PHYLOGENETIC RELATIONSHIPS IN SUBTRIBE ASCLEPIADINAE (APOCYNACEAE: ASCLEPIADOIDEAE)1 David Goyder,2 Ashley Nicholas,3 and Sigrid Liede-Schumann4 ABSTRACT A first approximation of evolutionary relationships within the Asclepiadinae (Apocynaceae: Asclepiadoideae) is inferred from two DNA data sets, from the nuclear ITS region, and from plastid trnT-L and trnL-F spacers and trnL intron. Both the subtribe as a whole and the phylogenetic radiation of largely herbaceous genera around Asclepias are monophyletic and well supported. While resolution within the Asclepias complex is limited, the lineages identified demonstrate that current generic delimitation is unsatisfactory. Asclepias should be either restricted to New World members of the subtribe or expanded to encompass the entire radiation. In the African species, it is shown that, although the corona is taxonomically important for recognizing species, it can be misleading as an indicator of phylogenetic relationship. Vegetative similarities often prove a more reliable guide, but more detailed studies are needed to refine these lineages and to detect morphological characters to aid in their practical recognition. Key words: Apocynaceae, Asclepiadaceae, Asclepiadinae, phylogenetics, ITS, trnT-L, trnL-F, trnL. Generic delimitation has long been problematic in the Asclepiadinae (Apocynaceae: Asclepiadoideae) sensu Liede (1997). Brown (1902: 299) commented in the Flora of Tropical Africa: ‘‘Undoubtedly Xysmalobium R. Br., Asclepias L. and Schizoglossum E. Mey. are but artificial divisions of one natural genus, since they cannot be separated by characters that do not break down at some point.’’ Nevertheless, he went on to assign species to the three genera on the basis of the form of the corona—cucullate in Asclepias, dorsally flattened in Schizoglossum, and thickened or laterally compressed in Xysmalobium. Such an artificial arrangement separated apparently allied species, such as A. lisianthoides (Decne.) N. E. Br., S. carsonii (N. E. Br.) N. E. Br., and X. ceciliae N. E. Br., all currently referred to Glossostelma Schltr. (Goyder, 1995). Subsequent morphological studies have proposed the use of non-coronal characters to reunite some of these natural groups. Bullock (1952, 1953a, b, 1955, 1956, 1957, 1961) started the process with a number of groups in tropical Africa but never articulated his reasoning beyond a statement that Asclepias should refer only to New World species and that he proposed to ‘‘exclude Asclepias Linn. from Africa, except as an adventive, and to divide further the species hitherto assigned to it into other genera’’ (1952: 406). Some of Bullock’s generic concepts have been developed by later workers—Aspidoglossum E. Mey., Miraglossum Kupicha, and Schizoglossum (Kupicha, 1984); Aspidonepsis Nicholas & Goyder (Nicholas & Goyder, 1992); Glossostelma (Goyder, 1995); Gomphocarpus R. Br. (Goyder, 2001b; Goyder & Nicholas, 2001); Margaretta Oliv. (Goyder, 2005); Pachycarpus E. Mey. (Smith, 1988; Goyder, 1998a); Stathmostelma K. Schum. (Goyder, 1998b); and Trachycalymma (K. Schum.) Bullock (Goyder, 2001a). In southern Africa, Bullock’s unpublished views on Xysmalobium sensu N. E. Br. (1902), as deduced from annotations of herbarium collections housed at K, formed the basis of a postgraduate thesis (Langley, 1980), but the generic changes proposed therein were never formally published. Likewise, unpublished theses by Nicholas (1982, 1999) provide a framework for the Asclepias complex in southern Africa, but a satisfactory synthesis of tropical and southern African taxa is still lacking. The present contribution is designed firstly to test the monophyly of the Asclepias generic complex, and secondly to see what lineages can be detected within this assemblage. 1 We are grateful to the directors and curators of K, MSUN, NH, NU, UBT, UDW, ULM, and UPS for access to their collections. Sigrid Liede-Schumann would like to acknowledge Ulrich Meve for his permanent support and Angelika Täuber for technical assistance. Jeff Ollerton kindly provided some of the ingroup samples. 2 Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3AB, United Kingdom. d.goyder@kew.org. 3 School of Biological and Conservation Science, Westville Campus, University of KwaZulu-Natal, Private Bag X54001, Durban 4000, South Africa. Nicholasa@ukzn.ac.za. 4 Department of Plant Systematics, University of Bayreuth, 95444 Bayreuth, Germany. Sigrid.Liede@uni-bayreuth.de. ANN. MISSOURI BOT. GARD. 94: 423–434. PUBLISHED ON 27 JULY 2007. 424 Annals of the Missouri Botanical Garden MATERIAL AND METHODS (trnT-L spacer and ITS) from our own material for A. syriaca, but not for A. tuberosa. The entire nuclear internal transcribed spacer region (ITS) including 5.8S of ribosomal DNA (rDNA) was amplified using the flanking primers ITS4 and ITS5 following a modified protocol based on Baldwin (1992) and described in Meve and Liede (2001). ITS sequences of 68 species were obtained for the present study, with one sequence studied previously. An interpretable ITS sequence could not be obtained for Trachycalymma pseudofimbriatum Goyder. Published or unpublished ITS sequences by the last author are available for all outgroup taxa except for Oxystelma bornouense. However, only those of O. esculentum and Solenostemma arghel could be aligned to the Asclepiadinae sequences, albeit with difficulties. TAXA Sixty-five species in 21 out of the 24 recognized Asclepiadinae genera and one undescribed genus constitute the ingroup (Table 1). Two accessions from different localities were analyzed in each of five species (Asclepias crispa P. J. Bergius, A. gibba (E. Mey.) Schltr., Aspidoglossum heterophyllum E. Mey., Gomphocarpus physocarpus, and Kanahia laniflora (Forssk.) R. Br.). In Asclepias, the only genus represented in both the Old World and the New World, 10 Old World species and six New World species were included. Marsdenia sylvestris (Retz.) P. I. Forst. (Marsdenieae) was used as the most distant outgroup. As closer outgroups, species were selected from the closely related subtribes of the Asclepiadeae, the Tylophorinae, and Cynanchinae, including Glossonema Decne. and Odontanthera Wight. Exemplars from the American Metastelminae–Gonolobinae–Oxypetalinae assemblage, sister to the Asclepiadinae– Cynanchinae–Tylophorinae clade (Rapini et al., 2003; Liede-Schumann et al., 2005; Goyder, 2006), were not included, because studies by Liede (2001) and Liede and Täuber (2002) had shown that sequences from this taxonomic group added numerous long indels but presented little additional information as to the directionality of site changes. Instead, taxa previously found not to fit in any of the recognized subtribes were included: Eustegia minuta (L. f.) N. E. Br. (Rapini et al., 2003); Calciphila galgalensis (Liede) Liede & Meve (Liede & Kunze, 2002; Liede-Schumann & Meve, 2006); Oxystelma bornouense R. Br. and O. esculentum (L. f.) Schult. (Liede & Täuber, 2000); and Solenostemma arghel (Delile) Hayne (Liede et al., 2002a—as S. oleifolium (Nectoux) Bullock & E. A. Bruce). DNA EXTRACTION AND PCR DNA was isolated from fresh or dried leaf tissue using the CTAB protocols (Doyle & Doyle, 1987). Polymerase chain reaction (PCR) primers and protocol for the plastid trnT-L and trnL-F spacers (primers ‘‘a,’’ ‘‘b’’ and ‘‘e,’’ ‘‘f,’’ respectively) and the trnL intron (primers ‘‘c’’ and ‘‘d’’) correspond exactly to Taberlet et al. (1991). Sequences were obtained on an ABI Prism Model 310 Version 3.0 sequencer (Applied Biosystems, Foster City, California, U.S.A.). Of the 80 accessions, 65 have been sequenced for this study; the remaining 15 were previously deposited at European Molecular Biology Laboratory (EMBL; Table 1). The trnL intron and trnL-F spacer of Asclepias syriaca L. and A. tuberosa L. were taken from EMBL, and we were able to supplement the missing partial sequences DATA ANALYSIS Sequences were pre-aligned with Sequence Navigator Version 1.0.1 (Perkin-Elmer, Inc., Waltham, Massachusetts, U.S.A.) and manually adjusted. Separate indel coding was not performed, because the number of phylogenetically informative indels was very low in the chloroplast DNA (cpDNA) data set, while the indels could not be coded unambiguously in the ITS data set. Both data sets were analyzed separately and in combination. Phylogenetic analysis and tests for clade support were performed, using PAUP version 4.0b8 (PPC) (Swofford, 1998). Phylogenies were generated using Fitch parsimony as implemented in PAUP* employing heuristic searches, with 1000 replicates, random stepwise addition, MULPARS off, and tree-bisection-reconnection (TBR) branch swapping. The resulting trees were then used as starting trees for a second round of search with MULPARS on. Our search strategy aimed at finding as many different islands of trees as possible. Sets of equally parsimonious trees recovered from each analysis were summarized by strict consensus. Decay indices (Bremer, 1988; Donoghue et al., 1992) and bootstrap values (Felsenstein, 1985) derived from 1000 replicates (maxtrees set to 100) were calculated as measures of support for individual clades. Decay analyses were performed with AutoDecay 4.0 (Eriksson, 1998) in combination with the reverse constraint option of PAUP*. A partition homogeneity test as implemented in PAUP* was conducted (10 addition sequence replicates) to test the congruence of the cpDNA and the ITS data sets. RESULTS The sequence alignment is available from the authors and from TreeBASE (study accession number Tribe, subtribe Marsdenieae Asclepiadeae, unplaced Asclepiadeae, Cynanchinae Taxon Marsdenia sylvestris (Retz.) P. I. Forst. Calciphila galgalensis (Liede) Liede & Meve Eustegia minuta (L. f.) N. E. Br. Oxystelma bornouense R. Br. DNA voucher trnL intron trnL-F spacer AJ402137 (F) AJ402142 (F) unpubl. Somalia: Bari, Thulin & Warfa 6205 (UPS) AJ492337 (B) AJ492338 (B) AJ492339 (B) [AJ492756 (B)] AJ410205 (C) AJ290883 (A) AJ410206 (C) AJ290882 (A) AJ410207 (C) AJ290881 (A) unpubl. — AJ290884 (A) AJ290885 (A) AJ290887 (A) AM396877 AJ428831 (D) AJ290845 (A) AJ428832 (D) AJ290846 (A) AJ428833 (D) AJ290847 (A) AM396878 [AJ320444 (E)] AJ428804 (D) AJ428805 (D) AJ428806 (D) [AJ492803 (B)] AJ428813 (D) AJ428814 (D) AJ428815 (D) [AJ492809 (B)] AJ290915 (A) AJ290916 (A) AJ190917 (A) [AJ320460 (E)] AM295719 AM295720 AM295721 AM396879 AM295724 AM295723 AM295722 AM396880 AM295725 AM295726 AM295727 AM396881 AM295730 AM295729 AM295728 AM396882 AM295731 AM295732 AM295733 AM396883 AM295734 — — AM396884 South Africa: Bruyns 4357 (K) Kenya: Tsavo East, Liede & Newton 3210 (UBT) Oxystelma esculentum (L. f.) Schult. Egypt: Elephantine Island, cult. Bayreuth ex Shirley s.n. (UBT) Solenostemma arghel (Delile) Hayne Egypt: Aziz s.n. (MO) Cynanchum ellipticum (Harv.) South Africa: Eastern Cape, Liede 2933 R. A. Dyer (UBT) Glossonema boveanum (Decne.) Kenya: Loyengalani, Liede & Newton Decne. 3239 (ULM) Odontanthera radians (Forssk.) Yemen: North, Müller-Hohenstein & Deil D. V. Field 1967 (UBT) Tylophora flexuosa R. Br. Philippines: Laguna, Liede 3252 (UBT) Asclepiadeae, Tylophorinae Asclepiadeae, Asclepias albens (E. Mey.) Schltr. Asclepiadinae Asclepias brevipes (Schltr.) Schltr. Asclepias crispa P. J. Bergius, acc. 1 Asclepias crispa P. J. Bergius, acc. 2 Asclepias cucullata (Schltr.) Schltr. Asclepias curassavica L., acc. 1 Asclepias curassavica L., acc. 2 Asclepias fascicularis Decne. South Africa: Eastern Cape, Nicholas 2813 (UDW) South Africa: Free State, Nicholas 2866 (UDW) South Africa: Eastern Cape, Nicholas 2814 (UDW) South Africa: Eastern Cape, Nicholas 2823 (UDW) South Africa: KwaZulu-Natal, Nicholas 2842 (UDW) Venezuela: Mérida, Liede & Meve 3316 (UBT) U.S.A.: Potgieter & Cahilly 243 (NY) U.S.A.: ex hort UC Davis — AM295735 AF102381 (G) AM295736 AF214156 (G) AM295737 ITS Goyder et al. Subtribe Asclepiadinae trnT-L spacer AJ402118 (F) Cameroon: Mokolo, Meve 919 (UBT) Volume 94, Number 2 2007 Table 1. Voucher specimens and EMBL accession numbers for the taxa investigated. Sequences already published: A 5 Liede and Täuber (2000); B 5 Liede and Kunze (2002); C 5 Liede (2001); D 5 Liede et al. (2002a); E 5 Liede et al. (2002b); F 5 Meve and Liede (2001); and G 5 Potgieter and Albert (2001). ITS sequence numbers in square brackets refer to published sequences that could not be fitted into the ITS alignment. — AM396885 425 426 Table 1. Continued. Tribe, subtribe Taxon Asclepias gibba (E. Mey.) Schltr., acc. 1 Asclepias gibba (E. Mey.) Schltr., acc. 2 Asclepias linaria Cav. Asclepias longifolia Michx. Asclepias praemorsa Schltr. Asclepias stellifera Schltr. Asclepias syriaca L. Fanninia caloglossa Harv. Glossostelma ceciliae (N. E. Br.) Goyder South Africa: Eastern Cape, Nicholas 2824 (UDW) South Africa: Eastern Cape, Nicholas 2838 (UDW) Mexico: Michoacán, Conrad 9305 (UBT) U.S.A.: Wyatt s.n. (UGA) South Africa: Transkei, Ollerton 214 (NU, spirit) South Africa: KwaZulu-Natal, Nicholas 2841 (UDW) ex hort. Münster, in cult. Münster (MSUN) U.S.A.: Potgieter & Cahilly 242 (NY) South Africa: KwaZulu-Natal, Ollerton 255 (NU, spirit) Tanzania: Nkansi District, Bidgood et al. 2583 (K) South Africa: KwaZulu-Natal, Nicholas 2877 (UDW) South Africa: Western Cape, Bruyns 6855 (K) South Africa: Eastern Cape, Nicholas 2816 (UDW) South Africa: KwaZulu-Natal, Ollerton 228 (NU, spirit) South Africa: KwaZulu-Natal, Ollerton 223 (NU, spirit) Gambia: Huber s.n. (UBT) South Africa: Mpumalanga, Nicholas 2867 (UDW) South Africa: Ollerton 217 (NU, in spirit) Tanzania: Ufipa District, Goyder et al. 3834 (K) trnT-L spacer trnL intron trnL-F spacer ITS AM295740 AM295739 AM295738 AM396886 AM295741 AM295742 AM295743 AM396887 AM295746 AM295747 AM295718 AM295745 AM295748 AM295717 AM295744 AM295749 AM295716 AM396888 AM396889 AM396890 AM295713 AM295714 AM295715 AM396891 AJ410178 (C) AJ410179 (C) AJ410180 (C) AM396892 — AM295712 AF214312 (G) AM295711 AF141158 (G) AM295710 — AM396893 AM295700 AM295699 AM295698 AM396894 AM295707 AM295708 AM295709 AM396895 AM295701 AM295702 AM295703 AM396896 AM295706 AM295705 AM295704 AM396897 AM295695 AM295696 AM295697 AM396898 AM295694 AM295693 AM295692 AM396899 AJ428795 (D) AJ428796 (D) AJ428797 (D) AM396900 AM295691 AM295690 AM295689 AM396901 AM295686 AM295687 AM295688 AM396902 AM295685 AM295769 AM295768 AM396903 Annals of the Missouri Botanical Garden Asclepias tuberosa L. Asclepias woodii (Schltr.) Schltr. Aspidoglossum connatum (N. E. Br.) Bullock Aspidoglossum delagoense (Schltr.) Kupicha Aspidoglossum heterophyllum E. Mey., acc. 1 Aspidoglossum heterophyllum E. Mey., acc. 2 Aspidonepsis diploglossa (Turcz.) Nicholas & Goyder Aspidonepsis flava (N. E. Br.) Nicholas & Goyder Calotropis procera (Aiton) W. T. Aiton Cordylogyne globosa E. Mey. DNA voucher Tribe, subtribe Taxon Gomphocarpus abyssinicus Decne. Margaretta rosea Oliv. subsp. corallina Goyder Miraglossum verticillare (Schltr.) Kupicha Pachycarpus appendiculatus E. Mey. Pachycarpus asperifolius Meisn. Pachycarpus concolor E. Mey. Pachycarpus coronarius E. Mey. Pachycarpus dealbatus E. Mey. South Africa: Eastern Cape, Nicholas 2796 (UDW) Zambia: Lusaka, Strid 2266 (K) South Africa: Eastern Cape, Nicholas 2798 (UDW) South Africa: Eastern Cape, Nicholas 2829 (NH) South Africa: KwaZulu-Natal, Nicholas 2878 (UDW) Kenya: Tsavo East, Liede & Newton 3211 (ULM) Yemen: Mangelsdorff 116 (UBT) South Africa: Eastern Cape, Nicholas 2836 (UDW) Tanzania: Ufipa District, Goyder et al. 3791 (K) South Africa: KwaZulu-Natal, Ollerton 244 (NU, spirit) South Africa: KwaZulu-Natal, Nicholas 2871 (UDW) South Africa: Limpopo, Nicholas 2858 (UDW) South Africa: KwaZulu-Natal, Nicholas 2856 (UDW) South Africa: Eastern Cape, Nicholas 2812 (UDW) South Africa: KwaZulu-Natal, Nicholas 2849 (UDW) Tanzania: Njombe District, Bidgood et al. 2106 (K) trnL intron trnL-F spacer ITS AM295752 AM295751 AM295750 AM396904 AM295765 AM295766 AM295767 AM396905 AM295764 AM295763 AM295762 AM396906 AM295759 AM295758 AM295760 AM295757 AM295761 AM295756 AM396907 AM396908 AJ290875 (A) AJ290876 (A) AJ290877 (A) AJ320446 (E) AM295753 AM295754 AM295755 AM396872 AM295577 AM295578 AM295579 AM396873 AM295582 AM295581 AM295580 AM396874 AM295583 AM295584 AM295585 AM396875 AM295588 AM295587 AM295586 AM396876 AM295589 AM295590 AM295591 AM396840 AM295594 AM295593 AM295592 AM396841 AM295595 AM295596 AM295597 AM396842 AM295600 AM295599 AM295598 AM396843 AM295601 AM295602 AM295603 AM396844 AM295606 AM295605 AM295604 AM396845 AM295607 AM295608 AM295609 AM396846 427 Pachycarpus goetzei (K. Schum.) Bullock Ethiopia: Shewa Region, Liede & Meve 3520 (UBT) South Africa: Drewe 534 (K) trnT-L spacer Goyder et al. Subtribe Asclepiadinae Gomphocarpus cancellatus (Burm. f.) Bruyns Gomphocarpus fruticosus (L.) W. T. Aiton Gomphocarpus glaucophyllus Schltr. Gomphocarpus physocarpus E. Mey., acc. 1 Gomphocarpus physocarpus E. Mey., acc. 2 Gomphocarpus tomentosus Burch. Kanahia laniflora (Forssk.) R. Br., acc. 1 Kanahia laniflora (Forssk.) R. Br., acc. 2 Lagarinthus multicaulis E. Mey. DNA voucher Volume 94, Number 2 2007 Table 1. Continued. 428 Table 1. Continued. Tribe, subtribe Taxon Pachycarpus grantii (Oliv.) Bullock subsp. marroninus Goyder Pachycarpus lineolatus (Decne.) Bullock Pachycarpus natalensis N. E. Br. Pachycarpus reflectens E. Mey. Pergularia daemia (Forssk.) Chiov. Stathmostelma pauciflorum (Klotzsch) K. Schum. Stathmostelma verdickii De Wild. Stenostelma corniculatum (E. Mey.) Bullock Trachycalymma buchwaldii (Schltr. & K. Schum.) Goyder Trachycalymma foliosum (K. Schum.) Goyder Trachycalymma pseudofimbriatum Goyder Woodia mucronata (Thunb.) N. E. Br. Xysmalobium fraternum N. E. Br. Tanzania: Njombe District, Goyder et al. 3886 (K) Tanzania: Ufipa District, Goyder et al. 3779 (K) South Africa: KwaZulu-Natal, Ollerton 254 (NU, spirit) South Africa: Eastern Cape, Nicholas 2801 (UDW) Tanzania: Arusha, Masinde 888 (in cult. Bayreuth) South Africa: Eastern Cape, Ollerton 216 (NU, spirit) South Africa: Ward s.n. (UDW) South Africa: North Western Province, Nicholas 2865 (UDW) South Africa: KwaZulu-Natal, Nicholas 2850 (UDW) Kenya: Narok, Liede & Newton 3221 (ULM) Zimbabwe: Karoi, Nicholas 2860 (UDW) Tanzania: Ufipa District, Bidgood et al. 2647 (K) South Africa: Mpumalanga, Balkwill 10908 (K) Tanzania: Iringa Distr., Goyder et al. 3924 (K) Tanzania: Ufipa District, Goldblatt et al. 8100 (K) Ethiopia: Sidamo, Haugen 1503 (K) South Africa: Eastern Cape, Nicholas 2809 (UDW) Tanzania: Sumbawanga District, Bidgood et al. 2390 (K) trnT-L spacer trnL intron trnL-F spacer ITS AM295612 AM295611 AM295610 AM396847 AM295613 AM295614 AM295615 AM396848 AM295618 AM295617 AM295616 AM396849 AM295619 AM295620 AM295621 AM396850 AJ290891 (A) AJ290892 (A) AJ290893 (A) AM396851 AM295624 AM295623 AM295622 AM396852 AM295625 AM295630 AM295626 AM295629 AM295627 AM295628 AM396853 AM396854 AM295631 AM295632 AM295633 AM396855 AM295636 AM295635 AM295634 AM396856 AM295637 AM295638 AM295639 AM396857 AM295642 AM295641 AM295640 AM396858 AM295643 AM295644 AM295645 AM396859 AM295648 AM295647 AM295646 AM396860 AM295684 AM295649 AM295650 AM396861 AM295653 AM295652 AM295651 — AM295654 AM295655 AM295656 AM396862 AM295659 AM295658 AM295657 AM396863 Annals of the Missouri Botanical Garden Schizoglossum atropurpureum E. Mey. Schizoglossum cordifolium E. Mey. Schizoglossum eustegioides (E. Mey.) Druce Schizoglossum filiforme (L. f.) Druce Stathmostelma diversifolium Goyder DNA voucher Volume 94, Number 2 2007 Table 1. Continued. Tribe, subtribe Taxon Xysmalobium gerrardii Scott-Elliott DNA voucher trnT-L spacer trnL intron trnL-F spacer ITS AM295660 AM295661 AM295662 AM396864 AM295665 AM295664 AM295663 AM396865 AM295666 AM295667 AM295668 AM396866 AM295671 AM295670 AM295669 AM396867 AM295672 AM295673 AM295674 AM396868 AM295677 AM295676 AM295675 AM396869 AM295678 AM295679 AM295680 AM396870 AM295683 AM295682 AM295681 AM396871 Goyder et al. Subtribe Asclepiadinae South Africa: KwaZulu-Natal, Ollerton 251 (NU, spirit) Xysmalobium heudelotianum Decne. Tanzania: Ufipa District, Goyder et al. 3830 (K) Xysmalobium involucratum South Africa: Eastern Cape, Nicholas (E. Mey.) Decne. 2802 (UDW) Xysmalobium kaessneri S. Moore Tanzania: Mbeya District, Lovett et al. 3980 (K) Xysmalobium parviflorum Harv. ex South Africa: KwaZulu-Natal, Nicholas Scott-Elliott 2840 (UDW) Xysmalobium tysonianum (Schltr.) South Africa: KwaZulu-Natal, Ollerton N. E. Br. 233 (NU, spirit) Xysmalobium undulatum (L.) W. T. South Africa: Eastern Cape, Nicholas Aiton 2830 (UDW) Gen. indet., aff. Asclepias Tanzania: Njombe District, Goyder et al. 3895 (K) 429 430 Annals of the Missouri Botanical Garden Table 2. Summary statistics of the data sets and of the trees resulting from parsimony analysis. CI 5 consistency index; RI 5 retention index; RC 5 rescaled consistency index. Parsimony informative characters Unknown data cells Missing (partial) sequences Outgroup No. of trees (limited by computer capacity) No. of steps CI RI RC cpDNA ITS Combined 153 208 277 222 (trnT-L spacer) 50 (trnL intron) 131 (trnL-F spacer) Asclepias tuberosa (trnT-L spacer) 28 Marsdenia sylvestris Eustegia minuta Calciphila galgalensis Cynanchum ellipticum Glossonema boveanum Odontanthera radians Tylophora flexuosa Oxystelma bornouense Oxystelma esculentum Solenostemma arghel 56,100 Asclepias tuberosa Trachycalymma pseudofimbriatum Oxystelma esculentum Oxystelma esculentum Solenostemma arghel Oxystelma esculentum Solenostemma arghel 56,500 62,000 650 0.5 0.5 0.32 778 0.52 0.66 0.34 262 0.71 0.81 0.57 5 S1600, matrix accession numbers 5 M2879 (cpDNA), M2880 (ITS), M2881 (combined); Sanderson et al., 1994) and comprises 80 taxa and 2847 characters (1088 characters in the trnT-L spacer (primers a–b), 530 characters in the trnL-intron (primers c–d), 429 characters in the trnL-F spacer (primers e–f), and 800 characters for ITS). Details of the data sets, as well as statistics of the trees resulting from parsimony analysis, are listed in Table 2. The partition homogeneity test shows no significant discordance between the cpDNA and the ITS data sets (P 5 0.09), and the retrieved clade structure in the cpDNA data set agrees largely with the one retrieved from the ITS data set, despite the poor resolution of the former. The analyses reveal that the Asclepiadinae as presently circumscribed are monophyletic. The small Old World genera Pergularia L. (P. daemia (Forssk.) Chiov.), Calotropis R. Br. (C. procera (Aiton) W. T. Aiton), and Kanahia R. Br. (K. laniflora (Forssk.) R. Br.) are sister to the remaining members of the subtribe, which themselves form a well-supported clade, referred to here as the Asclepias generic complex. Several smaller clades were detected, but overall resolution is limited. The genera in this complex will need to be recircumscribed following our results. DISCUSSION The core group of principally herbaceous genera around Asclepias contains most of the species diversity within the Asclepiadinae. This group is both monophyletic and well supported, resolving into two clades in the ITS (not shown) and combined analyses (Fig. 1). One clade consists of taxa found only in the New World (A. linaria Cav., A. curassavica L., A. fascicularis Decne., A. longifolia Michx., and A. syriaca L.); the other consists solely of Old World taxa. The trnT-L data set reveals no overall resolution into New or Old World lineages, but several minor Old World clades and one minor New World clade are retrieved (not shown). Asclepias, the oldest generic name in the complex (1753), must be applied either to the whole assemblage or, if other genera are recognized, to just the New World clade, as this contains A. syriaca, its type species. This result is consistent with Bullock’s (1952) assertion that Asclepias should be restricted to New World taxa. Within the African radiation, a number of clades were detected, which are well supported in one or more of the analyses. A clade comprising Xysmalobium heudelotianum Decne., X. fraternum N. E. Br., X. kaessneri S. Moore, Stathmostelma verdickii De Wild. (a species excluded Volume 94, Number 2 2007 Goyder et al. Subtribe Asclepiadinae 431 Figure 1. Strict consensus of 62,000 trees resulting from analysis of the combined data set (l 5 778 steps, excluding uninformative characters, CI 5 0.52, excluding uninformative characters, RI 5 0.66, RC 5 0.34). Numbers mapped on the tree indicate bootstrap percentages above 50% and decay values. 432 Annals of the Missouri Botanical Garden from Stathmostelma by Goyder (1998b) who suggested an affinity with Gomphocarpus longissimus K. Schum.), and Trachycalymma buchwaldii (Schltr. & K. Schum.) Goyder was detected in all analyses. These species are found in tropical African wooded savannas at moderate to high altitude. A second group of Xysmalobium species (X. parviflorum Harv. ex Scott-Elliott, X. gerrardii Scott-Elliott, and X. tysonianum (Schltr.) N. E. Br.) forms a clade with a group of broad-leaved Asclepias (A. albens (E. Mey.) Schltr. and A. crispa) with moderate support (60%) in all analyses. These species share a similar habit with many decumbent stems arising from a large woody tuber and have dense umbelliform pedunculate inflorescences. They occur in open grassland habitats in southeastern parts of tropical and subtropical Africa. The position of the apparently closely allied taxon, Lagarinthus multicaulis E. Mey. (Langley, 1980; 5 A. multicaulis (E. Mey.) Schltr. non A. multicaulis Vell.), is unresolved, however. Xysmalobium, clearly, is polyphyletic as currently circumscribed (Langley, 1980), and the type of the genus, X. undulatum (L.) W. T. Aiton, occurs in yet another clade separate from other exemplars of the genus. Asclepias woodii (Schltr.) Schltr. links with A. gibba in ITS and the combined analyses with strong support (99%). This is surprising, because they do not share obvious similarities of habit and vegetative or floral morphology (Brown, 1908). Asclepias gibba is much branched from the base, while A. woodii is an erect, single-stemmed herb; the corona of A. gibba terminates in a long drawn-out apical tooth much taller than the staminal column, while that of A. woodii is shorter and lacks the terminal appendage. The Schizoglossum/Aspidoglossum/Miraglossum complex of Kupicha (1984) is identified in all analyses (A. heterophyllum, A. connatum (N. E. Br.) Bullock, A. delagoense (Schltr.) Kupicha, M. verticillare (Schltr.) Kupicha, S. atropurpureum E. Mey., S. cordifolium E. Mey., and S. filiforme (L. f.) Druce; bootstrap 86% in the combined analysis). Kupicha (1984) noted the similarities of S. filiforme to species of Aspidoglossum but excluded it from the latter genus because of its pedunculate inflorescences and differing corona morphology. The results here link it to A. connatum and A. delagoense. Schizoglossum eustegioides (E. Mey.) Druce was excluded from the complex altogether by Kupicha (1984), who suggested its affinities might lie with Stenostelma Schltr.; ITS and the combined analyses support this suggestion, but there is no resolution with trnT-L. A small group of narrow-leaved Asclepias species in southern Africa—A. brevipes (Schltr.) Schltr., A. cucullata (Schltr.) Schltr., and A. stellifera Schltr.— occurs in all analyses. A group of short-lived Gomphocarpus species with narrow leaves (G. abyssinicus Decne., G. fruticosus (L.) W. T. Aiton, and G. physocarpus E. Mey.) forms a wellsupported clade (bootstrap 99%) with ITS and the combined analyses but is not resolved with trnT-L. Neither G. cancellatus (Burm. f.) Bruyns nor G. glaucophyllus Schltr., both broad-leaved species, form part of this core Gomphocarpus lineage. Gomphocarpus cancellatus is the only species of the generic complex with a predominantly winter-rainfall distribution in southern Africa and has some anomalous morphological features, such as the combination of broad leaves with perennial rather than annual stems and the more robust processes on the follicle. Gomphocarpus glaucophyllus appears in a predominantly tropical clade sister to Pachycarpus lineolatus (Decne.) Bullock and Trachycalymma foliosum (K. Schum.) Goyder. Gomphocarpus glaucophyllus is found in burnt grasslands and savanna woodlands of south tropical Africa, and its broad glaucous leaves are borne on annual shoots from a perennial woody rootstock. In contrast, the narrow-leaved species of Gomphocarpus occur in open disturbed habitats in temperate regions of southern Africa and are shrubby in habit, but are probably short-lived and generally lack thickened or tuberous roots. Stathmostelma diversifolium Goyder, S. pauciflorum (Klotzsch) K. Schum., Margaretta rosea Oliv. subsp. corallina Goyder, and the undescribed taxon, all from seasonally burned montane habitats in tropical Africa, also form a well-supported clade with ITS and the combined analysis but are unresolved with the chloroplast data set. Goyder (2005) pointed to similarities in corona morphology between Margaretta and Stathmostelma. The undescribed taxon is one of several from the highlands of southern Tanzania and northern Malawi that the first author has found difficult to assign to the genera currently recognized. Finally, a group of southern African Pachycarpus species (P. appendiculatus E. Mey., P. asperifolius Meissn., P. concolor E. Mey., P. dealbatus E. Mey., P. natalensis N. E. Br., and P. reflectens E. Mey.) is detected in trnT-L and the combined analyses. Two of these (P. concolor and P. dealbatus) form a minor clade in ITS, as do two of the tropical species (P. goetzei (K. Schum.) Bullock and P. grantii (Oliv.) Bullock subsp. marroninus Goyder). However, P. coronarius E. Mey. remains in an unresolved position in the combined analysis, and P. lineolatus, a widely distributed tropical species, links somewhat unexpectedly to the south tropical African species Gomphocarpus glaucophyllus, as already indicated. In the exemplars above, vegetative features seem a more reliable guide to evolutionary relationships than do the traditional floral characters. In particular, Volume 94, Number 2 2007 Goyder et al. Subtribe Asclepiadinae there has been a switch between the Xysmalobiumtype fleshy corona and Asclepias- or Gomphocarpustype cucullate corona both in tropical and southern African clades. Margaretta and Stathmostelma sensu Goyder (1998b) have much more showy and brightly colored flowers than the small white flowers of the undescribed taxon but are similar vegetatively with erect annual stems arising from a carrot-shaped tuber. The delimitation of Asclepias could be restricted to species from the Americas, and the Old World clade subdivided into genera with redefined limits. This would require a more intensively sampled survey and the use of additional genomes. Some of these lineages might correspond to the tropical group with erect, fusiform follicles (e.g., Xysmalobium heudelotianum); a more narrowly defined Gomphocarpus; an expanded Margaretta containing Stathmostelma and some tropical taxa currently unassigned to genus; Schizoglossum including Miraglossum and Aspidoglossum but excluding some anomalous taxa; and the broadleaved group of Xysmalobium and Asclepias from southern African grasslands. Goyder, D. J. 1995. Notes on the African genus Glossostelma (Asclepiadaceae). Kew Bull. 50: 527–555. ———. 1998a. A revision of Pachycarpus E. Mey. (Asclepiadaceae: Asclepiadeae) in tropical Africa with notes on the genus in southern Africa. Kew Bull. 53: 335–374. ———. 1998b. A revision of the African genus Stathmostelma K. Schum. (Apocynaceae: Asclepiadeae). Kew Bull. 53: 577–616. ———. 2001a. A revision of the tropical African genus Trachycalymma (K. Schum.) Bullock (Apocynaceae: Asclepiadoideae). Kew Bull. 56: 129–161. ———. 2001b. Gomphocarpus (Apocynaceae: Asclepiadeae) in an African and a global context—An outline of the problem. Biol. Skr. 54: 55–62. ———. 2005. Infraspecific variation in Margaretta rosea Oliv. (Apocynaceae: Asclepiadoideae). Kew Bull. 60: 87–94. ———. 2006. An overview of Asclepiad biogeography. Pp. 205–214 in S. A. Ghazanfar & H. J. Beentje (editors), Taxonomy and Ecology of African Plants, Their Conservation and Sustainable Use. Royal Botanic Gardens, Kew. ——— & A. Nicholas. 2001. A revision of Gomphocarpus R. Br. (Apocynaceae: Asclepiadeae). Kew Bull. 56: 769–836. Kupicha, F. K. 1984. Studies on African Asclepiadaceae. Kew Bull. 38: 599–672. Langley, R. W. 1980. Taxonomic Studies in the Asclepiadeae with Particular Reference to Xysmalobium R. Br. in Southern Africa. M.Sc. Thesis, University of Natal, Pietermaritzburg. Liede, S. 1997. Subtribes and genera of the tribe Asclepiadeae (Apocynaceae, Asclepiadoideae)—A synopsis. Taxon 46: 233–247. ———. 2001. Subtribe Astephaninae (Apocynaceae– Asclepiadoideae) reconsidered: New evidence based on cpDNA spacers. Ann. Missouri Bot. Gard. 88: 657–668. ——— & H. Kunze. 2002. Cynanchum and the Cynanchinae (Apocynaceae–Asclepiadoideae): A molecular, anatomical and latex triterpenoid study. Org. Divers. Evol. 2: 239–269. ——— & A. Täuber. 2000. Sarcostemma R. Br. (Apocynaceae–Asclepiadoideae)—A controversial generic circumscription reconsidered: Evidence from trnL-F spacers. Pl. Syst. Evol. 225: 133–140. ——— & ———. 2002. Circumscription of the genus Cynanchum (Apocynaceae–Asclepiadoideae). Syst. Bot. 27: 789–800. ———, U. Meve & A. Täuber. 2002a. What is subtribe Glossonematinae (Apocynaceae: Asclepiadoideae)? A phylogenetic study based on cpDNA spacer. Bot. J. Linn. Soc. 139: 145–158. ———, A. Täuber & J. Schneidt. 2002b. Molecular considerations on the Tylophorinae K. Schum. (Apocynaceae–Asclepiadoideae). Edinburgh J. Bot. 59: 377–403. Liede-Schumann, S. & U. Meve. 2006. Calciphila, a new genus in African Asclepiadeae (Apocynaceae, Asclepiadoideae), and taxonomic rectifications in Cynanchum. Novon 16: 368–373. ———, A. Rapini, D. J. Goyder & M. W. Chase. 2005. Phylogenetics of the New World subtribes of Asclepiadeae (Apocynaceae–Asclepiadoideae): Metastelmatinae, Oxypetalinae and Gonolobinae. Syst. Bot. 30: 184–195. Literature Cited Baldwin, B. G. 1992. Phylogenetic utility of the internal transcribed spacers of nuclear ribosomal DNA in plants: An example from the Compositae. Molec. Phylogen. Evol. 1: 3–16. Bremer, K. 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795–803. Brown, N. E. 1902–1903. Asclepiadeae. Pp. 231–503 in W. T. Thiselton-Dyer (editor), Flora of Tropical Africa, Vol. 4(1). Lovell Reeve & Co., London. ———. 1908. Asclepias. Pp. 663–714 in W. T. ThiseltonDyer (editor), Flora Capensis, Vol. 4(1). Lovell Reeve & Co., London. Bullock, A. A. 1952. Notes on African Asclepiadaceae. I. Kew Bull. 7: 405–426. ———. 1953a. Notes on African Asclepiadaceae. II. Kew Bull. 8: 51–67. ———. 1953b. Notes on African Asclepiadaceae. III. Kew Bull. 8: 329–362. ———. 1955. Notes on African Asclepiadaceae. V. Kew Bull. 9: 579–594. ———. 1956. Notes on African Asclepiadaceae. VII. Kew Bull. 10: 611–626. ———. 1957. Notes on African Asclepiadaceae. VIII. Kew Bull. 11: 503–522. ———. 1961. Notes on African Asclepiadaceae. IX. Kew Bull. 15: 193–206. Donoghue, M. J., R. G. Olmstead, J. F. Smith & J. D. Palmer. 1992. Phylogenetic relationships of Dipsacales based on rbcL sequences. Ann. Missouri Bot. Gard. 79: 333–345. Doyle, J. J. & J. L. Doyle. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. Bot. Soc. Amer. 19: 11–15. Eriksson, T. 1998. AutoDecay 4.0. Stockholm Univ., Stockholm. Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791. 433 434 Annals of the Missouri Botanical Garden Meve, U. & S. Liede. 2001. Reconsideration of the status of Lavrania, Larryleachia and Notechidnopsis (Asclepiadoideae–Ceropegieae). S. Afr. J. Bot. 67: 161–168. Nicholas, A. 1982. Taxonomic Studies in Asclepias L. with Particular Reference to the Narrow-leaved Species in Southern Africa. M.Sc. Thesis, University of Natal, Pietermaritzburg. ———. 1999. A Taxonomic Reassessment of the Subtribe Asclepiadinae (Asclepiadaceae) in Southern Africa. Vols. 1 & 2. Ph.D. Thesis, University of Durban-Westville, Durban. ——— & D. J. Goyder. 1992. Aspidonepsis (Asclepiadaceae), a new southern African genus. Bothalia 22: 23–35. Potgieter, K. & V. A. Albert. 2001. Phylogenetic relationships within Apocynaceae s.l. based on trnL intron and trnL-F spacer sequences and propagule characters. Ann. Missouri Bot. Gard. 88: 523–549. Rapini, A., M. W. Chase, D. J. Goyder & J. Griffiths. 2003. Asclepiadeae classification: Evaluating the phylogenetic relationships of New World Asclepiadoideae (Apocynaceae). Taxon 52: 33–50. Sanderson, M. J., M. J. Donoghue, W. Piel & T. Eriksson. 1994. TreeBASE: A prototype database of phylogenetic analyses and an interactive tool for browsing the phylogeny of life. Amer. J. Bot. 81: 183. Smith, D. M. N. 1988. A revision of the genus Pachycarpus in southern Africa. S. Afr. J. Bot. 54: 399–439. Swofford, D. L. 1998. PAUP*. Phylogenetic Analysis Using Parsimony (*and other methods), Version 4. Sinauer, Sunderland, Massachusetts. Taberlet, P., L. Gielly, G. Pautou & J. Bouvet. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Pl. Molec. Biol. 17: 1105–1109.