Cassia [Tournefort] Linnaeus, Sp. Pl. 376. 1753 & Gen. Pl. ed. 5, 178. 1754, minore ex parte, androecio descripto et syn. Senna Tournef. exclu- sis.—Sp. lectotypica (Gaertner, Fruct. sem. pl. 2: 313, t. 147, fig. 1. 1791, by implication; Britton & Brown, III. Fl. N. U.S. and Brit. Poss. ed. 2, 2: 335. 1913; accepted in Index Nom. Genericorum): C. fistula Linnaeus, a choice inappropriate because the androecium described in the protologue (Gen. Pl. ed. 5, .l.c.) is that of Senna, not that of Cassia sens, restr., but a happy one insofar as it preserves the name Cassia in its historical pre-Linnaean sense.—Cassiana Rafinesque, "Amer. Mon. Mag. Crit. Rev. 2: 266. 1818" fide Merrill, Index Rafin. 143. 1949, a superfluous substitute for Cassia Linnaeus.
Cathartocarpus Persoon, Syn. Pl. 1: 459. 1805.—Lectogeneritypus: C. fistula (Linnaeus) Persoon = Cassia fistula Linnaeus.
Bactyrilobium Willdenow, Enum. hort. berol. 439. 1809.—Generitypus: B. fistula (Linnaeus) Willdenow = Cassia fistula Linnaeus.
Cassia sect. Fistula DeCandolle ex Colladon, Hist. Cass. 83. 1816, nom. superfluum.—Sp. typica: C. fistula Linnaeus.—Cassia subgen. Fistula (DeCandolle) Bentham in Martius, Fl. Bras. 15 (2): 83, in Conspectu definitum, 92. 1870, nom. superfluum.
Cassia sensu Link, Handbuch 2: 138. 1831; Roxburgh, Fl. Ind. ed. Carey 2: 332. 1832; Britton & Rose, 1930, p. 229; Britton & Killip, 1936, p. 169.
Cassia subgen. vel sect. Fistula sensu Bentham, 1871, pp. 513-518. Cassia subgen. Cassia sensu De Wit, 1955, pp. 202-228.
Cassia sensu Irwin & Barneby in Polhill & Raven, 1981, p. 105.
Hypanthium solid, turbinate or slenderly vase-shaped, sometimes obscure externally, that of unfertilized fls disjointing from pedicel; calyx 5-merous, the sepals little graduated obtuse, reflexed at an thesis, deciduous; corolla zygomorphic, the petals commonly yellow or pink, less often red, white or heterochrome, the vexillum then of different color from the rest; petals obovate or elliptic-oblong, clawed or subsessile, subhomomorphic or the vexillum differentiated by crenulate margin, or by callosities at base of blade, or by laterally appendaged claw. Androecium (Fig. 1) zygomorphic 10-merous, the stamens irregularly accrescent toward abaxial side of fl, 2-cyclic: antesepalous (lower) cycle consisting of a) 3 long abaxial members, their filament sigmoidally bent proximally and incurved distally, often dilated (ribbonlike) or abruptly nodulose in the outer curve, their anther dorsifixed ascending, dehiscent ventrally above middle by vertical slits (absent in C. roxburghii, gaping in C. grandis) and by basal pores, and b) 2 adaxial much shorter, erect but apically recurved or coiled and the anther small, sterile or nearly so; antepetalous cycle consisting of a) 2 unequal pairs of stamens arising between and on either side of the 3 long sigmoid ones, their filament straight erect or gently bent forward, sometimes carnosulous-incrassate, their dorsifixed anther either ascending, or horizontal, or resupinate, the thecae dehiscent by basal pore; and b) 1 antevexillar stamen, much smaller and sterile. Ovary stipitate, linear incurved, 1-locular multiovulate, the poorly differentiated style terminating in a minute ciliolate stigmatic cavity looking in- and upward. Pod (Fig. 2) pendulous, shortly stipitate, linear in profile, terete or laterally subcompressed, the sutural ribs either slender immersed or thickened and externally raised as simple or double keels; valves woody, stiffly chartaceous or crusta- ceous, consisting of a thin epidermal layer transversely fissured when dry and a ± woody endocarp, this (except in C. hintoni) produced internally between ovules to form adventitious septa traversing usually the whole cavity, but in few spp. across only ± half the cavity, which is then divided into 2 shallowly inter- digitating files of locules; seed-locules packed with either a) a sweet or bitter, commonly malodorous, gelatinous pulp drying to a brittle blackish pitch free from the seed, or b) a fibrous-suberous coin-shaped envelope ultimately detached from the valves but enveloping the seed; seed-funicle filiform; aril 0; seeds obovoid- ellipsoid biconvex, in all but C. hintoni lying transverse, broadside to the septa, obcompressed, the punctiform hilum near base on one broad face, in C. hintoni basipetal, broadside to the valves, laterally compressed, the hilum sub-basal on one rim, the testa in all smooth lustrous exareolate; endosperm copious; cotyledons irregularly corrugated, epigeal in germination.—Trees, some precociously flowering as shrubs, some reaching large size. Phyllotaxy either spiral or distichous. Pubescence of simple and sometimes partly of minute thickened discolored trichomes. Stipules either simple or 2-lobed, usually caducous. Lvs paripinnate, the pulvinulate lfts either small and numerous (±10-25 pairs) or few and larger. Inflorescence racemose, basically determinate, terminal to either hornotinous leafy branchlets or to greatly shortened leafless branchlets arising (singly or geminate) from annotinous branches below current foliage, the latter appearing to be simple cauliflorous racemes. Pedicels laterally compressed, subtended by a bract and furnished at or shortly above base on either side by a similar but smaller bractlet, these all either caducous or persistent into anthesis. x = 12, 14.
We here formally adopt the restricted definition of genus Cassia that we proposed at the Legume Conference at Kew in 1978 (Irwin & Barneby, 1981, p. 105). The true cassias differ from the sennas, which formed the major element of the Linnaean genus Cassia and subsequently came to dominate botanists’ concept of that genus, in the basically determinate (not lateral axillary) inflorescence, bibracteolate pedicels and characteristic androecium. The androecia of Cassia and Senna are superficially similar in their bilateral symmetry and in the inequality of their members, which dwindle from front to back of the flower, but the organization of the stamens is different. In Cassia sens. str. the three long abaxial stamens are antesepalous, members of the lower whorl of five; in Senna the two longest are antepetalous, members of the upper whorl. In Cassia the four stamens of medium length belong to the antepetalous whorl, two of them standing erect from between the long sigmoid ones, whereas in Senna the four median ones stand together in pairs, one of each pair antepetalous and one antesepalous. The dorsally attached, versatile anthers of Cassia, which dehisce both ventrally by distal slit and basally by a pore, are grossly different from the basifixed anthers of Senna, which dehisce only terminally. These differences between Cassia and Senna are definitive and constant, forming a morphological discontinuity uniformly encountered on all continents. As a syndrome they equal or surpass any generic criterion widely accepted in the taxonomy of Leguminosae.
Cassia differs from Senna in the fruit also, but here and in the seeds the difference is not absolute. The pod of Cassia is indehiscent and with one exception (C. hintoni) multicellular by stiff adventitious septa between the seeds, which lie horizontal to the long axis of the pod and are obcompressed, with hilum at foot of one broad face. The pod of Senna commonly dehisces along one or both sutures, but the orientation and compression of the seeds is variable. In Cassia hintoni the interseminal septa are poorly developed and the seed is at once obliquely basipetal and laterally compressed, with hilum near base of one rim like that of some sennas; and in several members of Senna ser. Coluteoideae the pod is at once indehiscent and multilocular like that of Cassia. The long woody pipelike pod of Senna spectabilis superficially resembles that of some true cassias, but when ripe it splits open along the ventral suture and it arises from a flower wholly unlike that of any Cassia in its androecium and asymmetric corolla, features that remain in close harmony with related sennas characterized by piano- compressed pods. Cassia lacks petiolar glands, but while the gland characteristic of the majority of sennas is absent in some series of the genus it is infallibly present wherever the pod approaches that of Cassia in form or other attributes. An internally pulpy pod, common but not ubiquitous in Cassia, occurs also in some sennas (e.g. S. pendula, S. bacillaris); it is a specialization doubtless related to seed dispersal which we suppose to have developed independently in the two genera, as it has in some neotropical acacias. By itself, therefore, the pod does not provide a sure means of separating Cassia from Senna, although the few exceptions are not genuine obstacles but merely distractions due to convergent evolution. The fallibility of the pod deserves emphasis, nonetheless, because of its influence on the historical taxonomy of Cassiinae.
Cassia and Senna became first known to European botanists through the pods of C. fistula and of S. italica and S. alexandrina, articles of herbal medicine imported from the Orient by way of Egypt and Asia Minor. The fistular woody multilocular pulp-laden pod of Cassia appeared (and really is) fundamentally distinct from the bifacial papery unilocular dry pod of Senna, and it was in these terms that Tournefort (Inst. Rei Herb. 618, t. 390, 392. 1700) defined the genera. And it was to Tournefort’s durably influential work that the origin of confusion between Cassia and Senna can be traced. While Tournefort illustrated Cassia by pods of C. fistula and C. grandis and by a flower probably of the latter, he also listed as species of Cassia seven American sennas, mostly known to him only from antecedent accounts by Piso and Plumier. All of these were deliberately absorbed by Linnaeus (1753, 1754, 11. cc.), along with several sennas discovered after 1700 and the five then-known species of Chamaecrista, into the heterogeneous genus that survived into modern times. In this historically momentous decision Linnaeus adapted, in the high-handed manner that shocked many contemporary botanists, a familiar name to an unfamiliar usage. For the genus Cassia described in the fifth edition of Genera Plantarum is no longer Cassia of the ancients but explicitly Senna. The carefully described androecium, which is profoundly at odds with that of the genuine cassias or of the chamaecristas attributed to Cassia in Species Plantarum, is that of Senna as observed alive in the conservatories at Hartekamp and Upsala; for the androecium of Cassia sens, str., a genus of tropical trees that had not yet been brought to flower in cultivation, was evidently unknown to him. The world monographs of Cassia by DeCandolle in Colladon (1816), Vogel (1837, incorporating a wealth of Brazilian sennas discovered by Sellow) and Bentham (1871), while attaining progressively sophisticated understanding of relationships between natural groups of Cassiinae, reaffirmed the generic concept of Linnaeus, and stifled the voices of protest.
In fact the intratribal generic limits, as we understand them today, had been worked out intuitively as early as 1739 by Jacob Breyne (Prod. fasc. rar. pi. secund. 50), who first described the genera Chamaecrista and Chamaecassia, the latter equivalent to Senna P. Mill. Later on, piecemeal partitions of Linnaean Cassia were attempted by Persoon and Willdenow, who abstracted from it the true cassias as Cathartocarpus (1805) and Bactyrilobium (1809) respectively, and also by Link (1831) and Roxburgh (1832), who contradictorily expelled the sennas and chamaecristas and reserved a purified Cassia for the immediate kindred of C. fistula. When Britton and colleagues, in 1930-1936, revived Cassia sensu Link and Roxburgh, its just claim to generic status was devaluated by association with the false claims of a horde of ill-considered segregates from Senna. Nevertheless, Bentham’s often quoted remark (1871, p. 503) to the effect that "no tolerably fair botanist" had ever mistaken a cassia for anything else is true only provided that the concept of a subtribe Cassiinae be substituted for a comprehensive genus Cassia.
In his work on Caesalpiniaceae and Mimosaceae the generic criteria emphasized by Britton were primarily carpological and his taxonomy became as a consequence highly formalized, like that of Rydberg in the papilionate legumes. While the elastically dehiscent pod of Chamaecrista effectively and consistently, even by itself, separates that great genus from Cassia and Senna collectively, there is no one single feature of the pods of these two genera that will consistently separate one from the other. The terete indehiscent pulpy pod of Senna bicapsularis is morphologically very close to that of some genuine Cassiae, even in its biseriate seeds, and the chain of character states that has been found to lead from an idealized piano-compressed dehiscent pod in Senna to a terete indehiscent one in Cassia sens, restr. has made the case for the latter fatally vulnerable and given decisive support to Bentham’s inclusive concept. Our definition of Cassia is, however, no longer based on particulars of the fruit, but on an infallible syndrome of characters expressed in form of inflorescence, bracteoles and floral morphology that rigorously defines a genus much further removed from any one set of Sennae than any one group of these is from another.
In his masterly discussion of Cassia sect. Fistula, Bentham (1871, p. 507) found insuperable difficulty in arranging the species into subsidiary groups, owing to lack of correlation between spatial dispersal and phenetic characters, particularly size and number of leaflets, persistence of floral bracts, details of staminal and carpological structure, and substance (pithy or pulpy) enveloping the seeds. With the advantages of more complete materials and knowledge of several species discovered since 1871, we find on the contrary that it is possible to distinguish between the capricious effects of parallel or convergent evolution and clues to presumed genetic relationships that are confirmed by patterns of distribution on a continental scale. We have accordingly arranged the American cassias and those exotic ones of which we have detailed knowledge into named series, our purpose being to bring out relative degrees of relationship between groups and stress the isolation within the genus of several unique species. Following a key to these series, which encapsulates our basic classification of Cassia, will be found (p. 10) a practical key to species native and commonly cultivated in the Americas which is adapted for use both with flowering and fruiting specimens.