Introduction

The palaeotropical genus Sarcostemma R. Br. has long been considered as a difficult group in Apocynaceae-Asclepiadoideae-Asclepiadeae. In consequence deviating perceptions are widespread in the taxonomic literature (Liede & Meve 1992; Meve & Liede 1996; Bruyns 2003, 2011; Goyder 2008). In the last 20 years, the generic delimitation within Asclepiadeae saw much phylogenetic enlightenment. Molecular studies using cpDNA markers have convincingly refuted the enlarged concept of Sarcostemma introduced by Holm (1950), which was based mainly on morphological similarities of corona structure. It has been shown that neither Oxystelma R. Br., nor the American genera Funastrum E. Fourn., Pentacyphus Schltr. and Philibertia Kunth are congeneric with core Sarcostemma (Liede & Täuber 2000). All these taxa constitute well circumscribed genera in their own right, occupying widely divergent positions in the Asclepiadeae, so that their corona similarities have to be interpreted as convergences (Meve & Liede 1999).

Both cpDNA and (nuclear) ITS data have shown that Sarcostemma s.s., a group that accommodated smooth, non-warty and non-striate succulents spread over the Old World tropics, characterised by a ‘double’ corona of an outer ring and inner staminal corona parts (as opposed to the ‘simple’ ring-shaped corona of typical Cynanchum), is deeply nested in Cynanchum L., which would be paraphyletic if Sarcostemma was considered an independent genus (Liede & Kunze 2002; Liede & Täuber 2002). Sarcostemma was demonstrated to constitute a subclade of the large clade formed by all Madagascan stem succulent Cynanchinae, including the small genera Folotsia Costantin & Bois, Karimbolea Desc. and Platykeleba N. E. Br. This clade, again, constitutes the top clade of a Cynanchum clade consisting exclusively of Madagascan species. Meve & Liede (2002) concluded that the African mainland must have been originally colonised by Madagascan precursors of Sarcostemma viminale (L.) L. It is, however, still unclear, whether other regions of the distribution areas of Sarcostemma taxa, e.g. India, have been reached via mainland Africa (and Arabia), or independently and directly from Madagascar by long distance dispersal, and whether redistributions from these secondary areas back to Madagascar have taken place as well.

Following these results, the other small succulent genera have been merged with Cynanchum and their species have been transferred formally to Cynanchum by Liede & Meve (2001). However, for Sarcostemma, only those species needed for Flora treatments have hitherto been transferred to Cynanchum (Madagascar, Liede & Meve (2001); tropical Africa, Goyder 2008).

A wide variety of commonly used molecular markers has hitherto failed to produce any differences between morphologically well distinguishable units in Sarcostemma (Liede-Schumann, unpubl. results), leading to the conclusion that the genetic distances between the different Sarcostemma taxa are extraordinarily low. For this reason, the traditional taxonomic concept using morphological and karyological characters for taxon delimitation is followed to the present day (cf. Liede & Meve 1989, 1992, 1993, 1995; Meve & Liede 1996, 1997) and will also be applied here. Our taxonomic work in Sarcostemma is based on the concept that species are characterised by differences in floral structure, whereas subspecies are characterised by differences at least in habit (Liede & Meve 1993). Following this concept, a number of circumscribable entities have been described as new taxa, the last by Liede-Schumann & Meve (2005). The many subspecies of Cynanchum viminale (L.) Bassi typically represent ‘discernable natural groups’ which are often identical with a specific ‘geotype’, e.g., a distinct regional representative of the C. viminale complex (cf. Liede & Meve 1993; Meve & Liede 1996; Liede-Schumann & Meve 2005). However, this concept is still somewhat vague since recognition of the different subspecific taxa needs some experience in that group. Also, it is not possible to allocate a subspecific taxon name to every collection of C. viminale s.l. Specimens from western and central Africa are especially poor in specific characters that would allow classification beyond attribution to C. viminale s.l. With the new combinations proposed in the following, the genus Sarcostemma ceases to exist as an accepted taxon.

New combinations

Cynanchum arabicum (Bruyns & P. I. Forst.) Meve & Liede, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120418-1

Sarcostemma arabicum Bruyns & P. I. Forst., Edinburgh J. Bot. 48: 333 (Bruyns & Forster 1991). Type: Saudi Arabia, Hijila, 15 km E of Abha, Nasher IH 137 (holotype E).

Cynanchum areysianum (Bruyns) Meve & Liede, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120419-1

Sarcostemma areysianum Bruyns, S. African J. Bot. 77: 802 (2011). Type: Yemen, Abyan Governorate, Jabal Areys, 1,400 – 1,700 m alt., Jan. 2007, Bruyns 10287 (holotype B; isotype E).

Notes. First counts of the chromosome number of Cynanchum areysianum are here provided: 2n = 22. Vouchers: YEMEN. Tawr al-Bahah, 13°11'N 44°18'E, c. 250 m alt., 21 Nov. 2000, Lavranos, Mies & McCoy 31335 (UBT); above Khadifut in the mountains of Ra’s Fartaq, 15°39'N, 52°12'E, c. 150 m alt., 15 Nov. 2000, Lavranos, Mies & McCoy 31313 (UBT).

For further remarks see below ‘Notes on novelties from southern Yemen’.

Cynanchum brevipedicellatum (P. I. Forst.) Liede & Meve, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120420-1

Sarcostemma brevipedicellatum P. I. Forst., Austral. Syst. Bot. 5: 59 (Forster 1992). Type: Australia, Queensland, Gregory North Distr., Toko Tange, Glenormiston Stn., Dec. 1982, Forster 1420 (holotype BRI; isotypes AD, CANB, CBG, DNA, K!, MEL, PERTH, PRE, QRS).

Cynanchum daltonii (Decne. ex Webb) Liede & Meve , comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120421-1

Sarcostemma daltonii Decne. ex Webb, in Hook., Niger Flora: 149, t. 14. (Hooker 1849). Type: Cape Verde Islands, J. D. Hooker 1843 (lectotype K!, designated here).

Notes. This endemic of the Cape Verde Islands has been placed as a synonym of Sarcostemma viminale subsp. thunbergii (G. Don) Liede & Meve, a decision influenced by the tetraploid genomes found for both taxa (Liede & Meve 1993). However, we have since investigated more material of the complex and realised that S. daltonii should be separated from Cynanchum viminale based on the unique character composition of a non-twining, trailing to arching habit, long and short laterals (peduncles; cf. Gonçalves 2002), and a chromosome number of 2n = 44. Regarding floral characters, C. daltonii is morphologically most close to C. viminale subsp. viminale, though the follicles are much stouter. Also, subsp. viminale always possesses a diploid genome with 2n = 22 chromosomes (Liede & Meve 1995; Malla et al. 1978, Meve, unpubl. data).

Cynanchum forskaolianum (Schult.) Meve & Liede, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120422-1

Sarcostemma forskaolianum (as “forskålianum”) Schult. in Roem. & Schult., Syst. Veg. (ed. 15) 6: 117 (Schultes 1820). Type: Saudi Arabia, between Mudhaylif and Mahayl on the Jeddah-Gizan road, 500 ft., Collenette 1977 (neotype K!, designated by Meve & Liede (1996)).

Cynanchum mulanjense (Liede & Meve) Liede & Meve, stat. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120423-1

Sarcostemma mulanjense Liede & Meve, Novon 2: 223 (1992). Type: Malawi, Mulanje (Mt), outer slopes at the Chitakale stream, west branch, Chapman & Chapman 6892 (holotype MO!).

Cynanchum viminale (L.) L. subsp. mulanjense (Liede & Meve) Goyder (2008: 417).

Notes. Goyder (2008) considered Cynanchum mulanjense a subspecies of C. viminale. Applying the concept of species and subspecies as outlined above, namely, using variation in floral morphology as a criterion of species delimitation, C. mulanjense deserves specific status. It is characterised by nodding flowers, narrow, slightly twisted creamy petals and an elongated gynostegium, although the latter character breaks down in central and northern Malawi (cf. Goyder 2008). With this character combination, C. mulanjense seems to be more closely related to C. oresbium (Bruyns) Goyder than to C. viminale. These two species share both floral similarities and a preference for similar habitats, rocky outcrops in Malawi and Mozambique, respectively.

Cynanchum pearsonianum Liede & Meve, nom. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120424-1

Sarcostemma pearsonii N. E. Br., Bull. Misc. Inform., Kew 1913: 301 (Brown 1913), non Cynanchum pearsonii N. E. Br. (Brown 1914: 18). Type: Namibia, Great Karasberg, on stony plains SW of Krai Kloof, 1,600 m alt., 19 Jan. 1913, Pearson 8460 (holotype K!; isotype BOL!).

Notes . Sarcostemma pearsonii, a species that is confined to very dry and stony habitats in Namibia and NW South Africa, was neglected for a long time, although its erect, shrubby habit and the small and twisted corolla lobes make it one of the most easiest species to identify (Liede & Meve 1989). Brown (1914) assigned the name Cynanchum pearsonii to a shrubby species with small leaves, which is nowadays treated as a synonym of Cynanchum meyeri (Decne.) Schltr. (Liede 1993).

The necessary nomen novum proposed here, tries to resemble as much as possible that of the name Sarcostemma pearsonii, under which this species has started to become a well-known taxon within the last 20 years.

Cynanchum sarcomedium Meve & Liede, nom. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120425-1

Sarcostemma intermedium Decne., in DC., Prodr. 8: 538 (Decaisne 1844), non Cynanchum intermedium N. E. Br. (Brown 1908: 747). Type: India ([‘Peninsula Ind. orientalis’]), Herb. Wight 1556 (holotype P, not traced; isotype G-DC! (2 sheets)).

Notes . Medium-thick stems, small and acute corona lobes and an elongated style-head make this species readily distinguished from the other representatives of the Cynanchum viminale relationship in India. But since N. E. Brown (1908) used the name Cynanchum intermedium for a leafy South African twiner (of an uncertain status), a replacement name for Sarcostemma intermedium in Cynanchum became necesssary, and the name Cynanchum sarcomedium is proposed here.

Cynanchum socotranum (Lavranos) Meve & Liede, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120426-1

Sarcostemma socotranum Lavranos, Natl. Cact. Succ. J. 27(2): 37 (1972). Type: Yemen, Socotra, Hama-di-roh, on vertical limestone faces, c. 350 m alt., April 1967, Smith & Lavranos 309 (holotype K!).

Cynanchum viminale (L.) Bassi (1768: 17) subsp. viminale

Notes . The protologue of Cynanchum (Euphorbia) viminale (Linnaeus 1753) refers to an illustration of a plant from Egypt (Felfel Tavil) as the type and states explicitly ‘Habitat in Africae maritimis’. However, Indian taxa such as Sarcostemma acidum (Roxb.) Voigt and S. brevistigma Wight & Arn. were synonymised under S. viminale subsp. viminale by Ali (1983), based on material from Pakistan. Later, this synonymy has been adopted by Jagtap & Singh (1999) for India. Other regional Floras, such as the Flora of China and the Flora of Nepal (www.eFloras.org) have not followed this usage and recognised S. acidum and S. brevistigma as separate species. The present authors follow the view of Ali (1983) and consider these two species as synoynyms of C. viminale subsp. viminale, thus extending the area of the latter from West Africa and Egypt to India.

Cynanchum viminale (L.) Bassi subsp. australe (R. Br.) Meve & Liede, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120428-1

Sarcostemma australe R. Br., Prodr.: 463 (Brown 1810). Type: Australia, South Australia, Bird Island, Petral Bay, Isle St. Francis, 3 Feb. 1802, R. Brown (Iter Australiense 2872) (lectotype BM!, designated by Forster (1992)).

Sarcostemma viminale subsp. australe (R. Br.) P. I. Forst., Austral. Syst. Bot. 5: 64 (Forster 1992).

Cynanchum viminale (L.) Bassi subsp. brunonianum (Wight & Arn.) Meve & Liede, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120429-1

Sarcostemma brunonianum Wight & Arn., in Wight, Contr. Bot. India: 59 (1834). Type: India, [‘Peninsula Ind. orientalis’], Cody Cally, village of Autcherroccum, Herb. Wight 1557 (lectotype G-DC! (No. G00301010), designated here; isolectotypes E, G-DC! (No. G00136527)).

Sarcostemma viminale (L.) R. Br. subsp. brunonianum (Wight & Arn.) P. I. Forst. (Forster 1992: 63).

Cynanchum viminale (L.) Bassi subsp. orangeanum (Liede & Meve) Liede & Meve, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120430-1

Sarcostemma viminale subsp. orangeanum Liede & Meve, Bot. J. Linn. Soc. 112: 70 (1993). Type: South Africa, Northern Cape, 11.4 km E Kuruman, Liede & Meve 579 (holotype K!; isotype MSUN!).

Cynanchum viminale (L.) Bassi subsp. stocksii (Hook. f.) Meve & Liede, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120431-1

Sarcostemma stocksii Hook. f., Fl. Brit. India 4: 27 (Hooker 1883). Type: Pakistan, Sind, Stocks 509 (holotype K!).

Sarcostemma viminale subsp. stocksii (Hook. f.) Ali (1983: 31).

Notes . This subspecies seems to be indistinguishable from Cynanchum viminale subsp. viminale except for the stouter follicles that are much more divaricate and taper only insignificantly towards base and apex. The seeds are also thicker, and wingless, dark brown and rugulose (cf. Liede 2002: 235). A very good illustration of the type of Sarcostemma stocksii is available: Hooker (1852, plate 861), published under ‘S. brachystigma’, an erroneous name for S. brevistigma, now a synonym of C. viminale subsp. viminale (http://www.bio.uni-bayreuth.de/psysl/hooker/hooker.php?pg=0861).

Cynanchum viminale (L.) Bassi subsp. thunbergii (G. Don) Liede & Meve, comb. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120432-1

Sarcostemma thunbergii G. Don, Gen. Hist. 4: 156 (1838). Type: South Africa, Cape Province, Worcester, Karoo Botanic Garden, Bayer 68 (neotype NBG!, designated by Liede & Meve (1993).

Sarcostemma viminale subsp. thunbergii (G. Don) Liede & Meve (1993: 10).

Cynanchum viminale (L.) Bassi subsp. welwitschii (Hiern) Liede & Meve, comb. et stat. nov.

http://www.ipni.org/urn:lsid:ipni.org:names:77120433-1

Sarcostemma welwitschii Hiern, Cat. Afr. Pl. 1: 689 (1898). Type: Angola, Huilla, base of Morro de Lopollo, Welwitsch 4261 (holotype BM!; isotypes BR!, C).

Notes . This element described from southern Angola is still quite poorly known, though, undoubtedly, it is a member of the Cynanchum viminale complex that extends into Namibia. It is a remarkably glabrous, vigorously tangling scrambler and climber, which can produce voluminous plants. The inflorescences are mostly lateral and sessile, rarely on short laterals (peduncles) or terminal. The most conspicuous character seems to be the rather low number (two to five) of flowers per inflorescence. C. viminale subsp. welwitschii is a tetraploid with a chromosome number of 2n = 44, as the single count available suggests, obtained from material from northernmost Namibia (Voucher: Ruacana, Albers et al. 3617, UBT). With its polyploidy, subsp. welwitschii is distinguished from C. viminale subsp. stipitaceum (Forssk.) Meve & Liede and subsp. viminale, the taxa otherwise having most similarities with subsp. welwitschii. From the other tetraploid taxon of southern Africa, C. viminale subsp. thunbergii, C. viminale subsp. welwitschii differs in habit, and in stems that are thinner, softer and green instead of grey-green.

Figueiredo & Smith (2009) included two taxa in their list of succulents of Angola; Sarcostemma welwitschii Hiern and S. viminale subsp. viminale. These authors neither mention the second Angolan taxon described by Hiern, S. andongense Hiern, nor S. viminale subsp. stipitaceum into which Liede-Schumann & Meve (2005) have sunk S. andongense. More material, especially living plants, from Angola and adjacent regions need to be studied before a final taxonomic conclusion can be made.

Notes on novelties from southern Yemen

A floristic inventory by J. Lavranos, T. McCoy and B. Mies in southern Yemen in November 2000 yielded some additional morphological variants of stem-succulent Cynanchum of the Sarcostemma subgroup. Four living cuttings taken during this excursion were transferred into cultivation in the greenhouse of the Department of Plant Systematics, University of Bayreuth. All four accessions flowered in the following years. One accession from near Lawdar (Lavranos, McCoy & Mies 31372, UBT) was found to represent a rose- to white-flowered form of Cynanchum viminale subsp. stipitaceum (Fig. 1A), another one from Ra’s Fartaq (Lavranos, McCoy & Mies 31363, UBT) is vegetatively similar to subsp. stipitaceum, florally a mixture of this subspecies and C. areysianum or C. arabicum (Fig. 1B), but possesses a reddish to red-coloured corolla. It could represent a natural hybrid. Hybrids, however, are rare in Cynanchum, and molecular analysis to support this assumption has yet to be done. Finally, two accessions, one from Tawr al-Bahah (Lavranos, Mies & McCoy 31335, UBT), and one from above Khadifut in the mountains of Ra’s Fartaq (Lavranos, Mies & McCoy 31313, UBT), are vegetatively comparable to C. arabicum, but florally similar to the Socotran endemic C. socotranum (Figs 1B – E, 2). For this element we therefore prepared a new species description but later had to recall it from the manuscript when Bruyns (2011) described Sarcostemma areysianum, which exactly matches our new species. In addition to the brief differential diagnosis given by Bruyns (2011) we made further observations on the status of the three closely related taxa C. arabicum, C. areysianum and C. socotranum: Vegetatively, C. areysianum cannot be directly distinguished from C. arabicum, with its erect to arching (non-twining), thick (usually 6 – 7 mm diam.), greyish green and stout stems (Fig. 1D). Usually the stems and especially the nodes of C. areysianum are even thicker than in C. arabicum and additionally equipped with a corky ring; the internodes are mostly shorter (cf. Fig. 1D). C. areysianum therefore is really a stout plant totally different in habit from the delicate C. socotranum, whose stems are around 2.5 – 3.5 (– 4) mm thick and often constricted at the nodes (Fig. 1D, central shoot). Flowers in C. areysianum appear terminally or on short lateral peduncles (Fig. 1B, top, 1 C), as can be seen in most plants of C. arabicum, whereas the flowers in C. socotranum normally appear laterally in a (sub)sessile manner (Fig. 1E). With C. socotranum, however, C. areysianum shares the rather small flowers with membraneous, oblong, suberect to erect corolla lobes, although the flowers are a bit larger and the corolla lobes are much more twisted as in C. socotranum, and also creamish brown to rose instead of greenish yellow. Other shared features include the thickish, rounded and more or less erect staminal corona lobes. However, in contrast to C. socotranum the corona lobes are inflated, uneven and much surpassing the style-head (Fig. 2C). All three species lack a complete outer coronal ring formed by interstaminal and (outer) staminal parts as found in most other ‘Sarcostemma’ taxa. Instead, the ‘ring’ is not continuous and broken into ‘five plus five’ separated fringes or scales. Very significant for C. areysianum, one mostly lobular and emarginate scale is fused to the basal back of each staminal corona lobe (cf. Fig. 2B). This corona structure is unique in Cynanchum. In addition, the staminal corona lobes are of very delicate tissue and inflated. They are also larger, as in the two related species, and of an uneven surface and irregular outline. Furthermore, they cover and greatly exceed the conspicuously large, massive and strictly conical style-head (Fig. 2E). Details of the gynostegium look very similar to those found in C. socotranum, although it is larger: the gynostegium is clearly higher than broad (but as high and as broad as in C. arabicum, cf. Bruyns & Forster 1991: Fig. 1E), the guide rails are strictly pointed (not forming a ‘mouth’ as in C. arabicum), and the fleshy base of the gynostegium leaves considerable open space beneath the guide rail entrance (whereas the mouth/base of each guide rail is directly fused to the gynostegium without leaving an open space beneath in C. arabicum). In contrast to both species, the anther appendages are much longer than wide, triangular-lanceolate and rather acute, while they are triangular-deltate and more or less as long as wide in C. socotranum and C. arabicum. Finally, the style-head (Fig. 2E) is larger and higher than in the two related species, whereas the style itself is elongated up to 0.8 mm in length (not shown in Fig. 2), and reminiscent of C. socotranum. In C. arabicum the style is considerably shorter. C. areysianum indeed shares a number of significant features, especially of the flowers, with C. socotranum, which is therefore clearly identified as its sister-species.

Fig. 1
figure 1

A Cynanchum viminale subsp. stipitaceum, Yemen; B Cynanchum areysianum (left) together with Cynanchum arabicum, Yemen (right); C Cynanchum areysianum; D shoots of Cynanchum areysianum, C. socotranum and C. arabicum (from left to right); E Cynanchum socotranum. A from Lavranos et al. 31372; B from Lavranos et al. 31313 (left), Radcliffe-Smith & Henchie 4624 (right); C from Lavranos et al. 31335; D from Lavranos et al. 31335 (left), Mies et al. 1503 (middle), Radcliffe-Smith & Henchie 4624 (right); E from Mies et al. 1503 (all in UBT). photos: ulrich meve.

Full size image
Fig. 2
figure 2

Cynanchum areysianum. A flower bud; B flower in lateral view, with two petals removed; C gynostegium and corona, with parts of corona removed; D pollinarium; E style-head, with one pollinarium in natural position. All from Lavranos, Mies & McCoy 31335 (UBT). drawn by ulrich meve.

Full size image