Astragalus trichopodus

Commonly robust, sometimes relatively slender, erect or less often diffuse, bushy-branched perennial, with thick, woody taproot and at length knotty trunk, suffruticulose caudex, or at least basally indurated stems, finely strigulose or strigulose-villosulous with appressed and sometimes a few or many narrowly ascending, straight or largely straight hairs up to 0.3-0.7 mm. long, the stems commonly canescent distally, sometimes throughout, the herbage greenish, cinereous, or canescent, rarely (in age, or in dry situations) white-tomentulose, the leaflets either glabrous or pubescent above; stems several or very numerous, erect and ascending or (especially when hanging out from steep banks or bluffs) assurgent and trailing, (2) 2.5-8 (10) dm., rarely dwarfed and only 0.5-2 dm. long, branched or spurred at all or most nodes preceding the first peduncle, striate and usually hollow proximally, together forming handsome leafy and floriferous clumps; stipules thinly herbaceous becoming papery, fragile and (especially the lowest) deciduous in age leaving a scar or low collar on the stem, broadly deltoid-acuminate to lanceolate or lance-acuminate, (2) 3-7 mm. long, usually decurrent around half or less of the stem s circumference (exceptionally the lowest connate; cf. an insular form of var. lonchus discussed below), the blades commonly deflexed; leaves (2.5) 5-16 (20) cm. long, the lowest shortly petioled, the rest subsessile, with (15) 21-35 (39) ovate- or lance-oblong, oblong-elliptic or -oblanceolate, rarely (upward) linear- oblong or (downward) ovate- or obovate-cuneate, obtuse, obtuse and mucronulate, or emarginate, flat or loosely folded leaflets (2) 5-20 (25) cm. long; peduncles erect or incurved-ascending, (4) 6-26 (30) cm. long, all but the shortest upper ones surpassing the leaf; racemes loosely or a first rather densely (10) 15-50- flowered, the flowers early spreading and nodding at full anthesis, the axis nearly always elongating, (2) 3.5-13 (16) cm. long in fruit; bracts submembranous, ovate or lanceolate, 1-3.3 mm. long; pedicels at first ascending, straight, early arched out- or downward or divaricate in age, at anthesis 1.2-2.6 (4) mm., in fruit thickened, 1.4-5.2 mm. long, persistent; bracteoles commonly 2, sometimes rudimentary or 0; calyx 5-8.7 mm. long, thinly or sometimes only minutely strigulose or strigulose-villosulous with white, sometimes mixed with fuscous or black hairs, the oblique disc (0.8) 1-1.5 mm. deep, the pallid, campanulate tube 3.6-5.4 mm. long, (2.2) 2.5-4.2 (4.8) mm. in diameter, obliquely turbinate, rounded, or a little distended dorsally (but scarcely gibbous) at base, the subulate teeth (0.9) 1.1-3.7 mm. long, the orifice often oblique, the whole becoming papery, marcescent unruptured; petals greenish-white or cream-colored, rarely yellowish or white veined with pink-purple, the keel nearly always concolorous, the tip sometimes faintly maculate, the wing- and keel-claws exserted from the tube; banner recurved through 40-45° (or further in withering), rhombic-ovate, -elliptic, or -oblanceolate, shallowly notched, 11.3-19 mm. long, (5.4) 5.7-10 mm. wide; wings (10.1) 10.4-16.1 nun. long, the claws (4.6) 5-7.9 mm., the linear or narrowly oblanceolate, obtuse, erose, or rarely emarginate, straight or slightly incurved blades 6.3-9.7 mm. long, 1.6-3.2 mm. wide; keel 8.6—13.3 mm. long, the claws 4.5-7.8 mm., the obliquely triangular or broadly half-obovate blades 4.6—6.3 mm. long, (2.2) 2.4—3.4 mm, wide, abruptly incurved through 85—95° to the bluntly deltoid or sometimes subporrect and then sharply deltoid apex; anthers 0.5-0.8 (0.9) mm. long; pod pendulous, stipitate, continuous with the receptacle, the filiform, flexible, straight or downwardly arched, strigulose-villosulous, green or often brownish-purple stipe 5—17 mm. long, the body varying (according to variety) from linear-elliptic to broadly and plumply ovate or half-ovate in profile, laterally compressed when narrow, otherwise more or less greatly inflated, often bladdery, (1.3) 1.5-4 (4.5) cm. long, 4.8-21 mm. in diameter, the thin, pale green or purplish- tinged, glabrous or finely strigulose valves becoming papery, stramineous, lustrous, finely cross-reticulate, not inflexed, the funicular flange 0—0.6 mm. wide; dehiscence apical and downward through the ventral suture, sometimes its whole length and the valves then explanate in age; ovules (10) 12—30; seeds brown, dark chocolate-brown, or greenish, smooth, pitted, or wrinkled, dull, (1.8) 2—3.2 mm. long.

The complex of forms assembled here under the specific heading of A. trichopodus has traditionally been treated, if we ignore for the moment some inconsiderable segregates, as three distinct species, A. trichopodus (sens. strict.), A. leucopsis, and A. Antiselli, equivalent to our vars. trichopodus, lonchus, and phoxus respectively. Jepson at one time (1925, p. 572) reduced A. Antiselli to varietal rank, but in his summary Flora he again separated them. The wealth of material now available permits an assessment of the essential differences which prove to be much less numerous than supposed, existing only in the fruit. It is true that the hairs on the stems and leaves of var. lonchus are more often, in whole or part, of a more sinuous type than has been observed in the other varieties, but the vesture is quite variable, and some specimens of var. lonchus and var. phoxus are indistinguishable on this basis. Turning to the pod, we find an almost uninterrupted series of variants from the narrow, commonly symmetric, laterally flattened one characteristic of var. phoxus into the broadly and plumply inflated, strongly asymmetric one normal to var. lonchus, with the moderately inflated, either symmetric or asymmetric pod of var. trichopodus neatly intermediate between the extremes. The nearest approach to discontinuity in the series falls between the tumid pod of var. trichopodus and the flat, two-sided, and sharply bicarinate pod of var. phoxus; but the coastal form of the latter, which has been called A. gaviotus and has unusually broad and sometimes subtumescent fruits, very closely simulates the narrowest and least inflated states of var. trichopodus.

The geographic dispersal of the three varieties deserves special notice, because of the bearing it may have on the taxonomic situation. Typical var. trichopodus is comparatively rare, much more restricted and discontinuous in range than the other two varieties. In the first place it occupies a very short stretch of coastal territory in northern Ventura and southeastern Santa Barbara Counties lying precisely between (but not overlapping) the northernmost mainland outpost at Point Mugu of var. lonchus and the first appearance of var. phoxus on the coast a few miles west of Santa Barbara. The second and, so far as number of colonies and individuals are concerned, the most important area for var. trichopodus embraces the low Puente and Chino Hills, a region less than twenty miles in greatest diameter and exactly interposed between the southeastern outposts of var. phoxus in the foothills of the Verdugo Hills and the Pacific slope of the Santa Ana Mountains. It is precisely here that var. lonchus deserts the immediate shoreline and climbs a little inland along canyons running to the ocean. The third station for var. trichopodus is Catalina Island, where it is known only from the bay of Avalon; var. lonchus is found elsewhere on the island in wilder spots. Possibly var. trichopodus is introduced there. A glance at Map No. 112 will show the continuous range of var. lonchus from the Santa Maria Plains in Baja California north along the coast to Point Mugu and the Channel Islands, and that of var. phoxus, mostly back of the shoreline, from San Fernando Valley northwest to central San Luis Obispo County, with a few odd coastal stations between Santa Barbara and Point Conception. The position of var. trichopodus, intermediate morphologically in the pod and geographically (except on Catalina Island, where perhaps irrelevant) between var. lonchus and var. phoxus, suggests two hypotheses. The intermediate form might be interpreted as having arisen, possibly independently in its two main areas, through hybridization between two distinct species, A. Antiselli to the north and A. leucopsis to the south; or alternatively as a somewhat artificial and arbitrary selection of specimens taken out for convenience from too long a series of intergradient forms. Before considering the merits of either hypothesis it is necessary to introduce a fourth species, A. filipes, which may have been involved in the formation of the A. trichopodus complex.

Predominantly a species of the sagebrush deserts of the Great and Columbia Basins, A. filipes reappears strangely and disjunctly in the interior mountains of southern California and extreme northern Baja California, where its few scattered stations indicate a relictual occurrence. It is easily distinguished from A. trichopodus by its subterranean root-crown and connate lower stipules; but the pod, detached from the plant, is almost precisely that of var. phoxus in form even though it dehisces (as does that of related intermountain species) first from the base and through the stipe, not apically (as in all forms of A. trichopodus). While var. phoxus is characteristic of the chaparral belt below 3000 feet altitude, it extends occasionally, as along Piru and Sespe Creeks in Ventura County, upward to the 4000 feet contour where it very nearly approaches some outlying colonies of var. filipes on the flanks of Mount Pinos. The possibility of pollen exchange in this area cannot be overlooked; moreover, in view of its present disjunct dispersal, A. filipes, must have ranged more widely in California in the past, so other opportunities may have arisen in former times. Such hypothesis would go far to explain the astonishing similarity in the pods of vicariant species which do not appear otherwise at all closely related.

With extreme regret I am obliged to revive, at the varietal level, and at the expense of the familiar epithets leucopsis and Antiselli, two names which are not only odd in form but based originally on minor variants of no distinction. The argument for adopting the category of subspecies in place of the classical varietas here assumes a certain force.