Senna pallida var. geminiflora

Weak shrubs with terete albescent hornotinous and very slender pliant annotinous branchlets, at anthesis 0.5-2 m, pilosulous throughout or the ample, thin- textured lfts pilosulous beneath only and sometimes there only along midrib, the 2-fld racemes all axillary to coeval lvs and shorter than them, rarely becoming paniculate distally late in season.

Stipules setiform or narrowly linear-attenuate 2-6.5 x 0.1-0.3 mm.

Major lvs 7-13 cm; petiole (12-) 15-30 mm, ±1.5-2.5 times as long as first segment of rachis; rachis 2.5-7(-8) cm; gland 1 at or (adnate and apparently) above first pair of lfts 2-3.5 mm, the stipe often pubescent, the body 0.3-0.7 mm wide; lfts (3-)4-7 pairs, the distal ones elliptic, elliptic-oblanceolate or obovate, deltate- to triangular-acute or obtusely rounded at apex, (27-)30- 44(-50) x (9-) 10-19(-21) mm, 1.9-3.1 times as long as wide, the midrib with 7-11 pairs of major camptodrome and some intercalary secondary veins prominulous on both faces and giving rise on both to a finely prominulous 3- and 4-nary mesh of veinlets.

Peduncles 8-18(-27) mm; pedicels (12-) 16-25 mm, subtended by usual glands; long sepals 7-10 mm; long petals 17-28 mm; androecium of var. pallida, the 4 median anthers (their filaments free or rarely united at base) 3.3-5 mm, the 3 abaxial ones (4.5-)5-7.5, the beak 1.8-2.5 mm; ovary white-strigulose; style 0.7-1.4 mm.

Stipe of pod 5-9 mm, the body 7-13.5 x 0.42-0.55(-0.6) cm, the valves scarcely expressed over the seeds, the interseminal septa 2-2.5 mm apart; seeds (of var. pallida) 2.7-3 x 1.7-2.6 mm, the areole ±0.5 mm diam.-Collections: 42.

Tall subdeciduous forest, wooded bottomland, canyon slopes, becoming abundant in regenerating brush-woodland, 30-600 m, coastal lowlands and foothills of Sierras Madre Occidental and Sur, apparently common in Jalisco and Colima, n. interruptedly to Rio Mayo in s. Sonora and immediately adjoining Chihuahua, not recorded from coastal Michoacan but reappearing in s. Guerrero (mun. La Union) and s.-centr. Oaxaca; in Jalisco ascending e. and up to 1200 m along Rio Grande de Santiago to the barrancas of Guadalajara and the extreme s.-w. prong of Zacatecas; at its n. limit on Rio Mayo and n. to the middle Bavispe valley in Sonora represented by populations variably intergradient into var. shreveana.- Fl. (IX-)X-IV.

This elegant variety differs from sympatric but probably ecologically differentiated var. pallida in the ample, thin-textured leaflets delicately and intricately reticulate on both faces and is the only form of C. pallida so characterized. The lax habit of the very slender pliant branchlets and the ample thin foliage might suggest that var. geminiflora consists merely of a selection of shade-forms of var. pallida; but we have many examples of mesomorphic var. pallida from Central America with leaflets as large and thin-textured as those of var. geminiflora which nevertheless retain the reticulation normal for their kind. The variety is strongly characterized at the middle of its range and southeastward, but at its northern limit on Rios Mayo and Bavispe is represented by forms with leaf-reticulation increasingly faint and the leaves themselves subject to the seasonal dimorphism by means of which S. pallida sens. lat. is adapted to survive temporary droughts. Most of the collections from Alamos, Sonora, and vicinity are of this ambiguous type, seemingly passing into the yet more xeromorphic var. shreveana closely adjacent.

The variety is newly described here, the epithet being deliberately chosen to coincide with that of Cassia geminiflora Collad., in the belief that they are one and the same in a taxonomic sense. Colladon described his species from a drawing made in Mexico by an artist of the Nueva Espana expedition; we lectotypify it herein by the engraved reproduction in Histoire des Casses. A specimen that agrees very well with the picture in having 6-7 pairs of elliptic leaflets survives in the Sesse & Mocino herbarium and we suspect but cannot be absolutely certain that it is equivalent to a typotype. The specimen itself is labelled "C. biflora but cannot be the C. biflora described by Sesse (Fl. Mex. 101; Pl. Nov. Hisp. 59) from Apatzingan which had 8-9 pairs of leaflets and which must have been, for geographical reasons, our var. pallida, although the leaflet number is not fully in harmony with what is known of var. pallida in the Balsas valley. We hazard a guess that Sesse’s C. biflora (but not his specimen) may belong to var. quiedondilla, described from the same general region in Michoacan (near Ario de Rosales), but this is no more than speculation.