The nomenclatural confusion associated with this taxon reflects considerable biological diversity. The varietal entities recognized in the present treatment are distinct taxa which, on morphological and geographical grounds, are readily separable from the main mass of the C. tetraphylla complex. What remains is by no means homogeneous, but there is as yet insufficient evidence to favor recognition of further infra-specific taxa.
Bentham’s acceptance of C. desvauxii in preference to the earlier C. tetraphylla Desv. was in accordance with the belief of Colladon (1816) and de Candolle (1825) that C. tetraphylla Mill. was an earlier homonym. However, as indicated by Burkart (1957), there is no published record of C. tetraphylla Mill., and therefore this name must be considered a nomen nudum. For this reason, C. tetraphylla Desv. must be accepted as the earliest legitimate name.
Amshoff (1939) indicated, in her circumscription of C. tetraphylla in Surinam, that she examined the type specimen (unidentified as to collector) at Paris. However, the present author has been unable to locate this presumed type, nor did Bentham see it in connection with his monograph (1871). Desvaux (1814), in his original diagnosis of C. tetraphylla, recognized 2 forms. The first, "A. foliis obovatis obtusis," is taken to be the typical form; the second, "B. foliis lanceolatis acutis,’' is probably a form of C. tetraphylla var. brevipes (Benth.) Irwin with particularly long fructal pubescence: "Elle diftere de la Cassia lanceolata pars ses fruits veins, et par son port." These same variants were recognized by Colladon (1816). The material examined by Amshoff must, when her characterization of it is compared with the material at hand, have been singularly diminutive or depauperate in foliolar dimension, to wit. Leaflets.. . (up to) 1 cm long,’’ and, in consequence, "pedicels during flowering 1-2 times as long as the leaves." No material examined in the course of the present study had leaflets shorter than the figure Amshoff gives as a maximum, nor were any pedicels more than 1.2 times longer than the largest leaflets observed. L
The C. tetraphylla var. longifolia of Amshoff (1939) is, as she indicated, quite variable, especially in stipule form, and because this variant is in fact characterized by a degree of foliolar length which appears almost thoughout the range of C. tetraphylla, the recognition accorded the Surinam material seems unsound.
C. pulchra H. B. K., differing from C. tetraphylla by rounded leaflet apices, a character which, while fairly uniform in most material from Colombia, Venezuela, and the Guianas, is nonetheless common throughout the range of C. tetraphylla, similarly should not be segregated.
From close examination of exsiccatae available, the following forms, best left unnamed but perhaps of potential varietal status, should be noted:
(1) From stream banks in the vicinity of Mt. Auyan-tepuí, Edo. Bolívar, Venezuela, there have come 2 collections (K. D. Phelps & C. B. Hitchcock 307 in part; Vareschi & Folclats 4544) in which all parts except glands, corolla and the inner calycular surfaces are velvety-pubescent. In this respect, C. tetraphylla var. mollissima (Benth.) Irwin of South Brazil is suggested, but the Venezuelan variant, as far as it is known, has shorter stipules; leaflets narrower, both pubescent and glaucous beneath, gray-green (not yellow-green) above; bracteoles more nearly approximate; pod with longer but more sparingly distributed pubescence. Maguire & Maguire 35386 and 35471 from Serrania Yutaje in Edo. Amazonas, Venezuela, is similar except that in age the leaflets become glabrate.
(2) From the lower slopes of Mt. Roraima and in the high savannas to the south and west there is a form, most frequently seen along water courses, in which the petiole is glabrous and the inferior leaflets are strongly inflexed, the superior pair remaining quite or nearly straight. One specimen, Altsoti 538 (NY), bore an unpublished specific annotation. This variant approaches the Surinam and Pará material designated C. tetraphylla var. longifolia by Amshoff, the latter differing by its oblong stipules and pubescent petioles, while in the Gran Sabana, to the west of Roraima, there is intergradation with the typical var. tetraphylla. Maguire Sc Maguire 40046, from Serra Tepequém, southwest of Roraima in the Brazilian territory of Rio Branco, is similar to this form, but Maguire Sc Maguire 40043, from the same locality, is clearly referable to var. tetraphylla.
(3) In the mountains of central Minas Gerais and southern Goiás, where extensive variability has been noted in connection with other taxa, C. tetraphylla approaches C. langsdorffii, assuming characters suggestive of C. langsdorffii Kunth var. latifoliola Irwin of Paraguay. The lanceolate or narrowly ovate stipules are largely in the 1.0-1.5 cm range in length; petiolar glands are frequently slightly elevated and ovoid; the inferior pair of leaflets is commonly broader and occasionally longer than the superior pair; and flowers tend to be smaller. Pronounced glaucescence occurs in the vicinity of Diamantina (e.g. Irwin 2469). One population in southern Goiás (Irwin 2556a) is indeed remarkable, having, in addition to stipules ca. 2.0 cm long and 0.7 cm broad, and other characters of this form, a brilliant scarlet banner petal. This last condition, in less striking degree, was noted in C. chartacea Irwin var. chartacea (Irwin 2420) .
(4) The majority of material from Paraná and Paraguay which has been examined shows stem pubescence, stipule size, and leaflet form similar to the expression of these characters in C. tetraphylla var. mollissima (Benth.) Irwin. It would appear that in this area, the southern distributional limit of both var. tetraphylla and var. mollissima, considerable intergradation occurs between these taxa.
It is clear that in C. tetraphylla var. tetraphylla morphological variability differs from population to population, and is so varied that further designation of infraspecific taxa should not be attempted until this variable complex becomes better known over its entire geographical range. However, the varietal taxa which have been proposed seem distinct within their respective regions.