Chamaecrista fasciculata

Our concept of Ch. fasciculata, as formulated in the foregoing description and sinonymy, is coextensive with those of Cassia fasciculata presented by Turner (1959), Pullen (1963) and Isely (1975), but with these differences: 1) while recognizing the internal diversity of the species we find this of the same order, and no more extreme, than encountered in related "mimosoid" chamaecristas of the West Indies and Brazil, and in consequence we are unable to maintain formal subspecific categories; and 2) we transfer to Ch. chamaecristoides var. cruziana much of what has passed in Texas as C. fasciculata var. ferrisiae. As redefined, Ch. fasciculata embraces all those extratropical North American chamaecristas in which a monocarpic root coincides with subcentrically costate leaflets, adnate peduncles and relatively large, longistyled flowers. It is only feebly and artificially separated, on grounds of convenience, from 1) Ch. glandulosa by the annual lifespan, from 2) large-flowered races of monocarpic Ch. nictitans by the more nearly sessile petiolar gland combined with temperate (and fully discrete) dispersal; and from 3) Ch. chamaecristoides, with which it freely intergrades in southern Texas, by the greater symmetry of the leaflets. Of three sympatric species, the widespread Ch. nictitans (including its var. aspera) is instantly separable by the small, short-pedicelled 5-8-androus flowers, and the local, Floridian Ch. deeringinana and Ch. lineata var. keyensis by perennial life-form, the former rhizomatous, the latter tap-rooted and differing further in relatively few (5-9) pairs of leaflets.

In the past 80 years Ch. fasciculata sens. lat. has been repeatedly analyzed, alternately from synthetic and analytic viewpoints, but all of the more thorough and detailed studies—we regretfully exclude from their number Britton & Rose’s account in North American Flora—have treated the problems of internal diversity as though soluble without reference to the parental stock of tropical Chamaecrista, that hydra of which Ch. fasciculata itself is not more than an imperfectly severed head. For seen in relation to all of the kindred Neotropical chamaecristas the species emerges as merely a northern extension of a virtually continuous fabric of circum-Caribbean and Central American forms. The selective forces of a temperate climate have compressed the life-cycle of Ch. fasciculata into an obligately annual span, but have left the foliage, perianth and reproductive organs essentially unaltered, the flowers of Ch. glandulosa sens. lat., Ch. fasciculata sens. lat., Ch. chamaecristoides, Ch. rufa, and the largest-flowered races of Ch. nictitans var. disadena being structurally identical, as are the pods and seeds. We have much sympathy for the extreme synthetic position of Otto Kuntze (1894) who, following a voyage round the world and sundry encounters with supra- axillary chamaecristas in places as far apart as the Virgin Islands, New York, St. Louis, Trinidad, Bolivia, Annam and the Deccan, saw all of them as expressions of one multiracial species. The situation had been clearly perceived by Bentham (1871, p. 576) when he remarked that "...the whole group of the Chamaecristae verae seem to pass so gradually one into the other, that the delimitation of species here given must be admitted to be very unsatisfactory.’’ But a logical application of Kuntze’s principle to the taxonomy of the group leads to so cumbrous and elaborate an hierarchical system as to defeat the purposes of classification. The specific definition of Ch. fasciculata is as unsatisfactory to us as it was to Bentham but we think that in practice it is unavoidable.

Collation of the dense and extensive literature concerning Ch. fasciculata sens. lat. cited in our paragraphs of synonymy shows that really only a few variable qualities have ever been found taxonomically useful or significant, even though these qualities may occur in several character-states and thereby appear more numerous than they are. They fall into the categories of 1) growth-habit; 2) vesture; 3) petiolar glands; with, in particular cases, 4) leaflet-number; 5) color of anthers; and 6) width of pod. Each category deserves comment.

1) Growth-habit: The commonest form, prevalent throughout the range of the species, is virgate erect, simple when small or young, branched upward when vigorous or old, and extensively branched where mild climate permits a long season of growth. Less often the central axis is weak or inhibited and the growth- impulse is diverted into lateral branches, which may be either incurved-ascending or almost prostrate. Pullen (1963, p. 32) found that branching patterns are readily modified under cultivation; apparently they are not genetically controlled.

2) Vesture: A duplex vesture of shorter incurved or appressed and longer spreading hairs is characteristic not only of Ch. fasciculata but of all close Central American, South American and West Indian kindred, as also of the related Palaeotropical Ch. mimosoides (Brenan, 1967, p. 100-105). Throughout this group there is random emphasis now on one type of hair, now on the other, either preponderant or either (even both) lacking, the long hairs moreover sometimes lutescent but more often not. Long hairs (setae) may occur on all stem-axes and leaf-stalks as well as on the pedicels and sepals, but sometimes only on the latter. Short hairs may clothe only the stems (overall or in lines), leaf-stalks and pedicels, or may extend to one or both faces of the leaflets. The whole gamut of variation in length, orientation and dispersal of hairs is complex but essentially continuous and is very imperfectly correlated with other morphological characters or geography. Its visual impact is nevertheless often great, and has provoked description of several taxonomically inconsiderable forms. The full range of variation in pubescence is found mainly around the Gulf coastal plain between eastern Texas and western Florida. Northward and westward pilose stems and facially pubescent leaflets become independently rarer. Five of Pullen’s six varieties of C. fasciculata were characterized principally in terms of pubescence, even though it was admitted that two of these may occur closely associated within a given population and that intermediate forms difficult to assign to any one of them are frequent. Except in the case of var. puberula, in which the leaflets are pubescent, Isely (1975) abandoned this position, substituting numbered variants for named varieties, although somewhat paradoxically insisting on sparse pubescence, provided that this coincides, in Florida (but not elsewhere), with large petiolar glands, as a differential character of var. brachiata. The cause of the inherent instability of pubescence in sect. Chamaecrista is unknown; but once recognized as a virtually ubiquitous phenomenon in the group it falls into perspective and may be dismissed as a useless or meaningless taxonomic character.

3) Petiolar glands: The parameters of observed variation in the petiolar gland of Ch. fasciculata will be found in our description of the species; we consider here only their possible bearing on the intraspecific taxonomy, their significance, that is, as markers of genuine genetic differentiation. The petiolar gland is a specialized outgrowth of the leaf-stalk and as such it is found in very large measure to vary in size with the vigor of its parent leaf. Glands on the first cauline leaves of a young plant, which are generally smaller and simpler than adult ones, are predictably smaller than glands of the mature foliage. The larger glands, 1 mm or more in diameter, are associated with relatively robust leaves and coarse leaf-stalks, and this irrespective of the development of the stipe, which varies independently of the head s diameter both in length and thickness. A very small gland, 0.3-0.6 mm diameter, associated as one would expect with slender-stalked leaves and relatively few and small (±7-10-jugate) leaflets, has been considered diagnostic of Cassia fasciculata var. rostrata of southern and Panhandle Texas, at least so long as combined with glabrous foliage and red anthers; but the same gland in Massachusetts (MacKeever 770, NY), Florida (Tracy 6536, NY) or Mississippi (holotypus of Ch. mississippiensis) is passed over as insignificant, possibly because combined in the first case with partly yellow anthers, in the second with up to 12 pairs of leaflets, and in the last with puberulent foliage. Likewise a massive scutellate or trough-shaped gland, said to characterize, when linked in Florida with glabrescent foliage and stems, a robust endemic var. brachiata, can be matched readily, associated either with puberulent or with setose stems, in plants from as far north and west as Iowa and Missouri. If the petiolar gland is interpreted as a vegetative organ, subject to the sorts of emphasis and decline that are accepted as normal in the ontogeny of leaf-blades, its diagnostic value dwindles abruptly.

4)Leaflet number: As already mentioned in relation to the petiolar gland, the number of leaflets is, at least often, related to the leaf’s position on the parent axis, whether this be the primary stem or a branch of the second or later orders, the earlier leaves constantly being shorter and simpler. In unfavorable or marginally tolerable habitats the ontogenetic sequence from simpler shorter to longer plurifoliolate leaves fails to find expression, a neotenic effect which Vassil’cenko (1965) has very convincingly attributed to environmental stress. This seems to be as true of typical Ch. rostrata on the high Plains of western Kansas and Texas as of phenotypically similar forms occurring as far east as New England (Vail in 1888, NY). The so-called C. fasciculata var. rostrata of the lower Rio Grande valley contradictorily combines in one plant the few leaflets and small gland of the elsewhere juvenile leaf with a vigorous, freely branching habit of growth. Isely (1975, p. 96, map 38), accepting var. rostrata as this had been defined by Turner (1959, map 39) and Pullen (1963, p. 65), suggests that it is a polyphyletic concept consisting of "several distinct to ambiguous [geographically] peripheral phases of [C. fasciculata]" in which small neat glabrous foliage and small glands happen to coincide. And indeed the average plant from the Staked Plains differs from that of the south Texas lowlands in its diminished stature and broader pod. Plants of the lowland race cultivated at Austin (Irwin 1412, NY) were found incompatible with other races of C. fasciculata, but since Pullen obtained only 20 fruits from 340 artificial intraspecific crosses in C. fasciculata, more evidence of genuine sexual barriers within the complex is desirable.

5) Anther color: Variation in anther color, noted first by Torrey & Gray (1840, p. 395-396), has been capably analyzed for Ch. fasciculata by Isely (1975, p. 90), who found a general drift toward yellow eastward from the Mississippi valley and toward red or red-violet westward, but so many populations with anthers of mixed color as to render the character useless in practice. There is no known link between anther-color and insect pollinators, but some may nevertheless exist.

6) Pod: While the pod of Ch. fasciculata sens. lat. varies greatly in length (with ovule number) and breadth (with seed size), there seems to be no discoverable discontinuity in the whole range of variation and hardly more than a hint of correlation with dispersal. The longest pods are found in peninsular Florida connected with so-called var. brachiata, but pods hardly shorter have been collected here and there throughout the Middle West. The narrowest pods, already mentioned as characterizing most of the south Texan forms that have been referred to so-called var. rostrata, are duplicated in parallel but not identical slender phenotypes in coastal Mississippi and elsewhere. The remarkable broad pod and massive seeds of so-called var. macrosperma are borne on a plant which at anthesis cannot by any detail be distinguished from the common robust glabrescent oartridge pea of the Atlantic Coastal Plain. And among paratypes of var. macrosperma listed by Fernald in the protologue we encounter some fruits that can be matched for size in populations from far inland. Nevertheless it would be interesting to see this macrosperma phase planted outside its estuarine habitat and studied cytogenetically.

In summary it may be said that 1) intraspecific diversity in Ch. fasciculata is greatest in the warm temperate and subtropical crescent suspended between the Everglades in Florida and the lower Rio Grande in Texas, and shows some incipient geographic segregation toward each end of this arc; and that 2) diversity decreases north and westward, as though there had occurred a progressive elimination of biotypes and simplification of the gene-pool as the species expanded into newly available habitats. While we agree with Isely (1975, passim) that potentially weedy Ch. fasciculata doubtless does exploit the ever-increasing disturbance of its environment and that the chance of reconstructing its pre-colonial dispersal is gone forever, we see no necessity to invoke introgressive hybridization between formerly better defined entities as prime or efficient causes of the random and reticulate patterns of phenotypic variation that meet the eye today. Similar erratic variation patterns are encountered, under quite different, insular circumstances in Ch. lineata, and were discovered in Ch. flexuosa and Ch. serpens by the earliest botanical travellers in Brazil, long before any significant modification of the primeval vegetation had taken place. A student of tropical Chamaecrista has the choice of accepting as a fact of life that sort of intraspecific variability that inspired the dissection of Ch. fasciculata or of retreating to the position of Britton & Urban who, driven by logic rather than sense, found a species in every particular combination of gland and hairiness that fell in their way. But in passing over the observed variation as taxonomically insignificant it is well to bear in mind that its cause and its biological significance remain a mystery.

Each of the named segregates from Ch. fasciculata listed in the synonymy was based on a particular plant or small populational sample the majority of which, like figures in an oriental carpet, had a particular individuality even while repeating a stated theme. This account would be incomplete without a record of their real or supposed differential characters:

i. Cassia chamaecrista of herb. Linn.—Virgate, incurved-puberulent; petiolar gland obconic ±0.8 mm diam; leaflets to 10 pairs; long petals to 14 mm.

ii. C. triflora Jacqu.—Apparently exactly as the preceding, except leaflets 6-8 pairs. A plant cultivated at Wien, not seen by us, the low leaflet number suggesting C. lineata, but said to be monocarpic.

iii. C. fasciculata Michx.—Exactly = 1. Michaux (or Richard) interpreted C. chamaecrista as a larger-flowered form, either puberulent or hirsute (hb. Michx., P!).

iv. fma. pubicaulis O. Kze.—Diffuse, the stems and lf-stalks densely pilosulous; petiolar gland tack-shaped, ±0.4 mm diam; lfts to 9 pairs, glabrous ciliolate.

v. fma. hirsuta O. Kze.—Probably virgate, the stems, lf-stalks, pedicels and ovary spreading-pilose; gland squatly drum-shaped ±0.6 mm diam; lfts 9-10 pairs, glabrous ciliolate.

vi. var. robusta Poll.—Exactly as v. except the glands substipitate saucershaped, 0.8-1.5 mm diam and lfts to 11 pairs.

vii. C. mississippiensis Poll.—Slender, diffusely branched from base, the stems, lf-stalks and     dorsal face of lfts puberulent; gland either short-stipitate or subsessile, 0.3-0.5 mm diam; lfts 8-13 pairs, very small, to 4 5-5 5 mm, few-veined (secondaries 2-3 pairs); fls small, the long petal ±10 mm