is separated by the innovation-shoots being intravaginal and not clothed with leafless scales, to say nothing of the broader leaves, dense panicle, and spikelets with the empty glumes scabrid on the keel.
6. A. parviflora, R. Br. Prodr. 170.—Culms laxly tufted, very slender, weak, often decumbent or prostrate at the base, erect or ascending above, quite smooth, 6–18 in. long. Leaves chiefly towards the base of the culms, the lowermost soon withering, 2–6 in. long, usually narrow and often almost filiform, but in luxuriant specimens broader and sometimes 112–110 in. diam., flaccid, flat or involute, smooth or the margins minutely scabrid; sheaths long, grooved, quite smooth; ligules long, membranous, lacerate. Panicle varying in length from 2 to 6 in. or more, compound, very lax and slender, drooping; primary branches long, capillary, scaberulous, erect at first but soon spreading, trichotomously divided, lowermost in clusters of 4–6, upper in distant pairs; secondary branches from above the middle, again divided; pedicels thickened at the tips. Spikelets very minute, about 115 in. long, shining, pale-green, sometimes tinged with purple. Two outer glumes slightly unequal, lanceolate, acute, membranous, slightly scabrid on the keel, margins hyaline; 3rd or flowering glume about ¼ shorter, broad, truncate, hyaline, delicately 5-nerved, awnless. Palea wanting.—Hook. f. Fl. Nov. Zel. i. 296; Handb. N.Z. Fl. 328. A. scabra, Benth. Fl. Austral. vii. 576 (not of Willd.).
North Island: Inland Patea and shores of Cook Strait, Colenso! South Island: Pelorus Valley, J. Macmahon! near Westport, Townson! near Dunedin, Petrie!
I am greatly puzzled with this species, which can be recognised without much difficulty by the weak habit, very slender lax spreading panicle, and minute spikelets, which are smaller than those of any other New Zealand species. It was originally referred to A. parviflora by Hooker in the Flora, but does not quite match the plate of tbat species given in the "Flora Tasmanica" (t. 158), nor any Australian specimens that I have seen. Bentham referred the Australian plant to A. scabra, Willd., a North American species; but that is a larger and more erect plant, with a more copiously divided panicle, and with narrower spikelets, much more scabrid on the keel. Professor Hackel, who has examined my specimens, says, "Not easy to name. Surely not A. scabra, Willd., but very near the North American A. perennans, Tuck. It is most probably A. parviflora, R. Br., but without seeing one of Brown's types I cannot be quite sure of the identity." It should be mentioned that most of the specimens referred to A. parviflora by New Zealand botanists are nothing but small states of A. Dyeri, Petrie, (the A. canina of the Handbook), as, for instance, the plant figured as A. parviflora by Buchanan in his New Zealand Grasses, t. 20c. All such specimens can be at once distinguished by the strict habit, contracted panicle, and larger spikelets.
7. A. tenella, Petrie in Trans. N.Z. Inst. xxii. (1890) 442.—Apparently annual. Culms laxly tufted, erect, very slender, quite smooth and glabrous, 3–4-noded, 6–15 in. high. Leaves few, much shorter than the culms, erect, very narrow, filiform or setaceous, involute, finely striate; sheaths rather long, close, smooth;
