroviolacea". The smaller of the two 

 grasses figured in Suppl. Fl. Dan., T. 65 (P. stricta) may probably be 

 referred to the present variety. 



My specimens of var. atroviolacea f. prolifera as it may be called, 

 are from a rather limited area in the Goose Fjord, where it grew in 

 fields and slopes of stiff clay. 



Occurrence. F. ii/pica. Grinnell Land, Discovery Harbour (Hart !). 

 Hayes Sound region : mouth of Flagler Fjord, Twin Glacier Valley (Hart 

 867), Gape Rutherford (302), Fram Harbour (1105), Cocked Hat Is- 

 land (1262), Bedford Pirn Island (1256). South coast: Fram Fjord (1667); 

 Harbour Fjord, "green patch" (2157), slopes east of the anchorage 



