Variable in stature according to race, site, and season, usually quite slender, with slender or subfiliform taproot and 1—3 erect, ascending, decumbent and distally incurved, or truly prostrate primary stems 2.5-3.5 cm. long, either simple or diffusely branched near the base, thinly to quite densely strigulose or strigose-hirsutulous with straight or nearly straight, appressed or loosely ascending hairs up to (0.3) 0.4—0.75 mm. long, the herbage green, cinereous, or canescent, the leaflets pubescent on both sides or more commonly either glabrous or medially glabrescent above, the inflorescence densely villosulous or pilosulous; stipules thinly herbaceous becoming pallid and papery-membranous in age, (0.6) 1-5 mm. long, the short lower ones deltoid or ovate, the rest ovate-triangular to lance-acuminate, with spreading blades, all semi- or the lowest a little more than semiamplexicaul, glabrous or nearly so dorsally, ciliate, and the margins sometimes beset with a few minute, tack-shaped processes; leaves 0.8-7.5 cm. long, the lower ones slender- petioled, the rest very shortly so or subsessile, with (9) 11-17 linear, linear-oblong, -elliptic, oblanceolate, oblong-cuneate, or exactly cuneate, deeply retuse or (in some lower leaves) shorter and obcordate, flat or loosely folded leaflets 1.5-14 mm. long; peduncles slender, erect or incurved-ascending, (0.5) 1-8 cm. long, all shorter, all longer, or the lower ones shorter and the upper longer than the leaf; racemes densely (5) 7-30-flowered, the erect and ascending flowers crowded into ovoid, subglobose, or oblong-cylindric heads at full anthesis 6-15 mm. in diameter, the axis not or scarcely elongating, 2-20 (22) mm. long in fruit; bracts membranous, pallid, ovate- or lance-acuminate, 0.6-3 mm. long; pedicels erect, often very short or subobsolete, at anthesis 0.1-0.4 mm., in fruit 0.2-0.5 (0.6) mm. long, naturally persistent but not thickened and readily disjointing under pressure; bracteoles 0; calyx 3-5.4 mm. long, densely pilosulous or villosulous (villous) with spreading or loosely ascending, black, white, or mixed black and white hairs up to 0.55-1 (1.1) mm. long, the subsymmetric, campanulate or turbinate disc 0.3-0.5 (0.6) mm. deep, the campanulate or at length ovoid-campanulate, pallid, submembranous tube (1.6) 1.7-3.2 mm. long, the subulate or subulate-setaceous teeth varying from much shorter to a little longer, the whole becoming papery-scarious, at first distended but finally ruptured and marcescent about the lower half, or fully investing the whole pod; petals whitish tinged with lavender, or pink-purple with white or pallid wing-tips, the purple turning bluish when dry; banner recurved through ± 25°, oblanceolate, spatulate- or rhombic-oblanceolate, or obovate- cuneate, shallowly notched, about 4-9.5 mm. long; wings a little shorter, the blades linear-oblanceolate, narrowly obovate, or rarely lanceolate, obtuse, straight or slightly incurved; keel 3-7.2 mm. long, variable in shape; anthers 0.15-0.4 mm. long; pod erect or (due to crowding) spreading, sessile, plumply ovoid or subglobose, 2-4 mm. long, a little laterally compressed, carinate ventrally by the prominent, thick, straight or low-convex suture, gibbous-convex and deeply sulcate dorsally, the sulcus either open and narrow, or closed by the connivent lobes of the fruit, the thinly fleshy, green or purple-tinged, densely to quite thinly strigose-hir- sutulous or rarely glabrous valves becoming stiffly papery, brownish, transversely reticulate by 4—6 (7) prominent nerves ascending from the ventral suture and nearly always elevated at maturity into prominent, sharp- or blunt-edged corrugations, the obtuse lateral angles thus strongly rugulose, inflexed as a complete or sometimes incomplete septum 0.2—0.5 (0.8) mm. wide; seeds olivaceous, ochraceous, orange, or brown, more or less rugulose, dull or somewhat lustrous, 1.4-2.3 mm. long.
Although much more variable and polymorphic than A. Gambelianus, the two-seeded milk- vetch, A. didymocarpus, is less so than might be inferred from the number of segregates which have been proposed or from the synonymy listed in the following pages. It cannot be overemphasized that the pod remains essentially uniform throughout the species, whereas the differential characters of the varieties, which involve only slight modifications in flower-size, proportions of the calyx-tube and -teeth, shape of the keel-tip, habit of growth, and distribution or density of the vesture, only assume importance insofar as they are correlated with patterns of dispersal. I refer here to the macroscopic morphological features of the plants, those visible to the naked eye or under magnification of ten diameters. After experimental study and comparison of the anatomy and embryology of the Didymocarpi, James (1950, 1951, passim) recognized two species within the A. didymocarpus of these pages, separable in the following terms:
1. Early trifoliolate leaves usually 3 in number, rarely 2 or 1; vascular bundles in the petioles 5; n = 13 A. dispermus
1. Early trifoliolate leaves 0 to 2; vascular bundles 3; n = 12 A. didymocarpus
Since James’s papers appeared in print, no opportunity has arisen to reassemble the amount of material necessary for a systematic appraisal of the vascular anatomy of the petiole; but it seems to hold out little hope of clarifying the taxonomy if the coastal A. Milesianus, which so closely resembles A. didymocarpus in all but flower-size, is to be referred to A. dispermus because of its five bundles. The early trifoliolate leaves which immediately succeed the cotyledons wither quickly and are lacking from most herbarium specimens. Before any satisfactory division of the complex can be attained, it will be necessary to have chromosome counts of var. obispensis and var. Milesianus, as well as further data about populations which seem, on morphological grounds, to be intermediate between A. didymocarpus and A. dispermus. While James’s observations are of the greatest interest and point up the immensity of our ignorance of the finer points of anatomy in Astragalus, I must leave the assessment of her results to the future. And if no better key character can be found to distinguish these two supposed species than the color of the calyx-hairs (James, 1951, p. 65), which would lead to the most contradictory results in practice, the comprehensive view of a pluriracial A. didymocarpus presented here will probably find favor with the majority of botanists.