as it was collected at no other place. It doubtless is common along the Muddy. The region of the Muddy is a very barren desert, scarcely anything growing except along the streams or ditches. The water and the soil are quite salty and there are low hills of salt which border along the Muddy and the Virgin. The willow, however, seeks the stream banks and ditches, the little marshy places being too alkaline."

Salix Tweedyi (Bebb) n. comb. S. Barrattiana Tweedyi Bebb, Contrib. U. S. Nat. Herb. 3:572. 1896.—A shrub, with short stout divaricate branches; bark on the older twigs gray, somewhat shining, on younger twigs chestnut to deep reddish-brown, usually quite glabrous except the twigs of the season, which are commonly densely pubescent with spreading gray hairs: buds large, 6-10mm long, chestnut, thinly pubescent with long hairs, at length glabrate: leaves elliptical and acute at both ends, to oval, acute to somewhat obtuse at the apex, acute to rounded at the base, or subcordate in mature foliage, occasionally obovate-oval, rounded at the apex, especially when young, 5-7cm long, 2.5-4cm wide; glabrous on both sides or the young leaves sparsely pubescent above with long gray hairs, green above, paler but not glaucous beneath; margin finely and closely glandular-denticulate or serrate, the yellowish glands projecting at right angles to the apparently entire margins of the young leaves, in mature leaves the glands on the points of irregular, spinulose teeth, perpendicular or sometimes pointing toward the apex of the leaf; petioles 8-15mm long, frequently tomentose; stipules large, broadly reniform, somewhat clasping, 5-14mm long, glandular spinulose-denticulate: pistillate aments precocious, sessile, stout, lateral and terminal, spreading, often curved, 4-6cm long, 1.5-1.8cm wide in fruit: capsules green and glabrous, or sometimes finely pubescent near the apex, ovate-conical, pedicellate, 7-8mm long when mature; pedicel 1mm long, smooth; style 1.5-2.5mm long, green, stigmas divided, about 0.5mm long; scale obovate, acute or obtusish, 2-2.5mm long, black, clothed with long, straight, white hairs; gland about Imm long.-Plate XII, figs. 3-7.

WYOMING: head of Big Goose Creek, Big Horn Mountains, nos. 11 (the type) and 12, Frank Tweedy, July 15-24, 1893; "along streams in Teton Basin, July. With ovaries nearly smooth! Professor Porter." J. M. Coulter, 1872.

This last specimen is listed by COULTER' under the name of S. Barrattiana. In the U. S. National Herbarium the Hayden Survey specimen is ticketed "Trail R. Mts., Idaho Terr." Trail Creek, along which the party worked, is, however, in southwestern Montana, not far north of the Yellowstone Park. The sheet carries pistillate aments and both young and mature foliage. Though collected twenty-one years earlier than the type collection, it is not mentioned by BEBB, and was probably unknown to him.

The variety was named and described by BEBB in an article by J. N. ROSE, 1 Hayden, Ann. Rept. U. S. Geol. Surv. Terr. 6: 781.

based on a small collection of plants made by FRANK TWEEDY, of the U. S. Geological Survey. His collections were made in the Big Horn Mountains, "between longitude 107° and 107° 30′ and latitude 44° 30′ and 44° 40′, chiefly on the headwaters of the east and west forks of Big Goose Creek and on the high divide between these streams and Shell Creek."

The original description and the accompanying data are as follows:

"Leaves at first thinly overspread on the upper surface with floccose hairs, soon smooth and green both sides; capsules glabrous. In bogs and along mountain streams, altitude 2,460 to 3,080m (8,000 to 10,000 ft). Head of Big Goose Creek, Big Horn Mountains, July 15 (nos. 11 and 12).”

BEBB, after noting the original and few subsequent collections of the rare Salix Barrattiana and remarking on its silky leaves and capsules, continues:

"That there should be more or less variation in this vesture was to be expected from what is known of the two most nearly allied species. S. Hookeriana was first described as having 'very smooth' capsules, but subsequent observations have shown that they are more frequently tomentose, and a like variation, though in less degree, prevails in the case of S. Richardsoni; but in neither is this variation so pronounced as to lessen the surprise with which we find the one species of the group heretofore most conspicuous for its silky vesture, appearing, as in Mr. TWEEDY'S specimens, so markedly glabrate. The leaves, and in fact the aments as well, bear a very close and deceptive resemblance to some forms of S. Barclayi; but the aments are closely sessile, terminal as well as lateral, the styles long r and the stigmas bifid; the leaves alone could scarcely be distinguished one from the other."

A study of the type material and of later collections leaves no doubt that this willow is specifically distinct from S. Barrattiana. Not only do the nearly glabrous leaves and the glabrous capsules serve to distinguish it, but the leaf margin thickly set with conspicuous glands is a marked character.

The varietal name proposed by BEBB was "denudata," but as this was preoccupied by S. commutata denudata Bebb (1888) ROSE substituted the name "Tweedyi" and credited it to BEBB, whose name appears in the text as the author. COCKERELL has since contended that ROSE and not BEBB must be considered the author of the variety. This view, however, cannot be accepted.

Salix Wolfii Idahoensis, n. var.-Leaves usually densely silky-villous with shining hairs, giving a decided luster to the surface, scarcely blackening in drying: capsules 3-4mm long, thinly but permanently silky pubescent; style about 1mm long.

WYOMING: Mammoth Hot Springs, Yellowstone National Park, no. 5655, Aven and Elias Nelson, July 3, 1899; Swan Lake Valley, Yellowstone National Park, alt. 7300ft (2250m), F. H. Knowlton, July 9, 1888.

MONTANA: Madison Cañon, J. M. Coulter, Hayden Survey, 1872. The specimen so labelled in the US. National Herbarium (accession no. 253,723) bears on the label, "Madison Cañon, Idaho Ter.," which must certainly be an error, since the 2 BOT. GAZETTE 22:268. 1896.

Madison Cañon explored by the Hayden Survey in 1872 is in Montana, not far from the boundary of the Yellowstone Park. This specimen was probably listed under the name of Salix glauca L., which is credited by COULTER in his list 3 to the "Upper cañon of the Madison." S. Wolfii is, of course, not mentioned, because it was not published until six years later, in 1878.

IDAHO: Forks of Wood River, alt. 6000 (1800m), no. 3399 (type), L. F. Henderson, July 25, 1895.

OREGON: Banks of Wallowa River, mouth of Hurricane Creek, no. 2400, W. C. Cusick, June 11, 1900.

The relationships of this little willow are all with S. Wolfii Bebb, from the typical form of which it differs mainly in the silky-pubescent capsule and the rather more silky leaves. The different specimens show considerable variation in the amount of pubescence on the capsules. In general, the specimens from the more northwestern localities show the denser pubescence, both on leaves and capsules. In some cases they tend toward a glabrate condition in age. In S. Wolfii the capsules are glabrous even when young. This variety is found from the Yellowstone Park northwestward across Montana and Idaho to eastern Oregon. The species is found from central Colorado to Montana and Idaho.

Salix Nelsoni, n. sp.-A shrub, 1.2-3m high, freely branching, very leafy; twigs shining, the older gray with a reddish tinge, the younger bright chestnut or darker, and often drying dark: buds large, smooth, chestnut, 7-9mm long, beaked: leaves petioled, petioles 3-7mm long, smooth; stipules none; blades oblanceolate or rarely narrowly lanceolate, acute at both ends, cuneate at the base, 8-15mm wide, 3-5.5cm long, entire (except as noted below), the margin somewhat involute, especially toward the base; glabrous, or thinly silky villous when young, green and very shining above, pale and somewhat glaucous below; primary veins prominent, usually elevated and rather greenish-yellow on the upper surface, prominently reticulate below by the elevation of both primaries and secondaries;4 apical and subapical leaves on sterile shoots narrowly elliptical, the margin shallowly glandular crenate-serrate: pistillate aments sessile with sometimes a few small foliaceous bracts at the base, spreading, slender and cylindrical when young, very much thickened in fruit, 1.5-3cm long: capsules silky pubescent, sessile, 5-6mm long; styles entire or rarely bifid, 0.3-0.5mm long; stigmas smooth, 0.7-1.0mm long, dark; scales dark, ovate, acute, clothed on both sides with long white hairs: staminate aments not seen.-Plate XIII, figs. 8–11.

COLORADO: Little Fountain Creek (south of Pike's Peak), alt. 9200 ft, no. 12, J. C. Blumer, September 5, 1903; Mt. Lincoln, Park Co., alt. 12,000 ", J. M. Coulter, 3 Ann. Rept. U. S. Geol. Surv. Terr. 6:782. 1873. ·

4 In the type the primaries and sometimes also the secondaries on the upper surface are very broad and deep green in color, giving a strikingly beautiful effect, which is not so evident in other specimens.

July 9, 1873; Beaver Creek, Larimer Co., roadsides in pine woods, no. 1440, L. N. Goodding, July 4, 1903.

WYOMING: Centennial Valley, no. 1754, Aven Nelson, August 18, 1895; Centennial, Albany Co., bogs, no. 8822, Aven Nelson, August 7, 1902; Laramie Peak, Albany Co., along the creek, in clumps, 4-10ft high, no. 7580 (type), Aven Nelson, July 13, 1900; head of Big Goose Creek, Big Horn Mountains, no. 47, Frank Tweedy, July 15-24, 1893.

ALBERTA: Banff, low ground, side of road to Devil's Head Lake, alt. 4500", no. 2253, W. C. McCalla, June 8 and August, 1899.

This beautiful willow is found at high altitudes in the mountains of Colorado and northward. It is most closely related to S. chlorophylla Anderss. and the pistillate aments closely resemble those of that species. It is readily distinguished by the oblanceolate leaves which, when mature, are prominently nerved above and reticulated beneath. Authentic staminate material has not been seen.

I take pleasure in dedicating this species to Professor AVEN NELSON, of the Rocky Mountain Herbarium, who has made available so much valuable material in this difficult genus.-CARLETON R. BALL, U. S. Department of Agriculture, Washington, D. C.

EXPLANATION OF PLATES XII AND XIII.

PLATE XII.

FIGS. 1 and 2. S. Gooddingi Ball; fig. 1, fruiting branch, natural size; fig. 2, capsule and scale. X 10.

FIGS. 3-7. S. Tweed yi (Bebb) Ball. Fig. 3, sterile twig; fig. 4, pistillate ament; fig. 5, mature leaf showing stipule and bud; fig. 6, capsule (×10); fig. 7, outline of margin of young leaf showing glands. XI0.

PLATE XIII.

FIGS. 8-11. S. Nelsoni Ball; Fig. 8, Sterile twig, from type; fig. 9, fruiting branch; fig. 10, elliptical apical leaf; fig. 11, capsule, X 10.