Annuals and winter-annuals, mostly of 3-4 months duration, nearly always quite slender, sometimes diminutive and fugitive, varying from nearly glabrous below the thinly strigulose inflorescence to densely strigulose, villosulous, or pilose with appressed, incurved, or partly (rarely all) spreading, often lustrous hairs up to 0.4—1.35 mm. long, the herbage green, cinereous, or canescent, the leaflets either glabrous or pubescent above; stems (1) 2-30 (45) cm. long, 1-several (rarely over 15), when solitary erect, when 3 then the central one erect and the lateral ones incurved-ascending, or these freely branching from near the base and the branches either incurved or prostrately radiating in the form of loosely woven mats, either simple above the base or spurred (branched) at 1-3 nodes preceding the first peduncle, often floriferous from well below the middle; stipules (1) 1.5-6 (9) mm. long, the lowest ovate-triangular, early becoming papery, the upper ones triangular- or lance-acuminate to linear-caudate, mostly herbaceous, at least when young, all about semiamplexicaul, or the lowest fully so, but free; leaves 1-6.5, rarely 8 (9) cm. long, all slender-petioled or the upper ones often subsessile, with (5) 7-19 (23) leaflets varying from linear-elliptic and acute to obcordate, 2-14 (17) mm. long, nearly always flat, all of the same type throughout the plant or dimorphic, those of the lower leaves broader; peduncles slender or filiform (0.2) 0.5-9 (10) cm. long, erect or incurved, the early ones sometimes shorter than the later ones and only 1-2-flowered; racemes loosely but sometimes subcapitately (1) 2-10 (15), in one var. often 10-27-flowered, the axis not or not greatly elongating. 0-3 cm. (but rarely over 2 cm.) long in fruit, commonly produced beyond the last flower as a subulate appendage; bracts submembranous, ovate to lance-acuminate, 0.5-2.5 mm. long; pedicels at anthesis 0.4—1 mm., in fruit commonly arched outward, a trifle thickened, 0.5—1.6 mm. long; bracteoles 0 (exceptionally a minute scale); calyx 2.5-5.7 mm. long, varying from nearly glabrous to densely villosulous or (the teeth especially) pilose, the subsymmetrically campanulate or turbinate disc (0.4) 0.5—1.1 mm. deep, the tube 1.3—2.8 (3.1) mm. long, the teeth varying from subulate to narrowly lanceolate or linear- setaceous, 1-3.2 mm. long, the whole becoming papery, ruptured or not, marcescent; petals whitish, pinkish-lilac, pinkish- or amethystine-purple, when brighdy colored the banner with a large pale eye, the color turning bluish or violet on drying; banner recurved through 40—45°, ovate-, obovate-, or suborbicular-cuneate, shallowly notched, 3.7—10 (13) mm. long; wings 3.7—8.6 (10.7) mm. long, the blades oblanceolate to oblong-obovate, obtuse or emarginate, straight or slightly incurved; keel 3.7—6.8 (9.3) mm. long, the blades either half-obovate and bluntly deltoid or obliquely triangular to lunately half-ellipsoid, varying according to var. from very obtusely rounded to sharply triangular-acute and porrect; anthers 0.2-0.4 (0.6) mm. long: pod ascending, spreading, or declined, sessile or contracted at base into a short stipe or obscure stipelike neck up to 0.5—0.9 mm. long, linear or linear-oblanceolate in profile, nearly straight, incurved near the base and straight or nearly so thereafter, or less often evenly and gently incurved its whole length through ± ? or even ½ circle, when first formed triquetrously and a trifle laterally compressed, with nearly plane lateral and narrower, shallowly sulcate dorsal faces, the faces becoming flat or slightly convex as the ovules swell, subequally 3—sided or somewhat tetragonal when ripe, the thin, green or purple-tinged, glabrous, strigulose, or villosulous valves becoming papery, stramineous or brownish and ultimately black or nearly so, delicately cross-reticulate, inflexed as a complete septum up to about 2 mm. wide or the septum sometimes incomplete or subobsolete; dehiscence apical and downward through the ventral and part-way through the dorsal suture; ovules 10—18 (22); seeds brown, olivaceous, or ocher, often purple-dotted, smooth, pitted, or wrinkled, dull or sublustrous, 1.3-2.4 (2.7) mm. long, in the narrowest pods often distorted by crowding.
The small-flowered milk-vetch, A. Nuttallianus, embraces a considerable array of distinct forms which have been treated in the past as constituting a complex evidently polymorphic but susceptible or not, according to opinion, to subdivision into varieties, or alternatively as a series of small, closely related species. While maintaining a certain respect for the latter viewpoint, first formulated by Small and elaborated by Rydberg (1927, pp. 325-330; 1929, pp. 429-431), I here follow Gray (1864, p. 199) and Jones (1923, p. 269) in taking a wide and comprehensive view of the circumscription of the species, at the same time rejecting the contention of some contemporary botanists that the component forms are hopelessly confluent and beneath taxonomic notice.
The racial situation within A. Nuttallianus is very unusual, the existence of sympatric varieties which occupy the same environment being unique in American Astragalus. However, I must first mention the initial difficulties in attempting to organize the now very extensive material to be found in the larger herbaria. Most forms of the species are subject to a good deal of variation common in annual herbs everywhere, both within the limits of a population and from one colony to the next, due to accident of the soil, the season, and the weather. Thus plants growing in light porous soils or loose sands tend to be slender and erect, with a very slender root of short duration, while others nearby in hard-packed clay are comparatively robust, prostrate, and longer-lived. Then young plants flowering within a month to six weeks after germination of the seeds tend to have fewer leaflets to the leaf and few-flowered (often 1- flowered) racemes, and these juvenile characters tend to become perpetuated into the plant’s maturity, especially under arid conditions. The same types of variation occur independently in several forms distinguishable by other, sometimes less obviously apparent but more stable characters, and are likely to be mistaken for heritable when they are trivial and superficial in nature. Another difficulty arises from the fact that A. Nuttallianus finds a congenial home in sands and gravels along rights-of-way; and there is little doubt in my mind that the pristine patterns of dispersal have become disturbed and deranged during the past fifty years or so. As a result there are confusing records of some varieties occurring far from the known center of their abundance and sometimes within the territory properly belonging to a related form, giving the impression that no geographic rhyme or reason lies behind the polymorphic front. Apart from these exceptions A. Nuttallianus is ordinarily represented outside the state of Texas by only one variety in any one place. The vars. austrinus, imperfectus, cedrosensis, and micranthiformis are true vicariants, occupying more or less extensive areas of their own and intergrading along the common boundaries. They behave as do the component races of most multiracial astragali and there can be no question that they belong to one species. In western Texas the vars. macilentus and austrinus are at least partly sympatric, and in eastcentral Texas the former is joined by vars. trichocarpus, pleianthus, Nuttallianus, and the local var. zapatanus. It is astonishing that there is almost no intergradation, even though the vars. pleianthus and trichocarpus are morphlogically confluent when the whole range of variation is examined. In the herbarium I often find two different forms of small-flowered milk-vetch mounted together under one label, suggesting plasticity and random variation (which does occur in a few minor characters), but in the field I have learned to appreciate the uniformity among members of the average population or colony. In eastern Texas it is an easy matter to find two, and not rare to find three varieties of A. Nuttallianus growing together under identical conditions and in equal vigor on the same roadside bank or in the same fallow field, sometimes in separate patches but not uncommonly intermingled; yet the individual plants show no sign of hybridization and mass collections from such places can easily be sorted afterward into component categories. The question is bound to arise as to whether the so-called varieties have acquired as a by-product of the original mutation (or subsequently) the intersterility which is commonly accepted as good evidence of differentiation at the specific level. It would be rash to assume that this is so until a genetical study can be made of the whole complex, a study which will be required before a truly sound taxonomic treatment of the group can be had. Moreover, the nearly always small and often inconspicuously colored flowers of A. Nuttallianus are certainly capable, in most instances, of self-fertilization and in some forms are semicleistogamous, circumstances which must favor, if they do not actually ensure pure lines of inheritance. However this may turn out, the Texan material could logically be treated as three species, A. austrinus (including our vars. austrinus, pleianthus, trichocarpus, and zapatanus), the polymorphic A. macilentus (possibly further divisible), and A. Nuttallianus sens. strict.; or these entities, in Texas, are about equally distinct from one another as our var. Nuttallianus is from the nearly related and also sympatric A. leptocarpus. The characters useful in the taxonomy of the Texan forms are vesture of the leaves, calyx, and pod, shape of the leaflets, length of the fruiting raceme, size of the flowers, and shape of the keel-tip, are not of equal value in every case. When we extend the field of enquiry to the extra-Texan forms of the species and attempt to apply the same standards of comparison, we begin to encounter vexatious problems. The far western representatives of A. Nuttallianus differ one from another in similar ways to the Texan but, as already seen, they are vicariant and intergrade fully. Examining the whole complex, I find no sharp division between forms with acute and others with emarginate leaflets, with flowers of particular size-ranges, or with keel-tip at once very strongly incurved and bluntly rounded as opposed to porrect and triangular-acute. There seems to be no logical alternative to the comprehensive delimitation of A. Nuttallianus adopted below.
The centers of greatest complexity and population density of A. Nuttallianus coincide with the focus of speciation in subsect. Leptocarpi. It seems reasonable to suppose that the whole group is of Texan origin and the western forms of the small-flowered milk-vetch are of comparatively recent origin. Accordingly the varieties are presented in an order of modification starting with var. Nuttallianus and passing by small steps into the presumably derived vars. imperfectus and cedrosensis of the western deserts.
The small-flowered milk-vetch is the commonest annual astragalus with slender pods of the hamosoid type, which is native to the interior southwestern United States and Mexico. Apart from its close relatives in Texas, keyed out above, it is the only one in which the pod remains attached to the receptacle until dehiscence, by which time it has turned from green or purplish-green through stramineous to dark brown or almost black. It is not easy to distinguish all aspects of the species from flowering specimens of A. Emoryanus, but difficulty is likely to arise only with var. macilentus in which all leaflets are sometimes retuse and the ovary always glabrous. In the deserts of southern Nevada, southeastern California, and Baja California, A. Nuttallianus var. imperfectus and var. cedrosensis are sympatric and sometimes directly associated with A. acutirostris, a species formerly included within A. Nuttallianus which may be recognized, before the deciduous, commonly resupinate pod has formed, by the loose racemes and by leaflets obcordate or emarginate in all (even the uppermost) leaves. The rare A. nyensis, endemic to southern Nevada and also first described as a variety of A. Nuttallianus, is sympatric with var. imperfectus, but is more easily recognized at flowering time by its densely hirsutulous vesture and again by the emarginate leaflets.
It seems appropriate to mention here my indebtedness to the able Texan botanists Dr. Shinners (SMU), Dr. Turner (TEX) and Dr. Warnock (SRSC) for loans and gifts of pertinent material. Any merit which the present account of the A. Nuttallianus complex may possess must be attributed to the uniquely rich materials that they and their predecessors have assembled and that they so generously made available for leisured study.