A synopsis of Peristrophe ( Acanthaceae ) in southern Africa

A synopsis of Peristrophe Nees (Acanthaceae) in southern Africa is provided. Peristrophe comprises nine species, one with two subspecies in southern Africa. A key for identification and descriptions of species not included in recent literature are provided. Morphology of tertiary bracts, indumentum on stems and distribution is illustrated and diagnostic characters, distribution, habitat, flowering time and conservation status are discussed for each species.


I
undertook a revision of Peristrophe Nees in southern Africa (Balkwill 1985) and while doing so, published an account of the Peristrophe grandibracteata complex (Balkwill et al. 1988) and some new species (Balkwill et al. 1985;Balkwill & GetlifTe Norris 1989).I thought that the discussions under these species and that of Brummitt (Wood et al. 1983: 451) clarified the fate of previously recognised species, but from Welman (1993), it seems that this is not the case.At present, no satisfactory key to the southern African species is available in the literature and there is likely to be a long delay before further accounts of Acanthaceae are published in the Flora o f southern Africa series.I have therefore decided that it would be constructive to publish a synopsis to aid in the identifica tion of species of this genus.
The genus Peristrophe occurs in Africa and the East Indies and comprises about 25 species, of which nine oc cur in southern Africa.Peristrophe is very closely related to Dicliptera Juss.and differs only in the nature of the placental bases, which are inelastic in Peristrophe and elastic in Dicliptera.Some workers have tried to apply other characters (e.g.width of the tertiary bracts) as ge neric characters, and this has led to the incorrect clas sification of species such as P. angolensis (S.Moore) K.Balkwill, P. transvaalensis (C.B.Clarke) K.Balkwill and P. hereroensis (Schinz) K.Balkwill, all of which were described in the genus Dicliptera.Another species that has been confused with Dicliptera is P. bivalvis (L.) Men., which has been confused with D icliptera foetida (Forssk.)Blatt.(Wood et al. 1983).In southern Africa, the shape and venation of the tertiary bracts, if used si multaneously, can act as a guide to the genera.The ter tiary bracts of the southern African species of Peristrophe are either nanow and single-veined, or if broad, then the veins are pinnate and the secondary veins are reticulate near the margins (Figure 1), whereas the bracts of D i cliptera are 3-, 5-or even 7-veined from the base and the secondary veins do not form conspicuous reticulations near the margins.
Nees (1832,1847), recognised two sections in Peris trophe and subdivided the first of these into two subsec tions, but he did not name the sections or subsections.Since that time, three species of Peristrophe have been described from Namibia and these do not belong in the groups described by Nees.Thus, it seems that Nees's (1847) first section comprises three groups, but these three groups are sufficiently different to be recognised at the rank of section.The species of Peristrophe in southern Africa belong to two of these groups (referred to as sections 2 and 3 below).It would be premature to formally describe these sections here, before a mod ern and monographic account of the genus has been completed.
Section 1: inflorescence of two to seven monochasial cymes (inflorescence units) umbellately arranged, 'um bels' sometimes compounded.Tertiary bracts spathulate, ovate to obovate, with three, five or seven veins from the base, remaining green when mature.Capsule clavate.
Included species (none southern African): P. baphica (the type of the genus), P. montana Nees and P. speciosa Nees.
Section 2: differs from Section 1 by having only two inflorescence units in each 'umbel'; a single primary vein from the base of the tertiary bracts; tertiary bracts that have conspicuous secondary veins and become membranous when mature (Figure 1A-C); and capsules that are obtrullate in profile.
Section 3: differs from Sections 1 and 2 by the narrower, less conspicuous tertiary bracts, which are lanceolate to oblanceolate (not ovate to broadly ovate) and narrower than 4 mm (Figure 1D-I) and from Section 2 by the tertiary bracts that are green (not membranous and conspicuously veined) at maturity and by having clavate (not obtrullate) capsules.
Diagnostic characters: P. hereroensis can be distin guished from P. grandibracteata by its green stems with sparse eglandular hairs (Figure 2B) rather than white stems with dense ornamented eglandular hairs.It differs from P. namibiensis by the tertiary bracts that are widely cuneate at the base (Figure IB) (not reniform or cordate), secondary bracts that are longer than 7 mm (not shorter than 6 mm) and by the narrower leaves [length : breadth ratio greater than (not less than) 2.4: 1].Distribution: northern Namibia (Figure 3).Habitat: on various kinds of soils, usually in the shade of trees of Acacia, in Kalahari Acacia wooded grassland and deciduous bushland and the Kalahari/Karoo-Namib transition (White 1983).Flower ing time: between November and July, with a peak in March and April.Conservation status: P. hereroensis is known from 17 collections from a number of different localities.It is unlikely to be rare or threatened.Diagnostic characters: P. namibiensis differs from P. grandibracteata by its green (not white) stems and from P. grandibracteata and P. hereroensis by its much smaller secondary bracts that are not leaflike and by its cordate tertiary bracts (Figure 1C).3a.subsp namibiensis than 1.25 : 1; and pubescent capsules distinguish subsp.namibiensis.Distribution: southern half of Namibia (Fig ure 4).Habitat: on dolomite koppies (and possibly other habitats) in the Kalahari/Karoo-Namib transition and Namib Desert (White 1983).Flowering time: April to Au gust.Conservation status: this taxon is known from only five collections, suggesting that its conservation status re quires investigation.3b.subsp.brandbergensis K. Balkwill in Balkwill et al. in South African Journal of Botany 54: 52 (1988).Type: Namibia, Outjo Farm OU 516, Sandsteinruecken, Giess & Barnard 7921 (WIND, holo.!; NBG!, PRE!).

Justicia ligulata
The names P. paniculata and P bicalyculata were both validly published in publications bearing the date 1775.They are taxonomic synonyms but P. paniculata has pri ority (Wood et al. 1983) 5).This suggests that the species is an annual, although it has been said to reach 2 m high (fide notes on Smith 1293 in PRE).
Diagnostic characters: smaller flowers, less than 13.7 mm long, very shallow disc and small simple tap root system serve to separate P. paniculata from all other southern African species of Peristrophe.It has often been confused with P. decorticans and can be separated from the latter species by the mature bark, which is smooth and black in P. paniculata and white and peeling in P. decor ticans, by the ratio of length of larger tertiary bract to that of the shorter, which is (1.4-) 1.6-2.2(-2.6): 1 in P. pa n i culata and 1.2-1.4(-1.6): 1 in P. decorticans', and on the basis of distribution: P. paniculata grows in Namibia and in northwestern Botswana; P. decorticans in eastern Bo tswana and in Northern Province.The fruits of P. pa n i culata are more densely sericeous than those of any other species in southern Africa, but the small flowers are the most reliable character with which to distinguish P. p a n i culata.Distribution: P paniculata is a widely distributed species and ranges as far east as India, and as far west and south as Namibia (Figure 6).Habitat: in southern Af rica, it is found in the semi-arid savanna and semidesert areas.Flowering time: February to August.Conservation status: this widespread species is well represented in many herbaria.The first flowers formed are cleistogamous so that the seed bank is rapidly replenished after seeds ger minate.This species is neither rare nor threatened.

5.
Peristrophe Evergreen suffruticose perennial up to 1 m high.Leaves ovate to narrowly elliptic, acuminate, attenuate at base, 26-41(-42) x 8-13(-14) mm, with many multicellular eglandular trichom es and cystoliths; petiole (1.7-)2.0-4.3(^4.6)mm.Inflorescence of monochasial cymes (inflorescence units), (1 or 2)3( 4) umbellately ar ranged, sometimes compounded; inflorescence axis (4.0-)4.2-7.0(-7.4)mm long; longest peduncle of inflo rescence units (2.4->4.1-12.9mm long.Bracts: secondary bracts 2, free, lanceolate, (4.7-)5.1-6.6 x 0.6-1.0(-1.1)mm, pubescent, sessile; tertiary bracts oblanceolate, un equal, larger one (13.6-)13.8-16.1 (-16.4)x (0.4-)0.6-l.l(-1.5) mm, almost tomentose, trichomes multicellular and eglandular.Calyx: tube 1.5 mm deep; lobes, lanceolate, 5.5 mm long, margins membranous and ciliate, with small glandular trichomes on outer surface.Corolla: tube (7.3-) 7.5-9.0mm long, sericeous to tomentose; lip in lower position ovate, (7.1-)8.8(-9.0)x 2.0-2.7 mm; lip in upper position narrowly elliptic, 7.0-8.4(-8.8)It was necessary to transfer P. transvaalensis from the genus Dicliptera as the placentae are inelastic at the base.D iagnostic characters: the closest ally to P. transvaalensis is P. angolensis which has a characteristic inflorescence structure with many inflorescence units contracted into an axil, and an occasional inflorescence unit with a very long peduncle.In contrast, the inflorescence units of P. trans vaalensis are not contracted into the leaf axils.The broader tertiary bracts of these two species serve to sepa rate them from the other species of Peristrophe in southern Africa.P. transvaalensis has a marked vestiture of mul ticellular uniseriate eglandular trichomes, which imparts a grey-green colour to the leaves of living plants; this feature is very unusual in this section of the genus.Dis tribution: Botswana and southwestern Northern Province (Figure 6).Habitat: interface of Acocks's (1988) Sourish M ixed B ushveld andSour B ushveld [now M ixed Bushveld (Van Rooyen &Bredenkamp 1996a) and Waterberg Moist Mountain Bushveld (Van Rooyen & Bre denkamp 1996b)].Flowering time: in the field the plants were beginning to flower in January, and were flowering profusely when revisited in May, whereas the plants in cultivation were still in flower in August of the same year.It appears that the species flowers throughout the year, but more profusely in winter.Conservation status: P. transvaalensis is not a common species and is known from only six gatherings from a relatively restricted area and populations of the plant are sparse.The conservation status of this species requires urgent investigation.
Diagnostic characters: P. cliffordii can be separated from all other species of Peristrophe in southern Africa by its much smaller tertiary bracts (Figure IF) that are covered in rust-coloured glandular and eglandular hairs.In addition, it, P. gillilandiorum and P. transvaalensis have many grey hairs between the ridges on the stems (Figure 2E) which separates them from P. grandibracteata, which has dense white hairs on its stem, and all other species, which have no or very few hairs between the ridges on the stem.The broader leaves separate this species from P. gillilandiorum, and the much smaller tertiary bracts sepa rate it from P. transvaalensis.Distribution: highly local ised near Weipe, west of Messina in the Northern Province (Figure 6).Habitat: on Kalahari sands amongst Colophosperm um mopane in the frost-free Limpopo Valley.Flowering time: autumn and winter.Conservation status: this species is extremely rare and localised.Repeated droughts coupled with browsing do not augur well for it.
Diagnostic characters: the long, narrow leaves of this species are distinctive.For other differences, see under P. cliffordii.Distribution: very localised in southern Zim babwe (on Sentinel Ranch) and Northern Province (mainly on the Farm Schroda) (Figure 7).Habitat: rocky ridges and on deep clay amongst Hyphaene coriacea Gaertn.Flowering time: autumn and winter.Conservation status: P. gillilandiorum is known from only five speci mens and is highly localised.One of the habitats (heavy clay) is being ploughed and planted to cotton, making this species both rare and threatened.
Diagnostic characters: P. cernua is closely allied to P. decorticans as discussed under the latter.P. cernua can be separated from P. paniculata by the larger flowers; from P. transvaalensis by the shorter, narrower tertiary bracts (cf. Figure IE & H), and from P. cliffordii and P. gillilandiorum by the lack of curved eglandular trichomes between the ridges on the stem (Balkwill et al. 1986).Distribution: northern KwaZulu-Natal to Eastern Cape (Figure 7).Habitat: P. cernua is usually found in valley bushveld, and the few specimens that are found out of this veld type are found in coastal forest and thomveld communities.Flowering time: plants have been collected in flower throughout the year, but there is a very marked peak in collections in July and August.Conservation status: P. cernua is widespread and common where it oc curs; it is neither rare nor threatened.
Identification of the type of P. cernua has been con fused by the complexities of Ecklon & Zeyher's collec tions and numbering system.Zeyher collected and distributed material from the Uitenhage area before initi ating a collecting pact with Ecklon (Gunn & Codd 1981).Until then, Ecklon had not collected near Uitenhage and had distributed his specimens through a botanical ex change called Unio Itineraria.Shortly after their pact was initiated, Ecklon collected at Uitenhage and wrote 'A list of plants found in the district of Uitenhage between the months of July 1829 and February 1830', in which Jus ticia capensis is listed under Family 40 (Ecklon 1830).Later he and Zeyher collected in the area together and distributed specimens with either or both their names on the labels and after their pact lapsed, Zeyher collected from there again.Their numbering system was not one with consecutive numbers for consecutive gatherings, but rather the localities were coded in numbers on the labels.Because of the coding system, inconsistencies occurred during the distribution of material so that specimens from the same gathering may have different data on the labels and specimens with the same number, locality and col lecting data may be from different gatherings (Gunn & Codd 1981).Some specimens were distributed with per sonal numbers, some with combined numbers and some with locality numbers.It seems that the number 40 (prob ably the family number for Acanthaceae) is applied to any sheet they collected of P. cernua.Nees (1841) does not cite a herbarium for the type of P. cernua (and cites his own herbarium as well as others in 1847).The type was presumably in his own herbarium and distributed to GZU.There are two sheets at GZU that were collected by Eck lon, distributed through the Unio Itineraria and part of a mixed gathering labelled Justicia capensis, suggesting that they are part of the type gathering-these sheets are pre sumably the holotype.A sheet has been found at STE, bearing the number 556 (now crossed out), the name Jus ticia capensis, and an English version of the locality on the sheets at GZU-this sheet is most likely an isotype.One or both of the sheets presently housed at Kew, must represent the gathenng(s) to which Hooker (1840) referred when offering Zeyher's material for sale.Hooker's listing of the species antedates Nees's description, but as it was not accompanied by a description, diagnosis or precise locality, it is a nomen nudum.It is uncharacteristic of Nees not to quote previous references when they existed, and unlikely that both Nees and Hooker would have inde pendently decided on the same name for the same species, so that it appears that Hooker must have seen Nees's manuscript before it was published, or that Ecklon and/or Zeyher may have circulated the specimens with a name which was taken up by Nees.It would, however, have been characteristic of Nees to cite the source of a name, even if he had taken it up from a label.One of the sheets at Kew bears the name Justicia capensis and a locality in Ecklon's hand.The sheets at Kew bearing the number 40 are undoubtedly those to which Hooker (1840) referred, most likely those to which Nees referred in 1847, but not those to which he referred in 1841.The Kew sheets also bear the name Justicia capensis and this may indicate that these sheets were amongst those to which Ecklon (1830) referred.It is therefore possible that the material at Kew is part of the type gathering (although it has the wrong month on the label).There are two Ecklon & Zeyher specimens and one Zeyher specimen from the type local ity at SAM, all with the number 40, but none of these was collected in October and so are not considered part of the type gathering.
The size of the tertiary bracts and flowers of this spe cies displays clinal variation (Balkwill et al. 1994).The smaller structures are present in Eastern Cape and the larger in KwaZulu-Natal.It is likely that this cline ac counts for a number of the synonyms of P. cem ua.This species sometimes produces white individuals (Balkwill 445) and white populations (Brink 327).As the description of P. krebsii differs from P. cem ua mainly by the white corolla, it is possible that this name was applied to a white form of P. cem ua.Ross (1972) recorded P. hensii from KwaZulu-Natal, but this name refers to a tropical species.
. The species has only been col lected between February and August, and most often in May in southern Africa.The specimens collected in the earlier months are of younger and smaller plants than those collected in the later months.