Abstract

Based on morphological characters a phylogenetic analysis of 97 African and Madagascan Cynanchum species is conducted. Of these, 93 are members of sect. Cynanchum and their relationships are explored using sect. Rhodostegiella, sect. Roulinia, and Pentarrhinum as outgroups. Stable clades of two to five taxa are identified, and their relationships to the outgroups and to the remaining four African Cynanchum species are explored using Tylophora as outgroup. All analyses are performed with all characters considered as unordered and under the assumption that some characters are ordered. Monophyly of the leafless succulent taxa is tested by excluding characters directly related to succulence. The Madagascan taxa as a whole are not monophyletic. The small Madagascan succulent genus Folotsia should probably be included in Cynanchum. Three African species are excluded from sect. Cynanchum and transferred to the small African genus Pentarrhinum as P. balense, P. gonoloboides, and P. somaliense. The position of C. ledermannii, of which only incomplete material is available, cannot be resolved at present. After a long period of neglect, research in the Apocynaceae-Asclepiadoideae has intensified in recent years around the globe, mainly in the course of major flora projects (e.g., Li et al. 1995, for China; Forster 1996, for Australia; Morillo, in press, for the Venezuelan Guyanas). Parallel to these studies, analyses of important character complexes have been undertaken (e.g., Liede and Kunze 1993; Liede and Weberling 1995; Liede 1996a), and a review of tribal assignments of all genera has been presented (Liede and Albers 1994). However, phylogenetic studies have been few and mainly restricted to the subtribal and generic level (Liede 1994a, 1996b, 1996d). There are probably three main reasons for this lack of phylogenetic studies on species level. First, the subfamily contains very few mediumsized genera; genera are either large (more than 100 species, e.g., Hoya, Marsdenia, Tylophora) or very small (less than five species). In all the larger genera, there is still a vast number of species insuffiently known. Second, differences between species are rather subtle, difficult to quantify and polarize. Third, as so little is known about generic relationships, identification of suitable outgroups presents a serious problem. By examining part of the large genus Cynanchum, this paper explores the possibilities and limits of a phylogenetic study on species level in the Asclepiadoideae. In the most likely highly derived tribe Asclepiadeae (Liede 1996a, 1997b), which is characterized by hanging pollinia, Cynanchum constitutes the largest genus. It is usually defined by a corona of gynostegial origin consisting of at least partly fused staminal and interstaminal parts. Due to this rather vague generic definition, however, circumscription and size of Cynanchum vary widely between workers. In the narrow sense, Cynanchum comprises ca. 200 species living mainly in secondary growth in the semi-arid tropics. The genus comprises a wide array of growth forms; most species are medium-sized leafy twiners sprouting from a woody rootstock; however, erect herbs and succulent twiners with leaves reduced to papery scales are also found. After completing a revision of Cynanchum on the African mainland (31 species; Liede 1993a, 1994b, 1996e) and Madagascar (66 species; Liede 1993b, 1996c; Meve and Liede 1994; Liede and Meve 1996) this paper explores relationships of the species involved by phylogenetic analysis. Regarding the choice of ingroup taxa, two questions need to be addressed: 1) why not examine the whole genus?, and 2) are the species included really monophyletic? Analysis of the whole genus Cynanchum is premature because its subgeneric delimitation is still subject to debate. Liede (1996b) found conclusive morphological and chemical evidence to separate Vincetoxicum from Cynanchum, and Liede and Meve (1997), following Stevens (1988), argued to re-established Metastelma as a separate genus, distinguished from Cynanchum by its bearded corolla lobes and corona of free staminal parts. Inclusion of Ditassa in Cynanchum has been recognized as untenable even by its earlier supporters (compare Morillo 1986, 1989). Liede (1997a) sug-

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call