Revision of the genus Calpurnia ( Sophoreae : Leguminosae )

Taxa recognised in this revision are:  Calpurnia aurea (Aiton) Benth. subsp.  aurea ; C.  aurea (Aiton) Benth. subsp.  indica Brummitt; C. floribunda Harv ; C. glabrata Brummitt ; C. intrusa (R.Br in W.T.Aiton) E.Mey.; C. reflexus A.J.Beaumont sp nov.;  C. sericea Harv. and  C. woodii Schinz. A putative hybrid between  C. sericea and  C. woodii is recorded. With the exception of C aurea , all species are restricted to southern Africa.  C. reflexus is possibly extinct, and C.  woodii is considered rare. Characters examined in this revision are habitat, habit, vestiture, leaf morphology and anatomy; floral, pollen, fruit and seed morphology and anatomy; and seedling morphology. Illustrations and a key to taxa are provided.

Calpurnia E.Mey. is a small genus of papilionoid legumes, and belongs to the Sophora group of the primi tive tribe Sophoreae sensu Polhill (1981).According to Polhill (1981), Calpurnia is closely related to Maackia Rupr.& M axim, Cladrastis Raf. and Salweenia Baker f., of the Northern Hemisphere.Calpurnia species are slen der trees or small shrubs with pinnate leaves and attractive yellow papilionoid (lowers in racemes or panicles.All but one species are narrow endemics of southern Africa.C. aurea subsp.aurea extends north into Ethiopia and, together with C. aurea subsp.indica, into southern India. Early classifications placed Calpurnia species under Sopliora L., of the tribe Sophoreae (Aiton 1789), Robinia L., o f the tribe R o binieae (L'H éritier 1791), and Podalyria Willd.(W illdenow 1799) and Virgilia Poir.(Lamarck 1793), of the tribe Podalyrieae.Meyer (1836) transferred four species from Virgilia and published a new species to establish Calpurnia.The genus com m em orates the Roman poet C alpernicus Siculus who, in the middle of the First C entury A.D., wrote poetry similar to that of Virgilius.Superficially, species o f Calpurnia resemble those o f Virgilia (com m em orating Virgilius).Phillips (1917) recognised seven species of Cal purnia.Subsequent w orkers have contributed taxonomic notes on species and some have recognised subspecific taxa (Gillett 1965;Brumm itt 1967Brumm itt , 1970)).Yakovlev (1971) provides ihe most com prehensive recent account of the genus, notw ithstanding the correction of some erroneous nom enclature by Ross (1976).

I'ollen
Mature buds preserved in Ihe mixture: formalin: acetic acid: ethanol (FAA), and material from herbarium sheets were used.Buds from herbarium sheets were softened by gently boiling in water with a few drops of 'Teepol', before the anthers were removed.Anthers were sonicated in distilled water for one minute to release pollen.The suspensions were passed through 250 (jm mesh filters to remove anther debris.Two drops of lactic acid were added to pollen filtrates to prevent excessive expansion of the pollen grains.Pollen was acetolysed using a freshly prepared acetic anhydride and concen trated sulphuric acid mixture (9:1) for 2 minutes 45 sec onds in a water bath at 100°C.For scanning electron microscopy, acetolysed grains were pipetted onto glass cover slips and fixed onto brass view ing stubs.Specimens were viewed using a Hitachi S570 Scanning Electron M icroscope (SEM) at 10 kv.For transmission electron microscopy, acetolysed grains were embedded in Spurr's resin using conventional procedures.Sections slightly thicker than 0.1 (jm were stained and viewed with a Jeol T200 Transmission Electron M icroscope (TEM) at 80 kv.
The epithet floribunda refers to the profuse flowering of this species with, as Harvey (1862)
C. w oodii has a very restricted distribution in KwaZulu-Natal (Figure 4) and occurs at approximately 1 500 m above sea level, in the U nderberg region and at South Downs, near Estcourt, Weenen County.The south east-facing grassland slopes overlooking W agendrift Dam and the M oor Park Nature Reserve near Estcourt, support a locally abundant population o f this species.Here, plants form a stand in Southern Tall Grassveld (Acocks 1988) dom inated by Themeda triandra Forssk.C. woodii occurs on steep, dry slopes in loose, shaley soils undisturbed by cultivation or grazing.There appears lo be little that is unique to the area where C. woodii grows that could explain the scarcity of this species.The (lowers are prone to insect herbivory and beetles o f the order Bruchidae cat the seeds.The stems and leaves however, rem ain ungra/.cdeven by the local goats.C. woodii is a very appealing species, especially when in flower, and it is hoped that an interest in its cul tivation could be fostered, which might ensure the sur vival of this very restricted and little-known species.Flowering time: December to May. Vouchers: Beaum ont 72, 90 (NU); Rent?. 512 (K); Sidey 124 (M O); Wood 4377 (BOL, K).
Accordingly, we reject all Calpurnia synonyms associat ed with the Burm an fragm ent in the De Candolle Herbarium, G eneva (G-DC), and propose the new name Calpurnia reflexus.The specific epithet refers to the revolute nature ol the calyx teeth, which is unique in Calpurnia (Figure 9).
In addition to the revolute calyx teeth, the monadelphous stam ens and densely papillate base of the staminal sheath are unique characters for this species.Yakovlev (1971), citing calycine and staminal characters, suggest ed that this taxon m ight be worthy of recognition as a new genus.We prefer to recognise this taxon as a Calpurnia species because o f the sim ilarity of the remaining vegetative and floral characters between it and the other m em bers o f this genus.There is a paucity of representative herbarium material of C. reflexus.This plant appears to be very scarce and might be extinct.The first two authors visited the recorded Lesotho localities of the Leribe District, and west of the Thaba-Putsoa Range near R am abantas between Roma and Nyakosoba, in May 1990, but were unable to find the plant.The west ern lowlands o f Lesotho are heavily overgrazed, the flat land is cultivated or badly eroded, and consequently, indigenous flora survives only on isolated koppics.
The specific epithet refers to the intrusive nature of the calyx base.
Many workers have misapplied the name Calpurnia intrusa to specimens of C. sericea Harv.The confusion surrounding the identity of C. intrusa, has arisen because successive workers perpetuated a m isunderstanding of the original concept of this taxon.M eyer's (1836) con cept of C. intrusa is in fact, C. sericea Harv., and not the Western Cape endemic which Brown described (Aiton 1811).This taxonomic mistake was perpetuated (Phillips 1917;Burtt-Davy 1932;Henkel 1934) until Brummitt (1967) and Ross (1976) clarified the nomenclature.Brummitt (1967) describes two specimens at the British Museum of Natural History mounted on one sheet: Hort.Kew 1796 and Hort.Kew 1797, and identified as Virgilia intrusa.We examined these specimens and, although they do not bear Brown's handwriting, we believe that they represent the plant which he had in mind when he described Virgilia intrusa.Brummitt (1967) also notes that the plant illustrated in tablet 2617 in Curtis's Botanical Magazine 53 (1826), is in fact, C. aurea subsp.sylvatica, although Brown verified the illustration to be of Virgilia intrusa.Plants of C. aurea subsp.aurea resemble those of C. intrusa in some respects.Both have large flowers, and smaller specimens of C. aurea tend to be found in southern Africa, where they resemble C. intrusa with respect to organ size.Tablet 2617 in C urtis' s Botanical Magazine 53 (1826) clearly portrays the inflat ed base of the calyx above the hypanthium, typical of C. aurea, and unlike the deeply intrusive calyx of C. intrusa.Meyer (1836) in transferring V. intrusa to Cal purnia, cites a Dregc specimen (which is a specimen of  The leaflets are characteristically oblong and usually glabrescent above and pubescent below.Yakovlev (1971) established C. intrusa var.glabrata to distinguish speci mens with 'totally glabrous leaflets'.However, the leaflets of the type specim en o f C. intrusa var.glabrata are not glabrous.The upper surfaces o f the leaflets are usually glabrous but the rachis, petioles and lower leaflet surfaces are sparsely to m oderately villous.Some speci mens have leaflets which are equally pubescent on both surfaces (e.g.Hort.Kirstenbosch, sub BOL470050).Ovary pubescence, both in term s of distribution and den sity, is also highly variable, ranging from hairy through out.hairs confined to the sutures, to glabrous.We do not support Y akovlev's establishm ent of varieties in C. intrusa based on variation o f pubescence o f leaflets or of ovaries because these characters are so variable and char acter states are not clearly defined.
C. intrusa is endem ic to the Western Cape, and occurs south o f 33° S and between 20° and 23° E in the Groot Swartberge region (Figure 4).It occupies an island of alluvial sand and calcrete surrounded by conglomerate sandstone, in turn surrounded by quartzitic sandstone with shale and tillite.C. intrusa is found in gorges in relict forest patches.Flowering tim e: Decem ber to June.
The epithet aurea refers to the golden-yellow flowers.In South Africa it is known as Natal laburnum.
C. aurea subsp.aurea, the best known Calpurnia spe cies, is a frequent entry in African floras and botanical checklists.It is cultivated as an ornamental and as a shade tree in tea and coffee plantations (Brununitt 1967).
C. aurea subsp.aurea has the largest vegetative, flo ral and fruit characters in Calpurnia.It is the tallest species, often growing as a slender tree in forest clear ings and forest margins.In drier, open scrub and grass land habitats it grows as a shorter, much branched shrub (Figure 12).Another characteristic of this species is the base of the calyx tube which is slightly inflated around the apex of the hypanthium.W hen viewed laterally, the base of the calyx tube does not obscure the hypanthium, but the calyx tube and teeth conceal the petal claws and lower portions of the stamens (Figure 11H).Flowering time: throughout the year when mild to hot w eather pre vails, usually during spring and summer.
We recognise two subspecies in Calpurnia aurea: subsp.aurea and subsp.indica.Subsp.indica occurs in India; however, specimens from Indian herbaria were  Britto (1986) provides a clear account of the two sub species in India.He uses ovule number per pod to distin guish subsp.indica (with six to seven ovules) from subsp.aurea with 12 to 15 ovules.Brummitt (1967) cites differ ences in the relative lengths of the teeth to the campanulate part of the calyx, to distinguish the two subspecies.
We reject previous decisions by other authors to recognise subsp.sylvatica within C. aurea, or to recog nise it as a distinct species.Former taxonomic treatments distinguished such specimens by their glabrous leaflets and ovaries.Vestiture of leaves and ovaries is notorious ly variable among other members of this genus as well as among specimens of C. aurea sensu lato.The infraspe cific discrimination in C. aurea based upon degree of pubescence of these features has been quite arbitrary.Previous workers did not consider the pubescence of other structures.Vestiture on stipules, petioles, rachides, and petiolules do not necessarily correlate with the pubescent or glabrous nature of the leaflets of specimens.Features of the inflorescence (the peduncles, bracts, pedicels and calyx) may also be variously pubescent to glabrous, independent of the condition of the ovary.Plants in this genus are rarely, if ever, glabrous, and usu ally some pubescence is present, at least at the junctions of structures for example, bracts and pedicels and peti oles.Plants of C. aurea vary greatly in the distribution and density of their pubescence, and classification at the specific or subspecific levels using these criteria is impractical and controversial with respect to defining the limits between character states.
The epithet glabrata refers to the glabrescent nature of plants.
The most striking feature of C. glabrata is the almost glabrous nature of all parts.Presence or absence of hairs, however, is a character to be used with great caution when identifying Calpurnia specimens.In the absence of floral or fruit characters, vegetative specimens of C. glabrata may be confused with glabrescent specimens of C. aurea subsp.aurea occurring in the southern Cape.Flower size varies in C. glabrata (Figure 13E).Flowers may be similar in size to those of C. sericea (Figure 6E) or larger, resembling those of C. floribunda (Figure 3E).
The calyx of C. glabrata conceals the wing and keel petal claws, and the filament bases.These features are exposed in flowers of C. sericea, which C. glabrata oth erwise resembles.In general outline and size, flowers of C. glabrata may resemble those of C. floribunda, but the pubescence of the leaves, calyx and ovary of the latter is denser than in C. glabrata.
The fruits of C. glabrata most closely resemble those of C. sericea and C. reflexus, although the upper suture wing of C. glabrata fruits is slightly broader.

C.
glabrata (Figure 4) is concentrated in Swaziland but it also occurs in Mpumalanga between 30° and 3 1 1 E and in KwaZulu-Natal to 29° S. It occurs between 200 to 1 500 m.Flowering time: October to April.This species occurs predominantly in North Eastern Mountain Sourveld (Acocks 1988) in hillside grassland, Acacia caffra scrub, roadsides and riverbanks.In Swaziland it grows in disturbed roadside grassland with the weeds Pteridium aquilinum (bracken) and Psidium guajava (wild guava).
Taxa recognised in this revision are: Calpurnia aurea (Aiton) Benth.subsp.aurea; C. aurea (A iton) Benth.subsp.indica B rum m itt; C. floribunda Harv ; C. glabrata Brum m itt; C. intrusa (R.Br in W .T.Aiton) E.M ey.; C. reflex u s A .J.B eaum ont sp nov.; C. sericea Harv.and C. woodii Schinz.A putative hybrid between C. sericea and C. woodii is recorded.With the exception o f C aurea , all species are restricted to southern Africa.C. reflexus is possibly extinct, and C. woodii is consid ered rare.C haracters exam ined in this revision are habitat, habit, vestiture, leaf m orphology and anatom y; floral, pollen, fruit and seed m orphology and anatom y; and seedling morphology.Illustrations and a key to taxa are provided.IN TR O D U C T IO N

M
A TERIA LS A ND M ETH O D S RESULTS AND DISCU SSIO N Pollen Ferguson & Skvarla (1981) reviewed the pollen of the Sophoreae and Calpurnia conforms to the general m or phology of the tribe.The pollen of all species of Calpurnia is in monads, tricolporate, spheroidal to pro late, and apocolpate (Figure I A. B).Further, colpus m ar gins arc smooth and the texture of Ihe colpus furrow may be smooth or crustate in all species.Size variation is m in imal and most grains are about 20 x 15 mm.Com para tively fewer columcllae and a less continuous tectum dis tinguishes C. intrusa pollen (Figure 1C) and C. sericea (Figure ID) from that of oilier Calpurnia species.The pollen wall is between 0.5 111111 and 1.0 mm I hick in all species.The pollen of the putative hybrid, C. sericea x C.
Figure 8 illustrates the putative hybrid C. sericea x woodii.This is the first record of possible hybridisation in Calpurnia.Wc have only seen two hybrid plants, both by the river at the picnic site o f Moor Park Nature Reserve in KwaZulu-Natal (Figure 4), with C. sericea nearby, and C. woodii occurring about 2 km away.The hybrids are vigorous and express a com bination of char acters from the parents.The hybrid individuals are taller than plants of both C. sericea and C. w oodii, and com prise many close-set, sem i-erect, flexible stems which arc rem iniscent of those o f C. sericea.Vegetative charac ters which appear to be derived from C. woodii are the dense villous-tom entose pubescence o f the leaves and stipules, and the com paratively long stipules.The leaves of the hybrid plants also resem ble those of C. woodii in shape, they are, however, smaller.The inflorescences of the hybrid plants are long, lax, m any-flow ered racemes concentrated tow ards the ends o f branches.This arrange ment is sim ilar to the condition found in C. sericea, com pared to the robust, erect, densely-flow ered and shortly paniculate inflorescences o f C. woodii.Floral characters of the hybrid plants which resem ble C. sericea arc: the
FIG U RE 12.-Known distribution o f Calpurnia aurea subsp aurea in southern Africa.unavailable to us for study.Although limited material from other herbaria was examined, the present authors feel that more consultation of specimens of subsp.indica is needed in order to present a full descriptive account of this taxon.