McKenzie, R.J. &amp; Barker, N.P. (2008). Radiation and biogeography of the predominantly southern African subtribe Arctotidinae (Asteraceae: Arctotideae).

Nigel Barker

ABSTRACT Lagenophora (Astereae, Asteraceae) has 14 species in New Zealand, Australia, Asia, southern South America, Gough Island and Tristan da Cunha. Phylogenetic relationships in Lagenophora were inferred using nuclear and plastid DNA regions. Reconstruction of spatio-temporal evolution was estimated using parsimony, Bayesian inference and likelihood methods, a Bayesian relaxed molecular clock and ancestral area and habitat reconstructions. Our results support a narrow taxonomic concept of Lagenophora including only a core group of species with one clade diversifying in New Zealand and another in South America. The split between the New Zealand and South American Lagenophora dates from 11.2 Mya [6.1–17.4 95% highest posterior density (HPD)]. The inferred ancestral habitats were openings in beech forest and subalpine tussockland. The biogeographical analyses infer a complex ancestral area for Lagenophora involving New Zealand and southern South America. Thus, the estimated divergence times and biogeographical reconstructions provide circumstantial evidence that Antarctica may have served as a corridor for migration until the expansion of the continental ice during the late Cenozoic. The extant distribution of Lagenophora reflects a complex history that could also have involved direct long-distance dispersal across southern oceans.

Molecular Phylogenetics and Evolution 49 (2008) 1–16 Contents lists available at ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Radiation of southern African daisies: Biogeographic inferences for subtribe Arctotidinae (Asteraceae, Arctotideae) Robert J. McKenzie *, Nigel P. Barker Molecular Ecology and Systematics Group, Department of Botany, Rhodes University, P.O. Box 94, Grahamstown 6140, South Africa a r t i c l e i n f o Article history: Received 18 April 2007 Revised 29 June 2008 Accepted 9 July 2008 Available online 18 July 2008 Keywords: Arctotideae Arctotidineae Asteraceae Biogeography Cape floral region Compositae Dispersal–vicariance analysis Molecular phylogeny Molecular dating a b s t r a c t The majority of the approximately 80–90 species in subtribe Arctotidinae occur in southern Africa with the centre of diversity in the winter-rainfall region. Three species are restricted to afromontane eastern Africa and three species are endemic to Australia. To investigate biogeographic and phylogenetic relationships within Arctotidinae, sequence data from four cpDNA regions (psbA-trnH, trnT-trnL and trnL-trnF spacers and trnL intron) and the ITS nrDNA region for 59 Arctotidinae species were analyzed with parsimony and Bayesian-inference approaches. Eight well-supported major lineages were resolved. The earliest-diverging extant lineages are afromontane or inhabit mesic habitats, whereas almost all sampled taxa from the winter-rainfall and semi-arid areas have diverged more recently. Molecular dating estimated that the major clades diverged during the Miocene and Pliocene, which is coincident with the trend of increasing rainfall seasonality, aridification and vegetation changes in southwestern Africa. Trans-oceanic dispersal to Australia was estimated to have occurred during the Pliocene. Ó 2008 Elsevier Inc. All rights reserved. 1. Introduction Southern Africa is an important centre of floristic biodiversity. Approximately 24,000 species of vascular plants are recorded from the region, of which approximately 80% are endemic to southern Africa (Cowling and Hilton-Taylor, 1997). Almost 10% of this plant diversity is contributed by the Asteraceae (the daisy family), with approximately 250 genera and 2250 species recorded from the region (Koekemoer, 1996). Over 80% of these species are endemic to southern Africa (Cowling and Hilton-Taylor, 1997). The family has radiated in all of the recognized southern African biomes and is one of the more species-rich families in most, if not all, of these biomes. The diversity and endemism of Asteraceae in southern Africa is focused in eight putative centres (Koekemoer, 1996). These centres are generally consistent with centres of endemism indicated by the southern African flora as a whole (e.g. see van Wyk and Smith, 2001). Despite being a major component of the flora, the evolution and biogeography of southern African Asteraceae is poorly investigated and well-resolved phylogenetic reconstructions are required to elucidate the origins and radiation of the Asteraceae lineages in the region (Galley and Linder, 2006). The only major southern African lineage of Asteraceae that has been investigated to any extent is the tribe Arctotideae, which has been the subject of both molec* Corresponding author. Fax: +27 46 6229550. E-mail address: r.mckenzie@ru.ac.za (R.J. McKenzie). 1055-7903/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2008.07.007 ular (McKenzie et al., 2006a) and morphological (Karis 2006) phylogenetic investigations. However, the sampling in previous studies was not comprehensive and here we present the first well-sampled phylogenetic and biogeographic investigation of the predominantly southern African subtribe Arctotidinae to help to fill this gap in our knowledge. Arctotideae is a small tribe containing about 215 species in 17 genera with an almost exclusively African distribution (Karis, 2007). Two subtribes are recognized with some authors placing other lineages in the tribe (for a brief review of the taxonomy and relationships of the tribe, see Funk et al., 2004). The subtribe Arctotidinae contains approximately 80–90 species currently classified into five genera (Karis, 2007). One species, Haplocarpha scaposa, has a wide distribution in montane and temperate southeastern Africa (e.g. Hilliard, 1977; Pope, 1992). Three species are indigenous to the high-altitude East African mountains (Mesfin Tadesse, 2004) and three species are endemic to Australia (Holland and Funk, 2006). The remaining species occur in southern Africa, with the centre of diversity in the Cape and Succulent Karoo Centres of Floristic Endemism (see van Wyk and Smith, 2001). Both areas occur in the winter-rainfall zone (Cowling, 1992; Dean and Milton, 1999; see Fig. 1) and are combined into a ‘Greater Capensis’ floristic region by some workers (Born et al., 2007). Because more than 50% of the species occur in the Cape Floristic Region (CFR), the Arctotidinae, and especially Arctotis s.str., might represent a true ‘Cape clade’ sensu Linder (2003, 2005). 2 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Fig. 1. A topographical map of southern Africa with the approximate boundaries of different rainfall regimes indicated (based on Chase and Meadows, 2007). Understanding the factors responsible for the unusually high levels of floristic richness and endemism in the winter-rainfall region of southern Africa (and the CFR in particular) has attracted considerable interest (for reviews for the CFR, see Linder, 2003, 2005). Molecular dating has indicated that plant lineages have been recruited into the Cape flora since at least the Cretaceous and the onset of radiations in these lineages is indicated to range from the Oligocene to the Pliocene (Linder, 2005), so the extant diversity cannot be explained solely by vicariance or explosive speciation in response to one or more concurrent triggers. Furthermore, the southern African flora as a whole is indicated to have evolved from disparate sources, including descendants of Gondwanan ancestors (Goldblatt, 1978; Anderson et al., 1999), temperate southern African progenitors (Goldblatt, 1978), groups of Eurasian or tropical African origin (e.g. Meerow et al., 2003; Hurka et al., 2005; Mummenhoff et al., 2005), and Australasian ancestors (e.g. Linder et al., 2003). Therefore a thorough understanding of diversification processes in the southern African flora is likely to be best achieved by a taxon-focused, rather than a broad floristic, approach because of the individuality of taxon histories within a flora (Verboom et al., 2003). The above evidence indicates that transcontinental and transoceanic dispersal has been important in the compilation of the present-day southern African flora. Indeed, there is increasing evidence for the importance of long-distance dispersal in determining extant plant biogeographic patterns worldwide (e.g. Price and Clague, 2002; de Queiroz, 2005). Phylogenetic reconstructions based on molecular data, in combination with molecular calibrations to estimate the date of divergence events, have provided strong evidence that in the Southern Hemisphere, following the breakup of Gondwana, transoceanic dispersal has been important in a diversity of animals (e.g. Vences et al., 2003; Schwarz et al., 2006) and plant groups (e.g. Baum et al., 1998; Mummenhoff et al., 2004; Cook and Crisp, 2005; Linder and Barker, 2005; Barker et al., 2007). Vicariance appears to have been much stronger in animals than plant groups (Sanmartín and Ronquist, 2004). Of particular significance to the present study, plant taxa with a disjunct southern African–Australian distribution, such as the Arctotidinae, are rare (Good, 1964; Thorne, 1972; Goldblatt, 1978), and as yet few of these taxa have been investigated in a phylogenetic context. A previous molecular phylogenetic study of Arctotidinae examined sequence variation in five cpDNA regions (ndhF, psbA-trnH, rps16, trnS-trnfM, and trnT-trnF) and the ITS nrDNA region for 18 species representing the major morphological variation in the subtribe (McKenzie et al., 2006a). Strong incongruence between the prevailing taxonomy and evolutionary relationships were indicated. A clade comprising two Haplocarpha species (H. nervosa and H. rueppellii, both formerly segregated in Landtia) from the southern and East African mountain chain was sister to the rest of the subtribe, and the widely distributed H. scaposa was indicated to be an early divergence. The placement of the Australian Cymbonotus lawsonianus and East African H. schimperi in a well-supported clade along with south-eastern African Arctotis species provided evidence for dispersal from southern Africa to Australia and migration to East Africa, respectively. However, comprehensive sampling of extant lineages is required for confident phylogenetic reconstructions and to allow more accurate biogeographic inferences to be made. The primary objective of the present study was to reconstruct a species-level phylogenetic hypothesis for a comprehensive sample of Arctotidinae based on sequence data from four R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 cpDNA regions (the psbA-trnH, trnT-trnL and trnL-trnF intergenic spacers and trnL intron) and the ITS nrDNA region. The specific aims were to use the phylogeny to: (1) investigate biogeographic patterns at the continental and regional levels; (2) make a first attempt at dating the diversification of Arctotidinae; and (3) further explore congruence between species relationships and the taxonomic classification, focusing on the polyphyly of Arctotis and Haplocarpha. 2. Materials and methods 2.1. Taxon sampling DNA sequences were obtained for 71 accessions from 52 species of Arctotidinae. These data were supplemented with published sequences for single accessions of 18 Arctotidinae and four species from the sister subtribe Gorteriinae (McKenzie et al., 2006a). Thus the ingroup comprised 89 accessions from 59 species of Arctotidinae, representing the five currently accepted genera, and the outgroup consisted of single accessions of four species of the sister subtribe Gorteriinae (Table 1). It is essential to avoid nomenclatural and taxonomic confusion, and we thus adopt the following nomenclatural usage. Generic concepts as elucidated by Karis (2007) are adopted, and Lewin’s (1922) infrageneric classification of Arctotis and Beauverd’s (1915) infrageneric classification of Haplocarpha are followed. In addition, the following corrections of recent treatments by Beyers (2000) and names that appear in Germishuizen and Meyer (2003) are relevant. Arctotis decurrens is the correct name for A. merxmuelleri and A. scullyi (McKenzie et al., 2006b). Two specimens (McKenzie 797/3 and McKenzie 1124/1) belong to a newly recognized species, A. debensis (McKenzie et al., 2006c). Arctotis flaccida is the correct name for Arctotis sp. ‘1’ designated by Beyers (2000). Arctotis semipapposa (listed as a synonym of A. flaccida by Beyers, 2000) is a very different, morphologically distinct species (McKenzie et al., in press). Arctotis sp. ‘2’ and sp. ‘4’ were designated by Beyers (2000) and this usage is retained here. Five other undetermined Arctotis species herein are designated spp. ‘A’, ‘B’, ‘C’, ‘D’, and ‘E’ (see Table 1). A full revision of the taxonomy of southern African Arctotidinae by the first author is currently in progress. 3 (Bioline, London), 1 ll 0.1 lM solution of each forward and reverse primer, 0.2 ll BioTaqÒ DNA polymerase (5 U/ll, Bioline, London, UK) and 0.75–1.5 ll unquantified DNA extract. The volume of 50 mM MgCl2 varied from 0.75–1.5 ll (1.5–3 mM) for trnT-trnF and 2–3 ll (4–6 mM) for psbA-trnH. Some reaction solutions contained 1.5–2 ll 0.1% bovine serum albumen. For some ITS amplification reactions, the solution differed in containing 5 ll 5 Colorless GoTaqÒ reaction buffer, 0.25 ll GoTaqÒ DNA polymerase (5 U/ll, Promega, Madison, WI) and no additional MgCl2. The DNA regions were amplified using a Hybaid PCR Sprint thermal cycler. The following parameters in the amplification reactions were standard for all regions: in the first reaction cycle, denaturing 95 °C, 45 s; primer extension 72 °C, 3 min; and the final extension cycle 72 °C, 10 min. The number of amplification cycles and primer annealing temperature differed for each region as follows; psbA-trnH: 25 cycles, 52 °C, 45 s; trnT-trnL spacer (A and B): 27–30 cycles, 52 °C, 45 s; trnT-trnL (A and D): 35 cycles, 52 °C, 45 s; trnL-trnF intron and spacer (C and F): 30 cycles, 52 °C, 45 s; trnL-trnF spacer (E and F): 25–30 cycles, 52 °C, 45 s; ITS: 30–33 cycles, 48–52 °C, 45 s. The PCR products were purified using the WizardÒ SV Gel and PCR purification kit (Promega, Madison, Wisconsin) and resuspended in 25–30 ll nuclease-free water. PCR products were sequenced in both forward and reverse directions. Sequencing reactions of the PCR products were performed using the BigDyeÒ Terminator v. 3.1 cycle sequencing kit according to the manufacturer’s instructions (Applied Biosystems, Foster City, California). The cycle-sequencing products were precipitated using the sodium acetate/EDTA protocol, and electrophoresed and resolved on an ABI Prism 3100 Genetic Analyzer at the Rhodes University DNA sequencing facility. 2.3. Sequence alignment For each accession contiguous sequences were compiled with Sequencher 4.2.2 (Gene Codes Corp., Ann Arbor) and edited visually. All sequences were deposited in GenBank (http:// www.ncbi.nlm.nih.gov/; accession numbers listed in Table 1). Sequences were aligned manually in MacClade 4.06 (Maddison and Maddison, 2000). Gaps were inserted manually based on visual inspection of the sequences. Alignments are deposited in TreeBASE (http://www.treebase.org/). 2.2. DNA extraction, amplification, and sequencing 2.4. Phylogenetic analyses For most samples total genomic DNA was extracted from about 1 cm2 of either fresh leaf material or leaves desiccated in silica gel using a CTAB extraction protocol (Doyle and Doyle, 1987). For Cymbonotus maidenii DNA was extracted from a herbarium specimen. Two non-coding cpDNA regions, the psbA-trnH intergenic spacer and trnT-trnF region (the trnT-trnL and trnL-trnF intergenic spacers and trnL intron), and one non-coding nrDNA region (ITS) were amplified and sequenced. The following primers were used for amplifying and sequencing: the primers A, B, C, D, E, and F of Taberlet et al. (1991) and Arct-trnL-R (ATT WTA TCR TTT CTG TAT CSG; previously unpublished) for the trnT-trnF region; psbAF and trnH-R for the psbA-trnH intergenic spacer (Sang et al., 1997); ITS18 (Käss and Wink, 1997; modified by Beyra-Matos and Lavin, 1999), ITS26 (Käss and Wink, 1997), Chrysanth-5.8F and Chromo-5.8R (Barker et al., 2005), ITS1, ITS4, and ITS5 (White et al., 1990) for the ITS region. Attempts to amplify both cpDNA regions for the C. maidenii accession and the psbA-trnH spacer for one Arctotis decurrens sample (Mucina 170903/20) were unsuccessful. In some instances, each species is represented by multiple samples, in order to assess species monophyly. Each 25 ll Polymerase Chain Reaction (PCR) solution contained 2.5 ll 10 PCR buffer (Bioline, London, UK), 1 ll 20 mM dNTPs Because in flowering plants the chloroplast genome is usually inherited as a single unit without recombination (Soltis and Soltis, 1998), the cpDNA regions were combined into one data set. Three data sets were thus analyzed: the cpDNA data set; the ITS data set; and a ‘total evidence’ data set (Kluge, 1989; Nixon and Carpenter, 1996) with the cpDNA and ITS data sets combined. Congruence in the phylogenetic signal of the cpDNA and ITS data sets was assessed with the partition homogeneity (or ILD) test (Farris et al., 1994) using PAUP 4.0b10 (Swofford, 2002) with 750 replicates, a maximum tree limit of 500, tree bisection–reconnection (TBR) branch swapping, with uninformative characters excluded following Lee (2001). Although conflict between the cpDNA and ITS data sets was indicated to be significant, the results must be interpreted with caution as the ILD test is an unreliable measure of data-set congruence in some circumstances (Ramírez, 2006). Under certain conditions this test can indicate significant incongruence between congruent data (Yoder et al., 2001; Hipp et al., 2004) and non-significance in cases of high incongruence (‘‘hypercongruence” sensu Ramírez, 2006). As an alternative strategy, conflicting nodes were regarded to be incongruent if they each received parsimony bootstrap support (BS) > 70% (Mason-Gamer and Kellogg, 1996) or Bayesian posterior probabilities (PP) > 95% (Alfaro et al., 2003). 4 Table 1 Collection, voucher and GenBank accession information for sequences included in this study Provenance of sample Collector(s) and collection number Herbariumb Arctotheca calendula (L.) Levyns Arctotheca forbesiana (DC.) K. Lewin Arctotheca marginata Beyers Arctotheca populifolia (P.J. Bergius) Norl. Arctotheca prostrata (Salisb.) Britten Arctotis acaulis L. 1 Arctotis acaulis L. 2 Arctotis adpressa DC. Arctotis angustifolia L. var. latifolia Harv. Arctotis arctotoides (L.f.) O. Hoffm. 1 Arctotis arctotoides (L.f.) O. Hoffm. 2 Arctotis argentea Thunb. Arctotis aspera L. var. aspera 1 Arctotis aspera L. var. aspera 2 Arctotis aspera L. var. scabra P.J. Bergius Arctotis auriculata Jacq. Arctotis bellidifolia P.J. Bergius 1 Arctotis bellidifolia P.J. Bergius 2 Arctotis breviscapa Thunb. Arctotis campanulata DC. var. puberula DC. Arctotis canescens DC. Arctotis debensis R.J. McKenzie 1 Arctotis debensis R.J. McKenzie 2 Arctotis decurrens Jacq. 1 Arctotis decurrens Jacq. 2 Arctotis dregei Turcz. 1 Arctotis dregei Turcz. 2 Arctotis elongata Thunb. Arctotis erosa Harv. 1 Arctotis erosa Harv. 2 Arctotis erosa Harv. 3 Arctotis erosa Harv. 4 Arctotis fastuosa Jacq. 1 Arctotis fastuosa Jacq. 2 Arctotis flaccida Jacq. 1 Arctotis flaccida Jacq. 2 Arctotis graminea K. Lewin Arctotis hirsuta (Harv.) Beauverd Arctotis hispidula (Less.) Beauverd Arctotis incisa Thunb. 1 Arctotis incisa Thunb. 2 Arctotis laevis Jacq. Arctotis lanceolata Harv. 1 Arctotis lanceolata Harv. 2 Arctotis leiocarpa Harv. Arctotis microcephala (DC.) Beauverd 1 Arctotis microcephala (DC.) Beauverd 2 South Africa: Western Cape South Africa: Western Cape South Africa: Northern Cape South Africa: Eastern Cape South Africa: Eastern Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Eastern Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Northern Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Eastern Cape South Africa: Eastern Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Western Cape South Africa: Western Cape South Africa: Eastern Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Western Cape South Africa: Northern Cape Namibia: Karas South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Eastern Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Eastern Cape Namibia: Karas South Africa: Eastern Cape South Africa: Eastern Cape McKenzie 808/3 McKenzie 1215/1 Bosenburg 2 Barker 1766 Barker 1765 McKenzie 823/4 McKenzie 827/1 McKenzie 1365/1 McKenzie 839 Barker 1950 McKenzie 855/1 McKenzie 1250 McKenzie 844 Samuel 41 McKenzie 1372 McKenzie 1285/1 Barker 1856 McKenzie 1060/1 McKenzie 1383/1 McKenzie 1284 McKenzie 1311/1 McKenzie 797/3 McKenzie 1124/1 McKenzie 1302/1 Mucina 170903/20 McKenzie 815/2 McKenzie 834/4 McKenzie 1085/1 McKenzie 1282 McKenzie 1290 Samuel 15 Barker 1872 McKenzie 1351 Mannheimer 2492 McKenzie 1065/3 McKenzie 1363/2 McKenzie 824/1 McKenzie 1070/2 McKenzie 760/1 McKenzie 826/2 McKenzie 1380/1 McKenzie 1355/1 McKenzie 835 Ramdhani 616 Mucina 150704/7 McKenzie 784/3 McKenzie 1409 GRA GRA NBG GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA, WIND GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GenBank accession numbers ITS psbA–trnH trnT–trnF DQ444720 EU846328 EU846329 EU846330 EU846331 DQ444721 EU846344 EU846345 EU846346 EU846334 DQ444722 EU846347 DQ444723 EU846348 EU846349 EU846350 EU846351 EU846352 DQ444724 EU846353 EU846354 EU846335 EU846336 EU846356 EU846355 EU846332 DQ444725 EU846357 EU846358 EU846359 EU846360 EU846361 EU846362 EU846373 EU846363 EU846364 EU846365 EU846366 EU846337 EU846367 EU846368 EU846369 EU846370 EU846371 EU846372 EU846339 EU846340 DQ444764 EU846399 EU846400 EU846401 EU846402 DQ444765 EU846414 EU846415 EU846416 EU846405 DQ444766 EU846417 DQ444767 EU846418 EU846419 EU846420 EU846421 EU846422 DQ444768 EU846423 EU846424 EU846406 EU846407 EU846425 — EU846403 DQ444769 EU846426 EU846427 EU846428 EU846429 EU846430 EU846431 EU846442 EU846432 EU846433 EU846434 EU846435 EU846408 EU846436 EU846437 EU846438 EU846439 EU846440 EU846441 EU846409 EU846410 DQ444808 EU846468 EU846469 EU846470 EU846471 DQ444809 EU846483 EU846484 EU846485 EU846474 DQ444810 EU846486 DQ444811 EU846487 EU846488 EU846489 EU846490 EU846491 DQ444812 EU846492 EU846493 EU846475 EU846476 EU846495 EU846494 EU846472 DQ444813 EU846496 EU846497 EU846498 EU846499 EU846500 EU846501 EU846512 EU846502 EU846503 EU846504 EU846505 EU846477 EU846506 EU846507 EU846508 EU846509 EU846510 EU846511 EU846478 EU846479 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Speciesa South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: KwaZulu-Natal South Africa: Eastern Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Northern Cape South Africa: Free State South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Western Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Northern Cape South Africa: Northern Cape Australia: Australian Capital Territory Australia: New South Wales Australia: New South Wales South Africa: Western Cape ex hort. South Africa: Western Cape South Africa: Eastern Cape South Africa: Eastern Cape South Africa: Eastern Cape South Africa: Western Cape Ethiopia: Bale Kenya: Central Province South Africa: Eastern Cape South Africa: KwaZulu-Natal Ethiopia: Arsi Kenya: Central Province McKenzie 1059/1 McKenzie 822 McKenzie 825/1 McKenzie 777/1 McKenzie 1205/2 McKenzie 1071/2 Samuel 42 McKenzie 1389/1 Ramdhani 508 White 41 McKenzie 811/1 McKenzie 843/1 McKenzie 892/1 McKenzie 1074/1 McKenzie 1281/1 McKenzie 875/1 Barker s.n. McKenzie 1234/1 Barker 1865 McKenzie 1075 McKenzie 1296/1 McKenzie 1300/1 McKenzie 1336/1 Samuel 31 McKenzie 1339/1 Samuel 35 Samuel 20 Bayer ACT-05001 Weston 2486 Radunz s.n., 5.ii.1982 Barker 1780 McKenzie 845/1 Barker 1767 McKenzie 970 McKenzie 1192 Mucina 290805/13 Barker 1906 Namaganda, Abdillahi & Nakamatle 1746 Barker 1772 Ramdhani 519 Barker 1899 Namaganda, Abdillahi & Nakamatle 1748 GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA GRA NSW NSW GRA GRA GRA GRA GRA GRA ETH EA GRA GRA ETH EA, GRA EU846374 EU846341 DQ444726 EU846375 EU846376 EU846377 EU846378 EU846379 EU846342 EU846343 EU846380 EU846381 EU846382 EU846383 DQ444728 DQ444729 EU846384 EU846385 DQ444727 EU846386 EU846387 EU846388 EU846389 EU846390 EU846391 EU846392 EU846393 DQ444730 EU846333 EU846338 DQ444731 DQ444732 DQ444733 DQ444734 EU846323 EU846327 DQ444735 EU846324 DQ444736 EU846325 DQ444737 EU846326 EU846443 EU846411 DQ444770 EU846444 EU846445 EU846446 EU846447 EU846448 EU846412 EU846413 EU846449 EU846450 EU846451 EU846452 DQ444772 DQ444773 EU846453 EU846454 DQ444771 EU846455 EU846456 EU846457 EU846458 EU846459 EU846460 EU846461 EU846462 DQ444774 EU846404 — DQ444775 DQ444776 DQ444777 DQ444778 EU846394 EU846398 DQ444779 EU846395 DQ444780 EU846396 DQ444781 EU846397 EU846513 EU846480 DQ444814 EU846514 EU846515 EU846516 EU846517 EU846518 EU846481 EU846482 EU846519 EU846520 EU846521 EU846522 DQ444816 DQ444817 EU846523 EU846524 DQ444815 EU846525 EU846526 EU846527 EU846528 EU846529 EU846530 EU846531 EU846532 DQ444818 EU846473 — DQ444819 DQ444820 DQ444821 DQ444822 EU846463 EU846467 DQ444823 EU846464 DQ444824 EU846465 DQ444825 EU846466 Outgroup taxa: Berkheya carduoides (Less.) Hutch. Cuspidia cernua (L.f.) B.L. Burtt Gazania krebsiana Less. Hirpicium echinus Less. South South South South Barker 1924 Barker 1896 Barker s.n. McKenzie 861 GRA GRA GRA GRA DQ444716 DQ444717 DQ444718 DQ444719 DQ444760 DQ444761 DQ444762 DQ444763 DQ444804 DQ444805 DQ444806 DQ444807 Africa: Africa: Africa: Africa: Eastern Cape Eastern Cape Eastern Cape Northern Cape R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Arctotis muricata Thunb. Arctotis perfoliata (L.f.) O. Hoffm. 1 Arctotis perfoliata (L.f.) O. Hoffm. 2 Arctotis pinnatifida Thunb. Arctotis reptans Jacq. Arctotis revoluta Jacq. 1 Arctotis revoluta Jacq. 2 Arctotis rotundifolia K. Lewin Arctotis scapiformis Thell. 1 Arctotis scapiformis Thell. 2 Arctotis semipapposa (DC.) Beauverd 1 Arctotis semipapposa (DC.) Beauverd 2 Arctotis stoechadifolia P.J. Bergius 1 Arctotis stoechadifolia P.J. Bergius 2 Arctotis sulcocarpa K. Lewin Arctotis venusta Norl. Arctotis verbascifolia Harv. 1 Arctotis verbascifolia Harv. 2 Arctotis sp. ‘2’ Arctotis sp. ‘4’ Arctotis sp. ‘A’ Arctotis sp. ‘B’ Arctotis sp. ‘C’ 1 Arctotis sp. ‘C’ 2 Arctotis sp. ‘D’ 1 Arctotis sp. ‘D’ 2 Arctotis sp. ‘E’ Cymbonotus lawsonianus Gaudich. 1 Cymbonotus lawsonianus Gaudich. 2 Cymbonotus maidenii (Beauverd) A.E. Holland & V.A. Funk Dymondia margaretae Compton Haplocarpha lanata Less. Haplocarpha lyrata Harv. Haplocarpha nervosa (Thunb.) Beauverd 1 Haplocarpha nervosa (Thunb.) Beauverd 2 Haplocarpha parvifolia (Schltr.) Beauverd Haplocarpha rueppellii (Sch. Bip.) Beauverd 1 Haplocarpha rueppellii (Sch. Bip.) Beauverd 2 Haplocarpha scaposa Harv. 1 Haplocarpha scaposa Harv. 2 Haplocarpha schimperi (Sch. Bip.) Beauverd 1 Haplocarpha schimperi (Sch. Bip.) Beauverd 2 GenBank numbers in bold are new submissions. a Duplicate samples of a species are numbered sequentially. b EA, East African Herbarium, Nairobi; ETH, National Herbarium of Ethiopia, Addis Ababa; GRA, Selmar Schonland Herbarium, Grahamstown; NBG, Compton Herbarium, Cape Town; NSW, National Herbarium of New South Wales, Sydney; WIND, National Herbarium of Namibia, Windhoek. 5 6 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Parsimony analyses were performed using PAUP. For each data set, a heuristic search was conducted with 1000 simple taxonaddition replicates, TBR branch swapping, and the MULTREES option in effect in order to ensure that multiple islands of equally parsimonious trees were found (Maddison, 1991). Further heuristic searches employed 1000 random taxon-addition replicates, holding one tree at each step, TBR branch swapping, with the MULTREES and STEEPEST DESCENT options in effect, for each replicate saving at most one tree P the tree length from the simple heuristic search, followed by a complete heuristic search on the saved trees with a maximum tree limit of 100,000. In all analyses no shorter trees were recovered than from the simple-addition search. Uninformative characters were excluded before all analyses and all characters were equally weighted and unordered. Parsimony analyses were repeated with unambiguous insertion/deletion events (indels) greater than 2 bp in length recoded as a binary (presence/absence) character, following Bayer et al. (2002), using the ‘simple indel coding’ method of Simmons and Ochoterena (2000). Parsimony bootstrap assessments for each data set were carried out using 1000 replicates in PAUP. The ITS data set was analyzed twice. In an initial analysis, all sites were considered, while a second analysis excluded four ambiguously aligned regions (aligned bp 70–79, 208–232, 256–262, and 479–497). Relative branch support was assessed using nonparametric bootstrap analysis (Felsenstein, 1985) with 1000 replicates, TBR branch swapping and a maximum tree limit of 500. Bayesian inference (BI) analyses using Markov Chain Monte Carlo methods (Yang and Rannala, 1997) were performed using MrBayes 3.1.1 (Huelsenbeck and Ronquist, 2001). Prior to analysis, sequences were partitioned into coding, exon, intron, and intergenic spacer regions based on sequences from the chloroplast genomes of Lactuca sativa, Nicotiana sylvestris, and N. tabacum from GenBank, and the ITS secondary structure model for the Asteraceae of Goertzen et al. (2003). Eleven partitions were defined. The psbAtrnH sequence comprised the psbA-trnH spacer, 53 bp at the 30 end of the psbA gene, and 32 bp at the 50 end of the trnH gene. The trnTtrnF region comprised the trnL intron, the flanking trnL 50 and 30 exons, and the trnT-trnL and trnL-trnF intergenic spacers. The ITS region comprised the 5.8S gene and the flanking internal transcribed spacers ITS1 and ITS2 (Baldwin, 1992). An optimal nucleotide-substitution model for each partition (Table 2) was selected for use in the BI analyses using the Akaike information criterion with MrModeltest 2.2 (J.A.A. Nylander, Uppsala University, Uppsala). Two independent BI analyses each with four Markov chains, three heated and one cold, starting from a random tree were run simultaneously for 1 106 generations with trees sampled every 100 generations. The trees sampled prior to stabilization of the log-likelihood value were discarded as ‘burn-in’ samples; the burn-in varied from 23,000 generations for the ITS data analysis, 30,000 for the cpDNA analysis and 40,000 for the total-evidence analysis. The remaining trees from the simultaneous runs were combined and used to generate a 50% majority rule consensus tree and to determine the PP for each node. The consensus trees from the Bayesian inference analyses are deposited in TreeBASE (http://www.treebase.org/; submission No. SN3970). 2.5. Molecular dating Ideally, dating of molecular phylogenies requires either reliable calibration points or a known rate of molecular evolution, but unfortunately obtaining reliable calibration points to date phylogenetic reconstructions within Asteraceae is problematic due to the family’s poor fossil record. Consequently, divergence dates were estimated from the ITS data set utilizing ITS mutation rates published for other Asteraceae taxa following the methodology of S Table 2 Nucleotide-substitution model for each partition specified in the Bayesian inference analyses Partition Modela Rate variation across sites psbA psbA–trnH spacer trnH trnT–trnL spacer trnL 30 exon trnL trnL 50 exon trnL–trnF spacer ITS1 5.8S ITS2 JC GTR JC GTR JC GTR JC GTR SYM GTR SYM Equal Gamma Equal Gamma Equal Equal Equal Gamma Gamma Proportion invariable + gamma Gamma a GTR, General Time Reversible; JC, Jukes-Cantor; SYM, Symmetrical. Howis (unpublished data). Three mutation rates (‘slow’, ‘average’, and ‘fast’) were used. Mutation rates published for Asteraceae range from 2.51 10 9 substitutions per site per year (s s 1 y 1; ‘slow’) in Eupatorium (Schmidt and Schilling, 2000) to 7.83 10 9 s s 1 y 1 (‘fast’) in Robinsonia (Sang et al., 1995). These rates span much of the range of rates published for angiosperms as a whole (see Kay et al., 2006). An ‘average’ rate (4.58 10 9 s s 1 y 1) was calculated based on the rates for four Asteraceae taxa listed by Kay et al. (2006); this rate was virtually identical to the average rate (4.59 10 9 s s 1 y 1) for the 10 herbaceous angiosperm taxa listed by the same authors. The ITS data set and cladogram were reduced to 50 terminals to exclude duplicates of species with almost identical sequences. Log-likelihood ratio tests were performed to compare the likelihoods of obtaining the same topology with or without a clock assumption. An appropriate evolutionary model was selected with MrModeltest 2.2. Maximum likelihood analyses were performed with PAUP with and without a clock enforced and the difference in log-likelihoods was tested by means of a v2 test. As the assumption of a strict molecular clock was rejected, divergence dates were estimated using a relaxed clock with Bayesian inference and MCMC procedures implemented in BEAST 1.4.7 (Drummond and Rambaut, 2007). A relaxed clock model allows the among-branch evolution rate to vary. In addition, BEAST has the advantage of permitting specification of a number of evolutionary models in a single framework (Drummond and Rambaut, 2007). Consistent with the MrModeltest results, a General Time Reversible nucleotide-substitution model with gamma + invariant sites was used. A relaxed molecular clock with branch substitution rates drawn from a lognormal distribution, auto-optimization of the operators, no topological constraints and a constant speciation rate per lineage (i.e. Yule tree prior) with a uniform prior distribution was used. Rates were sampled every 1000 cycles from 50,00,000 MCMC steps with a burn-in of 500,000 cycles. Running the analysis for 100,00,000 steps had no impact on the estimated divergence dates (data not presented). For all analyses two independent runs were performed, the log files were combined to check for convergence on the same distribution and to ensure adequate sample sizes, and viewed using Tracer v.1.4 (distributed with BEAST). Divergence dates estimated when branch substitution rates were drawn from an exponential distribution were considerably older and with much wider 95% highest posterior density (HPD) limits than those derived from a lognormal distribution, and in some instances the lower HPD bound extended well beyond the probable origin of the Asteraceae of c. 60 Mya. Therefore analyses utilising a lognormal distribution were the more plausible and are only considered herein. 7 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 2.6. Dispersal–vicariance analysis To infer the ancestral areas of the basal and internal nodes, a dispersal–vicariance analysis was performed with DIVA 1.1 (Ronquist, 1997). A fully dichotomized tree (a requirement of DIVA) was created based on a neighbor-joining (NJ) tree produced using PAUP. As can be seen from Figs. 2 and 3, not all clades were fully resolved, so use of a NJ approach obviated a subjective resolution of species-level relationships. We acknowledge that there are many possible permutations of species relationships in poorly resolved clades. However, in this study lack of resolution affected primarily the ‘core Arctotis’ clade, all members of which are from the Fynbos or Succulent Karoo regions, and exploration of alternative relationships is unlikely to affect the overall result of the DIVA analysis. It is emphasised that the basal and internal nodes of the NJ tree, which were the nodes of interest, were identical to those in the maximum parsimony and Bayesian total-evidence phylogenies. The selection of unit areas of endemism used in DIVA analyses can be problematic, as different workers may not agree on boundaries of regions of endemism, and some areas might be viewed as subsets of larger areas. For this analysis, we mostly used areas that correspond to vegetation units, namely afromontane, Fynbos, Nama-Karoo, Succulent Karoo, and Savanna biomes, as well as the widely recognized Albany Centre of Floristic Endemism (Cowling and Hilton-Taylor, 1997). We thus chose to differentiate between the Fynbos and Succulent Karoo, and not consider them as part of a larger floristic region (as suggested by Born et al., 2007, for example). Australia was treated as a single area. Analyses were run without limiting the number of ancestral area optimizations. The exclusion of some unsampled species may affect the results of a DIVA analysis, just as they can affect phylogenetic interpretations. In the present study, the majority of the unsampled taxa are Arctotis species and have CFR distributions and, based on morphological data, it is predicted they will be members of the ‘core Arctotis’ clade. This would not affect the outcome of the DIVA analysis. 3. Results 3.1. Phylogenetic analyses The sequence characteristics and variability of each DNA region and the combined datasets are summarized in Table 3. For each data set parsimony and BI generated almost identical phylogenetic reconstructions. Consequently, in the following discussion, the method of analysis is only specified when a phylogenetic arrangement is in conflict. In all analyses, the ‘Landtia’ clade (comprising Haplocarpha nervosa, H. rueppellii, and H. parvifolia) was sister to the rest of the Arctotidinae with maximum BS and PP support. Haplocarpha scaposa and Dymondia margaretae were placed on subsequently diverging, well-supported monotypic branches. The remaining taxa were placed in four strongly supported clades: a large ‘core Arctotis’ clade; a ‘Cymbonotus’ clade (composed of Cymbonotus lawsonianus, H. schimperi and species from Arctotis sect. Austro-orientales); a clade consisting of the two sampled species from Arctotis sect. Anomalae; and an Arctotheca–‘core Haplo- A. cpDNA B. ITS Berkheya carduoides Cuspidia cernua 100 1 100 1 97 1 100 1 0.01 Gazania krebsiana Hirpicium echinus Dymondia margaretae Cymbonotus maidenii Arctotis scapiformis2 Arctotis scapiformis1 78 Arctotis microcephala1 1 Arctotis microcephala2 Haplocarpha schimperi1 76 Haplocarpha schimperi2 * 0.99 ‘Cymbonotus’ 72 Arctotis hispidula 59 * Arctotis arctotoides1 1 0.83 Arctotis arctotoides2 Arctotis debensis1 Arctotis debensis2 100 Cymbonotus lawsonianus2 Cymbonotus lawsonianus1 1 92 Haplocarpha lanata Haplocarpha lyrata 1 53 — Arctotheca forbesiana 1 Arctotheca marginata 100 0.53 Arctotheca + Arctotheca prostrata 1 Arctotheca calendula ‘core Haplocarpha’ Arctotheca populifolia * Arctotis dregei1 100 Arctotis dregei2 Arctotis sect. 1 Arctotis sulcocarpa Arctotis breviscapa — Anomalae Arctotis perfoliata1 100 0.67 Arctotis perfoliata2 1 Arctotis campanulata puberula Arctotis lanceolata2 Arctotis graminea Arctotis semipapposa1 Arctotis decurrens2 Arctotis decurrens1 Arctotis spE 66 Arctotis acaulis1 1 Arctotis acaulis2 Arctotis adpressa Arctotis lanceolata1 Arctotis argentea * Arctotis aspera scabra Arctotis verbascifolia1 Arctotis sp2 Arctotis spA 66 Arctotis aspera1 Arctotis incisa2 * 0.67 Arctotis stoechadifolia2 Arctotis angustifolia latifolia 97 Arctotis aspera2 1 Arctotis bellidifolia1 Arctotis bellidifolia2 Arctotis elongata 100 Arctotis reptans Arctotis revoluta1 1 ‘core Arctotis’ Arctotis rotundifolia Arctotis semipapposa2 Arctotis verbascifolia2 Arctotis pinnatifida Arctotis incisa1 Arctotis laevis * Arctotis revoluta2 Arctotis auriculata 92 Arctotis spD2 * 1 Arctotis spD1 Arctotis flaccida1 Arctotis canescens Arctotis flaccida2 Arctotis spC2 Arctotis erosa1 Arctotis spC 1 Arctotis erosa3 Arctotis muricata Arctotis fastuosa 2 87 Arctotis stoechadifolia1 Arctotis graminea 1 Arctotis sp4 Arctotis verbascifolia2 Arctotis leiocarpa Arctotis auriculata Arctotis erosa1 Arctotis bellidifolia2 87 Arctotis hirsuta Arctotis laevis Arctotis erosa3 Arctotis campanulata puberula 1 Arctotis erosa2 Arctotis spA Arctotis erosa4 Arctotis spB 95 Arctotis flaccida1 Arctotis spD2 Arctotis flaccida2 Arctotis spD1 1 Arctotis fastuosa 1 Arctotis canescens Arctotis rotundifolia Arctotis fastuosa 2 75 Arctotis venusta Arctotis incisa1 0.88 Arctotis spB Arctotis semipapposa2 100 Haplocarpha scaposa1 Haplocarpha scaposa1 Haplocarpha scaposa Haplocarpha scaposa2 Haplocarpha scaposa2 73 1 96 Haplocarpha parvifolia Haplocarpha nervosa1 0.99 Haplocarpha nervosa1 Haplocarpha nervosa2 100 1 Haplocarpha nervosa2 85 Haplocarpha parvifolia ‘Landtia’ 1 97 Haplocarpha rueppellii1 60 Haplocarpha rueppellii1 1 Haplocarpha rueppellii2 Haplocarpha rueppellii2 1 0.62 Gazania krebsiana Hirpicium echinus Dymondia margaretae Haplocarpha schimperi1 Haplocarpha schimperi2 Arctotis arctotoides1 Arctotis arctotoides2 Arctotis debensis1 87 Arctotis debensis2 Arctotis hispidula 0.98 Arctotis microcephala1 Arctotis microcephala2 Arctotis scapiformis2 56 Arctotis scapiformis1 0.96 59 Cymbonotus lawsonianus2 75 Cymbonotus lawsonianus1 1 1 100 Arctotis perfoliata1 1 Arctotis perfoliata2 Haplocarpha lanata 90 Arctotheca marginata 96 1 Arctotheca prostrata 1 Haplocarpha lyrata 97 Arctotheca calendula 1 Arctotheca forbesiana Arctotheca populifolia 100 Arctotis sulcocarpa 100 Arctotis dregei1 1 Arctotis dregei2 1 69 Arctotis breviscapa 0.99 Arctotis acaulis2 Arctotis adpressa Arctotis angustifolia latifolia Arctotis argentea Arctotis aspera2 Arctotis aspera1 Arctotis aspera scabra Arctotis bellidifolia1 Arctotis decurrens2 Arctotis decurrens1 Arctotis incisa2 Arctotis muricata 77 Arctotis reptans Arctotis revoluta2 0.96 Arctotis revoluta1 Arctotis semipapposa1 Arctotis stoechadifolia2 Arctotis sp2 Arctotis sp4 Arctotis spE Arctotis acaulis1 Arctotis pinnatifida Arctotis elongata Arctotis lanceolata2 Arctotis erosa4 Arctotis venusta 96 Arctotis erosa2 1 Arctotis spC2 Arctotis spC1 Arctotis lanceolata1 Arctotis fastuosa 1 Arctotis hirsuta Arctotis leiocarpa Arctotis stoechadifolia1 Arctotis verbascifolia1 98 1 100 1 Dymondia Berkheya carduoides Cuspidia cernua 100 1 — 1 100 1 0.1 Fig. 2. Phylograms obtained from Bayesian-inference analyses. (A) cpDNA data set; (B) ITS data set. Support values above the branches are bootstrap percentages, and below the branches are Bayesian posterior probabilities. For terminal clades, support values are only given for nodes receiving bootstrap percentages > 75% and Bayesian posterior probabilities > 0.95. 8 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Dymondia margaretae Haplocarpha lanata Haplocarpha lyrata 62 Arctotheca forbesiana 1 100 95 Arctotheca calendula — 1 1 Arctotheca populifolia 0.93 Arctotheca marginata 55 0.99 Arctotheca prostrata Arctotis sulcocarpa 100 100 Arctotis dregei1 1 1 Arctotis dregei2 Cymbonotus maidenii 54 Arctotis scapiformis2 0.97 Arctotis scapiformis1 81 Arctotis microcephala1 0.97 Arctotis microcephala2 61 Haplocarpha schimperi1 0.95 76 95 1 Haplocarpha schimperi2 0.88 Arctotis hispidula 80 1 79 Arctotis arctotoides1 0.97 Arctotis arctotoides2 65 Arctotis debensis1 0.99 Arctotis debensis2 Cymbonotus lawsonianus2 100 1 Cymbonotus lawsonianus1 Arctotis breviscapa Arctotis perfoliata1 100 1 Arctotis perfoliata2 Arctotis campanulata puberula Arctotis lanceolata2 Arctotis semipapposa1 Arctotis decurrens2 Arctotis decurrens1 Arctotis spE Arctotis acaulis2 56 0.94 Arctotis adpressa Arctotis argentea Arctotis aspera scabra Arctotis sp2 61 Arctotis spA 0.94 Arctotis acaulis1 — 0.92 Arctotis lanceolata1 90 0.98 Arctotis verbascifolia1 Arctotis aspera1 85 Arctotis incisa2 1 Arctotis stoechadifolia2 Arctotis angustifolia latifolia Arctotis aspera2 95 Arctotis bellidifolia1 1 Arctotis bellidifolia2 Arctotis reptans Arctotis revoluta1 — Arctotis rotundifolia 0.62 Arctotis pinnatifida — Arctotis elongata 0.68 77 Arctotis graminea 1 Arctotis verbascifolia2 — Arctotis incisa1 0.97 Arctotis semipapposa2 100 Arctotis laevis 67 1 0.99 Arctotis revoluta2 Arctotis auriculata 75 Arctotis spD2 — 1 Arctotis spD1 0.91 — Arctotis canescens 0.96 96 Arctotis spC2 1 Arctotis spC1 Arctotis muricata — Arctotis stoechadifolia1 0.82 — 0.83 Arctotis sp4 Arctotis spB 91 Arctotis venusta 1 — Arctotis fastuosa 2 0.81 83 Arctotis erosa2 1 Arctotis leiocarpa — 0.79 Arctotis fastuosa 1 — Arctotis hirsuta 0.72 Arctotis erosa3 — 58 Arctotis erosa1 0.76 1 61 Arctotis erosa4 1 — Arctotis flaccida1 0.54 Arctotis flaccida2 Haplocarpha scaposa1 Haplocarpha scaposa2 95 Haplocarpha nervosa1 1 Haplocarpha nervosa2 Haplocarpha parvifolia 74 Haplocarpha rueppellii1 95 0.97 1 Haplocarpha rueppellii2 93 1 94 1 86 1 — 0.76 94 1 92 1 57 0.98 — 0.84 100 1 100 1 100 1 100 1 0.01 Dymondia ‘core’ Haplocarpha Arctotheca Arctotis sect. Anomalae ‘Cymbonotus’ perennial ‘core Arctotis’ annual ‘core Arctotis’ Haplocarpha scaposa ‘Landtia’ Fig. 3. Phylogram obtained from Bayesian-inference analysis of the total evidence. Support values are as in Fig. 2. carpha’ clade (comprising all Arctotheca species, H. lanata and H. lyrata). Relationships amongst these four clades varied depending on the data set and method of analysis. The major clades in the cpDNA phylogeny were strongly supported by both phylogenetic methods, but within-clade relationships were poorly resolved (Fig. 2A). The interior nodes linking Dymondia margaretae, ‘core Arctotis’, an Arctotis sect. Anomalae– Arctotheca–‘core Haplocarpha’ clade, and a ‘Cymbonotus’–Arctotis perfoliata clade were weakly to moderately supported by parsimony bootstrap (BS = 59–77%), but well supported in the BI analysis (PP = 0.96–1). Arctotis perfoliata was placed sister to the ‘Cymbonotus’ clade with weak support (BS = 56%, PP = 0.96). Arcto- 9 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Table 3 Summary of statistics from parsimony analyses of each DNA fragment and the combined datasets Characteristic psbA-trnH trnT-trnF Combined cpDNA ITS Total evidence Length range (nucleotides) Aligned length (nucleotides) Number of indels recoded (number that are parsimony informative) Number of variable sites Number of parsimony informative sites Number of most parsimonious trees Tree length (steps) CI RI 330–495 563 20 (12) 83 (14.7%) 48 (8.5%) — — — — 1348–1463 1593 31 (20) 195 (12.2%) 94 (5.9%) — — — — 1716–1950 2155 51 (32) 278 (12.9%) 142 (6.6%) 100 000 268 0.743 0.928 632–653 688 6 (4) 275 (40.0%) 206 (29.9%) 92 524 611 0.622 0.865 – 2843 57 (36) 553 (19.5%) 348 (12.2%) 94 496 872 0.606 0.872 The statistics are from parsimony analyses in which indels were recoded. Recoded indels are excluded from the data for sequence length, variability and parsimony informativeness. tis breviscapa was placed sister to the ‘core Arctotis’ clade with moderate support (BS = 77%, PP = 0.96). Although the rest of the ‘core Arctotis’ clade was strongly supported (BS = 96%, PP = 1), relationships within the clade were poorly resolved. Arctotis sect. Anomalae, Arctotheca and ‘core Haplocarpha’ formed a strongly supported clade (BS = 97%, PP = 1). A ‘core Haplocarpha’ clade was not retrieved. Haplocarpha lanata was sister to the rest of the Arctotheca–‘core Haplocarpha’ clade, and H. lyrata was nested within the Arctotheca clade but with little support (BS = 58%, PP = 0.78). The ITS phylogeny (Fig. 2B) differed from the cpDNA phylogeny in the following respects: the internal branches linking the Arctotheca–‘core Haplocarpha’, Arctotis sect. Anomalae, ‘Cymbonotus’ and ‘core Arctotis’ clades were poorly supported or unresolved, but relationships within the major clades were somewhat better resolved. ‘Core Haplocarpha’ and Arctotheca formed strongly supported sister clades. Cymbonotus lawsonianus was sister to the rest of the ‘Cymbonotus’ clade (BS = 72%, PP = 1). The two Cymbonotus species formed a basal grade to the southern and eastern African species in this clade. Arctotis perfoliata was placed close to the base of the ‘core Arctotis’ clade but with poor support (BS = 66%, PP = 0.67). The ‘core Arctotis’ clade was differentiated into highly supported annual and perennial lineages (BS = 97%, PP = 1, and BS = 95%, PP = 1, respectively). Exclusion of four ambiguously aligned regions reduced the ITS data set to 632 bp, of which 238 bp (37.7%) were variable and 174 bp (27.5%) were parsimony informative. Parsimony analysis yielded 1526 MPTs of 487 steps (CI 0.616, RI 0.867). The strict consensus tree topology differed to that obtained from the complete ITS data analysis in that the Arctotis sect. Anomalae clade was placed sister to the ‘core Haplocarpha’ clade but with no BS support (BS < 50% from 100 replicates). The ILD test indicated significant conflict existed between the cpDNA and ITS data sets (p = 0.001342). Visual comparison of tree topologies revealed incongruence in the placement of A. perfoliata and H. lyrata between the cpDNA and ITS trees (Fig. 2A and B), but the placement of both species was weakly supported by one data set. The placement of A. perfoliata in the cpDNA phylogeny was strongly supported (BS = 100%, PP = 1), but only weakly supported in the ITS tree (BS = 66%, PP = 0.67). In contrast, the placement of H. lyrata was strongly supported (BS = 92%, PP = 1) by the ITS data, but its placement within the Arctotheca clade from the cpDNA data received conflicting support (BS = 58%, PP = 1). Placement of the ‘Cymbonotus’ clade also differed in the cpDNA and ITS trees, but overall support was low (BS = 59%, PP = 1 in the cpDNA phlyogeny; BS < 50%, PP = 0.53 in the ITS phylogeny) and the data sets were thus combined. Both analytical methods yielded phylogenies with improved resolution from the total evidence data set (Fig. 3) than was obtained in the analyses of the separate data sets (Fig. 2A and B). The basal nodes were well supported (BS P 86%, PP = 1). The place- ment of the ‘Cymbonotus’ clade sister to the Arctotheca–‘core Haplocarpha’–Arctotis sect. Anomalae clade, and the ‘core Arctotis’ node were poorly supported, but the other internal nodes were well supported. The placement of A. perfoliata within the ‘core Arctotis’ clade received little support (BS < 50%, PP = 0.84). Highly supported annual and perennial clades were resolved within ‘core Arctotis’ (BS = 99%, PP = 1). Exclusion of A. perfoliata and H. lyrata from total-evidence analyses had no impact on the remaining tree topology, but support for the ‘core Arctotis’ clade improved considerably (BS = 83% from 200 bootstrap replicates), whereas the ‘Cymbonotus’–Arctotis sect. Anomalae–Arctotheca–‘core Haplocarpha’ clade remained weakly supported (BS = 56%). 3.2. Molecular dating Divergence dates estimated with the ‘slow’ and ‘average’ mutation rates resulted in considerably older mean divergence dates with wider 95% HPD ranges than dates estimated with the ‘fast’ mutation rate (Table 4). Based on the mean estimated dates, the Landtia clade (node 1) was estimated to have diverged some time during the early Oligocene to late Miocene epochs, but the crown of extant species (node 15) is of much more recent origin (Fig. 4 and Table 4). The three major extant clades—‘core Arctotis’, ‘Cymbonotus’, and Arctotis sect. Anomalae–‘core Haplocarpha’–Arctotheca (nodes 4, 5, and 6, respectively)—were indicated to have diverged in a rapid radiation during the mid or late Miocene. Cymbonotus lawsonianus (node 9) was indicated to have diverged within the ‘Cymbonotus’ clade during the late Miocene to late Pliocene. Radiation of the Arctotis arctotoides species complex (node 10) was dated to the late Pliocene at the earliest. Within this complex, Arctotis microcephala and H. schimperi (node 11) were estimated to have diverged during the late Pliocene or Pleistocene. 3.3. Biogeography and dispersal–vicariance analysis The two basalmost lineages in Arctotidinae have a wide geographic distribution (Fig. 4). The ‘Landtia’ clade grows largely at higher altitudes, and is distributed from the CFR (H. parvifolia) to southeastern Africa (H. nervosa) and afromontane regions of the East African mountain chain (H. nervosa and H. rueppellii). Haplocarpha scaposa is widely distributed in submontane and montane southeastern Africa and tropical eastern Africa. DIVA analysis indicated that Arctotidinae might have an afromontane origin, but the possible ancestral area of the widespread H. scaposa was equivocal. Furthermore, the accurate assessment of the ancestral area requires a resolved phylogeny of the sister group and use of a more distant outgroup. The ancestral area of all other extant Arctotidinae was indicated to be in the fynbos (i.e., in the CFR). The Arctotheca– ‘core Haplocarpha’ clade was indicated to have originated in fynbos, consistent with the centre of diversity in the CFR, with subse- 10 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Table 4 Estimated divergence dates for nodes of interest from BEAST analysis using ITS mutation rates published for other Asteraceae taxa Node Slow ESS Average 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 32.6 ± 0.17 (21.7–44.0) 18.6 ± 0.14 (13.1–24.7) 17.3 ± 0.11 (12.7–22.5) 16.1 ± 0.1 (11.7–20.7) 16.1 ± 0.1 (11.8–20.9) 13.8 ± 0.1 (9.4–18.8) 9.4 ± 0.07 (5.7–13.6) 3.9 ± 0.03 (1.4–6.9) 9.6 ± 0.07 (5.5–14.0) 3.7 ± 0.03 (1.9–6.0) 2.3 ± 0.02 (0.7–4.4) 10.6 ± 0.06 (7.4–13.9) 9.4 ± 0.05 (6.5–12.5) 6.8 ± 0.04 (4.6–9.1) 9.7 ± 0.07 (5.4–14.1) 1128.463 479.403 565.099 545.097 556.745 658.419 947.15 3391.419 1051.392 1120.797 2013.626 801.202 918.283 1055.801 1125.109 27.2 ± 0.11 11.8 ± 0.06 10.7 ± 0.05 9.3 ± 0.04 9.3 ± 0.04 8.0 ± 0.05 5.3 ± 0.03 2.0 ± 0.01 5.4 ± 0.03 1.7 ± 0.01 1.1 ± 0.01 6.0 ± 0.03 5.0 ± 0.02 3.4 ± 0.01 5.6 ± 0.03 (17.6–38.6) (8.2–15.9) (7.5–14.0) (6.9–12.2) (6.7–12.2) (5.6–10.6) (3.1–7.5) (0.8–3.5) (3.0–8.0) (0.9–2.7) (0.3–2.2) (4.2–8.1) (3.5–6.9) (2.3–4.6) (3.5–8.1) ESS Fast 2894.552 1157.966 1087.701 1035.193 1077.755 823.279 1827.21 4844.536 1957.358 1258.726 2001.685 1102.589 1533.579 1800.349 1423.365 10.2 ± 0.05 5.9 ± 0.03 5.5 ± 0.03 5.1 ± 0.03 5.1 ± 0.03 4.4 ± 0.03 3.0 ± 0.02 1.2 ± 0.01 3.0 ± 0.02 1.2 ± 0.01 0.73 ± 0.01 3.4 ± 0.02 3.0 ± 0.02 2.2 ± 0.01 3.0 ± 0.02 ESS (6.8 13.9) (4.2–7.8) (4.0–7.1) (3.8–6.7) (3.6–6.6) (3.0–6.0) (1.8–4.3) (0.44–2.2) (1.7–4.4) (0.59–1.9) (0.20–1.4) (2.4–4.4) (2.0–4.0) (1.5–2.9) (1.7–4.4) 1128.886 1027.545 966.825 836.345 859.889 749.412 1453.303 2792.901 1478.561 1229.467 1606.189 987.303 1040.264 1043.336 974.539 Results are from two independent analyses combined. For the node numbers see Fig. 3. Values (in My) are the mean divergence date ± SD and the 95% highest posterior density range. Slow = 2.51 10 9 s s 1 y 1 (Eupatorium; Schmidt and Schilling, 2000); average = mean of rates for four Asteraceae taxa listed by Kay et al. (2006); fast = 7.83 10 9 s s 1 y 1 (Robinsonia; Sang et al. 1995). Effective sample size (ESS) = the number of effectively independent samples from the posterior distribution. quent extension into Namaqualand (A. calendula and A. marginata), the Nama-Karoo and Succulent Karoo (A. calendula) and the Albany hotspot in southeastern South Africa (A. populifolia, A. prostrata and H. lyrata). The Arctotis sect. Anomalae lineage is centred in the CFR and Succulent Karoo, and extends into the Nama-Karoo (an undescribed species) and the Albany region (A. dregei). The ‘Cymbonotus’ clade was indicated to have migrated from fynbos to the Albany hotspot, savanna and afromontane southeastern Africa (Arctotis sect. Austro-orientales p.p.) and to have dispersed to Australia (C. lawsonianus and C. maidenii). An origin in the fynbos is implicated for the ‘core Arctotis’ clade, with subsequent independent radiations in the Succulent Karoo in both the perennial and annual clades. 4. Discussion The eight major lineages retrieved in the present investigation are congruent with those obtained in a previous study (McKenzie et al., 2006a), which was based on a smaller sample of taxa. Our results confirm the polyphyly of Arctotis and Haplocarpha as the genera are presently circumscribed. Haplocarpha species are segregated between four clades and Arctotis species are placed in three well-resolved lineages. A molecular phylogenetic study by Funk et al. (2007) included sequence data for two species not included in the present study, namely Haplocarpha oocephala (the sample was incorrectly named H. lanata) and Cymbonotus preissianus, as well as both cpDNA (ndhF and trnL-F) and ITS data for C. maidenii. Haplocarpha oocephala was placed on a monotypic lineage diverging between the ‘Landtia’ and H. scaposa lineages, and C. preissianus and C. maidenii were in the same clade as C. lawsonianus. The inclusion of a comprehensive sample of taxa in the present study has clarified the phylogenetic placement of several species whose placement was previously considered equivocal (McKenzie et al., 2006a). Arctotheca calendula was shown to belong unequivocally in the Arctotheca clade. In all analyses Arctotis breviscapa was placed sister to the rest of the ‘core Arctotis’ clade. The sample of H. lyrata was indicated to possess a chloroplast haplotype shared with Arctotheca, whereas the ITS data resolved a monophyletic ‘core Haplocarpha’. The conflicting placement of two accessions of Arctotis perfoliata sister to the ‘Cymbonotus’ clade in the cpDNA phylogeny and near-basal within the ‘core Arctotis’ clade from ITS data confirms the findings of McKenzie et al. (2006a). Using the criteria of Mason-Gamer and Kellogg (1996) and Alfaro et al. (2003), ‘hard’ incongruence (the conflicting placement of a taxon or clade is strongly supported in phylogenies derived from different data sets) was not apparent for any taxon or clade in the present study. Although the placement of Arctotis perfoliata and Haplocarpha lyrata conflicted between the cpDNA and ITS phylogenies, support was high for one data set only (cpDNA for A. perfoliata, ITS for H. lyrata). Exclusion of both species from analyses did not affect the resulting tree topology (data not presented), indicating the overall topologies reflect strong phylogenetic signal in the data. Both A. perfoliata samples possessed a unique 93 bp deletion in the psbA-trnH intergenic spacer and an additional seven unique 1 bp substitutions (three each in the trnT-trnF and ITS1 spacers and one in the psbA-trnH spacer). In morphology A. perfoliata possesses a combination of distinctive traits of the ‘Cymbonotus’ and ‘core Arctotis’ clades (McKenzie et al., 2005; R.J. McKenzie, unpublished data), which might reflect an ancestral hybridization event occurring during the evolution of A. perfoliata. However, incongruence between phylogenetic hypotheses might also reflect lineage sorting of ancestral polymorphisms, paralogy, lateral gene transfer, or erroneous phylogenetic reconstruction (Sang and Zhong, 2000). Sampling of additional accessions and DNA regions is needed to further resolve the phylogenetic relationships and evolutionary histories of A. perfoliata and H. lyrata and to establish what factors might account for the phylogenetic uncertainty. The ITS data resolved highly supported annual and perennial clades within the ‘core Arctotis’ clade, in contrast to the cpDNA data, but this might reflect the fewer informative cpDNA characters or a need for additional cpDNA data, rather than hard incongruence in the phylogenetic signal. The annual and perennial Arctotis clades were well-supported in the total-evidence analyses, indicating any conflicting signal in the cpDNA data set was weak. 4.1. Phylogenetic relationships and taxonomic implications Correspondence between relationships suggested by morphological and molecular data in Arctotidinae has been discussed previously (McKenzie et al., 2005, 2006a). Therefore the following discussion is focused on findings from the present study that build on those of previous studies. The present study confirms the finding of McKenzie et al. (2006a) that H. nervosa and H. rueppellii are closely related and belong to the earliest-diverging extant clade (‘Landtia’), and provides evidence that H. parvifolia, which was not included in the previous study, is a member of the same clade. This supports Beauverd’s (1915) observation that H. parvifolia possesses similarities in mor- 11 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 FYN Dymondia margaretae Haplocarpha lanata FYN ALB Haplocarpha lyrata FYN Arctotheca forbesiana Arctotheca calendula FYN, KAR, SK Arctotheca populifolia ALB, FYN Arctotheca marginata SK ALB, FYN Arctotheca prostrata Arctotis sulcocarpa SK Arctotis dregei ALB, FYN, KAR Cymbonotus maidenii AUS FYN 7 8 FYN 6 FYN FYN 3 5 Arctotis scapiformis ALB 11 10 Arctotis arctotoides Arctotis hispidula 9 FYN 4 SK FYN 12 SK FYN AFR FYN ALB SAV 13 SK FYN 2 SK FYN 14 AFR 1 SK AFR 15 AFR AFR Haplocarpha schimperi Arctotis microcephala AFR, ALB, SAV ALB ALB ALB Arctotis debensis AUS Cymbonotus lawsonianus Arctotis breviscapa FYN FYN Arctotis perfoliata Arctotis campanulata puberula SK Arctotis lanceolata FYN Arctotis semipapposa FYN Arctotis decurrens SK SK Arctotis spE FYN Arctotis acaulis FYN Arctotis adpressa FYN Arctotis aspera scabra FYN Arctotis sp2 SK Arctotis spA Arctotis verbascifolia FYN FYN Arctotis angustifolia latifolia FYN Arctotis aspera FYN Arctotis bellidifolia Arctotis reptans FYN Arctotis revoluta FYN, SK FYN Arctotis rotundifolia ALB, FYN Arctotis pinnatifida FYN Arctotis elongata FYN Arctotis graminea FYN Arctotis incisa FYN Arctotis laevis SK Arctotis auriculata SK Arctotis spD Arctotis canescens SK SK Arctotis spC FYN Arctotis muricata FYN Arctotis stoechadifolia FYN Arctotis sp4 SK Arctotis spB KAR, SAV, SK Arctotis venusta SK Arctotis fastuosa 2 SK Arctotis fastuosa 1 KAR, SK Arctotis leiocarpa SK Arctotis erosa FYN Arctotis hirsuta SK Arctotis flaccida Haplocarpha scaposa AFR, ALB, SAV Haplocarpha nervosa AFR, ALB, FYN AFR Haplocarpha rueppellii FYN Haplocarpha parvifolia Fig. 4. Ancestral areas in Arctotidinae indicated by DIVA analysis, mapped on a simplified cladogram from Bayesian-inference analysis of the total evidence data set. Key to area codes: AFR, afromontane biome; ALB, Albany centre of endemism; AUS, Australia; FYN, fynbos biome; KAR, Nama-Karoo biome, SAV, savanna biome; SK, Succulent Karoo biome. These regions in black boxes represent extant distributions, those in grey represent ancestral areas. Numbers in circles and squares (calibration points) represent node numbers associated with node ages listed in Table 4. phology to H. subg. Landtia, and specifically H. rueppellii. Haplocarpha nervosa, H. rueppellii and species now placed in their synonymy have been segregated as Landtia in the past (e.g. Lessing, 1831, 1832; Harvey, 1865; Bentham, 1873; Phillips, 1951). The morphological characters previously used to diagnose Landtia have been deemed untenable (e.g. Hedberg, 1957) and, especially focusing 12 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 on fruit morphology (McKenzie et al., 2005), the placement of H. parvifolia in the same clade further complicates morphological characterization of the clade. Haplocarpha rueppellii and H. scaposa possess a cypsela anatomy distinct from the other Arctotidinae sampled by Reese (1989), which thus might offer a possible synapomorphy for the ‘Landtia’ clade. Given the similarity in external vegetative morphology in the ‘Landtia’ clade, stem or leaf anatomy might also yield informative characters. Similarities in the morphology of H. parvifolia and H. oocephala have been noted previously (McKenzie et al., 2005, 2006a), but Funk et al. (2007) found that a sample of H. oocephala (misidentified as H. lanata in their paper) was placed on a monotypic lineage diverging after the ‘Landtia’ clade. Placement of H. scaposa and Dymondia margaretae on monotypic branches in all analyses confirmed previous findings (McKenzie et al., 2006a). Segregation of H. scaposa into a monotypic genus is strongly supported by the molecular data, but the species shares certain distinctive morphological features (e.g. papillose filaments, pappus scales with long acute apices) with species in the ‘core Haplocarpha’ clade. Sampling of all five Arctotheca species in the present study resolved a well-supported Arctotheca clade and confirmed a sister relationship with ‘core Haplocarpha’ (McKenzie et al., 2006a), but with the inclusion of H. lyrata in the Arctotheca clade in cpDNA-derived phylogenies. In the total-evidence analyses, ‘core Haplocarpha’ formed a basal grade sister to Arctotheca. As discussed earlier, sampling of additional DNA regions and multiple accessions is needed to resolve the evolutionary history of H. lyrata. The two species of Arctotis sect. Anomalae sampled (A. dregei and A. sulcocarpa) formed a highly supported clade distinct from ‘core Arctotis’ in all analyses, in agreement with previous results (McKenzie et al., 2006a). Lewin (1922) distinguished sect. Anomalae from other sections of Arctotis by the possession of neuter ray florets (female in all other Arctotis sections), a trait also found in Arctotheca and H. parvifolia. The species of A. sect. Anomalae share some distinctive characteristics with ‘core Arctotis’, such as possession of cypselae with well-developed abaxial wings (McKenzie et al., 2005), but other features (e.g. possession of floral scent, papillose filaments) supports a sister relationship with ‘core Haplocarpha’–Arctotheca. Holland and Funk (2006) broadened the concept of Cymbonotus to encompass the three Australian-endemic species of Arctotidinae by formally transferring Arctotis maidenii. Our results demonstrate that the ‘Cymbonotus’ clade has an Afro-Australian distribution and also includes the East African Haplocarpha schimperi and southern African species currently placed in Arctotis sect. Austroorientales (McKenzie et al., 2006a; this study). Our data resolved a well-supported ‘Cymbonotus’ clade, but whether it is sister to ‘core Arctotis’ or the Arctotis sect. Anomalae —‘core Haplocarpha’— Arctotheca lineage remains uncertain, seemingly reflecting a rapid divergence of the three lineages. Cymbonotus maidenii was sister to C. lawsonianus and C. preissianus in the study of Funk et al. (2007), which utilized both cpDNA and ITS data, and was not monophyletic with C. lawsonianus in the present study, a finding consistent with disparities in the morphology of these species (McKenzie et al., 2005; R.J. McKenzie, unpublished data). A relationship between C. maidenii and Arctotis perfoliata, as indicated by cypsela morphology (McKenzie et al., 2005), is not supported by the molecular data. A sister relationship between Haplocarpha schimperi and Arctotis microcephala within the ‘Cymbonotus clade’ is indicated. Despite the limited divergence in non-coding DNA sequences between the two species, H. schimperi has undergone greater evolution in phenotypic traits, especially in cypsela morphology (McKenzie et al., 2005), and possesses at least four autapomorphic character states absent in other members of the ‘Cymbonotus’ clade (geotro- pic capitula, woolly tomentum on the ray limb abaxial surface, a coronate pappus, and cypsela abaxial ribs not developed into wings). The present study, in which 45 of the approximately 60 species of Arctotis were sampled, reinforces the conclusion of McKenzie et al. (2006a) that the sections Anomalae and Austro-orientales should be excluded from Arctotis in order to render the genus monophyletic. Arctotis sect. Austro-orientales (excluding A. erosa and A. perfoliata) has a close phylogenetic relationship with Cymbonotus, rather than with ‘core Arctotis’. Arctotis erosa, an annual species placed in sect. Austro-orientales by Lewin’s (1922), belongs in the ‘annual Arctotis’ clade in ‘core Arctotis’. As discussed earlier, further investigations are needed to resolve the phylogenetic affinities of Arctotis perfoliata. In all analyses Arctotis breviscapa, an annual species endemic to sandveld in the southwestern Cape, was placed sister to the rest of the ‘core Arctotis’ clade and is indicated to be an early divergence from the ‘core Arctotis’ lineage. This agrees with Lewin’s (1922) hypothesis of evolutionary relationships among his sections in Arctotis. Arctotis breviscapa is anomalous in possessing cypselae with two well-developed adaxial wings in addition to three abaxial wings (McKenzie et al., 2005), but is otherwise similar in morphology to ‘core Arctotis’. The resolution of annual and perennial ‘core Arctotis’ clades by the ITS data conflicts with Lewin’s (1922) infrageneric classification of Arctotis. These clades were not retrieved in the cpDNA phylogenies, which might reflect ancestral polymorphisms, hybridization and introgression, or insufficient informative sites in the cpDNA data. 4.2. Biogeography and timing of diversification of Arctotidinae The estimated divergence dates presented in this paper are a first attempt at dating Arctotidinae diversification and must be interpreted with caution. Resolution of the timing of divergence events preferentially requires either a complete and reliably dated fossil record or reliable molecular mutation rates, neither of which are presently available for Arctotideae. Owing to the poor fossil record, obtaining reliable calibration points to date phylogenetic reconstructions within Asteraceae is a widespread problem. Rather than using an indirect calibration point derived from a higher-level molecular-dating analysis, which may potentially compound the error in estimated divergence dates (Graur and Martin, 2004), we utilized ITS mutation rates, representing the fastest and slowest extremes and an average rate, published for other Asteraceae taxa. It is acknowledged that the divergence dates estimate the evolution of the ITS region and not necessarily speciation or cladogenetic events. Natural selection is likely to act differentially on mutations (and thus evolution rates will vary) between loci and lineages. The strength of concerted evolution (Hillis et al., 1991) on the ITS region may vary between lineages and in lineages of ancestral hybrid origin. In addition, the similarity of the rates used to the overall ITS mutation rate in Arctotidinae is presently unknown. Nevertheless, the results obtained permit the formulation of hypotheses regarding the biogeography and timing of diversification of Arctotidinae. Bearing in mind the likely Calyceraceae–Asteraceae divergence within the last 60 Mya based on the oldest fossil pollen evidence (Zavada and de Villiers, 2000), and that Arctotidinae comprise primarily annuals and fast-growing perennials that may flower in their first year after germination, divergence dates estimated with the ‘fast’ ITS mutation rate are considered the most plausible of the three mutation rates used in this study. Thus the following discussion pertains to the dates estimated with a ‘fast’ mutation rate. Furthermore, the dating of Arctotidinae diversification with a ‘fast’ ITS mutation rate is compatible with the results of a recent molecular-dating study across all major clades of Asteraceae by Kim et al. (2005). Their study focused on the basal nodes and in- R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 cluded only one representative of Arctotideae (Gazania krebsiana), which was indicated to have diverged within the Liabeae–Arctotideae–Vernonieae clade approximately 25 Mya. However, the estimated divergence dates reported by Kim et al. (2005) may be slight underestimates, as the oldest microfossil evidence (Zavada and de Villiers, 2000) currently available for Asteraceae was not considered. Nevertheless, the divergence and early radiation in Arctotidinae may have followed expansion of the Antarctic ice cap, which covered most of that continent by about 17 Mya and had far-reaching global climatic impacts. In southwestern Africa a variety of new arid and semi-arid environments arose at this time (Pickford, 2004). The ancestral area of Arctotidinae was indicated to be in our afromontane area. The present distribution of the two earliest-diverging extant lineages, the ‘Landtia’ clade and Haplocarpha oocephala (the latter based on Funk et al., 2007), comprises a chain of disjunct populations along the southern and eastern African mountains extending from the northern Cederberg area in the Western Cape province, South Africa to the Ethiopian highlands (Hedberg, 1957; Hilliard, 1977; Pope, 1992; Beyers, 2000; Mesfin Tadesse, 2004). The next lineage to diverge (Haplocarpha scaposa) is widespread in montane parts of southeastern Africa and the Zambezi River catchment area (Hilliard, 1977; Pope, 1992). These distributions roughly coincide with what has been termed the Afrotemperate Phytogeographical Region, which comprises the Cape Floral Region, the greater Drakensberg mountains, and the afromontane Centre (Linder, 1990; Galley et al., 2007). The more recent derivation of Arctotidinae lineages in presently semi-arid and arid areas is consistent with trends in rainfall regimes in southwestern Africa since the Oligocene (e.g. see Linder, 2003) and phylogenetic analyses of a diverse range of plant lineages endemic to the winter-rainfall region (Verboom et al., in press). The estimated divergence of the ‘Landtia’ clade in the late Miocene follows the change to a more humid climate in the early midMiocene about 16 Mya (Dingle and Hendey, 1984). Mesic conditions persisted in southern Africa for about 7–8 Mya until the late Miocene (Partridge, 1997). The extant taxa in the ‘Landtia’ clade, as well as H. oocephala and H. scaposa, prefer mesic or perennially/ seasonally wet habitats, including bogs, streambanks and seepages. Dymondia margaretae, the next-diverging extant lineage within Arctotidinae, is a component of the seasonally inundated vlei vegetation that is a significant component of the Agulhas Plain flora (Cowling et al., 1988). Thus a preference for mesic or hydric habitats might be plesiomorphic in Arctotidinae and reflect an ancestral niche. This pattern exhibits similarities to the radiations of Ehrharta (Verboom et al., 2003) and Thamnocortus (Linder and Hardy, 2005), in that the earliest-diverging extant species are found in mesic habitats and that xerophytic adaptations or invasion of semi-arid regions are indicated to be derived. Most of the extant diversity in Arctotidinae was resolved into three well-supported lineages—‘core Arctotis’, ‘Cymbonotus’ and Arctotis sect. Anomalae–‘core Haplocarpha’–Arctotheca clades—that were estimated to have diverged during a rapid radiation centred in southern Africa during the late Miocene or around the Miocene–Pliocene boundary. The present-day centre of Arctotidinae diversity is in the winter-rainfall region, but almost all of the species in this region were indicated to be derived. Based on our dating estimates, the extensive radiation in the winter-rainfall region coincided with the trend towards increasing rainfall seasonality and intensified aridification following increased glaciation in Antarctica from 14 Mya (Zachos et al., 2001), associated with strengthening of the South Atlantic high-pressure cell (Linder, 2005). In addition, the development of the Benguela Current about 11– 14 Mya and summer drought in southwestern Africa are closely associated (Linder, 2003). The lower species diversity in the summer-rainfall region might reflect either a shorter time period avail- 13 able for diversification, less environmental heterogeneity compared to the winter-rainfall region, or the extant diversity might be relictual and that many species radiated but have gone extinct. Our estimated dates for the radiation of Arctotidinae in the winter-rainfall region are consistent with estimates obtained for numerous other plant clades endemic or near-endemic to the region (see Linder, 2003, 2005; Verboom et al., in press). However, molecular-dating analyses indicate that the onset of radiations in the winter-rainfall region in different families has not been coordinated but has occurred over the course of at least the last 20 Mya (e.g. Linder et al., 2003, 2005; Schrire et al., 2003; Klak et al., 2003; Bakker et al., 2004; Mummenhoff et al., 2005; Verboom et al. in press). The ‘core Arctotis’ clade, in which most of the extant diversity occurs in the Fynbos and Succulent Karoo biomes, was estimated to have diverged around the Miocene–Pliocene boundary. This is consistent with the appearance of fynbos vegetation, and a change to a cooler and drier climate in southwestern Africa, around the Miocene–Pliocene boundary (Linder, 2003). Divergence of the annual and perennial clades within ‘core Arctotis’ was estimated to have occurred during the late Pliocene. This followed closure of the Panamanian seaway and abrupt strengthening of the Benguela Current about 3.2 Mya, probably related to increased Antarctic glaciation, which further enhanced the seasonal aridity in the northwestern Cape during the past 3 Mya (Marlow et al., 2000). A hypothesized Pliocene origin of the annual life history in Arctotis is coincident with estimates for development of annualness in southern African Nemesia (Datson et al., 2008). The ‘annual Arctotis’ clade is centreed in the semi-arid Succulent Karoo. The ‘perennial Arctotis’ clade is centred in the Fynbos with outlying species occurring in Namaqualand and the Albany hotspot, and DIVA resolved the ancestral area as being either or both the Fynbos or Succulent Karoo. It is evident that migration in both directions between the Fynbos and Succulent Karoo has occurred in both clades. This might reflect range expansion, contraction and refugial phases during Pleistocene climatic oscillations, which Midgley et al. (2001, 2005) suggested have promoted vicariance and allopatric speciation in the contiguous Fynbos and Succulent Karoo biomes. The Pleistocene climate is characterized by alternating cycles of short interglacial periods (10,000–20,000 years) and longer glacial periods (about 100,000 years; Petit et al., 1999). Expansion and contraction of the winter-rainfall zone during glacial and interglacial periods, respectively, is indicated (Chase and Meadows, 2007). During glacial cycles, much of the western coast of southern Africa and its adjacent interior probably experienced an increase in rainfall concentrated in the winter months (Tankard and Rogers, 1978). This would have facilitated expansion of the mesic and fire-adapted CFR taxa. Conversely, expansion of the semi-arid and drought-adapted Succulent Karoo taxa is hypothesized during the drier interglacial periods. In contrast to ‘core Arctotis’, the extant diversity in the ‘Cymbonotus’ and Arctotis sect. Anomalae–‘core Haplocarpha’–Arctotheca clades is more limited, particularly in the winter-rainfall region, yet they have undergone considerable range expansion in southern Africa and, in the case of ‘Cymbonotus’, successfully crossed the Indian Ocean to Australia. An ancestral area for the Arctotis sect. Anomalae–‘core Haplocarpha’–Arctotheca lineage in the CFR was indicated, implying range expansion followed by speciation in southeastern Africa (Arctotheca populifolia, A. prostrata and H. lyrata), the Nama-Karoo and Succulent Karoo (Arctotheca calendula, A. marginata, Arctotis dregei and A. sulcocarpa). Divergence of the winter-annual Arctotis sect. Anomalae clade possibly during the early Pliocene suggests adaptation to summer-drought conditions. An origin for the ‘Cymbonotus’ clade in southeastern Africa is indicated. The dispersal from southern Africa to Australia, possibly 14 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 during the Pliocene, occurred long after the severing of a land connection between Africa and Antarctica during the Cretaceous (McLoughlin, 2001). The geologic history of the Indian Ocean Basin since the breakup of Gondwana is well documented. There is no geological evidence for a land mass that might have acted as a ‘raft’ or continuous bridge between Africa and Australia during the last approximately 20 My. Dispersal could have been either direct or via the few small isolated islands, such as on the Kerguelen Plateau, which could have acted as terrestrial ‘stepping stones’ for some plants and animals (e.g. Linder et al., 2003; Schwarz et al., 2006). However, the cypselae of all species in the ‘Cymbonotus’ clade lack adaptations for long-distance dispersal (see McKenzie et al., 2005), so the mode of dispersal is difficult to envisage. Increasing evidence indicates that many other Southern Hemisphere plant biogeographic patterns are best explained by hypotheses incorporating dispersal rather than Gondwanan vicariance alone (e.g. Baum et al., 1998; Mummenhoff et al., 2004; Cook and Crisp, 2005; Linder and Barker, 2005; Barker et al., 2007). Bidirectional dispersal of plants is implicated between most of the formerly Gondwanan land masses (Sanmartín and Ronquist, 2004). Molecular dating estimates indicate that trans-Indian-Ocean dispersal has occurred within the last 5 Mya in a number of diverse plant groups (e.g. Bakker et al., 1998; Meerow et al., 2003; Linder and Barker, 2005), and trans-oceanic dispersal from Malesia is implicated for elements of Indian Ocean island floras (Renvoize, 1979), indicating that the Indian Ocean is not an insuperable barrier to angiosperm dispersal. The nesting of H. schimperi within the ‘Cymbonotus’ clade presents another biogeographic puzzle. Haplocarpha schimperi is distributed from Eritrea to Tanzania at moderate to high altitudes (Mesfin Tadesse, 2004). Its sister species, Arctotis microcephala, is distributed in southern Namibia, southern Botswana, northern South Africa and Lesotho, and occurs at high altitudes in the Eastern Cape and Drakensberg mountains, which thus provide a possible afromontane link between the two species. Northward migration of A. microcephala along the eastern African mountain corridor during more mesic Quaternary interglacial periods, followed by extinction in the intervening region in response to increasing aridity, or long-distance migration followed by speciation, are alternative explanations for the present-day distribution. Regardless, divergence of A. microcephala and H. schimperi is indicated to have been recent, possibly during the Pleistocene. In conclusion, this study presents a comprehensive molecular phylogenetic investigation into a group of southern African Asteraceae. Our findings correspond with previous studies on the Cape flora in indicating that radiation might be linked to climate changes around the late Miocene and Pliocene. Only one of the clades identified here (the ‘core Arctotis’ clade) corresponds to a ‘Cape clade’ sensu Linder (2003), and the origins, biogeography, taxonomy and morphology of the subtribe as a whole are rather more complex. The phylogenetic reconstructions resolved wellsupported clades and provide the framework for a revised taxonomic classification of Arctotidinae. Future studies focusing on the basalmost lineages and sister taxa are needed to further elucidate the origins of Arctotidinae. Acknowledgments This study was supported by the Rhodes University Joint Research Committee and the National Research Foundation of South Africa (grant numbers 2042600, 2046932 and 2053645 to N.P.B., and a Postdoctoral Fellowship to R.J.McK.). The authors gratefully thank Coleen Mannheimer, Siro Masinde, Ladislav Mucina, Syd Ramdhani, James Samuel, Peter Weston and Andrew White for collection of material for DNA extraction; the late Josephine Beyers for providing material of Arctotheca marginata; the NSW herbarium for permission to extract DNA from loaned specimens; Sandra Mitchell and Elizabeth Muller for experimental data; Seranne Howis, Rhodes University DNA Sequencing Unit for technical assistance; the late Anne-Marie Brutsch, Ralph Clark, Este Coetzee, Nicolas Devos, Graeme Ellis, Per Ola Karis, Caryl Logie, Roy Lubke, Ladislav Mucina, Syd Ramdhani, Frida Stångberg, Peter Teske and Gerardo Zardi for field-work assistance; Per Ola Karis for many interesting discussions; two anonymous reviewers for comments on the paper. References Alfaro, M.E., Zoller, S., Lutzoni, F., 2003. Bayes or bootstrap? A simulations study comparing the performance of Bayesian Markov Chain Monte Carlo sampling and bootstrapping in assessing phylogenetic confidence. Mol. Biol. Evol. 20, 255–266. Anderson, J.M., Anderson, H.M., Archangelsky, S., Bamford, M., Chandra, S., Dettmann, M.E., Hill, R.S., McLoughlin, S., Rösler, O., 1999. Patterns of Gondwana plant colonisation and diversification. J. Afr. Earth Sci. 28, 145–167. Bakker, F.T., Hellbrügge, D., Culham, A., Gibby, M., 1998. Phylogenetic relationships within Pelargonium sect. Peristera (Geraniaceae) inferred from nrDNA and cpDNA sequence comparisons. Plant Syst. Evol. 211, 273–287. Bakker, F.T., Culham, A., Hettiarachi, P., Touloumenidou, T., Gibby, M., 2004. Phylogeny of Pelargonium (Geraniaceae) based on DNA sequences from three genomes. Taxon 53, 17–28. Baldwin, B.G., 1992. Phylogenetic utility of the internal transcribed spacers of nuclear ribosomal DNA in plants: an example from the Compositae. Mol. Phylogenet. Evol. 1, 3–16. Barker, N.P., von Senger, I., Howis, S., Zachariades, C., Ripley, B.S., 2005. The use of the Internal Transcribed Spacer (ITS) sequence data for plant phylogeography: two examples from the Asteraceae. In: Bakker, F.T., Chatrou, L.W., Gravendeel, B., Pelser, P.B. (Eds.), Plant Species-level Systematics: New Perspectives on Pattern and Process. Regnum Veg. 143. Koeltz, Königstein, pp. 217–244. Barker, N.P., Weston, P.H., Rutschmann, F., Sauquet, H., 2007. Molecular dating of the ‘‘Gondwanan” plant family Proteaceae is only partially congruent with the timing of Gondwanan break-up. J. Biogeog. 34, 2012–2027. Baum, D.A., Small, R.L., Wendel, J.F., 1998. Biogeography and floral evolution of baobabs (Adansonia, Bombacaceae) as inferred from multiple data sets. Syst. Biol. 47, 181–207. Bayer, R.J., Greber, D.G., Bagnall, N.H., 2002. Phylogeny of Australian Gnaphalieae (Asteraceae) based on chloroplast and nuclear sequences, the trnL intron, trnLtrnF intergenic spacer, matk, and ETS. Syst. Bot. 27, 801–814. Beauverd, G., 1915. Contribution à l’étude des Composées (suite X). Bull. Soc. Bot. Genève (sér. 2) 7, 21–56. Bentham, G., 1873. Compositae. In: Bentham, G., Hooker, J.D. (Eds.), Genera plantarum ad exemplaria imprimis in herbariis kewensibus servata definata, vol. 2 (1). Lovell Reeve, London, pp. 163–533. Beyers, J.B.P., 2000. Arctotis and Haplocarpha. In: Goldblatt, P., Manning, J.C. (Eds.), Cape Plants: a Conspectus of the Cape Flora of South Africa. Strelitzia 9. National Botanical Institute, Pretoria, and Missouri Botanical Garden, St. Louis, pp. 303– 306. 329. Beyra-Matos, A., Lavin, M., 1999. A monograph of Pictetia (Papilionoideae; Leguminosae) and review of the Aeschynomeneae. Syst. Bot. Monogr. 56, 1–93. Born, J., Linder, H.P., Desmet, P., 2007. The Greater Cape floristic region. J. Biogeog. 34, 147–162. Chase, B.M., Meadows, M.E., 2007. Late Quaternary dynamics of southern Africa’s winter rainfall zone. Earth-Sci. Rev. 84, 103–138. Cook, L.G., Crisp, M.D., 2005. Not so ancient: the extant crown group of Nothofagus represents a post-Gondwanan radiation. Proc. R. Soc. Lond. B 272, 2535–2544. Cowling, R.M. (Ed.), 1992. The Ecology of Fynbos: Nutrients, Fire and Diversity. Oxford University Press, Cape Town. Cowling, R.M., Campbell, B.M., Mustart, P., McDonald, D.J., Jarman, M.L., Moll, E.J., 1988. Vegetation classification in a floristically complex area: the Agulhas Plain. South Afr. J. Bot. 54, 290–300. Cowling, R.M., Hilton-Taylor, C., 1997. Phytogeography, flora and endemism. In: Cowling, R.M., Richardson, D.M., Pierce, S.M. (Eds.), Vegetation of Southern Africa. Cambridge University Press, Cambridge, pp. 43–61. Datson, P.M., Murray, B.G., Steiner, K.E., 2008. Climate and the evolution of annual/ perennial life-histories in Nemesia (Scrophulariaceae). Plant Syst. Evol. 270, 39– 57. Dean, W.R.J., Milton, S.J. (Eds.), 1999. The Karoo: Ecological Patterns and Processes. Cambridge University Press, Cambridge. de Queiroz, A., 2005. The resurrection of oceanic dispersal in historical biogeography. Trends Ecol. Evol. 20, 68–73. Dingle, R.V., Hendey, Q.B., 1984. Late Mesozoic and Tertiary sediment input into the eastern Cape basin (S.E. Atlantic) and palaeodrainage systems in south western Africa. Marine Geol. 56, 13–26. Doyle, J.J., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of fresh leaf material. Phytochem. Bull. 19, 11–15. Drummond, A., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol. Biol. 7, 214. R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Farris, J.S., Källersjö, M., Kluge, A.G., Bult, C., 1994. Testing significance of congruence. Cladistics 10, 315–319. Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–791. Funk, V.A., Chan, R., Holland, A., 2007. Cymbonotus (Compositae: Arctotideae, Arctotidinae): an endemic Australian genus embedded in a southern African clade. Bot. J. Linn. Soc. 153, 1–8. Funk, V.A., Chan, R., Keeley, S.C., 2004. Insights into the evolution of the tribe Arctoteae (Compositae: subfamily Cichoroideae s.s.) using trnL-F, ndhF and ITS. Taxon 53, 637–655. Galley, C., Bytebier, B., Bellstedt, D.U., Linder, H.P., 2007. The Cape element in the Afrotemperate flora: from Cape to Cairo? Proc. R. Soc. Lond. B 274, 535–543. Galley, C., Linder, H.P., 2006. Geographical affinities of the Cape flora, South Africa. J. Biogeog. 33, 236–250. Germishuizen, G., Meyer, N.L. (Eds.), 2003. Plants of Southern Africa: an Annotated Checklist. Strelitzia 14. National Botanical Institute, Pretoria. Goertzen, L.R., Cannone, J.J., Gutell, R.R., Jansen, R.K., 2003. ITS secondary structure derived from comparative analysis: implications for sequence alignment and phylogeny of the Asteraceae. Mol. Phylogenet. Evol. 29, 216–234. Goldblatt, P., 1978. An analysis of the flora of southern Africa: its characteristics, relationships, and origins. Ann. Missouri Bot. Gard. 65, 369–436. Good, R., 1964. The Geography of the Flowering Plants. Longmans Green and Co., London. Third edition. Graur, D., Martin, W., 2004. Reading the entrails of chickens: molecular timescales and the illusion of precision. Trends Genet. 20, 80–86. Harvey, W.H., 1865. Compositae. In: Harvey, W.H., Sonder, O.W. (Eds.), Flora Capensis, Being a Systematic Description of the Plants of the Cape Colony, Caffraria and Port Natal, vol. 3. L. Reeve, London, pp. 44–530. Hedberg, O., 1957. Afroalpine vascular plants: a taxonomic revision. Symbol. Bot. Upsal. 15, 1–412. Hilliard, O.M., 1977. Compositae in Natal. University of Natal Press, Pietermaritzburg. Hillis, D.M., Moritz, C., Porter, C.A., Baker, R.J., 1991. Evidence for biased gene conversion in concerted evolution of ribosomal DNA. Science 251, 308–310. Hipp, A.L., Hall, J.C., Sytsma, K.J., 2004. Congruence versus phylogenetic accuracy: revisiting the incongruence length difference test. Syst. Biol. 53, 81–89. Holland, A.E., Funk, V.A., 2006. A revision of Cymbonotus (Compositae: Arctotideae, Arctotidinae). Telopea 11, 266–275. Huelsenbeck, J.P., Ronquist, F., 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. Hurka, H., Paetsch, M., Bleeker, W., Neuffer, B., 2005. Evolution within the Brassicaceae. Nova Acta Leopold. N. F. 92, 113–127. Karis, P.O., 2006. Morphological data indicates two major clades of the subtribe Gorteriinae (Asteraceae–Arctotideae). Cladistics 22, 199–221. Karis, P.O., 2007. Tribe Arctotideae Cass. (1819). In: Kadereit, J.W., Jeffrey, C. (Eds.), The Families and Genera of Vascular Plants, vol. 8. Springer-Verlag, Berlin, pp. 200–206. Käss, E., Wink, M., 1997. Phylogenetic relationships in the Papilionideae (family Leguminosae) based on nucleotide sequences of cpDNA (rbcL) and nrDNA (ITS 1 and 2). Mol. Phylogenet. Evol. 8, 65–88. Kay, K.M., Whittall, J.B., Hodges, S.A., 2006. A survey of ribosomal internal transcribed spacer substitution rates across angiosperms: an approximate molecular clock with life history rates. BMC Evol. Biol. 6, 36. Kim, K.J., Choi, K.S., Jansen, R.K., 2005. Two chloroplast DNA inversions originated simultaneously during the early evolution of the sunflower family (Asteraceae). Mol. Biol. Evol. 22, 1783–1792. Klak, C., Reeves, G., Hedderson, T., 2003. Unmatched tempo of evolution in southern African semi-desert ice plants. Nature 427, 63–65. Kluge, A.G., 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates (Boidae, Serpentes). Syst. Zool. 38, 7–25. Koekemoer, M., 1996. An overview of the Asteraceae of southern Africa. In: Hind, D.J.N., Beentje, H.J. (Eds.), Compositae: Systematics. Royal Botanic Gardens, Kew, pp. 95–110. Lee, M.S.Y., 2001. Uninformative characters and apparent conflict between molecules and morphology. Mol. Biol. Evol. 18, 676–680. Lessing, C.F., 1831. Synanthereae Rich. In: Chamisso, L.C.A. von, Schlechtendal, D.F.L. von (Eds.), De plantis in expeditione speculatoria Romanzoffiana observatis rationem dicunt, pp. 83-170. Linnaea 6, 76–170. Lessing, C.F., 1832. Synopsis generum compositarum earumque dispositionis novae tentamen monographiis multarum capensium interjectis. Duncker and Humblot, Berlin. Lewin, K., 1922. Systematische Gliederung und geographische Verbreitung der Arctotideae–Arctotidinae. Repert. spec. nov. regni veg., Beih. 11, 1–75. Linder, H.P., 1990. On the relationship between the vegetation and floras of the afromontane and the Cape regions of Africa. Mitt. Inst. Allg. Bot. Hamburg 23b, 777–790. Linder, H.P., 2003. The radiation of the Cape flora, southern Africa. Biol. Rev. 78, 597–638. Linder, H.P., 2005. Evolution of diversity: The Cape flora. Trends Plant Sci. 10, 536– 541. Linder, H.P., Barker, N.P., 2005. From Nees to now—changing questions in the systematics of the grass subfamily Danthonioideae. Nova Acta Leopold. N. F. 92, 29–44. Linder, H.P., Eldenäs, P., Briggs, B.G., 2003. Contrasting patterns of radiation in African and Australian Restionaceae. Evolution 57, 2688–2702. 15 Linder, H.P., Hardy, C.R., 2005. Species richness in the Cape flora: a macroevolutionary and macroecological perspective. In: Bakker, F.T., Chatrou, L.W., Gravendeel, B., Pelser, P.B. (Eds.), Plant Species-level Systematics: New Perspectives on Patterns and Process. Koeltz, Königstein, pp. 47–73. Linder, H.P., Hardy, C.R., Rutschmann, F., 2005. Taxon sampling effects in molecular clock dating: an example from the African Restionaceae. Mol. Phylogenet. Evol. 35, 569–582. Maddison, D.R., 1991. The discovery and importance of multiple islands of mostparsimonious trees. Syst. Biol. 40, 315–328. Maddison, W.P., Maddison, D.R., 2000. MacClade 4: Analysis of Phylogeny and Character Evolution. Sinauer Associates, Sunderland. Marlow, J.R., Lange, C.B., Wefer, G., Rosell-Melé, A., 2000. Upwelling intensification as part of the Pliocene-Pleistocene climate transition. Science 290, 2288–2291. Mason-Gamer, R.J., Kellogg, E.A., 1996. Testing for phylogenetic conflict among molecular data sets in the tribe Triticeae (Gramineae). Syst. Biol. 45, 524–545. McKenzie, R.J., Herman, P.P.J., Barker, N.P., 2006b. Arctotis decurrens (Arctotideae), the correct name for A. merxmuelleri and A. scullyi. Bothalia 36, 171–173. McKenzie, R.J., Hjertson, M., Barker, N.P. in press. Typification of Arctotis plantaginea and names in the Arctotis semipapposa species complex (Asteraceae, Arctotideae). Taxon.. McKenzie, R.J., Mitchell, S.D., Barker, N.P., 2006c. A new species of Arctotis (Compositae, Arctotideae) from kommetjie grassland in Eastern Cape Province, South Africa. Bot. J. Linn. Soc. 151, 581–588. McKenzie, R.J., Muller, E.M., Skinner, A.K.W., Karis, P.O., Barker, N.P., 2006a. Phylogenetic relationships and generic delimitation in subtribe Arctotidinae (Asteraceae: Arctotideae) inferred by DNA sequence data from ITS and five chloroplast regions. Am. J. Bot. 93, 1222–1235. McKenzie, R.J., Samuel, J., Muller, E.M., Skinner, A.K.W., Barker, N.P., 2005. Morphology of cypselae in subtribe Arctotidinae (Compositae–Arctotideae) and its taxonomic implications. Ann. Missouri Bot. Gard. 92, 569–594. McLoughlin, S., 2001. The breakup history of Gondwana and its impact on preCenozoic floristic provincialism. Austral. J. Bot. 49, 271–300. Meerow, A.W., Lehmiller, D.J., Clayton, J.R., 2003. Phylogeny and biogeography of Crinum L. (Amaryllidaceae) inferred from nuclear and limited plastid noncoding DNA sequences. Bot. J. Linn. Soc 141, 349–363. Mesfin Tadesse, 2004. Asteraceae (Compositae). In: Hedberg, I., Friis, I., Edwards, S. (Eds.), Flora of Ethiopia and Eritrea, vol. 4 (2). National Herbarium, Addis Ababa University, Addis Ababa and Department of Systematic Botany, Uppsala University, Uppsala, pp. 1–408. Midgley, G.F., Hannah, L., Roberts, R., Macdonald, D.J., Allsopp, J., 2001. Have Pleistocene climatic cycles influenced species richness patterns in the Greater Cape Mediterranean Region? J. Mediterr. Ecol. 2, 137–144. Midgley, G.F., Reeves, G., Klak, C., 2005. Late Tertiary and Quaternary climate change and centres of endemism in the southern African flora. In: Purvis, A., Gittleman, J., Brooks, T. (Eds.), Phylogeny and Conservation. Cambridge University Press, Cambridge, pp. 230–242. Mummenhoff, K., Al-Shehbaz, I.A., Bakker, F.T., Linder, H.P., Mühlhausene, A., 2005. Phylogeny, morphological evolution, and speciation of endemic Brassicaceae genera in the Cape flora of southern Africa. Ann. Missouri Bot. Gard. 92, 400– 424. Mummenhoff, K., Linder, H.P., Friesen, N., Bowman, J.L., Lee, J.Y., Franzke, A., 2004. Molecular evidence for bicontinental hybridogenous genomic constitution in Lepidium sensu stricto (Brassicaceae) species from Australia and New Zealand. Am. J. Bot. 91, 254–261. Nixon, K.C., Carpenter, J.M., 1996. On simultaneous analysis. Cladistics 12, 221–241. Partridge, T.C., 1997. Evolution of landscapes. In: Cowling, R.M., Richardson, D.M., Pierce, S.M. (Eds.), Vegetation of Southern Africa. Cambridge University Press, Cambridge, pp. 5–20. Petit, J.R., Jouzel, J., Raynaud, D., Barkov, N.I., Barnola, J.M., Basile, I., Bender, M., Chappellaz, J., Davis, M., Delayque, G., Delmotte, M., Kotlyakov, V.M., Legrand, M., Lipenkov, V.Y., Lorius, C., Pépin, L., Ritz, C., Saltzman, E., Stievenard, M., 1999. Climate and atmospheric history of the past 420 000 years from the Vostok ice core, Antarctica. Nature 399, 429–436. Phillips, E.P., 1951. The Genera of South African Flowering Plants. Department of Agriculture, Pretoria. Pickford, M., 2004. Southern Africa: a cradle of evolution. South Afr. J. Sci. 100, 205– 214. Pope, G.V., 1992. Compositae. In: Pope, G.V. (Ed.), Flora Zambesiaca: Mozambique, Malawi, Zambia, Zimbabwe & Botswana, vol. 6 (1). Royal Botanic Gardens, Kew, pp. 1–264. Price, J.P., Clague, D.A., 2002. How old is the Hawaiian biota? Geology and phylogeny suggest recent exchange. Proc. R. Soc. Lond. B 269, 2429–2435. Ramírez, M.J., 2006. Further problems with the incongruence length difference test: ‘‘hypercongruence” effect and multiple comparisons. Cladistics 22, 289–295. Reese, H., 1989. Die Entwicklung von Pericarp und Testa bei Calenduleae und Arctotideae (Asteraceae)–ein Beitrag zur Systematik. Bot. Jahrb. Syst. 110, 325– 419. Renvoize, S.A., 1979. The origins of Indian Ocean island floras. In: Bramwell, D. (Ed.), Plants and Islands. Academic Press, London, pp. 107–129. Ronquist, F., 1997. Dispersal–vicariance analysis: a new approach to the quantification of historical biogeography. Syst. Biol. 46, 193–201. Sang, T., Crawford, D.J., Stuessy, T.F., 1997. Chloroplast DNA phylogeny, reticulate evolution and biogeography of Paeonia (Paeoniaceae). Am. J. Bot. 84, 1120– 1136. Sang, T., Crawford, D.J., Stuessy, T.F., Solva, O.M., 1995. ITS sequences and the phylogeny of the genus Robinsonia (Asteraceae). Syst. Bot. 20, 55–64. 16 R.J. McKenzie, N.P. Barker / Molecular Phylogenetics and Evolution 49 (2008) 1–16 Sang, T., Zhong, Y., 2000. Testing hybridization hypotheses based on incongruent gene trees. Syst. Biol. 49, 422–434. Sanmartín, I., Ronquist, F., 2004. Southern Hemisphere biogeography inferred by event-based models: plant versus animal patterns. Syst. Biol. 53, 216–243. Schmidt, G.J., Schilling, E.E., 2000. Phylogeny and biogeography of Eupatorium (Asteraceae: Eupatorieae) based on nuclear ITS sequence data. Am. J. Bot. 87, 716–726. Schwarz, M.P., Fuller, S., Tierney, S.M., Cooper, S.J.B., 2006. Molecular phylogenetics of the exoneurine allodapine bees reveal an ancient and puzzling dispersal from Africa to Australia. Syst. Biol. 55, 31–45. Schrire, B.D., Lavin, M., Barker, N.P., Cortes-Burns, H., von Senger, I., Kim, J.H., 2003. Towards a phylogeny of Indigofera (Leguminosae—Papilionoideae): identification of major clades and relative ages. Adv. Legume Syst. 10, 269–302. Simmons, M.P., Ochoterena, H., 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49, 369–381. Soltis, D.E., Soltis, P.S., 1998. Choosing an approach and an appropriate gene for phylogenetic analysis. In: Soltis, D.E., Soltis, P.S., Doyle, J.J. (Eds.), Molecular Systematics of Plants II: DNA Sequencing. Kluwer, Boston, pp. 1–42. Swofford, D.L., 2002. PAUP. Phylogenetic Analysis Using Parsimony (* and Other Methods), v. 4.0 beta 10. Sinauer Associates, Sunderland. Taberlet, P., Gielly, L., Pautou, G., Bouvet, J., 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Mol. Biol. 17, 1105–1109. Tankard, A.J., Rogers, J., 1978. Late Cenozoic palaeoenvironments and palaeoclimates of southern Africa. South Afr. J. Sci. 95, 194–201. Thorne, R.F., 1972. Major disjunctions in the geographic ranges of seed plants. Q. Rev. Biol. 47, 365–411. van Wyk, A.E., Smith, G.F., 2001. Regions of Floristic Endemism in Southern Africa. Umdaus Press, Hatfield. Vences, M., Vieites, D.R., Glaw, F., Brinkmann, H., Kosuch, J., Veith, M., Meyer, A., 2003. Multiple overseas dispersal in amphibians. Proc. R. Soc. Lond. B 270, 2435–2442. Verboom, G.A., Archibald, J.K., Bakker, F.K., Bellstedt, D.U., Conrad, F., Dreyer, L.L., Forest, F., Galley, C., Goldblatt, P., Henning, J.F., Mummenhoff, K., Linder, H.P., Muasya, A.M., Oberlander, K.C., Savolainen, V., Snijman, D.A., van der Niet, T., Nowell, T.L., 2008. Origin and diversification of the Greater Cape flora: Ancient species repository, hot-bed of recent radiation, or both? Mol. Phylogenet. Evol., in press, doi: 10.1016/j.ympev.2008.01.037.. Verboom, G.A., Linder, H.P., Stock, W.D., 2003. Phylogenetics of the grass genus Ehrharta: evidence for radiation in the summer-arid zone of the South African Cape. Evolution 57, 1008–1021. White, T.J., Bruns, T., Lee, S., Taylor, J., 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelfand, D.H., Sninsky, J.J., White, T.J. (Eds.), PCR Protocols: a Guide to Methods and Applications. Academic Press, San Diego, pp. 315–324. Yang, Z., Rannala, B., 1997. Bayesian phylogenetic inference using DNA sequences: a Markov Chain Monte Carlo method. Mol. Biol. Evol. 14, 717–724. Yoder, A.D., Irwin, J.A., Payseur, B.A., 2001. Failure of the ILD to determine data combinability for slow loris phylogeny. Syst. Biol. 50, 408–424. Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686–693. Zavada, M.S., de Villiers, S.E., 2000. Pollen of the Asteraceae from the PaleoceneEocene of South Africa. Grana 39, 39–45.

RELATED PAPERS

RELATED TOPICS

Log In

McKenzie, R.J. & Barker, N.P. (2008). Radiation and biogeography of the predominantly southern African subtribe Arctotidinae (Asteraceae: Arctotideae).

McKenzie, R.J. & Barker, N.P. (2008). Radiation and biogeography of the predominantly southern African subtribe Arctotidinae (Asteraceae: Arctotideae).

Related Papers

RELATED PAPERS

RELATED TOPICS