Loosely or densely tufted, perennial, with a woody taproot and knotty root-crown or shortly forking caudex, strigulose with appressed hairs up to 0.2-0.4 mm. long, the herbage greenish, cinereous, or subcanescent, the leaflets glabrous or nearly so above; stems several or numerous, prostrate or decumbent and incurved- ascending, (0.3) 1-3.5 (4) dm. long, simple or branched at 1-3 nodes preceding the first peduncle, floriferous from near or above the middle; stipules ± dimorphic, the lowest early becoming papery, fully amplexicaul, free or obscurely united opposite the petiole, the upper ones semiamplexicaul, herbaceous, with deltoid or lanceolate, commonly recurved blades 1.5-4.5 mm. long; leaves shortly petioled or the uppermost subsessile, (1) 1.5-6 (9) cm. long, with (9) 13-25 oblanceolate to linear-oblanceolate, truncate, obtuse, or obtuse and mucronulate, flat or folded leaflets (2) 4-13 mm. long; peduncles incurved-ascending at anthesis, prostrate in fruit, (2.5) 4-11.5 cm. long; racemes at first rather dense, mostly 10-20-flow- ered, the flowers early spreading and declined, the axis little elongating, (1) 1.5-6 cm. long in fruit; bracts submembranous, ovate or lanceolate, 1-3 mm. long; pedicels ascending, at anthesis 0.4-1.2 mm., in fruit thickened, persistent, 0.7-2.8 mm. long; bracteoles commonly 2, often minute; calyx 4-5.7 mm. long, strigulose with mostly dark, rarely all or mostly white hairs, the somewhat oblique disc 0.5-0.8 mm. deep, the tube 2.5-3.3 mm. long, 2-2.3 (2.5) mm. in diameter, the subulate teeth (1.3) 1.5-2.8 mm. long; petals purple, the color sometimes fugitive, or livid greenish-purple; banner recurved through 45-75°, ovate- or obovate-cuneate to broadly rhombic, shallowly notched, 7-9.6 mm. long, 3.6-5.8 mm. wide; wings nearly as long or up to 1.2 mm. longer, 7.2-9.8 mm. long, the claws 2.3-3.3 mm., the narrowly oblong or oblanceolate to nearly Unear, obtuse or emarginate, straight or gently incurved blades 4.7—7.1 mm. long, 1.5—2.5 mm. wide; keel 5.1—6.8 mm. long, the claws 2.3-3.4 mm., the half-obovate blades 3-3.7 (4) mm. long, 1.8-2.5 mm. wide, incurved through 90—100° to the bluntly deltoid apex; anthers 0.4—0.6 mm. long; pod ascending, spreading, or (when crowded) declined, sessile on the conic receptacle, plumply oblong-ellipsoid to subglobose or sometimes a little broader than long and then transversely oblong, inflated or strongly tumescent but firm and scarcely bladdery, retuse at both ends, bearing a minute subulate beak in the distal sinus, sulcate along both sutures, 5-9 mm. long, 5.5-7.5 mm. in diameter, a trifle obcompressed but essentially didymous, the green but red-cheeked, thinly strigulose valves becoming stiffly papery, stramineous or ultimately blackish, prominently and closely cross-reticulate, inflexed as a complete septum ± 2-3 mm. wide; seeds olivaceous or dark brown, sometimes purple-speckled, smooth but scarcely lustrous, 1.8—2.1 mm. long.—Collections: 12 (iii); representative: Lundell 5123 (ARIZ, K, TEX, US); Pringle 1753 (G, K, ND, NY, P, PH, POM, fragm., US); Palmer 394 in 1906 (GH, NY); Waterfall 15,559, 15,565 (NY); Ripley & Barneby 13,460 (CAS, MEXU, NY, RSA, US).
Plains and stony hillsides, in arid grassland or in piñon-juniper woodland, 5600—8500 feet, locally plentiful, southern Durango to northcentral San Luis Potosi and southern Zacatecas.—Map No. 159.—July to November.
Astragalus diphacus (with pod formed of twin bladders) Wats. in Proc. Amer. Acad. 17: 342. 1882. In the San Miguelito Mountains (816 Schaffner, mainly)."—Holotypus, collected in 1876, GH! isotypi, K, NY (dated "1879" and numbered 612/816), PH!—Hesperastragalus diphacus (Wats.) Rydb. in Bull. Torr. Club 53: 166. 1926.
Astragalus diphacus var. peonis (of the peasant laborer) Jones, Contrib. West. Bot. 10: 65. 1902.—"Zacatecas, Mexico, Jones, [May 16,] 1892."—Holotypus, POM!—Hesperastragalus peonis (Jones) Rydb. in N. Amer. Fl. 24 : 457. 1929. Astragalus peonis (Jones) L. E. James in Contrib. Dudl. Herb. 4: 66. 1951.
The twin pod milk-vetch, A. diphacus, is a lowly astragalus with narrow leaflets, small flowers, and no special individuality when in bloom; but the pod is highly modified and remarkable. Like that of A. crassicarpus on a greatly reduced scale, it varies from globose to plumply ovoid-oblong in dorsiventral profile or is so much shortened as to become broader than long and thus transversely oblong. Being recessed at both ends and bearing a minute, conic beak in the deeper, distal sinus, it might be likened to the silicle of some species of Physaria Gray; but although the cavity is much larger than needed to accommodate the small seeds, the valve-walls are of such firmly chartaceous texture that the pod can hardly be described as bladdery. The fruit resembles that of some Diphysi in form and texture but lacks a well-defined, laterally flattened beak and differs in the prominent transverse reticulation of the valves.
The identity of the so-called A. peonis, here reduced to A. diphacus, will remain controversial until better material showing flowers and fruits in organic association is acquired, for without it the immature and fragmentary typus cannot be evaluated with assurance. Rydberg (1929, l.c.) distinguished Hesperastragalus peonis from H. diphacus by canescent foliage and yellow petals contrasted with green foliage and white petals; the characters are poorly chosen. There is no difference in the type of pubescence, only in the density of it, and the petals of A. diphacus are normally purple, not white. It was the petals of H. peonis that were supposedly "white" (according to the protolog), although Jones’s field note which accompanies the typus has them "yellow" (probably ochroleucous when fresh). James (1950, 1951) introduced a new specific character, having found five vascular bundles in the petiole of A. peonis and only three in that of A. diphacus. In 1961 when I examined the typus of var. peonis, it consisted of three stems cut off above the base, two of them each bearing one raceme of small, ochroleucous flowers and one a raceme bearing two very young forming pods. I do not hesitate to associate this latter piece with a fruiting collection (Ripley & Barneby 13,452, NY, RSA) of what I interpret as a condensed phase of A. diphacus, common (but in November past flowering) on stony hills surrounding the city of Zacatecas at elevations up to 8500 feet. In this singularly bleak and wind-swept habitat, the plants are densely cespitose, with short, prostrate stems and neat, silvery foliage; the pod is not only identical when young with that of the original var. peonis, but also identical when ripe with that of typical, greenish, and loosely tufted, purple- flowered A. diphacus such as Pringle (No. 1753, cited supra) collected on "hills at Zacatecas" in late October, 1888. It would appear, therefore, that if there are really two distinct astragali of the A. diphacus type, they are virtually sympatric and differ only in color (and a trifle in size) of the corolla. The question of petal-color may be altogether irrelevant, however. Very possibly the flowering stems on the type-sheet of var. peonis are not conspecific with the fruiting stem, but represent some other small- and genuinely ochroleucous-flowered astragalus (perhaps A. hypoleucus) native to the Zacatecas hills. If such is the case, it may be conjectured that it was this species, and not the fruiting stem, which contributed the five vascular bundles noted by James. I hesitate to suggest that so alert and experienced an observer as Jones could have been deceived by a mixture of material, but as I have noted a similar case involving the sympatric A. amphioxys and A. cymboides on the Colorado Plateau, an explanation of "A. peonis"" in such terms is far from impossible.