Chamaecrista glandulosa var. flavicoma

Highly variable in stature, flowering sometimes precociously as a singlestemmed suffruticulose herb but shrubby when adult and in favorable circumstances arborescent, at anthesis (5-)7-40 dm, the usually erect, rarely diffusely assurgent branches, lf-stalks, and inflorescence most commonly both puberulent with short incurved hairs and setulose-pilose with incurved-ascending, horizontally patent, or subretrorse, usually yellowish or golden hairs up to 0.7-1.5 mm, but either type of vesture predominant or either absent, both variable in density, the densely and usually amply leafy annotinous branchlets varying from densely shaggy-setose-pilose to thinly pilosulous or glabrate, the concolorous foliage likewise variably pubescent, the membranous lfts pilosulous on both faces, or only beneath, or glabrate on the faces and ciliolate, exceptionally glabrous.

Stipules ovate-acuminate to narrowly lance-attenuate or linear-acuminate, those subtending major cauline lvs up to (3—)4—15 x (0.6-)l-3 mm, the thin-textured blades green often becoming dry and brown, from the slightly asymmetric to shallowly semi-cordate (exceptionally obliquely cordate-amplexicaul) base prominently (on Sa. Nevada de Sta. Marta subobsoletely) (5-)9-19-nerved, persistent (in s. Mexico exceptionally deciduous).

Major cauline lvs up to (3.5—)6—12(-14.5) cm, the petiole (including often strongly dilated pulvinus) 2-7(-10) mm, at middle 0.5-1.3 mm diam; gland l(-3) situated at, above or below middle of petiole proper, urceolate or shallowly saucer-shaped, round or elliptic (0.35-)0.4-1.1 mm diam, stipitate or subsessile, the glabrous or exceptionally puberulent stipe very variable in length and thickness, linear or dilated upward, the whole in profile usually club- or tack-shaped but when short-stipitate becoming squatly obconic and when very long-stipitate pin-shaped, (0.3-)0.4-2.8(-3.5) mm tall, 0.3 mm shorter to 2.3(-2.9) mm longer than diam of head, smaller glands often present on some distal segments of rachis; lfts of major cauline lvs up to (10-)12-22(-24) pairs, varying in outline from broadly oblong to oblong-elliptic, narrowly lance-oblong, or sublinear, mostly obtuse mucronate or aristulate, sometimes subacute, (7.5-)12-22(-25) x (1.6-)2-5.5(-6) mm, the blades plane or often subrevolute, from base 5—9-nerved by the slender centric or moderately displaced, straight or gently incurved midrib with on the narrow side 0-2, on the broad one 3-6 primary nerves, the secondary pinnate venulation (4—)5—11 pairs of major secondaries with, in each interval, 0—2 minor ones, these sometimes, however, almost as strong as the major ones, the lfts then becoming multistriate, the venulation varying in prominence, sometimes sharper above than beneath, sometimes the reverse.

Peduncles usually solitary, sometimes 2 superposed from each axil, together with often well-developed raceme-axis (1.5-)3-20(-22) mm, adnate through 2—14(—18) mm, (l-)2-5(-7)-fld; pedicels mostly 12-27(-50, the longest commonly at least 15) mm, some late ones or those distal to a several-fld raceme shorter; sepals (7—)8— 14(— 16) mm; long abaxial petal (9—)11—17(—19) mm; ovary strigulose or loosely white-pilose; style filiform 3-5(-7) mm; ovules 9— 16(— 17).

Pod (30-)35-65(-70) x (4.5-)5-7.5 (in n. Colombia sometimes 8-9) mm, the valves either pilosulous or subappressed-puberulent; seeds (3.2-)3.4-4.7 (in n. Colombia sometimes up to 5-5.8) mm, the testa dull ocher-brown, the pits concolorous or often fuscous, or fuscous overall.—Collections: 117.

Open rocky or grassy hillsides, roadside hedges, thickets, fallow fields and rough mountain pastures, sometimes a weed of cafetal, in wooded country only in natural glades, cut-over forest, or on stream-banks or sand-bars, mostly between 900 and 2250 m but in n. Colombia up to 3100 m, interruptedly widespread through the Andes from Sa. Nevada de Sta. Marta, Sa. de Perijá and Cordillera de Mérida in n. Colombia and n.-w. Venezuela s. through the Cauca valley and Cordillera Occidental s.-ward from Antioquia, highland Ecuador, and inter-Andean valleys of the Amazon slope in Peru to the Upper Urubamba valley and, perhaps disjunctly, to the Yungas valleys in deptos. Larecaja and La Paz, Bolivia; reappearing in slightly differentiated form (=C. pavoniana) on the Pacific slope of centr. and n. Peru, on open hillsides and dry streambeds of the coastal lomas and Andean foothills mostly below 1000 m, in valleys of Ríos Chancay, Chillón, Rimac and Lurín in Dept. Lima (prov. Chancay, Lima, Huarochirí) and Dept. Cajamarca (prov. Hualgayoc), this ecotype apparently naturalized, at least formerly, in the Canary Islands (Palma); in Colombia absent from Magdalena Valley, Cordillera Central and e. slope of Cordillera Oriental, where replaced by subsp. tristicula; remotely disjunct in s. Mexico and adj. Guatemala; in pine-oak woodland near 1700 m in Sa. Madre del Sur, Guerrero; and on karst limestones and in scrub forest below 150 m in Yucatán, Quintana Roo, Campeche, and Petén.- Fl. in Venezuela and Colombia mostly (VII-)VIII-III(-IV), from Ecuador s.-ward X-VI, in Guerrero IX-II, in Yucatán IX, VI.

Bv expanding our concept of var. flavicoma to embrace all (except the far- southern var. andicola) large-flowered, long-pedicelled chamaecristas with subcentrically costate leaflets known to occur in the Andes between Sa. Nevada de Marta and Bolivia, we sweep into one net a polymorphic shoal of minor races which Bentham would have or (1871, and in annotation of specimens) actually did distribute between the ineffectively differentiated C. flavicoma, C. glandulosa and C. virgata. For Bentham, as later for Britton, C. flavicoma was characterized by a dense yellow pilosulous vesture, but an entity defined by this one feature would now be as heterogeneous in stature, stipules, glands, and width or venulation of the leaflets as the parallel, often sympatric analogues with gray or white vesture, whether sparse or dense. The large-flowered Andean chamaecristas are clearly in a stage of evolutionary ferment and present a taxonomically baffling series of minutely, or sometimes grossly, distinguishable forms, most of which can be precisely characterized only if we narrow them down to one or few populations, these not necessarily correlated with credible geographic patterns. The alternative to a comprehensive var. flavicoma is a return to the Brittonian method of applying names to individual populations.

Analysis of the petiolar glands of all available Andean var. flavicoma has brought to light a great range in size and form, this running sometimes in harmony, sometimes in discord with other modalitites of variation, but nevertheless showing some geographic sorting. In Venezuela, Colombia and northern Ecuador (Pichincha) the prevailing gland is stoutly stipitate, in profile claviform and mostly 0.8-1.5 mm tall. From southern Ecuador (Chimborazo) south to northern Cuzco (prov. Conventión) in Peru, the average gland is shorter or, if no shorter, then slenderly stipitate, in profile tack-shaped or obconical; but the typus of C. stipulata and plants resembling it from the headwaters of Río Huallaga have the coarse glands of the northern region. Southward from the upper Urubamba valley in Cuzco to the Yungas in Bolivia the gland becomes rather abruptly longer and slender-stipitate, often pin-shaped in profile and 1-3.5 mm tall with head only 0.4-0.7 mm diameter. However in these three segments of the Andean range the other phenetic characters are extremely flexible in their expression, the mutual proportions of the duplex vesture, the color of hairs and of anthers (these red or yellow) the number of leaflets, the prominence of the secondary venulation, and the size of the pod each varying independently of the rest.

The most notable Andean variant known to us is that found on Sa. Nevada de Sta. Marta and Sa. de Perijá, in Magdalena and northern Norte de Santander in Colombia (to be expected in adjoining Venezuela), unlike all others in the massive pod 8-9 mm wide and correspondingly immense seeds about 5-6 mm long. But even this is not internally uniform, the plants of Sta. Marta having uniquely thin-textured stipules with scarcely visible venation, whereas those of Sa. de Perijá, which are Chamaecrista arborescens sens. str., have the ordinary striate stipules of var. flavicoma. The latter variant differs from the now well-known representative of var. flavicoma on Cordillera de Mérida immediately to the east, which has been segregated (if arborescent, but not when shrubby) as Ch. dendrina, in the fruit alone (pod valves 6-7.5, not 8-9 mm wide; seeds proportionately smaller) and could only be sustained as a distinct taxon by special license.

The cismontane Andean form of var. flavicoma segregated long ago as C. pavoniana G. Don is often recognizable, at least to the experienced eye, as distinct from most transmontane ones in the syndrome of thinly puberulent, distantly leafy stems, relatively small 6-9-nerved stipules mostly 4-9(-10) mm long, and narrow acutish, often falcately incurved leaflets up to 7.5-15 x 1.6-2.8(-3.2) mm. This form is taxonomically embarrassing, for it has no single gross morphological feature that cannot be matched, not only somewhere in var. flavicoma, but also somewhere within Antillean var. swartzii. In view of its relatively local range, mostly within a radius of 70 km of Callao, a port linked since European settlement of Peru with Spanish possessions in the Caribbean, it is possibly not truly indigenous, even though we have records going back to the late XVIII century, ut actually is naturalized var. swartzii. The fact that what appears to e 1 en ically the same form of C. glandulosa was found naturalized in the Canary s an s a century ago (Salzmann in 1869, P) is evidence of its weedy potential. Nevertheless we have found no single individual plant of West Indian var. swartzii that matches simultaneously at every point this Peruvian expression of C. glandulosa and therefore, with hesitancy and lack of conviction, refer it provisionally to var. flavicoma.

This seems the appropriate place to mention three collections from Peru which we are at present unable to account for: without locality, hb. Pavón. (BM); without locality, Mathews 918 (K); Huambos, prov. Chota, dept. Cajamarca, 2000 m, 11.IX.56 (fl, fr), Soukup 4529 (US). Collectively these plants closely resemble some Caribbean var. swartzii, differing from all var. flavicoma in having short stipules, abbreviated leaves with no more than 9-12 small leaflets in the larger leaves. The material is insufficient to show the full stature, which appears to be, however, distinctly fruticose. Possibly these represent ecotypes of var. flavicoma from unfavorable microhabitats, but this is pure surmise. Bentham identified Mathews 918 as C. glandulosa and we agree that it must represent a form of that species.

The disjunct Mexican populations here referred to var. flavicoma pose a real problem in classification, being not precisely matched by any particular Andean one, and being themselves internally heterogeneous. The plants of Yucatán and Campeche differ collectively from nearly all Andean var. flavicoma by the very narrow linear-attenuate stipules (6-11 x 0.6-1.1 mm), but have two types of foliage, one (Ch. yucatana, sens. str.) that compares favorably with some Andean plants with broad leaflets, the other (e.g. Lundell 1427, GH, NY, 7244, US) with narrow and fewer leaflets, forcibly reminiscent of Caribbean var. swartzii, and possibly more nearly related in that direction than southward. One of the two known collections of var. flavicoma from Sa. Madre del Sur in Guerrero (Coalcomán distr., Hinton 12708, G, K, NY, US) is visually arresting for the syndrome of long narrow (x±2 mm) leaflets, extremely long (to 5 cm) pedicels, and extraordinary stipules, which are asymmetrically cordate-amplexicaul and early deciduous. We would certainly have described this if we had not seen Hinton 10698 (G, K, NY, US) from district Minas a short distance to the east, in which the leaflets are the same but the persistent stipules and the pedicels comparable with Peruvian or Bolivian var. flavicoma. For the present we interpret Hinton 12708 as an exceptional individual variant, one which, like the also profusely large- flowered variant from Santa Marta, mentioned above, would be a welcome addition to the list of chamaecristas in tropical cultivation.

Our statement of dispersal for var. flavicoma deliberately omits as implausible the supposed type-locality on the Orinoco. As Humboldt and Bonpland were generally specific and precise in their locality data, the vagueness of the phrase "near the Orinoco River" is itself negatively suggestive. The parts of the type collection that we have seen match most closely in general aspect, and in details of gland, vesture, leaflets, and venation, recent and contemporary collections of var flavicoma from the middle and upper Cauca valley in Colombia, especially those from the environs of Popayán. We believe that C. flavicoma must have been collected first in the western cordillera of Colombia, not on the Orinoco.