TAXON 59 (1) • February 2010: 57–67 Devos & al. • Phylogeny and biogeography of Euryops A multi-locus phylogeny of Euryops (Asteraceae, Senecioneae) augments support for the “Cape to Cairo” hypothesis of floral migrations in Africa NicolasD evos,1,4 Nigel P. Barker,1 Bertil Nordenstam2 & Ladislav Mucina3,5 1 Molecular Ecology and Systematics Group, Department of Botany, Rhodes University, P.O. Box 94, Grahamstown, 6140, South Africa 2 Department of Phanerogamic Botany, Swedish Museum of Natural History, Box 50007, 104 05 Stockholm, Sweden 3 Department of Botany & Zoology, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa 4 Institute of Botany, University of Liege, B-22 Sart Tilman, 4000 Liege, Belgium 5 Department of Environmental and Aquatic Sciences, Curtin University of Technology, GPO Box U1987, Perth, WA 6945, Australia Author for correspondence: Nicolas Devos, ndevos@ulg.ac.be Abstract With about 100 species, Euryops (Cass.) Cass. ranks among the most speciose genera of the tribe Senecioneae (Asteraceae). The genus has its greatest diversity in South Africa, and displays an interesting disjunct distribution with most of the taxa found in southern Africa and a group of eight endemic species confined to the mountains of tropical East Africa and northeastern Africa. Molecular phylogenetic analyses of DNA sequence data from three chloroplast fragments and the nuclear ITS region were used to reconstruct the evolutionary history of 41 Euryops species in order to unravel species relationships and to determine the origin of the disjunct Afromontane taxa. Our results show a lack of support and resolution in the internal structure of the trees, but also reveal strong incongruence between the ITS and cpDNA datasets as assessed by Bayes Factors. We hypothesise that this is a consequence of the isolation and divergence of many populations over a short time period at some point in the history of the genus. Molecular dating based on our phylogenetic tree suggests that the genus diversified in South Africa around four million years ago. The origin of the East African species, dated at 1.9 Ma, well after the uplift of the East African mountains, is consistent with a scenario of a single dispersal event from South Africa northwards into the tropical East African mountains where diversification occurred, creating a monophyletic group of regional Afromontane endemics. Keywords Asteraceae; Bayes Factors; chloroplast DNA; disjunct distribution; Euryops; ITS; lineage sorting: migration routes; molecular dating; radiation; topological incongruence INTRODUCTION The genus Euryops (Cass.) Cass., revised by Nordenstam (1968a), comprises 97 described species plus another five awaiting description or recently described (Nordenstam & al., 2009; Nordenstam & al., in prep.). Along with Othonna L., it is the largest genus (in terms of number of species and lowerrank taxa) in the subtribe Othonninae (Asteraceae, Senecioneae), which also includes smaller genera such as Gymnodiscus Less. (two species, only found in South Africa), Hertia Less. (about ten species, found in Africa and Southwest Asia, five of which occur in South Africa), and Lopholaena DC. (18 species, found in southern Africa). With the exception of one annual herb (E. annuus Compt.), all members of Euryops are perennial shrubs characterised by coriaceous leaves and yellow or orange-flowered capitula on simple peduncles usually devoid of leaves or bracts. A comprehensive view of the evolutionary relationships within the genus is still lacking. However, based on various vegetative and floral morphological characters, the genus was divided into six sections (Euryops sects. Angustifoliae, Euryops, Chrysops, Brachypus, Psilosteum, Leptorrhiza)—all believed to represent groups having monophyletic origins (Nordenstam, 1968a). Euryops is restricted to Africa, except for one species, E. arabicus Steud., which is found in Arabia and Socotra as well as in the Horn of Africa (Nordenstam, 1968a, 1969; Fig. 1). The greatest diversity of the genus occurs in southern Africa, mostly restricted to South Africa. While several species are confined to fynbos shrublands of the Cape Floristic Region (CFR), the genus cannot be considered as a true “Cape clade” sensu Linder (2003) as it does not comprise at least 50% of Cape endemics. Chorological analysis of the genus indicates that the greatest concentration of species is found in the region of the highest peaks and mountain blocks along the Great Escarpment, spanning the Mpumalanga escarpment in the northeast of South Africa, the Drakensberg, the Sneeuberg in the Eastern Cape through to the Hantam mountains straddling the border of the Western and Northern Cape Provinces (Fig. 1). Not all species are found at high altitudes, and many of the low-altitude taxa are quite widespread, such as those found in the Albany Centre of Endemism (Fig. 1; Nordenstam, 1968a, 1969). While the genus is an important component of the Cape flora, it generally appears to be restricted to more peripheral mountain regions of the CFR close to or linked to the Great Escarpment. Eight species of Euryops are confined to the high mountains of tropical East Africa, Ethiopia, Somalia and southwestern Arabia (Fig. 1; Nordenstam, 1968a, 1969). The East African and Ethiopian species form part of the Afromontane flora. They are found only at high altitudes in one or a few 57 Devos & al. • Phylogeny and biogeography of Euryops of the high mountain ranges of Tanzania, Kenya and Ethiopia. Euryops elgonensis Mattf. is endemic on Mount Elgon (Kenya), E. dacrydioides Oliv. in Hook. is found only on Mount Kilimanjaro (Tanzania), E. brownei S. Moore occurs on four mountains (Kenya, Aberdare, Meru, Cherangani), E. arabicus Steud. occurs on high mountains in Somalia, Socotra and southern Arabia, E. pinifolius A. Rich. is found on the high mountains of Ethiopia, E. prostratus B. Nord. is endemic to the Bale mountains (Ethiopia), E. antinorii S. Moore is endemic to the Soha mountains (Ethiopia), whereas E. jacksonii S. Moore is restricted to the Aberdares and the Mau escarpment (Kenya). The ecology of these high-altitude taxa resembles that of the endemic taxa of high mountains of South Africa such as the Drakensberg and Sneeuberg. Because of this similarity in ecology and morphology, these two disjunct species groups are believed to be closely related (Nordenstam, 1968a, 1969). The floristic affinities between the CFR, the Greater Drakensberg (incl. Northern Escarpment, Southern Drakensberg, Sneeuberg and possibly also Amathole Mountains) and the East African Afromontane centre have been well documented (Weimarck, 1941; Killick, 1963, 1978; Wild, 1968; White, 1978; Linder, 1990; Carbutt & Edwards, 2004; Galley & Linder, 2006; Galley & al., 2007; Clark & al., 2009). Apart TAXON 59 (1) • February 2010: 57–67 from Euryops, many other plant lineages are disjunctly distributed between those three endemic centres. However, the historical origin of those disjunctions remains difficult to understand, and three competing hypotheses have been proposed towards their explanation. Two of these hypotheses invoke migration, while the third invokes vicariance. The first migration scenario suggests a northward expansion, with the Drakensberg and other Great Escarpment mountain blocks playing an important role as ‘stepping-stones’ for lineages migrating from the Cape northwards into the tropical Afromontane region (Weimarck, 1941; Holland, 1978; Linder, 1994; Galley & Linder, 2006; Galley & al., 2007). The second migration scenario implies opposite processes—southward migration from tropical Africa (Levyns, 1938, 1952, 1964; Axelrod & Raven, 1978) followed by radiation in the CFR. In contrast to these two migration (dispersal) hypotheses, some authors suggested a vicariance scenario where the current floras in each of the centres of diversity and endemism represent relics of a once more widespread floristic stock which underwent fragmentation due to climate change (Adamson, 1958; Wild, 1968). Recent reviews of the available literature (Galley & Linder, 2006; Galley & al., 2007) suggested that for many of the so-called “Cape clades” a northward migration took place—a claim Fig. 1. Satellite image of Africa, with the distribution of Euryops indicated by the black outlines. The inset map of southern Africa indicates the chorology of Euryops, modified from Nordenstam (1969: map 4) and PRECIS data. Isochores are in 3-species intervals, with highest numbers for each centre or region provided. Letters in upper case indicate centres of endemism, named following Nordenstam (1969: map 111): A, Albany; B, Barberton; C, Caledon; D, Drakensberg; G, Gariep; S, Sneeuberg; WUK, Western Upper Karoo. Lower case letters indicate centres as named by Koekemoer (1996): h, Hantam; m, Middelburg. 58 TAXON 59 (1) • February 2010: 57–67 now being considered the prevailing hypothesis explaining the distribution of many taxa shared between the Cape, the Drakensberg and other Afromontane regions. Euryops is, however, regarded as an old and widespread member of the African flora which differentiated in the Paleogene (Nordenstam, 1969), and therefore may not share the same biogeographic history common with the typically evolutionary young Cape clades which served as the basis of the biogeographic syntheses by Galley & Linder (2006) and Galley & al. (2007). The objectives of our work were: (1) to reconstruct a species-level phylogeny of Euryops using sequence data from four regions: the nuclear ribosomal internal transcribed spacer (ITS) region, the trnL-trnF and trnT-trnL spacers of the chloroplast genome, and the chloroplast rps16 region; (2) to assess the phylogenetic relationships of the high-altitude Euryops species in order to better understand the origin of the disjunct distribution displayed by Euryops; and (3) to shed light on the time of origin of Euryops using molecular dating techniques taking into account phylogenetic uncertainty and variation in rate of evolution between lineages in the tree. MATERIALS AND METHODS Plant material. — Forty-one of the ninety-seven Euryops species recognised by Nordenstam (1968a) were sampled for this study. Multiple accessions were sampled whenever possible so as to assess species monophyly. This sampling strategy increased the number of terminal accessions for the ingroup to 65. Several representatives of each one of the six sections recognised within Euryops were included in our sample. Based on the Senecioninae phylogeny of Pelser & al. (2007), Gymnodiscus capillaris Less., Othonna sedifolia DC., O. carnosa Less., O. euphorbioides Hutch., O. eriocarpa (DC.) Sch. Bip. and Hertia pallens (Benth. & Hook. f.) Kuntze were selected as outgroups. The vast majority of the samples used in this study were collected from the field in South Africa and preserved in silica gel for subsequent DNA extraction (Chase & Hills, 1991). Herbarium vouchers for these specimens are deposited at the Selmar Schönland Herbarium (GRA) and at the Herbarium of Stellenbosch University (STEU). These samples were supplemented by leaf material from further herbarium sheets housed at GRA, Uppsala Herbarium (UPS) and the East African Herbarium (EA) (see Appendix for details). DNA extraction, amplification, and sequencing. — Total genomic DNA was isolated from about 1 cm² of dried leaf tissue using a CTAB extraction protocol (Doyle & Doyle, 1987), but without the RNAase treatment. Three chloroplast DNA regions (rps16, trnT-trnL, trnL-trnF) and the ITS region of the nuclear ribosomal DNA were amplified and sequenced. The primers rps16-F and rps16-2R (Oxelman & al., 1997) were used to amplify and sequence the rps16 gene. The primers trnA and trnB (Taberlet & al., 1991) and, trnC and trnF (Taberlet & al., 1991) were used to amplifly and sequence the trnT-trnL and trnL-trnF loci, respectively. The ITS region was amplified and sequenced using the ITS4 (White & al., 1990) and ITS5 Devos & al. • Phylogeny and biogeography of Euryops (Stanford & al., 2000) primers. The PCR products obtained were purified using the PCR purification kit (Promega, Madison, Wisconsin, U.S.A.). Sequencing reactions were performed in both directions using the BigDye® Terminator v.3.1 cycle sequencing kit (Applied Biosystems, Foster City, California, U.S.A.) and the same primers as used for PCR amplifications. It is well known that the ITS region in plants often harbors extensive sequence variation, due to processes such as duplication, pseudogene decay or incomplete concerted evolution (Alvarez & Wendel, 2003). As a result, ITS sequences obtained from different species could in fact represent multiple paralogues that would produce a wrong estimate of the phylogenetic relationships between the species sampled when used in a phylogenetic analysis. We cannot be sure that all our sequences are orthologues. However, every single PCR reaction produced only one ITS amplicon and consequently, no cloning was required before sequencing. Direct sequencing of those amplicons produced clean chromatographs and no double peaks were found. This suggests that the universal ITS primers used were amplifying a single ITS copy. We do not have any reason to believe that those universal primers are actually preferentially amplifying different paralogues in the different accessions sequenced for this study. Sequence alignment. — Forward and reverse sequences were assembled and edited using Sequencher v.4.8 (Gene Codes Corporation, 1998). Contigs were aligned manually using MacClade 4.1 (Maddison & Maddison, 1992), with gaps inserted where necessary to preserve positional homology. Positions that were ambiguously aligned were excluded from the analyses. All sequences determined for this study were deposited in GenBank (see Appendix for accession numbers) and sequence alignments were deposited in TreeBase (http://www .treebase.org) under matrix accession number M4908. Phylogenetic a nalyses. — Phylogenetic relationships were assessed from the data using Bayesian inference (BI) as implemented by MrBayes v.3.1 (Ronquist & Huelsenbeck, 2003). The models for nucleotide substitutions that best fitted each of the datasets being analyzed (either the ITS or cpDNA) were selected using the Akaike Information Criterion (AIC; Akaike, 1974). This was done using MrModeltest v.2.2 (Nylander, 2004) in conjunction with PAUP* (Swofford, 2002). The prior distributions for the parameters in those models were set as follows: a flat Dirichlet prior was used for both stationary state frequencies and nucleotide substitution, a uniform prior was used on the shape parameter of the gamma distribution of rate variation (uniform [0.0, 200.00]), and an exponential prior, Exp(10), was used for branch lengths, and all trees were assumed to be equally likely. The nuclear and chloroplast partitions were analyzed separately to identify possible incongruence between partitions. Potential conflict in the signals displayed by the different partitions was searched for by comparing both topologies and posterior probabilities obtained from each one of the two datasets. Taxa that seemed to have an incongruent position supported by more than 0.95 posterior probabilities in the topologies being compared were pruned prior to combining the ITS and cpDNA datasets. For each dataset (ITS, cpDNA and combined), three 59 Devos & al. • Phylogeny and biogeography of Euryops independent runs were carried out using MrBayes v.3.1 (Ronquist & Huelsenbeck, 2003). Each run consisted of one chain that ran from a random starting tree for a total of three million generations. Trees were sampled every 10,000th generation. For the combined dataset, model parameters (with the exception of branch lengths and topology) were unlinked using the “unlink” command in MrBayes such that each partition in the combined analysis had its own set of parameters. Convergence of parameters between the three runs was confirmed by checking that standard deviation of split frequencies was below 0.01 and that the potential scale reduction factor was close to 1 for all parameters estimated. After discarding the burn-in, outputs from the three runs were combined for final inference of posterior probabilities of both trees and model parameters. Incongruence and hypotheses testing. — Congruence between the ITS and cpDNA datasets was statistically tested with the parsimony-based ILD test of Farris & al. (1994) as implemented in PAUP* v.4.0b10 using the heuristic search option with random sequence addition (100 random replications) and TBR branch-swapping. This ILD-test, also known as the partition homogeneity test in PAUP* (Swofford, 2002), has been commonly used to assess topological congruence and more generally combinability between partitions prior to phylogenetic analysis of large heterogeneous datasets. However, the ILD test has been strongly criticised as a measure of dataset heterogeneity and, consequently, as a test for combinability (Cunningham, 1997; Yoder & al., 2001; Barker & Lutzoni, 2002). As an alternative, the use of Bayes Factors within a Bayesian framework has been recently proposed to test for congruence between partitions (Irestedt & al., 2004; Nylander & al., 2004). The Bayesian approach of phylogenetic inference provides, through the calculation of Bayes Factors (hereafter BF), a convenient way for comparing how well two models describe the processes generating a dataset X (Kass & Raftery, 1995). Unlike a hierarchical likelihood ratio test, the models compared by mean of BF do not need to be hierarchically nested. BF is calculated as the ratio of the marginal likelihoods f(X|Mi) of the models being compared, B12 = f(X|M1) / f(X|M2). The marginal likelihood of a model is difficult to calculate accurately but can be roughly estimated as the harmonic mean of the likelihood values of the MCMC samples (Newton & Raftery, 1994). A value > 10 for 2 log B12 has been suggested as a very strong evidence that model 1 is far superior than the alternative model at describing the processes generating the dataset X (Kass & Raftery, 1995). This Bayes Factor model selection approach is used here to explore the potential incongruence between the ITS and the cpDNA partitions either with or without the taxa identified by eye as incongruent. The two models being compared when testing for incongruence between partitions are: (1) a “onetree” model under which the two partitions were constrained to evolve on the same topology; and (2) a “two-tree” model under which the two partitions are allowed to explore the treespace separately (“unlink” command in MrBayes). In both analyses, model parameters and branch-lengths were unlinked across the partitions. For both models, three independent runs with one chain each were run from a random starting tree for 60 TAXON 59 (1) • February 2010: 57–67 a total of three million generations. Model parameters and tree topologies were sampled every 10,000 generations. After convergence had been confirmed and the burn-in discarded, harmonic means were calculated over all three runs using the “sump” command in MrBayes. Twice the difference between the harmonic mean of the log-likelihoods of the “one-tree” model and the “two-tree” model was then computed. Molecular dating. — Both the ITS and cpDNA datasets were analyzed using a relaxed Bayesian approach as implemented in BEAST v.1.4.8 (Drummond & al., 2006; Drummond & Rambaut, 2007). In contrast to other dating methods, BEAST uses a Bayesian Markov chain Monte Carlo method to simultaneously estimate topology along with the node ages (Drummond & al., 2002). All BEAST analyses were run in the absence of topological constraints under a general time reversible model of nucleotide substitution with rate variation among sites modeled using a gamma shape parameter with four rate categories. Molecular evolution model parameters used flat priors, whilst tree priors were modeled according to a Yule speciation process. The divergence dates estimated were integrated over the various tree topologies sampled throughout the MCMC analysis, and weighted in proportion to their posterior probabilities. The MCMC chains were each run for 40 million generations, and parameters were sampled every 1000th generation. Inspection of the results using Tracer v.1.4 (Rambaut & Drummond, 2007) confirmed that stationarity and acceptable mixing of the sampled parameters was achieved. Trees were summarised as maximum clade credibility trees after removing, as a burn-in, the first 20% of samples. In the absence of fossil evidence in Euryops, all analyses were calibrated by setting a prior on the mean rate of nucleotide substitution under the uncorrelated log-normal relaxed molecular clock (ucld.mean parameter in BEAST). For the ITS dataset, estimates of absolute rate of molecular evolution found in the literature were used as prior and uncertainties around those estimates were factored using a uniform distribution characterised by a maximum and minimum rate of 0.0078 and 0.003 substitutions per site per million years, respectively. Richardson & al. (2001a) lists a variety of ITS divergence rates which generally fall into a range of 1.72 × 10 –3 substitutions per site per million years (s/s/My) in the Saxifragaceae (Vargas & al., 1998) to a more rapid 7.83 × 10 –3 s/s/My in the Asteraceae (Sang & al., 1995). The average Asteraceae mutation rate can be calculated as 5.21 × 10 –3 s/s/My (based on an average calculated from Richardson & al.’s list of Asteraceae mutation rates) with a lower extreme of 3 × 10 –3 s/s/My (from the Hawaiian silverswords; Baldwin & Sanderson, 1998), and a higher extreme of 7.83 × 10 –3 s/s/My (from Robinsonia DC.; Sang & al., 1995). Based on estimates presented in Palmer (1991), the average absolute rate of nucleotide substitution across a large number of chloroplast genes is 5 × 10 –4 s/s/My. As this estimate has been calculated across a large variety of land plants and various chloroplast genes, it has to be used with caution. It indeed includes both synonymous and non-synonymous substitutions and is likely to underestimate substitution rates for noncoding regions. We however factored uncertainties around TAXON 59 (1) • February 2010: 57–67 Devos & al. • Phylogeny and biogeography of Euryops that estimate by using as a prior a normal distribution with a standard deviation of 1 × 10 –4 centred at 5 × 10 –4 s/s/My. This prior distribution was placed, as for the ITS analysis, on the mean rate of nucleotide substitution parameter (ucld.mean parameter in BEAST). RESULTS A total of 662 aligned base pairs was obtained from the ITS1 + 5.8S + ITS2 regions, 762 bp from the rps16 region, 841 bp from the trnL-trnF region and 346 bp from part of the trnT-trnL region. Of the 1949 characters included in the complete cpDNA dataset, 1694 were constant, 146 were parsimony informative and 388 were identified by the Bayesian analysis as unique site patterns under a GTR + Γ model of nucleotide substitution. Of the 662 aligned base pairs from the ITS dataset, 407 were constant, 179 were identified as parsimony informative and 270 were identified bye the Bayesian analysis as unique site patterns under a GTR + Γ model of nucleotide substitution. According to the Akaike information criterion (AIC), the parameter-rich GTR + Γ model of nucleotide substitution was the best fit for both the ITS and cpDNA partitions. After discarding a burn-in of 20 trees for each one of the three independent runs, the posterior distribution of topologies was inferred and is presented for each analysis as a 50% majority rule consensus tree (Figs. 2, 3). The trees obtained from the analysis of the individual partitions (ITS and cpDNA, including rps16, trnL-F and trnT-L) differ in topology and degree of resolution. The parsimonybased ILD test revealed significant incongruence between the ITS and cpDNA datasets (P = 0.01). Bayes Factors also showed extensive incongruence between both partitions. Comparison of a model which allow the ITS and cpDNA datasets to evolve on separate topology (i.e., independent evolutionary histories) to a model in which both partitions share the same evolutionary history (same topology) gave a 2logBF of 358.24. This value strongly suggests that an unlinked model (“two-tree” model) is superior to the model assuming a common evolutionary history for both partitions and as a consequence indicates strong conflict between the nuclear and chloroplast DNA evolutionary histories. 1 Fig. 2. Phylogenetic relationships of Euryops inferred from sequences of three chloroplast DNA regions (trnL-trnF, trnT-trnL, rps16). Fifty percent majority rule consensus tree of 840 trees sampled by a MCMC chain using the program MrBayes. Posterior probabilities are indicated above branches. Marked clades are discussed within the text. Mountain or alpine species endemic to the Drakensberg and Sneeuberg Mountains of the Great Escarpment are in bold. East African species of Euryops are indicated by a grey box. All other species are from non-mountainous region of South Africa. Divergence dates for the major nodes in million years (Ma) are also given and discussed in the text. 2.2 Ma (0.8–4.7) 5.0 Ma (2.3–9.5) 9.9 Ma (4.7–17.7) 1 18.34 Ma (9.1–33.5) 21.3 Ma (10.6–37.8) 1 1 1 1 1 Clade A E. montanus E. decumbens E. petraeus E. annae E. lateriflorus #1 E. lateriflorus #2 1 1 E. multifidus E. subcarnosus E. dregeanus #1 E. dregeanus #2 E. othonnoides #1 E. brevilobus 1 E. othonnoides #2 E. abrotanifolius #1 E. abrotanifolius #2 1 1 1 0.95 E. speciosissimus #1 E. speciosissimus #2 1 0.51 E. abrotanifolius #3 E. abrotanifolius #4 E. abrotanifolius #5 E. abrotanifolius #6 E. pectinatus E. tenuissimus #1 E. tenuissimus #1 E. linifolius E. brownei Clade B E. arabicus #1 1 E. arabicus #2 1 E. arabicus #3 E. prostratus East Africa 0.99 E. dacrydioides E. dacrydioides E. pinifolius E. trilobus 1 E. lateriflorus #3 E. lateriflorus #4 E. empetrifolius 0.55 E. evansii #1 0.56 E. evansii #2 E. evansii #3 E. acraeus Clade C E. munitus E. euryopoides #1 E. brachypodus #1 0.67 E. brachypodus #2 E. hypnoides 0.99 E. spathaceus #1 1 E. spathaceus #2 0.99 E. anthemoides #1 E. anthemoides #2 E. hebecarpus E. ericoides E. ericifolius E. ursinoides E. euryopoides #2 E. candollei 0.97 Clade D E. algoensis E. virgineus #1 E. virgineus #2 E. galpinii E. tysonii E. annuus Hertia pallens Othonna sedifolia #1 O. sedifolia #2 O. carnosa Othonna sp. nov. 1 O. eriocarpa O. euphorbioides Gymnodiscus capillaris #1 G. capillaris #2 61 TAXON 59 (1) • February 2010: 57–67 Devos & al. • Phylogeny and biogeography of Euryops Thirty-two terminals with conflicting positions between the ITS and cpDNA trees were pruned before combining the datasets. Those terminals were identified by eye and considered as having an incongruent position only if their respective positions in both trees were supported by posterior probabilities of 0.95 or above. Despite the fact that many taxa with an incongruent position were removed, both datasets remained incongruent: a Bayes Factor of 25.6 indicating that a model in which each partition is allowed to have its own topology is significantly better than a model in which both partitions are evolving along the same topology. It is worth noting however that this value is 14 times smaller than the Bayes Factor obtained when no taxa were pruned. Both the ITS and cpDNA tree resolved Euryops as a wellsupported (posterior probability of 1) monophyletic group with the only annual species (E. annuus) sister to a well supported (posterior probability of 1) clade representing the perennial species of the genus (Figs. 2, 3). This clade was divided into a series of well-supported lineages indicated in Figs. 2 and 3 by letters A–E. Clades are named identically between trees in Figs. 2 and 3, as each one of those clades has some elements of commonality in terms of species membership. The Bayesian inference invariably failed to recover the relationships between these well-supported lineages. Species relationships within these lineages are mainly unresolved and strongly conflicting between the ITS and cpDNA tree topologies. Clade D includes three species in the cpDNA tree (Fig. 2) while the same lineage D includes five species in the ITS tree (Fig. 3). Both these lineages are supported by 0.97 and 0.99 posterior probabilities in the cpDNA and ITS trees, respectively. Clade C includes ten species in the cpDNA tree (Fig. 2) and is supported by a posterior probability of 1. Species belonging to this lineage in the cpDNA tree are distributed in three different lineages in the ITS tree (clades A, C and E; Fig. 3), all of them supported by high posterior probabilities values. Two species (E. spathaceus DC. and E. hebecarpus (DC.) B. Nord.) are found in clade A, three species (E. munitus (L. f.) B. Nord., E. euryopoides (DC.) B. Nord., E. ursinoides B. Nord.) are found in clade C, while the remaining five species form clade E in the ITS tree (Fig. 3). Another major conflict concerns five species (E. trilobus Harv., E. lateriflorus (L. f.) DC., E. empetrifolius DC., E. evansii Schltr., E. acraeus M.D. Hend.) that the Fig. 3. Phylogenetic relationships of Euryops inferred from sequences of the internal transcribed spacers of the nuclear ribosomal DNA. Fifty percent majority rule consensus tree of 840 trees sampled by a MCMC chain using the program MrBayes. For further details, see Fig. 2. 6.9 Ma (3.7–11.8) 9.7 Ma (5.69–16.00) 1 1 0.61 0.99 1 1 Gymnodiscus capillaris #1 G. capillaris #2 62 Clade A E. montanus E. decumbens E. multifidus E. petraeus E. annae 1 0.99 E. lateriflorus #1 1 E. lateriflorus #2 0.97 E. lateriflorus #4 E. empetrifolius 0.95 E. lateriflorus #3 0.54 0.90 1 E. spathaceus #1 E. spathaceus #2 1 E. trilobus 3.8 Ma E. evansii #1 E. tenuissimus #1 (2.16–6.41) 1 E. tenuissimus #1 0.79 E. linifolius E. hebecarpus 1 E. evansii #2 E. evansii #3 E. othonnoides #1 0.91 E. brevilobus E. othonnoides #2 0.95 E. pectinatus E. speciosissimus #1 E. speciosissimus #2 1 E. abrotanifolius #1 E. abrotanifolius #2 1 E. subcarnosus 1 E. dregeanus #1 0.99 E. dregeanus #2 E. abrotanifolius #5 1 E. abrotanifolius #6 4.3 Ma 1 E. abrotanifolius #3 0.97 E. abrotanifolius #4 (2.46–7.36) E. brownei Clade B E. arabicus #1 0.98 E. arabicus #2 East E. arabicus #3 0.98 1 E. dacrydioides Africa E. dacrydioides E. pinifolius 0.74 E. prostratus E. munitus Clade C 0.97 E. euryopoides #1 E. euryopoides #2 E. ursinoides 0.96 Clade D E. candollei 0.80 E. tysonii 1.9 Ma E. algoensis 0.99 (0.8–3.7) E. virgineus #1 E. virgineus #2 E. acraeus 1 E. brachypodus #1 Clade E E. brachypodus #2 1 E. anthemoides #1 1 E. anthemoides #2 E. ericoides E. ericifolius E. hypnoides E. galpinii E. annuus Othonna eriocarpa O. euphorbioides Hertia pallens O. sedifolia #1 O. sedifolia #2 O. carnosa Othonna sp. nov. 1 1 0.97 TAXON 59 (1) • February 2010: 57–67 cpDNA tree places, with a posterior probability of 1, within clade B together with a monophyletic group of East African high-altitude species (Fig. 2). The ITS tree suggests instead, that those species are nested within lineages A and D (with a posterior probability of 0.99; Fig. 2). It is worth noting that the taxa mentioned above do not represent an exhaustive list of the species with an apparent incongruent position. They are just those that display an incongruent position between our labeled clades. Incongruence is in fact also observed within those labeled clades. All the species endemic to the mountains of east and northeast Africa were resolved as a well-supported monophyletic group, irrespective of dataset (clade B in Figs. 2 and 3). While those species do not seem to show any close relationship with the endemic high-altitude species of southern Africa in the ITS tree, they do show a close relationship with E. evansii and E. acraeus in the cpDNA tree. These latter two species are endemic to the highest peaks of the Drakensberg (South Africa/Lesotho). The remaining South African high-altitude species earlier suggested as close relatives to the Afromontane species of East Africa (Nordenstam, 1968a) are spread in multiple lineages, and consequently do not seem to have a single origin (species in bold, Figs. 2, 3). The BEAST analyses recovered a posterior distribution of trees similar to the posterior distribution of trees sampled by the MrBayes analysis of both the ITS and cpDNA datasets, respectively. The 50% majority rule consensus trees and the posterior probability values were congruent between both analyses (data not shown). The divergence time for the major nodes are given in Fig. 3 for the ITS and in Fig. 2 for the cpDNA dataset. Based on the ITS dataset, the age of the node corresponding to the putative basal radiation (crown group, excluding E. annuus) was estimated at 4.33 Ma (95% highest posterior density: 2.46–7.36; hereafter 95% HPD), while the age of the most recent common ancestor of the Afromontane species of East Africa was estimated at 1.9 Ma (95% HPD: 0.81–3.7; Fig. 3). Age estimates based on the cpDNA dataset, although slightly older than the age estimates obtained from the analysis of the ITS dataset, had a 95% HPD overlapping the 95% HPD obtained with the ITS dataset. The age of the node corresponding to the putative basal radiation was estimated at 9.9 Ma (95% HPD: 4.7–17.7), while the age of the most recent common ancestor of the Afromontane species of East Africa was estimated at 2.2 Ma (95% HPD: 0.81–4.7; Fig. 2). DISCUSSION Phylogenetic relationships and incongruence. — Despite the use of four loci (2611 bp and 325 parsimony informative sites), the phylogenies obtained for the individual (ITS and cpDNA) datasets were largely unresolved. Rapid diversifications are typically characterised in a phylogeny by very short or nonexistent branches that are extremely difficult to resolve due to the paucity of shared derived characters (Richardson & al., 2001a; McCracken & Sorenson, 2005). Unresolved and poorly supported short branches have also been attributed to Devos & al. • Phylogeny and biogeography of Euryops the use of insufficient or inappropriate data for the level of divergence that is being analysed, and often can be resolved by including more characters (Maddison, 1989; DeSalle & al., 1994). However, the loci sequenced here have been used before in many studies dealing with species-level phylogenies. Therefore, we do not believe that an inappropriate choice of, or insufficient, markers would have caused this lack of resolution. Rather, we suggest that the lack of support and resolution in the internal structure of the trees could reflect the signature of the isolation and divergence of many ancestral populations in a short time frame. Phylogenetic reconstruction of rapidly radiating species or lineages is expected to produce strong incongruence among different gene trees (e.g., Shaw, 2002; Hughes & Vogler, 2004; McCracken & Sorensen, 2005). Much of this incongruence is the result of lineage sorting, a random process of retention and extinction of alleles in a lineage over time (e.g., Pamilo & Nei, 1988; Maddison, 1997; Nichols, 2001). The ILD test found a significant incongruence (P = 0.01) between the nrDNA and cpDNA gene partitions in Euryops suggesting the presence of strongly supported character conflict between the two partitions. However, presence of significant heterogeneity in the data does not necessarily imply that the partitions under investigation have different underlying topologies (Dolphin & al., 2000). When rate differences between partitions are too different, the ILD test could indeed suggest significant heterogeneity despite the partitions having similar underlying evolutionary histories (Dolphin & al., 2000). Nevertheless, in a Bayesian framework, the use of Bayes Factors to test for incongruence between gene partitions also suggested strong incongruence between the evolutionary histories of the nuclear and chloroplast DNA. Allowing the gene partitions to have separate topology clearly had a better fit to the data than a model in which the data partitions are forced to evolve on the same topology. The strong incongruence between the ITS and cpDNA datasets argues in favour of a rapid diversification of Euryops. However, other biological processes such as hybridisation, introgression or paralogy in the ITS dataset are known to produce incongruence between the phylogenetic trees estimated from different loci (Maddison, 1997; Nichols, 2001; Seehausen, 2004). The occurrence of such processes in Euryops cannot be completely ruled out. Distinguishing between hypotheses of reticulation (hybridisation, introgression and allopolyploidisation) and that of lineage sorting is extremely difficult because both processes are known to produce similar phylogenetic patterns (Holder & al., 2001). Many taxa had to be pruned in order to reduce the incongruence observed between the ITS and cpDNA datasets. It is unlikely that every one of those taxa has originated through hybridisation, allopolyploidisation or introgression. Spontaneous hybridisation and introgression have been rarely observed in Euryops and polyploidy seems to have played only a minor role in the diversification of the genus (Nordenstam, 1968a,b). Only four polyploid taxa have been identified based on chromosomes counts of 37 taxa (Nordenstam, 1968b). In two species, viz., E. oligoglossus DC. and E. lateriflorus, tetraploids were found as well as diploids (2n = 20, 40). One of these species 63 Devos & al. • Phylogeny and biogeography of Euryops (E. lateriflorus) was included in our dataset. Removing it from the analysis did not reduce the degree of incongruence between our two partitions. Incomplete lineage sorting due to putatively recent and fast diversification in the genus is more likely to be the underlying cause of the strong incongruence observed between the nuclear and cpDNA datasets. The putative radiation in Euryops is estimated to have occurred sometime between four millon years ago (based on estimates from the ITS dataset) and nine millon years ago (based on estimates from the cpDNA dataset). Whether that time is sufficient for lineage sorting to be completed is difficult to say because it is highly dependent on the effective population size at the time of diversification. However, even if four to nine million years is sufficient for lineage sorting to become completed, incongruence between gene trees is still likely to be observed. Ancestral polymorphism will eventually disappear as each descendant lineage becomes fixed for a different set of alleles or haplotypes, but the incongruences between gene trees and species trees that ancestral polymorphisms have generated would persist (McCracken & Sorensen, 2005). Our age estimates (i.e 4.33 [2.46–7.36] million years for the ITS and 9.9 [4.7–17.7] million years for the cpDNA) place the diversification of Euryops at the end of the Miocene, early Pliocene. This timing coincides with the strong upwelling of the cold Benguela Current waters along the west coast of South Africa and as a consequence the onset of aridity in the region (Marlow & al., 2000; Linder, 2003; de Menocal, 2004). “Sudden” aridification and climatic change during the late Miocene and early Pliocene is assumed to have initiated such explosive speciation on a range of continents (e.g., Richardson & al., 2001a; Hughes & Eastwood, 2006; Moore & Jansen, 2006, Linder, 2008) and especially in the Cape Floristic Region (e.g., Linder & al., 1992; Richardson & al., 2001b; Linder, 2003; Verboom & al., 2003; Klak & al., 2004). The lack of resolution and the incongruence between gene trees challenge the interpretation of the evolutionary history of Euryops. Nevertheless, it is worth noting that both datasets resolve the only annual species of Euryops as sister to the remainder of the genus (i.e., a lineage of ca. 90 perennial species). The long branch leading to this annual species could be explained by high rate of nucleotide substitution. It has been shown in numerous comparative studies that annual plants display a higher rate of molecular evolution compared with perennials (Andreasen & Baldwin, 2001). This pattern has been partialy explained by differences in generation times, efficiency of DNA repair or replication, and difference in population sizes between annual and perennial plants. The lack of diversification in this annual lineage is however puzzling especially if annuals are indeed harboring faster rate of molecular evolution. High exctinction rate on the other hand could explain the long branch length and would help explain the apparent absence of radiation in this lineage. Biogeographic patterns. — Although the inferred relationships are incongruent across loci and not well resolved, some interesting biogeographic patterns become evident. The most clear of these is the monophyletic origin of all the Euryops species endemic to the high mountains of East Africa. 64 TAXON 59 (1) • February 2010: 57–67 Many taxa that are most diverse in southern Africa and more specifically in the CFR have, like Euryops, outliers that form part of the Afromontane flora in northeastern tropical Africa (Weimarck, 1933, 1936). The interpretation of molecular phylogenies produced for the lineages displaying disjunct distribution between southern Africa and the uplands of tropical East Africa supports the idea that the Afromontane taxa are derived from the Cape clades from one or several migrations northwards, from the CFR (Galley & Linder, 2006; Galley & al., 2007). The cpDNA dataset indicates that in Euryops, the Afromontane species form a very well supported monophyletic group, nested within an otherwise southern clade. In fact, the outgroups used represent genera that are only found in Southern Africa and have been identified by Pelser & al. (2007) as the closest relatives of Euryops. The phylogenetic tree produced by Pelser & al. (2007) includes most of the genera of the tribe Senecioneae and leaves no doubt that the whole Othoninnae subtribe is a group that originated and diversified in southern Africa. The position of the Afromontane species is thus consistent with a scenario of a radiation from a single immigration into the East African mountains and corroborates Galley & al.’s (2007) interpretation of the same disjunction observed in other taxa such as Restionaceae and Pentaschistis (Nees) Spach (Poaceae). Moreover, the crown group age of this monophyletic East African lineage (1.9 million years based on the ITS dataset and 2.2 million years based on the cpDNA dataset) is consistent with a scenario of a recent migration event as opposed to a scenario of vicariance of a once more widespread flora in Africa that receded due to climatic changes, combined with the uplift of the East African mountains, which occurred much earlier (middle-late Miocene) during the development of the East African Rift System (Chorowicz, 2005). The estimated age of 1.9–2.2 million years for the East African clade is generally somewhat younger than the ages of East African clades and dispersal or migration events estimated by Galley & al. (2007: table 1). It would be tempting to suggest that these tropical Afromontane species originated from an ancestral species closely related to one of the extant high-altitude species from South Africa. Nordenstam (1968a) noted a close morphological affinity between E. prostratus B. Nord. from Ethiopia and the two alpine endemics (E. decumbens B. Nord., E. montanus Schltr.) from the Drakensberg in South Africa. In both the ITS and cpDNA gene trees, E. decumbens and E. montanus are closely related to each other with strong support. However, they do not show any close relationship with E. prostratus nor with any of the other East African Afromontane species. Nordenstam (1968a) also noted the close affinity between E. brownei and E. acraeus, endemics of East Africa and Drakensberg, respectively. Interestingly, the cpDNA gene tree resolved E. acraeus and E. evansii, another endemic from the Drakensberg, within the Afromontane East African clade (clade B, Fig. 2) that includes E. brownei. The Afromontane species of tropical Africa could thus have originated from a northward migration of an ancient high-altitude ancestral species related to the extant alpine endemics E. acraeus and E. evansii. TAXON 59 (1) • February 2010: 57–67 The Great Escarpment of South Africa (including both the Drakensberg and Sneeuberg) is an important centre of diversity for Euryops and other high-altitude taxa. Nordenstam (1969) noted that there were 16 endemic species in this region, and a further two new species have been recently discovered from the Sneeuberg (Nordenstam & al., 2009). However, unlike the species of East Africa, the species found in the Drakensberg and Sneeuberg mountains do not seem to be closely related. This suggests that local diversification (i.e., diversification from a single ancestor) has not occurred in these regions, but that the diversity is a consequence of multiple migrations or a consequence of these mountains acting as refugial habitats. The absence of local diversification in the Drakensberg has also been found in both the Restionaceae and Pentaschistis (Poaceae; Galley & al., 2007). In contrast, the diversity of the genus Disa Bergius (Orchidaceae) and the tribe Irideae (Iridaceae) in the Drakensberg is a consequence of in situ diversification following multiple migration events (Galley & al., 2007). These authors implicate the role of pollinator specificity as a factor driving in situ speciation: the windpollinated grass genus Pentaschistis and Restionaceae have not radiated in situ, whereas the biotic (and probably specialist) pollination syndromes of Disa and Irideae may have driven speciation following successful colonisation. Euryops, with its presumably typical asteraceous generalist pollination syndrome would, like the wind-pollinated examples cited above, probably not speciate as readily following migration into a new region. A more-intensively sampled and completely resolved species phylogeny will however be needed in order to reconstruct ancestral distribution ranges from which directionality of migration events can be rigorously addressed. CONCLUSIONS The strong topological incongruence between our gene trees, combined with the lack of support and resolution in the internal structure of both the ITS and cpDNA trees led us to suggest that the genus Euryops has radiated in southern Africa within a short time period. Molecular dating of both ITS and cpDNA trees suggest that the radiation has taken place during late Miocene, early Pliocene. This period coincides with drastic climatic changes in South Africa and has been shown to have initiated explosive species radiation in various plant groups such as Ruschioideae (Klak & al., 2004), Gazania (Howis & al., 2009), Phylica (Richardson & al., 2001b) and Zygophyllum (Bellstedt & al., 2008). Despite the lack of resolution and the incongruence between our gene trees, it was possible to resolve some aspects of the phylogenetic relationships within Euryops and interpret those within a plausible biogeographic framework. In particular, the Afromontane species have radiated following a single immigration event into the East African mountains, which corroborates Galley & al.’s (2007) interpretation of the same disjunction observed in other taxa such as Restionaceae and Pentaschistis (Poaceae) and augments support for the “Cape to Cairo” hypothesis of floral migrations in Africa. Devos & al. • Phylogeny and biogeography of Euryops ACKNOWLEDGMENTS The authors gratefully acknowledge the National Research Foundation of South Africa for financial support (through grant 2069059 to N.P.B., grant 2069036 to L.M. and a Postdoctoral Fellowship to N. Devos). Seranne Howis provided technical assistance, Robert McKenzie and Peter Teske helped during the field work. We thank the South African National Biodiversity Institute for providing access to distribution data of Euryops from the PRECIS system. Mats Thulin of Uppsala University Herbarium kindly supplied leaf material of Euryops arabicus and E. dacrydioides, and Siro Masinde of the National Herbarium in Nairobi, Kenya, provided material of some East African taxa of the genus. LITERATURE CITED Adamson, R.S. 1958. The Cape as an ancient African flora. Advancem. Sci. 58: 1–10. Akaike, H. 1974. A new look at the statistical model identification. IEEE Trans. Autom. Control 19: 716–723. Alvarez, I. & Wendel, J.F. 2003. Ribosomal ITS sequences and plant phylogenetic inference. Molec. Phylog. Evol. 29: 417–434. Andreasen, K. & Baldwin, B.G. 2001. Unequal evolutionary rates between annual and perennial lineages of checker mallows (Sidalcea, Malvaceae): Evidence from 18S-26S rDNA internal and external transcribed spacers. Molec. Biol. Evol. 18: 936–944. Axelrod, D.I. & Raven, P.H. 1978. Late Cretaceous and Tertiary vegetation history of Africa. Pp. 79–130 in: Werger, M.J.A. (ed.), Biogeography and ecology of southern Africa. The Hague: Dr. W. Junk. Baldwin, B.G. & Sanderson, M.J. 1998. Age and rate of diversification of the Hawaiian silversword alliance (Compositae). Proc. Natl. Acad. Sci. U.S.A. 95: 9402–9406. Barker, F.K. & Lutzoni, F. 2002. The utility of the incongruence length difference test. Syst. Biol. 51: 625–637. Bellstedt, D.U., van Zyl, L., Marais, E.M., Bytebier, B., de Villiers, C.A., Makwarela, A.M. & Dreyer, L.L. 2008. Phylogenetic relationships, character evolution and biogeography of southern African members of Zygophyllum (Zygophyllaceae) based on three plastid regions. Molec. Phylog. Evol. 47: 932–949. Carbutt, C. & Edwards, T.J. 2004. Cape elements on high-altitude corridors and edaphic islands: Historical aspects and preliminary phytogeography. Syst. Geogr. Pl. 71: 1033–1061. Chase, M.W. & Hills, H.H. 1991. Silica gel: An ideal material for field preservation of leaf samples for DNA studies. Taxon 40: 215–220. Chorowicz, J. 2005. The East African rift system. J. Afr. Earth Sci. 43: 379–410. Clark, V.R., Barker, N.P. & Mucina, L. 2009. The Sneeuberg: A new centre of floristic endemism on the Great Escarpment, South Africa. S. African J. Bot. 75: 196–238. Cunningham, C.W. 1997. Is congruence between data partitions a reliable predictor of phylogenetic accuracy? Empirically testing an iterative procedure for choosing among phylogenetic methods. Syst. Biol. 46: 464–478. De Menocal, P.B. 2004. African climate change and faunal evolution during the Pliocene–Pleistocene. Earth Planet. Sci. Lett. 220: 3–24. DeSalle, R., Absher, R. & Amato, G. 1994. Speciation and phylogenetic resolution. Trends Ecol. Evol. 9: 297–298. Dolphin, K., Belshaw, R., Orme, C.D.L. & Quicke, D.L.J. 2000. Noise and incongruence: Interpreting results of the incongruence length difference test. Molec. Phylog. Evol. 17: 401–406. Doyle, J.J. & Doyle, J.L. 1987. Preservation of plant samples for DNA restriction endonuclease analysis. Taxon 36: 715–722. Drummond, A.J., Ho, S.Y.W., Phillips, M.J. & Rambaut, A. 2006. 65 Devos & al. • Phylogeny and biogeography of Euryops Relaxed phylogenetics and dating with confidence. PLoS Biol 4: e88. doi:10.1371/journal.pbio.0040088. Drummond, A.J., Nicholls, G.K., Rodrigo, A.G. & Solomon, W. 2002. Estimating mutation parameters, population history and genealogy simultaneously from temporally spaced sequence data. Genetics 161: 1307–1320. Drummond, A.J. & Rambaut, A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol. Biol. 7: 214. doi:10.1186/14712148-7-214. Farris, J.S., Källersjö, M., Kluge, A.G. & Bult, C. 1994. Testing significance of incongruence. Cladistics 10: 315–319. Galley, C., Bytebier, B., Bellstedt, D.U. & Linder, H.P. 2007. The Cape element in the Afrotemperate flora: From Cape to Cairo? Proc. Roy. Soc. London, Ser. B, Biol. Sci. 274: 535–543. Galley, C. & Linder, H.P. 2006. Geographical affinities of the Cape flora, South Africa. J. Biogeogr. 33: 236–250. Gene Codes Corporation. 1998. Sequencher 4.01 reference: Advanced, user friendly software tools for DNA sequencing. Madison, Wisconsin: Gene Codes Corporation. Holder, M.T., Anderson, J.A. & Holloway, A.K. 2001. Difficulties in detecting hybridization. Syst. Biol. 50: 978–982. Holland, P.G. 1978. An evolutionary biogeography of the genus Aloe. J. Biogeogr. 5: 213–226. Howis, S., Barker, N.P. & Mucina, L. 2009. Globally grown, but poorly known: Species limits and biogeography of Gazania Gaert. (Asteraceae) inferred from chloroplast and nuclear DNA sequence data. Taxon 58: 871–882. Hughes, C. & Eastwood, R. 2006. Island radiation on a continental scale: Exceptional rates of plant diversification after uplift of the Andes. Proc. Natl. Acad. Sci. U.S.A. 103: 10334–10339. Hughes, J. & Vogler, A.P. 2004. The phylogeny of acorn weevils (genus Curculio) from mitochondrial and nuclear DNA sequences: The problem of incomplete data. Molec. Phylog. Evol. 32: 601–615. Irestedt, M., Fjeldså, J., Nylander, J.A.A. & Ericson, P.G.P. 2004. Phylogenetic relationships of typical antbirds (Thamnophilidae) and test of incongruence based on Bayes factors. BMC Evol. Biol. 4: 23. doi:10.1186/1471-2148-4-23. Kass, R.E. & Raftery, A.E. 1995. Bayes factors. J. Amer. Statist. Assoc. 90: 773–795. Killick, D.J.B. 1963. An account of the plant ecology of the Cathedral peak area on the Natal Drakensberg. Mem. Bot. Surv. South Africa 31: 1–178. Killick, D.J.B. 1978. The Afro-alpine region. Pp. 515–560 in: Werger, M.J.A. (ed.), Biogeography and ecology of southern Africa. The Hague: Dr. W. Junk. Klak, C., Reeves, G. & Hedderson, T. 2004. Unmatched tempo of evolution in Southern African semi-desert ice plants. Nature 427: 63–65. Koekemoer, M. 1996. An overview of the Asteraceae of southern Africa. Pp. 95–110 in: Hind, D.J.N. & Beentje, H.J. (eds.), Proceedings of the International Compositae Conference, Kew, 1994, vol. 1, Compositae: Systematics. Kew: Royal Botanic Gardens. Levyns, M.R. 1938. Some evidence bearing on the past history of the Cape flora. Trans. Roy. Soc. South Africa 26: 404–424. Levyns, M.R. 1952. Clues to the past in the Cape flora of today. S. African J. Sci. 49: 155–164. Levyns, M.R. 1964. Presidential address, migrations and origin of the Cape flora. Trans. Roy. Soc. South Africa 37: 85–107. 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. 1994. Afrotemperate phytogeography: Implications of cladistic biogeographical analysis. Pp. 913–930 in: Seyani, J.H. & Chikuni, A.C. (eds.), Proceedings of the 13th Plenary Meeting of AETFAT, Zomba, Malawi, 2–11 April 1991. Zomba: National Herbarium and Botanic Gardens of Malawi. Linder, H.P. 2003. The radiation of the Cape flora, southern Africa. Biol. Rev. 78: 597–638. 66 TAXON 59 (1) • February 2010: 57–67 Linder, H.P. 2008. Plant species radiations: Where, when, why? Philos. Trans., Ser. B. 363: 3097–3105. Linder, H.P., Meadows, M.E. & Cowling, R.M. 1992. History of the Cape Flora. Pp. 113–134 in: Cowling, R.M. (ed.), Fynbos: Nutrients, fire and diversity. Cape Town: Oxford Univ. Press. Maddison, W. 1989. Reconstructing character evolution on polytomous cladograms. Cladistics 5: 365–377. Maddison, W.P. 1997. Gene trees in species trees. Syst. Biol. 46: 523–536. Maddison, W.P. & Maddison, D.R. 1992. MacClade: Analysis of phylogeny and character evolution, version 4.0. Sunderland, Massachusetts: Sinauer. Marlow, J.R., Lange, C.B., Walter, G. & Rosell-Melé, A. 2000. Upwelling intensification as part of the Pliocene–Pleistocene climate transition. Science 290: 2288–2291. McCracken, K.G. & Sorensen, M.D. 2005. Is homoplasy or lineage sorting the source of incongruent mtDNA and nuclear gene trees in the stiff-tailed ducks (Nomonyx oxyura)? Syst. Biol. 54: 35–55. Moore, M.J. & Jansen, R.K. 2006. Molecular evidence for the age, origin, and evolutionary history of the American desert plant genus Tiquilia (Boraginaceae). Molec. Phylog. Evol. 39: 668–687. Newton, M.A. & Raftery, A.E. 1994. Approximate Bayesian inference by the weighted likelihood bootstrap (with discussion). J. Roy. Statist. Soc., Ser. B 56: 3–48. Nichols, R. 2001. Gene trees and species trees are not the same. Trends Ecol. Evol. 16: 358–364. Nordenstam, B. 1968a. The genus Euryops. Part I. Taxonomy. Opera Bot. 20: 1–409. Nordenstam, B. 1968b. The genus Euryops. Part II. Aspects of morphology and cytology. Bot. Not. 121: 209–232. Nordenstam, B. 1969. Phytogeography of the genus Euryops (Compositae). A contribution to the phytogeography of southern Africa. Opera Bot. 23: 1–77. Nordenstam, B., Clark, V.R., Devos, N. & Barker, N.P. 2009. Two new species of Euryops (Asteraceae: Senecioneae) from the Sneeuberg, Eastern Cape Province, South Africa. S. African J. Bot. 75: 144–152. Nylander, J.A.A. 2004. MrModeltest, version 2. Evolutionary Biology Centre, Uppsala University: Program distributed by the author. Nylander, J.A.A., Ronquist, F., Huelsenbeck, J.P. & Nieves-Aldrey, J.L. 2004. Bayesian phylogenetic analysis of combined data. Syst. Biol. 53: 47–67. Oxelman, B., Lidén, M. & Berglund, D. 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Pl. Syst. Evol. 206: 393–410. Palmer, J.D. 1991. Plastid chromosome: Structure and evolution. Pp. 5–53 in: Bogorad, L. & Vasil, I.K. (eds.), The molecular biology of plastids. San Diego: Academic Press. Pamilo, P. & Nei, M. 1988. Relationships between gene trees and species trees. Molec. Biol. Evol. 5: 568–583. Pelser, P.B., Nordenstam, B., Kadereit, J.W. & Watson, L.E. 2007. An ITS phylogeny of tribe Senecioneae (Asteraceae) and a new delimitation of Senecio L. Taxon 56: 1077–1104. Rambaut, A. & Drummond, A.J. 2007. Tracer, version 1.4. http:// tree.bio.ed.ac.uk/software/tracer/. Richardson, J.E., Pennington, R.T., Pennington, T.D. & Hollingsworth, P.M. 2001a. Rapid diversification of a species-rich genus of neotropical rain forest trees. Science 293: 2242–2245. Richardson, J.E., Weitz, F.M., Fay, M.F., Cronk, Q.C.B., Linder, H.P., Reeves, G. & Chase, M.W. 2001b. Rapid and recent origin of species richness in the Cape flora of South Africa. Nature 412: 181–183. Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. Sang, T., Crawford, D.J., Stuessy, T.F. & Silva-O., M. 1995. ITS sequences and the phylogeny of the genus Robinsonia (Asteraceae). Syst. Bot. 20: 55–64. Seehausen, O. 2004. Hybridization and adaptive radiation. Trends Ecol. Evol. 19: 198–207. TAXON 59 (1) • February 2010: 57–67 Shaw, K.L. 2002. Conflict between nuclear and mitochondrial DNA phylogenies of a recent species radiation: What mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. Proc. Natl. Acad. Sci. U.S.A. 99: 16122–16127. Stanford, A.M., Harden, R. & Parks, C.R. 2000. Phylogeny and biogeography of Juglans (Juglandaceae) based on matK and ITS sequence data. Amer. J. Bot. 87: 872–882. Swofford, D.L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4. Sunderland, Massachusetts: Sinauer. Taberlet, P., Gielly, L., Pautou, G. & Bouvet, J. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Pl. Molec. Biol. 17: 1105–1109. Vargas, P., Baldwin, B.G. & Constance, L. 1998. Nuclear ribosomal DNA evidence for a western North American origin of Hawaiian and South American species of Sanicula (Apiaceae). Proc. Natl. Acad. Sci. U.S.A. 95: 235–240. Verboom, G.A., Linder, H.P. & Stock, W.D. 2003. Phylogenetics of the grass genus Ehrharta Thunb.: Evidence for radiation in the summerarid zone of the South African Cape. Evolution 57: 1008–1021. Devos & al. • Phylogeny and biogeography of Euryops Weimarck, H. 1933. Die Verbreitung einiger afrikanisch-montanen Pflanzengruppen. I–II. Svensk Bot. Tidskr. 27: 400–419. Weimarck, H. 1936. Die Verbreitung einiger afrikanisch-montanen Pflanzengruppen. III–IV. Svensk Bot. Tidskr. 30: 36–56. Weimarck, H. 1941. Phytogeographical groups, centres and intervals within the Cape flora. Lunds Univ. Arsskrift 37: 1–143. White, F. 1978. The Afromontane region. Pp. 463–513 in: Werger, M.J.A. (ed.), Biogeography and ecology of southern Africa. The Hague: Dr. W. Junk. White, T.J., Bruns, T., Lee, S. & Taylor, J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pp. 315–324 in: Innis, M.A., Gelfand, D.H., Sninsky, J.J. & White, T.J. (eds.), PCR protocols: A guide to methods and applications. San Diego: Academic Press. Wild, H. 1968. Phytogeography in South Central Africa. Kirkia 6: 197–222. 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. Appendix. Voucher information and GenBank accession numbers for specimens used in this study. Voucher specimens are deposited in the Selmar Schonland Herbarium, Grahamstown, South Africa. Species are grouped according to their affiliation to the Euryops sections described by Nordenstam (1968a). Section: species: voucher (collector’s name, collector’s number and herbarium), GenBank accessions no. ITS, rps16, trnL-trnF, trnT-trnL E. sect. Angustifoliae: Euryops acraeus M.D. Hend.: N. Devos CH05 (GRA), EU667472, EU667538, EU670096, EU670158; Euryops algoensis DC.: N. Barker NB1908 (GRA), EU667483, EU667547, EU670105, EU670167; Euryops annae E. Phillips: R. Clark CDM4 (GRA), EU667470, EU667537, EU670094, EU670156; Euryops arabicus Steud.: M. Thulin & A.N. Gifri 8850 (UPS), EU667464, n/a, EU670088, n/a; Euryops arabicus Steud.: Mats Thulin, Abdi Dahir & Ahmed Osman 9438 (UPS), EU667465, n/a, EU670089, EU670152; Euryops arabicus Steud.: Thulin, Eriksson, Gifri & Langstrom 8134 (UPS), EU667463, n/a, EU670087, EU670151; Euryops brownei S. Moore: n/a 248 (EA), EU667477, n/a, n/a, n/a; Euryops candollei Harv.: M. Cunningham CC352 (GRA), EU667467, EU667534, EU670091, EU670153; Euryops dacrydioides Oliv. in Hook: Majda Zumer Z18 (UPS), EU667529, EU667591, EU670147, EU670206; Euryops dacrydioides Oliv. in Hook: Olov Hedberg 4377 (UPS), EU667528, EU667590, EU670146, n/a; Euryops dregeanus Schltr.: N. Devos ND090905_6 (GRA), EU667497, EU667560, EU670118, EU670180; Euryops dregeanus Schltr.: N. Devos ND090905_7 (GRA), EU667498, EU667561, EU670119, EU670181; Euryops empetrifolius DC.: Laidler 529 (PRE), EU667520, EU667582, EU670139, EU670201; Euryops lateriflorus (L. f.) B. Nord.: N. Barker NB1487 (GRA), EU667482, EU667546, EU670104, EU670166; Euryops lateriflorus (L. f.) B. Nord.: N. Devos ND040905_4 (GRA), EU667491, EU667555, n/a, EU670175; Euryops lateriflorus (L. f.) B. Nord.: N. Devos ND060905_11 (GRA), EU667493, EU667557, EU670114, EU670177; Euryops lateriflorus (L. f.) B. Nord.: N. Devos ND130206 (GRA), EU667504, EU667567, EU670124, EU670187; Euryops linifolius (L.) DC.: N. Devos ND140705_2 (GRA), EU667507, EU667570, EU670127, EU670190; Euryops multifidus Schltr.: N. Devos ND040905_3 (GRA), EU667490, EU667554, EU670112, EU670174; Euryops petraeus B. Nord.: V. Clark CDM2 (GRA), EU667468, EU667535, EU670092, EU670154; Euryops pinifolius A. Rich.: M. Thulin & A. Hunde 3930 (UPS), EU667530, EU667592, EU670148, EU670207; Euryops spathaceus DC.: N. Devos ND120605_2 (GRA), EU667500, EU667563, EU670121, EU670183; Euryops spathaceus DC.: N. Devos ND120605_4 (GRA), EU667502, EU667565, n/a, EU670185; Euryops subcarnosus subsp. vulgaris DC.: N. Devos ND080905_1 (GRA), EU667495, EU667558, EU670116, EU670178; Euryops tenuissimus subsp. tenuissimus (L.) DC.: N. Devos ND020905_4 (GRA), EU667489, EU667553, EU670111, EU670173; Euryops tenuissimus subsp. tenuissimus (L.) DC.: N. Devos ND130705_1 (GRA), EU667505, EU667568, EU670125, EU670188; Euryops tysonii Phill.: M. Cunningham CH08 (GRA), EU667473, EU667539, EU670097, EU670159; Euryops virgineus (L. f.) DC.: N. Devos ND260805 (GRA), EU667512, EU667575, EU670132, EU670195; Euryops virgineus (L. f.) DC.: R. McKenzie RM1023 (GRA), EU667521, EU667583, EU670140, EU670202. E. sect. Brachypus: Euryops decumbens B. Nord.: M. Cunningham CH09 (GRA), EU667474, EU667540, EU670098, EU670160; Euryops galpinii Bol.: N. Devos ND081005_4 (GRA), EU667496, EU667559, EU670117, EU670179; Euryops montanus Schltr.: M. Cunningham 1XT5314 (GRA), EU667462, EU667532, n/a, EU670150; Euryops prostratus B. Nord.: Anderberg 1708 (S), EU667478, EU667542, EU670100, EU670162. E. sect. Chrysops: Euryops abrotanifolius (L.) DC.: R. McKenzie RM1056 (GRA), EU667522, EU667584, EU670141, EU670203; Euryops abrotanifolius (L.) DC.: N. Devos ND130705_2C (GRA), EU667506, EU667569, EU670126, EU670189; Euryops abrotanifolius (L.) DC.: N. Devos ND160705_5 (GRA), EU667510, EU667573, EU670130, EU670193; Euryops abrotanifolius (L.) DC.: N. Devos ND280805_1 (GRA), EU667513, EU667576, EU670133, EU670196; Euryops abrotanifolius (L.) DC.: N. Devos ND310805_1 (GRA), EU667516, EU667579, EU670136, EU670198; Euryops abrotanifolius (L.) DC.: N. Devos ND310805_6 (GRA), EU667518, EU667581, EU670138, EU670200; Euryops evansii subsp. evansii Schltr.: M. Cunningham CH04 (GRA), EU667471, n/a, EU670095, EU670157; Euryops evansii subsp. parvus Schltr.: N. Devos ND190106 (GRA), EU667511, EU667574, EU670131, EU670194; Euryops evansii subsp. parvus Schltr.: M. Cunningham CH11 (GRA), EU667475, EU667541, EU670099, EU670161; Euryops trilobus Harv.: V. Clark CDM3 (GRA), EU667469, EU667536, EU670093, EU670155. E. sect. Euryops: Euryops brevilobus Compt.: N. Devos ND010905_6 (GRA), EU667488, EU667552, EU670110, EU670172; Euryops othonnoides (DC.) B. Nord.: N. Devos ND010905_1 (GRA), EU667486, EU667550, EU670108, EU670170; Euryops othonnoides (DC.) B. Nord.: N. Devos ND120905_6 (GRA), EU667503, EU667566, EU670123, EU670186; Euryops pectinatus subsp. pectinatus (L.) Cass.: N. Devos ND290805_2 (GRA), EU667514, EU667577, EU670134, n/a; Euryops speciosissimus DC.: N. Devos ND310805_4 (GRA), EU667517, EU667580, EU670137, EU670199; Euryops speciosissimus DC.: R. McKenzie RM1058 (GRA), EU667523, EU667585, EU670142, EU670204. E. sect. Leptorrhiza: Euryops annuus Compt.: N. Devos ND010905_4 (GRA), EU667487, EU667551, EU670109, EU670171. E. sect. Psilosteum: Euryops anthemoides B. Nord.: N. Devos ND120605_3 (GRA), EU667501, EU667564, EU670122, EU670184; Euryops anthemoides B. Nord.: T. Dold TD452 (GRA), EU667525, EU667587, n/a, n/a; Euryops brachypodus (DC.) B. Nord.: N. Devos ND01 (GRA), EU667485, EU667549, EU670107, EU670169; Euryops brachypodus (DC.) B. Nord.: R. McKenzie RM1108 (GRA), EU667524, EU667586, EU670143, EU670205; Euryops ericifolius (Belang.) B. Nord.: Palmer PALMER3934 (GRA), EU667519, n/a, n/a, n/a; Euryops ericoides (L. f.) B. Nord.: N. Devos ND160705_3 (GRA), EU667509, EU667572, EU670129, EU670192; Euryops euryopoides (DC.) B. Nord.: Doubell 29 (GRA), EU667476, n/a, n/a, n/a; Euryops euryopoides (DC.) B. Nord.: Youthed 751 (GRA), EU667531, EU667593, EU670149, n/a; Euryops hebecarpus (DC.) B. Nord.: N. Devos ND160705_1 (GRA), EU667508, EU667571, EU670128, EU670191; Euryops hypnoides B. Nord.: T. Dold TD4700 (GRA), EU667527, EU667589, EU670145, n/a; Euryops munitus (L. f.) B. Nord.: Brown 130898 (GRA), EU667466, EU667533, EU670090, n/a; Euryops ursinoides B. Nord.: T. Dold TD4649 (GRA), EU667526, EU667588, EU670144, n/a. Outgroups: Gymnodiscus capillaris (L. f.) Less.: N. Devos ND300805_1 (GRA), EU667515, EU667578, EU670135, EU670197; Gymnodiscus capillaris (L. f.) Less.: L. Mucina LM110805_3 (GRA), EU667480, EU667544, EU670102, EU670164; Othonna carnosa Less.: N. Devos ND120605_1 (GRA), EU667499, EU667562, EU670120, EU670182; Othonna eriocarpa (DC.) Sch. Bip.: N. Barker NB1938 (GRA), EU667484, EU667548, EU670106, EU670168; Othonna euphorbioides Hutch.: N. Devos ND060905_3 (GRA), EU667494, n/a, EU670115, n/a; Othonna sedifolia DC.: N. Devos ND050905_1 (GRA), EU667492, EU667556, EU670113, EU670176; Othonna sedifolia DC.: L. Mucina LM060605_13 (GRA), EU667479, EU667543, EU670101, EU670163; Othonna sp. nov.: L. Mucina LM210903_6 (GRA), EU667481, EU667545, EU670103, EU670165. 67 View publication stats