Chamaecrista ramosa var. mollissima

Either erect or diffuse, repeatedly, often stiffly ramified subshrubs at anthesis 3-12(-15)dm, the annotinous branchlets and pedicels puberulent or pilosulous, the stipules, lfts, sepals and ovary varying (independently) from glabrous or glabrous ciliolate to densely minutely pilosulous, exceptionally glabrous throughout; stipules ovate-cordate-acuminate, the larger up to 2.5-5 mm, ± as long, rarely shorter than internode; lf-stalks abruptly divaricate or widely ascending 1-2 mm, the petiole 0.6-1.3 mm, the rachis 0.2-0.8 mm; gland 0.1-0.8 mm diam, sometimes highly variable from lf to lf on a branchlet, either subsessile or shortly stipitate, the stipe coarse or slender, in profile shortly pin-, peg-, drum- or inversely coneshaped, (0. l-)0.2-0.5 mm tall, 0.3 mm longer to 0.3(-0.5) mm shorter than diam of head; lfts obliquely, dimidiately or falcately obovate or oblanceolate, the distal pair ±½-1/3 longer than the proximal, these of larger lvs 2.5-6.5(-7.5) x 1.2-2.7 mm, the midrib of all, but especially of the shorter proximal pair, incurved, the apex strongly oblique, mucronulate; pedicels 2—10(—13.5) mm; sepals 8-11.5 mm; longer petals up to 9—11.5(—13) mm; pod 26-36 x 4.5-7.5 mm; ovules 8-14.

Campo and cerrado of interior uplands and sand-flats, restinga and stony or sandy savannas of the coastal plain and lowlands, 10-925 m, interruptedly widespread over e. Brazil between 2°N and 14°S, from the lower Amazon valley downstream from Rio Trombetas to the delta region in Amapá and Pará, e. to interior Ceará, s. to coastal n. and interior Bahia, there s. along Espigão Mestre to both slopes of the Rio Corrente-Paraná divide in s.-w. Bahia and e.-centr. Goiás; reports from Minas Gerais based wholly on collections of Martius supposedly from near Diamantina and Serro have not been confirmed by the many modern collectors visiting this region.

The most highly reduced and xeromorphic of the microphyllous series in sect. Xerocalyx, notable for the minute, sometimes almost ericoid foliage which is, however, only a trifle smaller than that of some genuine var. ramosa but then different in the more oblique, dimidiate or falcate, obliquely mucronulate leaflets combined with shorter pedicels. The more and less densely pubescent states, corresponding with var. curvifolia and var. mollissima (sens. str.) of Bentham (1870) and Irwin (1964), which are sympatric in Ceará, Maranhão and Bahia, can no longer be recognized taxonomically. The variety passes by insensible steps eastward across the Planalto into the coarser and slightly larger- (but similarly oblique-) leaved var. lucida, and through this is connected by intergrades with var. parvifoliola and with Ch. desvauxii var. langsdorfii.

The collections of Martius, supposedly from Diamantina (Tejuco) and Serro Frio, cited by Irwin (1964) as typi of var. curvifolia and var. mollissima, are here rejected as such. Cassia curvifolia was based wholly on a unicate collection of Sellow, now lost. The var. mollissima was based on several well-documented collections in Bentham’s own herbarium in addition to that of Martius (M). Since the locality-data of the latter are in all probability inexact, we feel it appropriate to select a new lectotypus (q.v. supra).