PLATE 426. 



EiiAdRosTis ArnnRSTONKT, Stapf. (Fl. Cap., Vol. VII., p. 007). 



Nat. Order GramlnesB. 



Perennial. — Culms erect, simple or with 1-2 branches from the upper nodes, 

 over 2 feet long, wiry, glabrous, smooth, about 3-noded, lowest lengthened, inter- 

 nodes about 4 foot long, like the upper long exserted, with a ring of more or less 

 distinct glands below the nodes ; sheaths tight, glabrous, smooth, substriate, inter- 

 mediate 1 to 1 J inch long ; ligule a fringe of minute hairs ; blades narrow, linear, 

 tapering to a fine point, 6 to 8 inches, by 1 to l^ line, more or less involute or 

 convolute, rather firm, scaberulous on the upper side, otherwise smooth. 



Panicle oblong, erect, lax, 5 to 6 inches, by 1 to l^ line ; axis filiform, 

 smooth ; branches whorled, the lower in whorls of 4 to 7, obliquely erect or slightly 

 spreading, loosely divided from near the base, longest i to 1 J inch long, capillary 

 or subcapillary, straw-coloured, smooth, bearded at the axils ; branchlets capillary, 

 subdivaricate, short, 3-1-spiculate, smooth or scaberulous; lateral pedicels about 1 

 line long. 



Spikelets scattered, lanceolate, 1 to 2 lines long, grey, 2 to 4-flowered ; 

 rhachilla subpersistent. 



Glumes subequal, lanceolate in profile, acute or acuminate, very delicate, 

 faintly 1-nerved, up to 1 line long, or the lower slightly shorter. Valves oblongs 

 subobtuse, | line long, thinly membranous, smooth or almost sp, side-nerves short, 

 faint. Pales equal to the valves, keels scaberulous and fine. Anthers | line long. 

 Grain pallid, translucent, oblong-ellipsoid, over ^ fine long. 



Edbitdt : Natal. Without precise locality, Buchanan 277. 



By an unfortunate error in the Flora Capensis, Natal is credited with the 

 grass Eragrostis Lehmanniana, which was really gathered in the Orange lliver 

 Colony, and is there quoted as Buchanan 267, the quotation : " Natal, without 

 precise locality, Buchanan 277 " should have come under E. Atherstonei. The 

 specimen figured from was Buchanan's 277, which is the only one in the Herbarium, 

 and we find that the culms are 2-3-branched, and the leaves f line wide, and only 

 4 inches long, the sheath cises also are long-bearded. Dr. Stapf, who has pointed 

 out these differences, is of opinion that, with the scanty material at our disposal, 

 the separation of the Natal plant as a variety is not at present advisable. 



Fig 1, Lower glume ; 2, upper glume ; 3, valve ; 4, pale ; 5, pistil, stamens and lodicules. 

 All rnlia-t/efl. 



