TAXON 60 (6) • December 2011: 1667–1677 Oberlander & al. • Phylogeny of southern African Oxalis Molecular phylogenetics and origins of southern African Oxalis Kenneth C. Oberlander,1 Léanne L. Dreyer1 & Dirk U. Bellstedt2 1 Department of Botany and Zoology, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa 2 Department of Biochemistry, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa Author for correspondence: Kenneth C. Oberlander, kco@sun.ac.za Abstract Despite being a speciose component of the Cape Floristic Region at the southern tip of Africa, the southern African Oxalis lineage is systematically poorly understood. Palynological and preliminary phylogenetic studies of the group contrast with the current taxonomy, and indicate the need for further research. Here we present a large-scale phylogenetic analysis comprising 150 of ± 210 southern African species, sampled for plastid trnL-F and trnS-G and nuclear ITS markers, using three different inference methods. Secondly, we explore the origins of southern African Oxalis as a potential Cape clade. Despite substantial localised incongruence between plastid and nuclear datasets, analyses agreed on a monophyletic southern African clade, which consists of two species-poor and one extremely species-rich lineages. The current taxonomy is shown to be artificial, with not one section retrieved as monophyletic. Our topologies are consistent with previous palynological and phylogenetic studies, and provide a backbone for future systematic work. Reconstructions of ancestral area for the southern African lineage support an origin within or close to the Cape Floristic Region. Keywords BayesTraits; Cape clade; Cape Floristic Region; Greater Cape Floristic Region; incongruence; Cernuae subsect. Lividae; Oxalis pes-caprae; southern African Oxalis Supplementary Material Figures S1 and S2 (both in the Electronic Supplement) and the alignment are available in the Supplementary Data section of the online version of this article (http://ingentaconnect.com/content/iapt/tax). INTRODUCTION The Cape Floristic Region (CFR) at the southern tip of Africa is one of 25 global biodiversity hotspots (Myers & al., 2000) and thus ideally suited to test theories of plant evolution and diversification. Research on the CFR has progressed beyond the reconstruction of phylogenies for taxonomic purposes, towards the utilization of phylogenies to infer speciation and extinction processes (Van der Niet & al., 2006), key innovations (Klak & al., 2004), pollinator shifts (Bakker & al., 2005), substrate and habitat preferences (Verboom & al., 2004, 2009; Linder & Hardy, 2005), migration patterns (Galley & Linder, 2006; Galley & al., 2007), historical effects of fire (Bytebier & al., 2011), and ages of Cape clades (Richardson & al., 2001; Edwards & Hawkins, 2007; Forest & al., 2007; Verboom & al., 2009). Despite this interest, the phylogenetic investigation of Cape lineages has been uneven. For example, charismatic taxa such as Proteaceae (Rourke 1998; Barker & al., 2004, 2007; Barraclough & Reeves 2005; Sauquet & al., 2008) and Restionaceae (Eldenas & Linder, 2000; Linder & al., 2003, 2005; Hardy & al., 2008) have been well-studied, yet other lineages have not received the same phylogenetic attention. This is true for some of the largest genera in the CFR, such as Erica (Ericaceae) and Oxalis (Oxalidaceae). Oxalis L. (ca. 500 spp.) is a major component of the New World and CFR floras (Lourteig, 1994; 2000; Salter, 1944). South America (ca. 250 spp.) is the centre of morphological diversity (semi-succulent to woody shrubs, herbs, vines, geophytes) and putative place of origin of Oxalis. Southern African taxa (hereafter referred to as SA Oxalis) constitute ca. 40% (± 210 spp.) of the genus, and share a bulbous habit with above-ground plant parts borne on seasonal stems. The current taxonomy for SA Oxalis (Salter, 1944) is in need of revision and limited previous work has shown that it is at least partly artificial (Dreyer, 1996; Oberlander & al., 2004). Prior to the molecular age all SA Oxalis were assumed to be closely related (Knuth, 1930), but their relation to the rest of Oxalis was poorly understood. Although phylogenies to date have addressed some of these issues, the genus-level phylogeny, origin, age, and phylogenetic relations of SA Oxalis remain poorly known. Oxalis is both the seventh-largest genus (Goldblatt & Manning, 2000) and the largest geophytic genus in the CFR (Procheş & al., 2006). Despite being such a prominent group in the CFR, it is still not clear whether Oxalis is a Cape clade sensu Linder (Linder, 2003). Apart from Salter (1944), the most-recent major research on SA Oxalis entailed a species-level palynological review by Dreyer (1996). She reported extensive palynological diversity, and identified four major pollen types that are incongruent with the Salter (1944) taxonomy. Oberlander & al. (2004) used plastid DNA data (non-coding plastid trnL intron and trnLtrnF spacer, hereafter referred to as trnL-F) to reconstruct relationships in the large, loosely defined Oxalis sect. Angustatae subsect. Lineares. This study included less than 40% of southern African species and two putative outgroup taxa. Although the resultant trees were poorly resolved, the study clearly showed the artificiality of this subsection and substantial congruence with the Dreyer (1996) palynological classification. A molecular phylogeny used to study bulb evolution in 1667 Oberlander & al. • Phylogeny of southern African Oxalis the genus (Oberlander & al., 2009b) had better outgroup sampling, but only included 30 SA Oxalis species. Results showed a monophyletic SA Oxalis, and were consistent with results of both Dreyer (1996) and Oberlander & al. (2004). Other more recent systematic publications on SA Oxalis addressed minor classification issues (Bayer, 1992; Dreyer & Van Wyk, 1996) or described new species (Ornduff, 1973; Oliver, 1993; Williamson, 1999; Kumwenda & al., 2004; Manning & Goldblatt, 2008; Dreyer & al., 2009; Oberlander & al., 2009a). The need for a comprehensive, large-scale, systematic analysis of SA Oxalis is thus still wanting. The aims of this study were thus twofold: to provide wellsampled phylogenetic reconstructions of SA Oxalis and to use the produced phylogenies to test whether SA Oxalis are a Cape clade or not. To address the first aim, one nuclear (the internal transcribed spacer [ITS] of the nuclear ribosomal DNA repeat, including the ITS1 and ITS2 spacers and the 5.8 ribosomal gene) and two plastid markers (the trnL-F region, including the trnL intron and the trnL-trnF spacer; the trnS-G spacer) were sequenced for 75% of southern African species. These data were subjected to parsimony, likelihood, and Bayesian inference, and the resultant trees were explored in the context of previous work on SA Oxalis. To address our second aim, we used ancestral state reconstruction for potential areas of origin of SA Oxalis and other relevant clades. MATERIALS AND METHODS Species collection and sampling. — Most species were collected in the field during the 2001 to 2007 Oxalis flowering seasons. Sampling was supplemented by herbarium accessions (BOL), and living material from the Stellenbosch University and Karoo National Botanical Gardens. Voucher specimens are housed at STEU. Genus-level outgroup samples were provided by B. Gravendeel. The Salter collections in BOL and NBG herbaria were used for species identification. For outgroup sampling, a clade comprised of Averrhoa L., Biophytum DC., Dapania Korth, and Sarcotheca Blume (all Oxalidaceae) was used as monophyletic sister to Oxalis. This placement is well-supported in higher-level analyses of trnL-F data across the Oxalidales (K. Oberlander, unpub. data). Fourteen Oxalis species from seven New World and European sections were used as within-genus outgroups. DNA extraction and sequencing. — DNA extraction of most silica-dried samples followed a modified 2× CTAB procedure (Doyle & Doyle, 1987), described in Oberlander & al. (2004). In some cases PCR required dilutions of extracted samples. Attempts to increase species numbers using herbarium material failed, except for O. dines R. Ornduff (MO653) and O. fourcadei Salter (MO664). These two specimens were the youngest material tested (both less than 10 years old). Herbarium samples were extracted using a Qiagen DNeasy plant DNA extraction kit (Qiagen Sciences, Valencia, California, U.S.A.) according to the manufacturer’s protocols. For species complexes, multiple samples from a broad geographical or morphological range were included to test for species monophyly. 1668 TAXON 60 (6) • December 2011: 1667–1677 Amplification of trnL-F was performed using the Taberlet & al. (1991) primers c and f, while the trnS-G region was amplified using primers trnS and trnG described by Hamilton (1999). The ITS region was amplified with primers AB101 and AB102 (Sun & al., 1994). Several samples required the use of internal trnL-F (d and e) and ITS (ITS2 and ITS3) primers, due to poor-quality DNA. Each PCR reaction contained the following reagents: 0.2 mM each dNTP, 1× Supertherm buffer and 2.5 mM MgCl2, 0.5 µM of forward and reverse primers, 0.25 U of Supertherm Taq polymerase, 20–50 ng of template DNA. Ultra-distilled H2O was added to a final volume of 25 μl. Temperature protocols were a 94°C denaturation step for 60 s, a 60 s annealing step (52°C for all plastid primers, 55°C for ITS), and a 90 s extension step at 72°C repeated 30 (plastid) or 35 (ITS) times, followed by a 7 min extension step at 72°C. PCR products were electrophoresed on a 2% agarose gel to confirm band size and to check for potential multiple bands. All PCR products, including excised gel slices, were purified using Wizard DNA Prep purification kits (Promega, Madison, Wisconsin, U.S.A.) according to manufacturer’s protocols. Sequencing reactions consisted of 1 μl Big Dye Terminator Mix RR (Applied Biosystems, Foster City, California, U.S.A.), 3 μl sequence dilution buffer, 1 μl of primer (3 pmol/μl) and 5 μl of purified PCR product. Sequencing reactions used PCR primers. Both forward and reverse reads were sequenced for all markers, except trnS-G which is short enough to confidently read through in one direction. Sequencing protocols were: 10 s at 96°C, 30 s at 52°C (ITS 55°C), 4 min at 60°C, for 35 cycles. Sequencing products were analysed on an ABI 377 sequencer (Applied Biosystems) at Stellenbosch University’s Central Analytical Facility. Confirmation of base calling in sequence chromatograms was conducted manually in Chromas v.2.3 (http://www.techne lysium.com.au). Contig assembly was performed in BioEdit v.7.0.0 (Hall, 1999) using the embedded ClustalW function (Thompson & al., 1994) for alignment with subsequent manual optimization. Sequences were checked against GenBank (NCBI) submissions using BLAST searches to screen for potential contamination. Phylogenetic analysis. — Each individual marker was analysed separately, as well as combined plastid and total datasets, using both parsimony and model-based approaches. Parsimony analyses were performed in PAUP* v.4.0 beta 10 (Swofford, 2003). Likelihood analyses were performed using a genetic algorithm procedure implemented under the optimal model in GARLI v.0.951 and v.0.96 (Zwickl, 2006). Bayesian inference was performed using MrBayes v.3.1.2 (Ronquist & Huelsenbeck, 2003). Gaps were not coded. For parsimony analyses, heuristic searches employed 1000 replicates using random-addition-generated starting trees and TBR branch swapping to find multiple most parsimonious trees, using all characters with equal weighting and unordered. All other options were kept as default. A maximum of 10 best trees per replicate were saved to more efficiently explore tree space. Parsimony bootstrap (PB: 1000 replicates, starting trees generated by simple addition, 10 trees saved per replicate, TBR branch swapping) was implemented as a measure of clade support. TAXON 60 (6) • December 2011: 1667–1677 For maximum likelihood (ML) analyses, the Akaike information criterion (AIC) was used to choose an optimal model of sequence evolution in Modeltest v.3.6 (Posada & Crandall, 1998), using likelihood values obtained by PAUP*. Likelihood searches in GARLI v.0.951 and v.0.96 (Zwickl, 2006) used default parameters under the chosen model. Non-parametric bootstrap support measures (LB: 100 replicates) were constructed using the bootstrap function of the same. For Bayesian inference, MrModeltest v.2.2 (Nylander, 2004) was used to determine the optimal model of DNA evolution for the data via the AIC. Two concurrent analyses were run for 5 × 106 generations per dataset, utilizing Metropoliscoupled Markov Chain Monte Carlo (MCMCMC), sampling every 500 generations. Apart from the implemented model and generation time, all other settings were kept at default values. Visual inspection of the posterior distribution, together with MrBayes’ own convergence diagnostics, were used to judge stationarity and extent of the burn-in (10% of total run length in all cases). Consensus topology and branch lengths (excluding burn-in) were calculated using the sumt command. Clade support was estimated using posterior probability values (PP). The Incongruence Length Difference (ILD) test (Farris & al., 1995; implemented as the Partition Homogeneity Test in PAUP*; 100 replicates) was used to assess congruence of signal between the various datasets. This was also judged visually by inspection for incongruent nodes based on bootstrap values. Nodes were judged strongly incongruent if they supported different bipartitions of taxa with parsimony/likelihood bootstrap values greater than 70% in each segregate dataset. Ancestral area reconstruction. — For SA Oxalis, three geographical regions could serve as ancestral area: the New World (particularly South America), the CFR, or Africa outside the CFR (hereafter referred to as summer-rainfall Africa). For these reconstructions, the definition of the CFR was expanded to the Greater Cape Floristic Region (GCFR) as defined by Born & al. (2006). Although this region is larger than the traditional CFR, the GCFR was chosen because: (1) the flora of this region forms a distinctive unit that corresponds closely to the winter-rainfall region of southern Africa (Born & al., 2006); (2) Oxalis is well-represented and diverse in both the CFR and the non-CFR areas of the GCFR, but diversity levels drop off steeply at the boundaries of the GCFR, making character coding easier. No species are indigenous to both the New World and the African continent. The few Oxalis shared between the GCFR and summer-rainfall Africa were given the code of the region in which their range is largest. To test whether SA Oxalis is a Cape clade or not, we used a reversible jump Markov Chain Monte Carlo (rjMCMC) approach in the program MultiState of the BayesTraits v.1.0 package (Pagel & Meade, 2006). This approach searches among models of character evolution (in this case, models of change in geographical area) and visits different models and parameter values in proportion to their posterior probability. By providing a sample of trees from the Bayesian analyses of the combined data, error in phylogenetic estimation is also incorporated into the search. A sample of 1600 trees (every 20th tree, including different resolutions of clades 2–4) from the combined Bayesian Oberlander & al. • Phylogeny of southern African Oxalis analyses was trimmed to reduce the represented areas to the above-mentioned three. These three areas gave rise to six rates: rNew World→GCFR rNew World→summer-rainfall Africa rGCFR→summer-rainfall Africa and their reverses. All rjMCMC analyses were run for 5.05 × 107 generations, sampling every 1000th generation after a 5 × 105 generation burn-in period, under a gamma hyperprior with alpha and beta values chosen from a uniform distribution between 0 and 10. As suggested by the BayesTraits manual, a rate deviation value was chosen from initial analyses to give chain acceptance values between 20% and 40%. Each analysis was repeated three times. We tested the fit of different ancestral areas to the data using the “fossil” command, where nodes of interest can be fixed to different character states, and assessed against each other using Bayes Factors. The nodes of interest are those subtending SA Oxalis (clade 1) and its sister clade (Oxalis sect. Ionoxalis in these analyses), SA Oxalis itself, and its daughter clades 2, 3, and 4. By fixing the character states at these nodes, we could directly test the support provided by phylogeny and current distribution patterns for these ancestral area hypotheses. All analyses were checked for convergence on the same posterior distribution in Tracer (Rambaut & Drummond, 2007). Harmonic means provided by MultiState were averaged across three runs. Bayes Factors were estimated as twice the difference of the harmonic mean of the ln likelihood values between competing hypotheses (Kass & Raftery, 1995). RESULTS Species collection and sampling. — In total 150 species comprising ca. 75% of currently recognized SA Oxalis (Dreyer & al., 2009) were included (Appendix). All sections and subsections sensu Salter (1944) were sampled, except Oxalis sect. Cernuae subsect. Goetzea (1 sp.). Multiple accessions of species complexes either coalesced as monophyletic lineages with strong support or were close relatives, except in O. stenopetala (see below). Due to these results and computation time constraints, species complexes were reduced to single accessions. Plastid sequences of O. stenopetala grouped together strongly, but one ITS sequence proved highly divergent. Re-extraction and amplification yielded the same sequence. The ITS accession that yielded consistent positions in both ITS and plastid topologies was used to represent this species. DNA extraction and sequencing. — PCR of plastid markers routinely yielded single bands. PCR of ITS often gave multiple, well-spaced bands and high background even at higher annealing temperatures, so all ITS products were excised from gels. Sequence sampling was complete, except for six species (O. burkei, O. dines, O. fourcadei, O. lanata, O. livida, O. minuta), which lacked trnS-G. Part of the ITS1 spacer could not be sequenced for O. copiosa due to length polymorphisms. Oxalis oculifera had an unusual ITS sequence (many indels and substitutions, substitutions in highly conserved areas), but this was retained as the sole nuclear marker for this species, and because ITS analyses with and without O. oculifera were very 1669 Oberlander & al. • Phylogeny of southern African Oxalis TAXON 60 (6) • December 2011: 1667–1677 similar (data not shown). Alignments for each marker are available in the Supplementary Data section to the online version of this article. Properties of ITS and plastid data are shown in Table 1. Average (range in bp) sequence length for each dataset was: ITS: 772.2 (708–800); trnL-F: 917.7 (796–949); trnS-G: 716.3 (597–775; based on 165 taxa). Alignment of trnL-F was unproblematic. The trnS-G spacer was relatively easy to align across Oxalis, but outgroup alignment was more complex. ITS sequence data were very variable, to the point where portions of the ITS1 spacer were not confidently alignable to other genera in Oxalidaceae. Different alignments of this marker were attempted, using a variety of gap-opening procedures and manual alterations. In all cases the topology of produced trees was very similar. The final ITS alignment was automatically aligned for SA Oxalis, with successively more distant outgroups added and aligned manually. Phylogenetic analysis. — The trnL-F and trnS-G partitions showed no significant incongruence under the ILD test (P = 0.31) and were analysed as a single plastid dataset. However, the ILD test showed substantial disagreement between ITS and combined plastid data partitions (P = 0.01). Excluding obviously incongruent taxa did not increase congruence between partitions, even with more than 33% of taxa excluded (62 of 171 taxa, ILD test: P < 0.05). Incongruence results from many violations of basic phylogenetic assumptions. However, the backbone structures of the plastid and ITS trees are similar for SA Oxalis (see below). Given that the systematics of SA Oxalis is so understudied, we favor the results from the combined analyses here, and highlight incongruent taxa in the text for future scrutiny. Plastid trees were generally better-resolved, with more well-supported nodes than ITS-based trees (i.e., bootstrap values > 80%, PP values > 0.95; Table 1). Trees produced by combined data had better resolution and support than either individual dataset. Parsimony and likelihood analyses trees showed comparable levels of resolved and well-supported nodes, but Bayesian inference trees had substantially greater resolution and support. Model-based methods better resolved the spine of the topology, and were congruent with one another and with parsimony analyses of plastid and combined data. Oxalis sect. Corniculatae and O. barrelieri were always among the first lineages to separate within Oxalis, followed by successive branches leading to O. acetosella, the clade containing the O. tuberosa alliance (de Azkue & Martinez, 1990; Emshwiller, 2002) and sect. Ionoxalis, which was generally resolved as sister lineage to an SA Oxalis clade (clade 1, Fig. 1). Within SA Oxalis, three major clades were strongly supported (Fig. 1; Table 2): a clade containing the weed O. pes-caprae (8 spp.: clade 2), sect. Cernuae subsect. Lividae (3 spp.: clade 3), and a core clade (139 spp.: clade 4). The relationships between clades 2 to 4 differed between datasets (Fig. 1; Figs. S1–S2 in the Electronic Supplement). Clade 4 was poorly resolved despite strong support for its monophyly. Major internal clades included an unexpected, weakly supported clade of O. commutata and O. orbicularis, together with clades 5, 6, and 9. Various sub-clades within clade 6 (clades 7 and 8) and clade 9 (clades 10–13) were variously supported but all have potential morphological or palynological synapomorphies. Ancestral area reconstruction. — Support for ancestral area hypotheses as judged by Bayes Factors are presented in – 109 64.5 72 42.6 0.663/0.813 – 108 63.9 60 35.5 ITS 882 0.00 41.0 20 2218 0.379/0.703 – 78 46.2 46 27.2 Combined – – – – – – GTR + I + Γ 113 66.9 75 44.4 Likelihood Plastid Bayesian inference % strongly supported nodes % internal nodes resolved 0.499/0.733 2124 Model Nodes > 80% bootstrap and > 0.95 PP No. internal nodes resolved 4504 250 Consistency index/ Retention index 230 22.5 Tree length No. trees 27.1 1.90 Combined 3516 Parsimony Plastid % missing data 1.43 2634 Aligned characters % parsimonyinformative Table 1. Properties of marker datasets and trees for the ITS, plastid, and combined datasets. Node resolution is represented by number of nodes present in the parsimony or likelihood bootstrap consensus tree, or in the 50% majority-rule consensus Bayesian tree. – – – – – – TVM + I + Γ 102 60.4 66 39.1 ITS – – – – – – GTR + I + Γ 90 53.3 53 31.4 Combined – – – – – – GTR + I + Γ 142 84.0 101 59.8 Plastid – – – – – – GTR + I + Γ 122 72.2 89 52.7 ITS – – – – – – GTR + I + Γ 110 65.1 73 43.2 Fig. 1. Bayesian 50% majority-rule consensus tree of the combined ITS, trnL-F, and trnS-G dataset. Thickened branches have Bayesian posterior probability values (PP) > 0.95. Likelihood bootstrap and parsimony bootstrap values are indicated above and below branches, respectively. Outgroup genera Averrhoa, Biophytum, Dapania, and Sarcotheca are not shown. Numbered clades correspond to clades discussed in the text: clade 1 is SA Oxalis. For species included in Salter (1944), shapes on the terminal branches follow his sectional classification as indicated in the legend. 1670 TAXON 60 (6) • December 2011: 1667–1677 100 92 100 100 100 100 100 100 100 100 100 100 100 100 75 83 99 97 59 57 100 99 100 100 93 92 2 100 99 3 78 54 63 100 100 100 100 64 51 99 99 100 99 59 64 100 5 1 79 89 100 89 92 93 96 52 61 72 79 75 7 57 6 59 59 64 77 78 97 87 86 83 99 100 75 66 91 92 69 98 93 60 54 62 100 95 4 100 100 95 94 8 90 87 100 100100 100 88 85 93 96 58 100 100 96 92 100 99 99 97 91 81 83 73 51 94 94 75 70 100 100 96 100 88 85 Oxalis barrelieri Oxalis corniculata Oxalis dillenii Oxalis stricta Oxalis acetosella Oxalis cf. pachyrrhiza Oxalis vulcanicola Oxalis valdiviensis Oxalis hypsophila Oxalis perdicaria Oxalis brasiliensis Oxalis latifolia Oxalis cf. lasiandra Oxalis tetraphylla Oxalis pseudo-cernua Oxalis purpurascens Oxalis pes-caprae Oxalis cf. compressa Oxalis knuthiana Oxalis sp. nov. Oxalis haedulipes Oxalis copiosa Oxalis dentata Oxalis livida Oxalis lateriflora Oxalis commutata Oxalis orbicularis Oxalis imbricata Oxalis bowiei Oxalis cf. tragopoda Oxalis stellata Oxalis psilopoda Oxalis caprina Oxalis dichotoma Oxalis comosa Oxalis oculifera Oxalis obliquifolia Oxalis punctata Oxalis setosa Oxalis attaquana Oxalis pulchella Oxalis bullulata Oxalis obtusa Oxalis lichenoides Oxalis annae Oxalis densa Oxalis melanosticta Oxalis purpurea Oxalis adenodes Oxalis luteola Oxalis ambigua Oxalis grammopetala Oxalis fourcadei Oxalis cf. fergusoniae Oxalis convexula Oxalis nortieri Oxalis pocockiae Oxalis depressa Oxalis dilatata Oxalis salteri Oxalis namaquana Oxalis disticha Oxalis dines Oxalis dregei Oxalis cathara Oxalis monophylla Oxalis cf. pulvinata Oxalis flava Oxalis fabaefolia Oxalis cf. sonderiana Oxalis inconspicua Oxalis cf. canaliculata Oxalis campylorrhiza Oxalis sp. affin. campylorrhiza Oxalis furcillata Oxalis viscosa Oxalis crocea Oxalis clavifolia Oxalis stenoptera Oxalis uliginosa Oxalis aurea Oxalis adspersa Oxalis louisae Oxalis flaviuscula Oxalis suavis Oxalis pillansiana Oxalis deserticola Oxalis argillacea 9 78 81 Ionoxalis 100 100 99 100 Corniculatae Oberlander & al. • Phylogeny of southern African Oxalis 100 99 99 99 61 51 91 94 78 90100 99 13 85 87 87 75 84 80 75 77 100 100 98 100 100 100 85 79 11 81 88 81 63 100 100 62 54 88 70 85 63 84 79 10 82 73 76 73 73 66 93 81 62 56 100 100 60 52 100 97 12 99 92 54 97 89 77 74 96 68 96 97 57 100 99 88 90 60 53 59 53 Cernuae Oxalis zeekoevleyensis Oxalis virginea Oxalis orthopoda Oxalis incarnata Oxalis strigosa Oxalis truncatula Oxalis lanata Oxalis smithiana Oxalis bifurca Oxalis bifida Oxalis pendulifolia Oxalis heterophylla Oxalis duriuscula Oxalis palmifrons Oxalis tomentosa Oxalis oligophylla Oxalis hygrophila Oxalis massoniana Oxalis grammophylla Oxalis confertifolia Oxalis camelopardalis Oxalis capillacea Oxalis leptogramma Oxalis eckloniana Oxalis nidulans Oxalis minuta Oxalis fibrosa Oxalis engleriana Oxalis exserta Oxalis primuloides Oxalis reclinata Oxalis cuneata Oxalis kamiesbergensis Oxalis hirta Oxalis linearis Oxalis callosa Oxalis blastorrhiza Oxalis gracilis Oxalis ciliaris Oxalis oreophila Oxalis aridicola Oxalis tenella Oxalis xantha Oxalis campicola Oxalis burkei Oxalis zeyheri Oxalis suteroides Oxalis amblyosepala Oxalis stenopetala Oxalis tenuipes Oxalis phloxidiflora Oxalis comptonii Oxalis ebracteata Oxalis tenuis Oxalis porphyriosiphon Oxalis giftbergensis Oxalis gracilipes Oxalis meisneri Oxalis falcatula Oxalis burtoniae Oxalis polyphylla Oxalis urbaniana Oxalis glabra Oxalis leptocalyx Oxalis goniorrhiza Oxalis amblyodonta Oxalis pallens Oxalis versicolor Oxalis recticaulis Oxalis pusilla Oxalis argyrophylla Oxalis droseroides Oxalis natans Oxalis stictocheila Oxalis tenuifolia Oxalis multicaulis Oppositae Foveolatae Campanulatae Crassulae Latifoliolatae Sagittatae Stictophyllae Angustatae 1671 Oberlander & al. • Phylogeny of southern African Oxalis TAXON 60 (6) • December 2011: 1667–1677 Table 2. Support values for clades, by partition and inference method. Support for parsimony and likelihood are bootstrap values, support for Bayesian inference are posterior probabilities. Clades correspond to those mentioned in the text and figures. ITS Parsimony Clade Clade 1 (SA Oxalis) Clade 2 (O. pes-caprae and relatives) Likelihood Plastid Bayesian inference Parsimony Likelihood Combined Bayesian inference – 69 1.00 95 96 1.00 Parsimony Likelihood Bayesian inference 99 100 1.00 89 100 1.00 86 91 1.00 99 100 1.00 100 100 1.00 96 97 1.00 100 100 1.00 Clade 4 (core SA Oxalis) – 60 0.98 82 90 1.00 95 100 1.00 Clades 2 + 3 – – – 52 73 0.82 – – – Clade 3 (sect. Cernuae subsect. Lividae) Clades 3 + 4 – – 0.65 – – – 72 61 – 98 99 1.00 89a 84a 1.00a 99 99 1.00 Clade 6 – – – 51 67 0.94 – 57 1.00 Clade 7 (O. purpurea and relatives) – – – – – – 59 59 1.00 Clade 8 (O. flava and relatives) – 62 0.99 – – – – – 0.99 Clade 9 – – – – – – – – 1.00 Clade 10 63b 65b 0.90 – – 0.73d 54 62 1.00 Clade 11 (O. hirta and relatives) 99 99 1.00 99 100 1.00 100 100 1.00 – – 0.81 89 91 1.00 97 100 1.00 99 98 1.00 99 89 1.00 100 100 1.00 Clade 5 (O. stellata and relatives) Clade 12 (O. glabra and relatives) Clade 13 (sect. Angustatae subsect. Pardales) c a Without O. tragopoda. b Includes three species of sect. Sagittatae in large polytomy. c Three species of sect. Sagittatae sister to this clade. d Includes O. smithiana and O. bifurca. Table 3. Mean proportional likelihoods (Pr Lh) for unfixed analyses and Bayes Factors (BF) for fixed analyses of ancestral area reconstructions for SA Oxalis and other clades of interest. The highest Pr Lh are bolded. For clades, all Bayes Factors are computed against the hypothesis with the highest harmonic mean likelihood: positive Bayes Factors favour the best hypothesis. Following Kass & Raftery (1995), 0 < BF < 2 is barely worth a mention, 2 < BF < 6 is considered evidence in favour of the best hypothesis; BF > 6 is considered strong evidence in favour of the best hypothesis. The mean ln likelihood for the unfixed analyses is –41.11. Unfixed Clade BF New World Pr Lh SA Oxalis + Ionoxalis 0.156 0.936 GCFR Summer-rainfall Africa Pr Lh BF(best/fixed GCFR origin) Pr Lh best 0.015 11.962 0.049 6.548 BF(best/fixed New World origin) BF(best/summer-rainfall Africa origin) 0.078 0.002 8.093 0.924 best 0.074 4.164 Clade 2 best 0.004 13.035 0.281 0.300 0.715 2.314 Clade 3 0.118 0.000 23.070 1.000 best 0.000 16.177 Clade 4 0.080 0.001 15.727 0.986 best 0.013 5.560 SA Oxalis Table 3. All repeated MultiState runs were judged to have converged on the same stationary distribution. The rjMCMC procedure overwhelmingly favoured simpler rate models (> 99.1% 1- and 2-rate categories for all analyses) over more complex models with three to six different rates. All unfixed analyses strongly favoured a New World origin for the stem SA Oxalis lineage and a GCFR origin for SA Oxalis and clades 3 and 4. The proportional likelihoods favoured a summer-rainfall African origin for clade 2, but a GCFR origin also contributed substantially. Analyses fixed for ancestral area showed convincing evidence for a New World origin for the stem SA Oxalis lineage, and supported a GCFR origin for crown SA Oxalis 1672 (Table 3). The daughter clades 3 and 4 were reconstructed as GCFR lineages with moderate to strong support in fixed analyses, whilst clade 2 was weakly supported as a GCFR lineage. DISCUSSION Phylogenetic analysis. — Although outgroup sampling was sparse, outgroup taxa spanned the range of current taxonomic and morphological diversity in Oxalis (Appendix 1; Oberlander & al., 2009b). Ingroup size and taxonomic uncertainty forced us to compromise on choice and number of TAXON 60 (6) • December 2011: 1667–1677 outgroups. Despite this, outgroup topology and support values agreed with those from previous studies (Oberlander & al., 2009b). The placement of the only sampled member of Oxalis subg. Thamnoxys (O. barrelieri) and sect. Corniculatae remain the most uncertain (Oberlander & al., 2004, 2009b). As these taxa are only distantly related to SA Oxalis, this should have no significant effect on our results. Most analyses confirmed previous support for a sister-relationship between SA Oxalis and the New World section Ionoxalis. This needs further verification through inclusion of unsampled geophytic New World lineages sensu Lourteig (2000), such as O. sect. Articulatae R. Knuth, sect. Palmatifoliae DC. sensu Reiche, and sect. Pseudobulbosae Norl. Plastid, ITS, and combined analyses generally agreed on a monophyletic SA Oxalis. In trnL-F–produced trees, clade 3 was weakly supported as sister to section Ionoxalis (data not shown). Strong support for a monophyletic SA Oxalis in the plastid dataset thus comes entirely from trnS-G. For ITS, model-based inference methods also supported SA Oxalis, whilst parsimony analyses produced a polytomy with various outgroups (Fig. S2). This was most likely due to high levels of homoplasy in this dataset. All analyses (except parsimony analyses of ITS) supported three main lineages within SA Oxalis: the species-poor clades 2 and 3, and the core SA clade (clade 4), which contains most of the species (Table 2). The relationships between these three lineages remain uncertain. Plastid datasets weakly supported a (4 (2, 3)) topology with all three inference methods. Trees generated with ITS data were more dependent on inference method. Parsimony and likelihood produced unresolved trees, while Bayesian inference weakly supported a (2 (3, 4)) relationship. For the combined dataset, likelihood and parsimony analyses weakly supported a (2 (3, 4)) relationship, whilst Bayesian results showed a trichotomy. The relationship between these three clades requires independent corroboration from unlinked nuclear markers before more confident assessment of topology can be made. The core SA Oxalis clade (clade 4) was poorly resolved in all analyses. Even adding three plastid markers (the 3′ half of ndhF; the trnK intron, including matK; the trnT-trnL spacer) for a select group of 24 taxa did not appreciably increase resolution in this clade, despite the combined application of more than 7000 bp of information (although support values for clades 6 and 9 did increase; data not shown). In addition, the internal topology of this core clade differs substantially from previous classifications (Knuth, 1930; Salter, 1944). The short branches and poor support values within this clade may be the result of rapid radiation, as in other CFR lineages (Richardson & al., 2001; Klak & al., 2004; Mummenhoff & al., 2005). Of the four main lineages found, only one was strongly supported by all analysis methods and data partitions (clade 5, Table 2, with the exception of O. tragopoda, see below), and the particular arrangement of these four clades varies greatly from analysis to analysis. The large number of species, the variety of produced topologies, and discrepancies between previous work and this study precludes an in-depth discussion of relationships here. Comprehensive analyses of morphological vs. sequence-based Oberlander & al. • Phylogeny of southern African Oxalis datasets and updated taxonomic treatments are required before this can be done with any confidence. However, because potential morphological or palynological synapomorphies exist for many of these clades, we mention them here as a reference point for future morphology-based studies. Comparison to previous work. — Results of this study agree with results from Oberlander & al. (2004) that questioned the Salter (1944) taxonomy. Not one Salter (1944) section was monophyletic (Fig. 1), while all but three subsections (sect. Cernuae subsect. Lividae [clade 3: 3 spp.], sect. Cernuae subsect. Costatae [2 spp.], and sect. Angustatae subsect. Pardales [clade 13: 11 spp.]), are artificial based on DNA evidence. Comparison between the palynological classification of Dreyer (1996) and our analyses shows substantial congruence. The very different supra-areolate pollen type of clade 11 was highlighted by Oberlander & al. (2004), and this finding is corroborated by strong support for this lineage in our analyses. Incongruence. — The basic phylogenetic structure of SA Oxalis is complicated at finer scales by substantial incongruence between nuclear and plastid markers. Although inspection of topologies from separate analyses of nuclear and chloroplast partitions can reveal conflict, discriminating between these causes of conflict can be difficult. Hybridisation and introgression can be considered as likely causes, particularly if documented cases of hybridisation are known to occur in the genus. Although Salter (1944) found no evidence for hybridisation in the field, he mentioned a single potential hybrid between O. macra Schltr. and O. creaseyii Salter in garden collections. Nothing more is known about hybridization between SA Oxalis species. The octoploid Andean crop O. tuberosa Molina, however, is thought to have arisen through allopolyploid hybridization between O. picchensis Knuth or O. chicligastensis Knuth, and an undescribed species (Emshwiller, 2002; Emshwiller & Doyle, 2002; Emshwiller & al., 2009). Several studies have reported considerable variation in ploidy levels within SA Oxalis (summarized in Dreyer & Johnson, 2000). This is corroborated by preliminary research into Oxalis genomic C-value studies that suggest that polyploidy is very common in southern African taxa (Doležel, 1997; J. Suda, unpub. data), particularly within species complexes. It is currently unclear whether this is caused by auto- or allopolyploidy. Other sources of incongruence, such as incomplete lineage sorting, must also be considered. The placement of incongruent species in plastid vs. ITS phylogenies is unusual. In several cases where one sequence is strongly placed, the corresponding sequence from the other genomic compartment is not closely related to any other sampled Oxalis. For example, O. tragopoda (sect. Cernuae subsect. Stellatae) is clearly sister to O. bowiei in clade 5 based on ITS sequence data, a placement which is congruent with this species summer-rainfall distribution and with morphological characters (Figs. S1–S2). The plastid sequence for this species is unresolved in clade 4 in parsimony and likelihood analyses, and weakly supported as sister to clade 6 under Bayesian inference. Similarly, O. palmifrons (sect. Angustatae subsect. Multifoliolatae) is sister to O. tomentosa on the basis of ITS data, but is unresolved in clades 4 or 9 based on plastid data. 1673 Oberlander & al. • Phylogeny of southern African Oxalis The divergent ITS sequence of O. stenopetala shows no close similarity to other SA Oxalis, despite clearly being a member of that clade. In non-recombinant plastid genomes at least, these divergent sequences hint at substantial unsampled diversity in chloroplast lineages within SA Oxalis. Ancestral area reconstruction. — Linder (2003) provided two criteria for Cape Clade status: (1) more than 50% of clade diversity should be present in the CFR, and (2) the clade could be shown to have originated within the CFR. Even with the traditional delimitation of the CFR, SA Oxalis passes criterion 1 (Goldblatt & Manning, 2000); with an expanded GCFR concept, this percentage increases to over 90% of approx. 210 recognised species. In our sampling, clade 2 includes four species from Namibia/the Great Karoo, clade 3 is found exclusively in the GCFR, and clade 4 features six species that extend from the eastern GCFR up the eastern African seaboard as far as Ethiopia. The remaining criterion, of in situ CFR origin, is explicitly addressed by our analyses. It is shown that the SA lineage clearly had an origin in the New World, but extant southern African Oxalis diversity can be directly traced back to a GCFR origin, if not the smaller CFR. The situation for the traditionally defined CFR was not assessed, but it is clear that the origins of SA Oxalis are clearly tied to the CFR region. A critical caveat is sampling. Approximately 60 SA Oxalis species were not sampled for this study, but the overwhelming majority are native to the GCFR, and their addition would strengthen the GCFR origin hypothesis whatever their phylogenetic position. The major test of a GCFR origin for all southern African Oxalis would rely on the phylogenetic placement of the few unsampled extra-GCFR taxa, particularly tropical African and Malagasy taxa. Apart from summer-rainfall members of clade 2, other nonGCFR taxa are all deeply embedded within otherwise GCFR lineages, thus making them escapees from the GCFR. This pattern adds to a growing body of evidence that the CFR has contributed significantly to tropical African floras: examples of other Cape clades that show this pattern include Protea (Barraclough & Reeves, 2005) and Leucadendron (Barker & al., 2004), Cliffortia (Whitehouse, 2002), Disa, Irideae pro parte, Pentaschistis and the African Restionaceae (Galley & al., 2007). From a larger biogeographic perspective, a clear origin in the New World is unusual for a CFR taxon, yet this is the clear pattern for SA Oxalis. This finding is unlikely to be influenced by incomplete taxon sampling, as all unsampled potential relatives of SA Oxalis are New World, primarily southern South American taxa (sects. Articulatae, Palmatifoliae, Pseudobulbosae). Oxalis is one of the few CFR taxa that share a trans-South Atlantic Ocean relationship, together with Prionium, Drosera and Haemadoraceae, and possibly Bruniaceae (summarized in Galley & Linder, 2006). Furthermore, Oxalis is the second genus after Drosera for which directionality of dispersal can be inferred (from South America to Africa). The actual manner in which the ancestor to SA Oxalis crossed the Atlantic remains conjectural, and requires further research on the timing of this event and on seedling and bulb dormancy times. 1674 TAXON 60 (6) • December 2011: 1667–1677 ACKNOWLEDGEMENTS The authors thank the National Research Foundation, South Africa (GUN nr. 2053585) for research funding, and Western and Northern Cape Nature Conservation Depts (South Africa) for collection permits. We thank M.B. Bayer, B. Gravendeel, N. Helme, A. le Roux, L. Mucina, F. Roets, I. Till-Botrand, J. Vlok, J. Walter, and others for providing plant material, DNA samples and/or locality data. Three anonymous reviewers are also thanked for comments and improvements on the manuscript. LITERATURE CITED Bakker, F., Culham, A., Marais, E. & Gibby, M. 2005. Nested radiation in Cape Pelargonium. Pp. 75–100 in: Bakker, F., Chatrou, L., Gravendeel, B. & Pelser, P. (eds.), Plant species-level systematics: New perspectives on pattern and process. Ruggell, Liechtenstein: Gantner. Barker, N.P., Vanderpoorten, A., Morton, C.M. & Rourke, J.P. 2004. Phylogeny, biogeography, and the evolution of life-history traits in Leucadendron (Proteaceae). Molec. Phylogenet. Evol. 33: 845–860. 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 the break-up of Gondwana. J. Biogeogr. 34: 2012–2027. Barraclough, T. & Reeves, G. 2005. The causes of speciation in flowering plant lineages, species-level DNA trees in the African genus Protea. Pp. 31–46 in: Bakker, F., Chatrou, L., Gravendeel, B. & Pelser, P. (eds.), Plant species-level systematics: New perspectives on pattern and process. Ruggell, Liechtenstein: Gantner. Bayer, M.B. 1992. Salter’s revision of South African Oxalis (Oxalidaceae) and some new combinations. Herbertia 48: 58–69. Born, J., Linder, H.P. & Desmet, P. 2007. The Greater Cape Floristic Region. J. Biogeogr. 34: 147–162. Bytebier, B., Antonelli A., Bellstedt D.U. & Linder, H.P. 2011. Estimating the age of fire in the Cape flora of South Africa from an orchid phylogeny. Proc. Roy. Soc. B., Biol. Sci. 278: 188–195. De Azkue, D.& Martinez, A. 1990. DNA content and chromosome evolution in the shrubby Oxalis. Genome 30: 52–57. Doležel, J. 1997. Application of flow cytometry for the study of plant genomes. J. Appl. Genet. 38: 285–302. Doyle, J. & Doyle, J. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19: 11–15. Dreyer, L.L. 1996. A palynological review of Oxalis (Oxalidaceae) in southern Africa. Ph.D dissertation, University of Pretoria, Pretoria, South Africa. Dreyer, L.L. & Johnson, C. 2000. New chromosome number records of South African Oxalis species. S. African J. Bot. 66: 130–132. Dreyer, L.L. & Van Wyk, A.E. 1996. Taxonomic delimitation of Oxalis engleriana. Bothalia 28: 65. Dreyer, L.L., Roets, F. & Oberlander, K.C. 2009. Oxalis saltusbelli: A new Oxalis (Oxalidaceae) species from the Oorlogskloof Nature Reserve, Nieuwoudtville, South Africa. S. African J. Bot. 75: 110–116. Edwards, D. & Hawkins, J.A. 2007. Are Cape floral clades the same age? Contemporaneous origins of two lineages in the genistoids s.l. (Fabaceae). Molec. Phylogenet. Evol. 45: 952–970. Eldenas, P. & Linder, H.P. 2000. Congruence and complementarity of morphological and trnL-trnF sequence data and the phylogeny of the African Restionaceae. Syst. Bot. 25: 692–707. Emshwiller, E. 2002. Ploidy levels among species of the “Oxalis tuberosa Alliance” as inferred by flow cytometry. Ann. Bot. 89: 741–753. Emshwiller, E. & Doyle, J.J. 2002. Origins of domestication and TAXON 60 (6) • December 2011: 1667–1677 polyploidy in oca (Oxalis tuberosa: Oxalidaceae) 2: Chloroplastexpressed glutamine synthetase data. Amer. J. Bot. 89: 1042–1056. Emshwiller, E., Theim, T., Grau, A., Nina V. & Terrazas, F. 2009. Origins of domestication and polyploidy in oca (Oxalis tuberosa; Oxalidaceae) 3: AFLP data of oca and four wild, tuber-bearing taxa. Amer. J. Bot. 96: 1839–1848. Farris, J.S., Källersjö, M., Kluge, A.G. & Bult, C. 1995. Testing significance of incongruence. Cladistics 10: 315–319. Forest, F., Nänni, N., Chase, M., Crane, P. & Hawkins, J.A. 2007. Diversification of a large genus in a continental biodiversity hotspot: Temporal and spatial origin of Muraltia (Polygalaceae) in the Cape of South Africa. Molec. Phylogenet. Evol. 43: 60–74. Galley, C. & Linder, H.P. 2006. Geographical affinities of the Cape flora, South Africa. J. Biogeogr. 33: 236–250. 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. B, Biol. Sci. 274: 535–543. Goldblatt, P. & Manning, J. 2000. Cape plants—A conspectus of the Cape flora of South Africa. Pretoria: National Botanic Institute. Hall, T.A. 1999. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids. Symp. Ser. 41: 95–98. Hamilton, M.B. 1999. Four primer pairs for the amplification of chloroplast intergenic regions with intraspecific variation. Molec. Ecol. 8: 521–525. Hardy, C.R., Moline, P. & Linder, H.P. 2008. A phylogeny for the African Restionaceae and new perspectives on morphology’s role in generating complete species phylogenies for large clades. Int. J. Pl. Sci. 169: 377–390. Kass, R.E. & Raftery, A.E. 1995. Bayes Factors. J. Amer. Statist. Assoc. 90: 773–795. Klak, C., Reeves, G. & Hedderson, T. 2004. Unmatched tempo of evolution in southern African semi-desert ice plants. Nature 427: 63–65. Knuth, R. 1930. “Oxalidaceae”. Pp. 1–481 in: Engler, A. (ed.), Das Pflanzenreich, vol. 95 IV, 130. Leipzig: Engelman. Kumwenda, M.W., Dreyer, L.L. & Marais, E.M. 2004. A taxonomic reassessment of the varieties of Oxalis minuta (Oxalidaceae) and the change of O. minuta var. callosa to specific rank as O. hygrophila. S. African J. Bot. 70: 259–264. Linder, H.P. 2003. The radiation of the Cape flora, southern Africa. Biol. Rev. (Cambridge) 78: 597–638. Linder, H.P. & Hardy, C.R. 2005. Species richness in the Cape Flora: A macroevolutionary and macroecological perspective. Pp. 47–74 in: Bakker, F., Chatrou, L., Gravendeel, B. & Pelser, P. (eds.), Plant species-level systematics: New perspectives on pattern and process. Ruggell, Liechtenstein: Gantner. Linder, H.P., Eldenas, P. & Briggs, B. 2003. Contrasting patterns of African and Australian Restionaceae. Evolution 57: 2688–2702. Linder, H.P., Hardy, C.R. & Rutschmann F. 2005. Taxon sampling effects in molecular clock dating: An example from the African Restionaceae. Molec. Phylogenet. Evol. 35: 569–582. Lourteig, A. 1994. Oxalis L. subgenera Thamnoxys (Endl.) Reiche emend. Lourteig. Bradea 7: 1–199. Lourteig, A. 2000. Oxalis L. subgenera Monoxalis (Small) Lourteig, Oxalis y Trifidus Lourteig. Bradea 7: 201–629. Manning, J.C. & Goldblatt, P. 2008. Oxalidaceae: A new species of Oxalis from the Hantam-Roggeveld plateau, Northern Cape, South Africa. Bothalia 38: 75–78. Mummenhoff, K., Al-Shehbaz, I.A., Bakker, F.T., Linder, H.P. & Mühlhausen, 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. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B. & Kent, J. 2000. Biodiversity hotspots for conservation priorities. Nature 403: 853–858. Nylander, J.A. 2004. MrModeltest, version 2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University. Oberlander & al. • Phylogeny of southern African Oxalis Oberlander, K.C., Dreyer, L.L., Bellstedt, D.U. & Reeves, G. 2004. Systematic relationships in southern African Oxalis L. (Oxalidaceae): Congruence between palynological and plastid trnL-F evidence. Taxon 53: 977–985. Oberlander, K.C., Dreyer, L.L. & Curran, H.R. 2009a. An unusual new species of Oxalis (Oxalidaceae) from the Knersvlakte, South Africa. S. African J. Bot. 75: 239–245. Oberlander, K.C., Emshwiller, E., Bellstedt, D.U. & Dreyer, L.L. 2009b. A model of bulb evolution in the eudicot genus Oxalis (Oxalidaceae). Molec. Phylogenet. Evol. 51: 54–63. Oliver, E.G.H. 1993. Oxalidaceae—a new species of Oxalis from the Western Cape. Bothalia 23: 72–74. Ornduff, R. 1973. Oxalis dines, a new species from the Western Cape. J. S. African Bot. 39: 201–203. Pagel, M. & Meade, A. 2006. Bayesian analysis of correlated evolution of discrete characters by reversible-jump Markov chain Monte Carlo. Amer. Naturalist 167: 808–825. Posada, D. & Crandall, K.A. 1998. Modeltest: Testing the model of DNA substitution. Bioinformatics 14: 817–818. Procheş, S., Cowling, R.M., Goldblatt, P., Manning, J.C. & Snijman, D.A. 2006. An overview of the Cape geophytes. Bot. J. Linn. Soc. 87: 27–43. Rambaut A. & Drummond A.J. 2007. Tracer, version 1.5. http://beast .bio.ed.ac.uk/Tracer. Richardson, J.E., Weitz, F.M., Fay, M.F., Cronk, Q.C.B., Linder, H.P., Reeves, G. & Chase, M.W. 2001. 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. Rourke, J.P. 1998. A review of the systematics and phylogeny of the African Proteaceae. Austral. Syst. Bot. 11: 267–285. Salter, T.M. 1944. The genus Oxalis in South Africa: A taxonomic revision. J. S. African Bot., suppl. vol. 1: 1–355. Sauquet, H., Weston, P.H., Barker, N.P. Anderson, C.L., Cantrill, D.J. & Savolainen, V. 2009. Using fossils and molecular data to reveal the origins of the Cape proteas (subfamily Proteoideae). Molec. Phylogenet.Evol. 51: 31–43. Sun, Y., Skinner, D.Z., Liang, G.H. & Hulbert, S.H. 1994. Phylogenetic analysis of Sorghum and related taxa using internal transcribed spacers of nuclear ribosomal DNA. Theor. Appl. Genet. 89: 26–32. Swofford, D.L. 2003. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.0b10, 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. Thompson, J.D., Higgins, D.G. & Gibson, T.J. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position specific gap penalties and weight matrix choice. Nucl. Acids Res. 22: 4673–4680. Van der Niet, T., Johnson, S.J. & Linder, H.P. 2006. Macroevolutionary data suggest a role for reinforcement in pollination system shifts. Evolution 60: 1596–1601. Verboom, G.A., Archibald, J.K., Bakker, F.T., 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. 2009. Origin and diversification of the Greater Cape flora: Ancient species repository, hot-bed of recent radiation, or both? Molec. Phylogenet. Evol. 51: 44–53. Verboom, G.A., Linder, H.P. & Stock, W.D. 2004. Testing the adaptive nature of radiations: Growth form and life history divergence in the African grass genus Ehrharta (Poaceae: Ehrhartoideae). Amer. J. Bot. 91: 1364–1370. Whitehouse, C.M. 2002. Systematics of the genus Cliffortia L. 1675 Oberlander & al. • Phylogeny of southern African Oxalis (Rosaceae). Ph.D. dissertation, University of Cape Town, Cape Town, South Africa. Williamson, G. 1999. A new succulent Oxalis (Oxalidaceae) from the Richtersveld. Aloe 36: 68–70. TAXON 60 (6) • December 2011: 1667–1677 Zwickl, D.J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. dissertation, the University of Texas at Austin, Texas, U.S.A. Appendix. Species used in this study, presented as follows: Species Project reference number [Section-Subsection] Voucher specimen and locality: ITS GenBank accession; trnL-F GenBank accession; trnS-G GenBank accession. Taxonomic information for Oxalis is presented sensu Salter (1944) for southern African taxa and sensu Lourteig (1994, 2000) for other Oxalis. Species described since Salter (1944) or undescribed species are assigned a placement in inverted commas, based on the original description or on a presumed placement according to morphology. Averrhoa bilimbi L. AVBIL [–] (Ex Hort.): EU436862; AJ582291; EU437151. Averrhoa carambola L. AVCAR [–] (Ex Hort.): EU436863; EU437032; EU437152. Biophytum abyssinicum Steud. ex. A. Rich. BIOAB [–] (Ex Hort.): EU436864; AJ582292; EU437153. Biophytum sp. BIOPHIL [–] (Ex Hort.): EU436865; AJ582293; EU437154. Dapania pentandra Capuron DAPEN [–] (Ex Hort.): EU436861; EU437031; EU437150. Oxalis sp. no. MO620 [‘Cernuae R. Knuth-Costatae T.M. Salter’] Mannheimer 2591 (Sperrgebied, Namibia): EU436885; EU437047; EU437174. Oxalis “canaliculata” (manuscript name, M.B. Bayer) MO583 [‘Crassulae T.M. Salter’] (Ex Hort.): EU436946; EU437088; EU437232. Oxalis acetosella L. MO665 [Oxalis A. Lourteig] J. Walter 7182 (Niederösterreich, Austria): EU436870; EU437036; EU437159. Oxalis adenodes Sond. MO147 [Oppositae T.M. Salter-Subintegrae T.M. Salter] L. Dreyer 716 (Northern Cape, South Africa): EU436903; EU437060; EU437191. Oxalis adspersa E. & Z. MO66 [Angustatae T.M. Salter-Xanthotrichae T.M. Salter]; L. Dreyer 635 (Western Cape, South Africa): EU436936; AJ582290; EU437223. Oxalis ambigua Jacq. MO555 [Oppositae-Subintegrae] K. Oberlander 0140 (Northern Cape, South Africa): EU436905; EU437061; EU437193. Oxalis amblyodonta Salter MO597 [Angustatae-Multifoliolatae R. Knuth] M.B. Bayer 7397 (Western Cape, South Africa): EU437007; EU437116; EU437293. Oxalis amblyosepala Schltr. MO721 [Latifoliolatae T.M. Salter] K. Oberlander 0168 (Western Cape, South Africa): EU436982; EU437105; EU437268. Oxalis annae Bolus f. MO436 [Foveolatae T.M. Salter] L. Mucina 170903/3 (Western Cape, South Africa): EU436920; EU437072; EU437208. Oxalis argillacea Bolus f. MO282 [Angustatae-Xanthotrichae] L. Dreyer 786 (Western Cape, South Africa): EU436940; AJ582294; EU437227. Oxalis argyrophylla Salter MO313 [Angustatae-Lineares T.M. Salter] K. Oberlander 0040 (Western Cape, South Africa): EU437005; AJ582295; EU437291. Oxalis aridicola Salter MO103 [Latifoliolatae] L. Dreyer 672 (Western Cape, South Africa): EU436977; AJ582296; EU437263. Oxalis attaquana Salter MO45 [Foveolatae] M.B. Bayer 7007 (Western Cape, South Africa): EU436918; EU437071; EU437206. Oxalis aurea Schltr. MO267 [Latifoliolatae] L. Dreyer 771 (Western Cape, South Africa): EU436935; EU437080; EU437222. Oxalis barrelieri L. MO454 [Thamnoxys (Endl.) Reiche emend. A. Lourteig-Thamnoxys A. Lourteig] DUB 809 (Southwest Province, Cameroon); EU436866; EU437033; EU437155. Oxalis bifida Thunb. MO19 [Oppositae-Bifurcatae R. Knuth] L. Dreyer 608 (Western Cape, South Africa): EU437017; AJ582297; EU437303. Oxalis bifurca Lodd. MO324 [Oppositae-Bifurcatae] DUB 764 (Eastern Cape, South Africa): EU436965; EU437100; EU437251. Oxalis blastorrhiza Salter MO284 [Angustatae-Lineares] L. Dreyer 788 (Western Cape, South Africa): EU436974; AJ582298; EU437260. Oxalis bowiei Lindl. MO502 [Cernuae-Purpuratae T. M. Salter] (Ex Hort.): EU436895; EU437055; EU437183. Oxalis brasiliensis Lodd. MO391 [Ionoxalis (Small) R. Knuth] J. Holmes 0008 (Ex Hort.): EU436879; EU437044; EU437168. Oxalis bullulata Salter MO288 [Foveolatae] L. Dreyer 792 (Western Cape, South Africa): EU436908; EU437063; EU437196. Oxalis burkei Sond. MO29 [Angustatae-Lineares] L. Dreyer 618 (Western Cape, South Africa): EU437029; AJ582299; –. Oxalis burtoniae Salter MO304 [Angustatae-Lineares] K. Oberlander 0031 (Western Cape, South Africa): EU437008; AJ582300; EU437294. Oxalis callosa R. Knuth MO532 [Latifoliolatae] K. Oberlander 0120 (Northern Cape, South Africa): EU436973; EU437103; EU437259. Oxalis camelopardalis Salter MO371 [Angustatae-Pardales T.M. Salter] K. Oberlander 0071 (Western Cape, South Africa): EU437012; EU437119; EU437298. Oxalis campicola Salter MO276 [Angustatae-Lineares] L. Dreyer 780 (Northern Cape, South Africa): EU436981; AJ582301; EU437267. Oxalis campylorrhiza Salter MO127 [Latifoliolatae] L. Dreyer 696 (Northern Cape, South Africa): EU436947; AJ582302; EU437233. Oxalis sp. aff. campylorrhiza Salter MO120 [‘Latifoliolatae’] L. Dreyer 689 (Northern Cape, South Africa): EU436948; EU437089; EU437234. Oxalis capillacea E. Mey. ex Sond. MO36 [Angustatae-Pardales] L. Dreyer 625 (Western Cape, South Africa): EU437013; AJ582304; EU437299. Oxalis caprina L. MO7 [Cernuae-Stellatae R. Knuth] L. Dreyer & K. Oberlander 0001 (Western Cape, South Africa): EU436892; AJ582305; EU437180. Oxalis cathara Salter MO582 [Crassulae] (Ex Hort.): EU436931; EU437077; EU437218. Oxalis ciliaris Jacq. MO24 [Angustatae-Lineares] L. Dreyer 613 (Western Cape, South Africa): EU436979; AJ582307; EU437265. Oxalis clavifolia Sond. MO556 [Angustatae-Glandulosae T.M. Salter] K. Oberlander 0141 (Northern Cape, South Africa): EU436951; EU437090; EU437237. Oxalis commutata Sond. MO17 [Latifoliolatae] L. Dreyer 606 (Western Cape, South Africa): EU436898; AJ582308; EU437186. Oxalis comosa E. Mey. ex Sond. MO238 [Oppositae-Bifurcatae] L. Mucina 220901/30A (Northern Cape, South Africa): EU436954; EU437091; EU437240. Oxalis compressa L. f. MO519 [Cernuae-Eu-Cernuae R. Knuth] Bayer and Oberlander 0009 (Western Cape, South Africa): EU436883; EU437046; EU437172. Oxalis comptonii Salter MO298 [Angustatae-Lineares] L. Dreyer 802 (Western Cape, South Africa): EU436985; AJ582309; EU437272. Oxalis confertifolia (O. Kuntze) R. Knuth MO358 [Angustatae-Pardales] K. Oberlander 0075 (Western Cape, South Africa): EU437011; EU437118; EU437297. Oxalis convexula Jacq. MO209 [Foveolatae] K. Oberlander 0014 (Western Cape, South Africa): EU436913; EU437067; EU437201. Oxalis copiosa Bolus f. MO227 [Cernuae-Eu-Cernuae] K. Oberlander 0021 (Northern Cape, South Africa): EU436887; EU437049; EU437176. Oxalis corniculata L. MO256 [Corniculatae DC.] K. Oberlander 0025 (Western Cape, South Africa): EU436867; AJ582310; EU437156. Oxalis crocea Salter MO124 [Angustatae-Sessilifoliatae R. Knuth] L. Dreyer 693 (Northern Cape, South Africa): EU436950; AJ582311; EU437236. Oxalis cuneata Jacq. MO287 [Angustatae-Lineares] L. Dreyer 791 (Western Cape, South Africa): EU436969; EU437101; EU437255. Oxalis densa N.E. Br. MO783 [Foveolatae] K. Oberlander 0169 (Northern Cape, South Africa): EU436921; EU437073; EU437209. Oxalis dentata Jacq. MO49 [Cernuae-Lividae T.M. Salter] C. Cilliers 0004 (Western Cape, South Africa): EU436888; AJ582312; EU437177. Oxalis depressa E. & Z. MO464 [Foveolatae] K. Oberlander 0085 (Western Cape, South Africa): EU436909; EU437064; EU437197. Oxalis deserticola Salter MO526 [Angustatae-Glandulosae] K. Oberlander 0114 (Western Cape, South Africa): EU436939; EU437083; EU437226. Oxalis dichotoma Salter MO333 [Cernuae-Stellatae] K. Oberlander 0052 (Western Cape, South Africa): EU436893; EU437053; EU437181. Oxalis dilatata L. Bolus MO524 [Foveolatae] Bayer & Oberlander 0014 (Western Cape, South Africa): EU436910; EU437065 ; EU437198. Oxalis dillenii Jacq. MO667 [Corniculatae] J. Walter 6565 (Niederösterreich, Austria): EU436868; EU437034; EU437157. Oxalis dines Ornduff MO653 [‘Campanulatae T.M. Salter’] N.A. Helme 1355 (Western Cape, South Africa): EU436943; EU437085; –. Oxalis disticha Jacq. MO596 [Campanulatae] K. Oberlander 0159 (Western Cape, South Africa): EU436942; EU437084; EU437229. Oxalis dregei Sond. MO447 [Campanulatae] L. Mucina 210903/11 (Northern Cape, South Africa): EU436930; AJ582313; EU437217. Oxalis droseroides E. Mey. ex Sond. MO362 [Angustatae-Glandulosae] K. Oberlander 0079 (Western Cape, South Africa): EU436994; EU437111; EU437280. Oxalis duriuscula Schltr. MO461 [Oppositae-Subintegrae] K. Oberlander 0082 (Western Cape, South Africa): EU437020; EU437123; EU437306. Oxalis ebracteata Savign. MO75 [Angustatae-Glandulosae] L. Dreyer 644 (Western Cape, South Africa): EU436990; AJ582314; EU437276. Oxalis eckloniana Presl. MO39 [Sagittatae T.M. Salter] L. Dreyer 628 (Western Cape, South Africa): EU437025; AJ582315; EU437311. Oxalis engleriana Schltr. MO195 [Angustatae-Multifoliolatae R. Knuth] K. Oberlander 0010 (Western Cape, South Africa): EU436961; AJ582316; EU437247. Oxalis exserta Salter MO117 [Angustatae-Lineares] L. Dreyer 686 (Northern Cape, South Africa): EU436967; AJ582317; EU437253. Oxalis fabaefolia Jacq. MO152 [Crassulae] L. Dreyer 721/A (Western Cape, South Africa): EU436925; EU437075; EU437212. Oxalis falcatula Salter MO476 [Angustatae-Lineares] K. Oberlander 0096 (Western Cape, South Africa): EU436992; EU437109; EU437278. Oxalis fergusoniae Salter MO501 [Foveolatae] K. Oberlander 0111 (Western Cape, South Africa): EU436912; EU437066; EU437200. Oxalis fibrosa Bolus f. MO332 [Sagittatae] K. Oberlander 0051 (Western Cape, South Africa): EU436958; EU437095; EU437244. Oxalis flava L. MO25 [Crassulae] L. Dreyer 614 (Western Cape, South Africa): EU436924; AJ582318; EU437211. Oxalis flaviuscula Salter MO132 [Crassulae] L. Dreyer 701 (Northern Cape, South Africa): EU436929; AJ582319; EU437216. Oxalis fourcadei Salter MO664 [Foveolatae] P.A. Bean & J.H. Vlok 2046 (Eastern Cape, South Africa): EU436923; EU437074; –. Oxalis furcillata Salter MO228 [Foveolatae] K. Oberlander 0022 (Northern Cape, South Africa): EU436952; AJ582320; EU437238. Oxalis giftbergensis Salter MO292 [Angustatae-Sessilifoliatae] L. Dreyer 796 (Western Cape, South Africa): EU436988; AJ582321; EU437274. Oxalis glabra Th. MO155 [Angustatae-Lineares] L. Mucina 170601/2 (Western Cape, South Africa): EU437000; AJ582322; EU437286. Oxalis goniorrhiza E. & Z. MO594 [Angustatae-Lineares] K. Oberlander 0157 (Western Cape, South Africa): EU437002; EU437114; EU437288. Oxalis gracilipes Schltr. MO76 [Angustatae-Lineares] L. Dreyer 645 (Western Cape, South Africa): EU436991; EU437108; EU437277. Oxalis gracilis Jacq. MO535 [Angustatae-Lineares] K. Oberlander 123 (Western Cape, South Africa): EU436976; EU437104; EU437262. Oxalis grammopetala Sond. MO571 [Foveolatae] K. Oberlander 0156 (Northern Cape, South Africa): EU436906; EU437062; EU437194. Oxalis grammophylla Salter MO101 [Angustatae-Pardales] L. Dreyer 670 1676 TAXON 60 (6) • December 2011: 1667–1677 Oberlander & al. • Phylogeny of southern African Oxalis Appendix. Continued.. (Western Cape, South Africa): EU437016; AJ582323; EU437302. Oxalis haedulipes Salter MO434 [Cernuae-Eu-Cernuae] L. Mucina 180903/9 (Northern Cape, South Africa): EU436886; EU437048; EU437175. Oxalis heterophylla D.C. MO518 [Oppositae-Bifurcatae] Bayer and Oberlander 0008 (Western Cape, South Africa): EU437018; EU437121; EU437304. Oxalis hirta L. MO77 [Angustatae-Sessilifoliatae] L. Dreyer 646 (Western Cape, South Africa): EU436971; AJ582324; EU437257. Oxalis hygrophila Kumwenda & Dreyer MO230 [‘Latifoliolatae’] Kumwenda and Dreyer 0001 (Western Cape, South Africa): EU437024; AJ582333; EU437310. Oxalis hypsophila Phil. MO456 [Alpinae Reiche] L. Dreyer 805 (Ex Hort.): EU436874; EU437040 ; EU437163. Oxalis imbricata E. & Z. MO345 [Oppositae-Subintegrae] K. Oberlander 0062 (Western Cape, South Africa): EU436896; EU437056; EU437184. Oxalis incarnata L. MO346 [Oppositae-Subintegrae] K. Oberlander 0063 (Western Cape, South Africa): EU436957; EU437094; EU437243. Oxalis inconspicua Salter MO138 [Latifoliolatae] L. Dreyer 707 (Northern Cape, South Africa): EU436944; EU437086; EU437230. Oxalis kamiesbergensis Salter MO592 [Angustatae-Lineares] (Ex Hort.): EU436970; EU437102; EU437256. Oxalis knuthiana Salter MO153 [Cernuae-Eu-Cernuae] L. Dreyer 721 (Northern Cape, South Africa): EU436884; AJ582325; EU437173. Oxalis lanata L. f. MO211 [Oppositae-Subintegrae] L. Dreyer 751 (Western Cape, South Africa): EU437027; EU437125; –. Oxalis lasiandra Zucc. MO455 [Ionoxalis] (Ex Hort.): EU436876; EU437041; EU437165. Oxalis lateriflora Jacq. MO82 [Cernuae-Lividae] L. Dreyer 651 (Western Cape, South Africa): EU436890; EU437051; –. Oxalis lateriflora Jacq. MO83 [Cernuae-Lividae] L. Dreyer 652 (Western Cape, South Africa): –; –; EU437178. Oxalis latifolia Knuth MO254 [Ionoxalis] K. Oberlander 0023 (Western Cape, South Africa): EU436875; AJ582326; EU437164. Oxalis leptocalyx Sond. MO386 [Angustatae-Lineares] K. Oberlander 0076 (Western Cape, South Africa): EU437001; EU437113; EU437287. Oxalis leptogramma Salter MO28 [Angustatae-Pardales] L. Dreyer 617 (Western Cape, South Africa): EU437014; AJ582327; EU437300. Oxalis lichenoides Salter MO281 [Foveolatae] L. Dreyer 785 (Western Cape, South Africa): EU436919; AJ582328; EU437207. Oxalis linearis Jacq. MO110 [Angustatae-Lineares] L. Dreyer 679 (Northern Cape, South Africa): EU436972; AJ582329; EU437258. Oxalis livida Jacq. MO361 [Cernuae-Lividae] K. Oberlander 0078 (Western Cape, South Africa): EU436889; EU437050; –. Oxalis louisae Salter MO139 [Crassulae] L. Dreyer 708 (Northern Cape, South Africa): EU436928; AJ582330; EU437215. Oxalis luteola Jacq. MO257 [OppositaeSubintegrae] K. Oberlander 0026 (Western Cape, South Africa): EU436904; AJ582331; EU437192. Oxalis massoniana Salter MO399 [Angustatae-Pardales] J. Holmes 0016 (Ex Hort.): EU437015; EU437120; EU437301. Oxalis meisneri Sond. MO468 [Angustatae-Sessilifoliatae] K. Oberlander 0088 (Western Cape, South Africa): EU436993; EU437110; EU437279. Oxalis melanosticta Sond. MO33 [Stictophyllae T.M. Salter] L. Dreyer 622 (Western Cape, South Africa): EU436901; AJ582332; EU437189. Oxalis minuta Thunb. MO163 [Sagittatae] L. Dreyer 727 (Western Cape, South Africa): EU437028; AJ582334; –. Oxalis monophylla L. f. MO9 [–] L. Dreyer and K. Oberlander 0003 (Western Cape, South Africa): EU436927; AJ582335; EU437214. Oxalis multicaulis E. & Z. MO390 [Angustatae-Sessilifoliatae] K. Oberlander 0080 (Western Cape, South Africa): EU436998; EU437112; EU437284. Oxalis namaquana Sond. MO144 [Crassulae] L. Dreyer 713 (Northern Cape, South Africa): EU436941; AJ582336; EU437228. Oxalis natans L. f. MO604 [Campanulatae] K. Oberlander 0160 (Western Cape, South Africa): EU437009; EU437117; EU437295. Oxalis nidulans E. & Z. MO212 [Sagittatae] L. Dreyer 752 (Western Cape, South Africa): EU437026; AJ582337; EU437312. Oxalis nortieri Salter MO503 [Foveolatae] K. Oberlander 0112 (Western Cape, South Africa): EU436914; EU437068; EU437202. Oxalis obliquifolia Steud. ex Rich. MO323 [Foveolatae] DUB 711 (Eastern Cape, South Africa): EU436915; EU437069; EU437203. Oxalis obtusa Jacq. MO122 [Oppositae-Subintegrae] L. Dreyer 691 (Western Cape, South Africa): EU436922; AJ582338; EU437210. Oxalis oculifera E.G.H Oliver MO295 [‘Latifoliolatae’] L. Dreyer 799 (Western Cape, South Africa): EU436953; AJ582339; EU437239. Oxalis oligophylla Salter MO293 [Angustatae-Lineares] L. Dreyer 797 (Western Cape, South Africa): EU437023; AJ582340; EU437309. Oxalis orbicularis Salter MO366 [OppositaeSubintegrae] K. Oberlander 0066 (Western Cape, South Africa): EU436899; EU437058; EU437187. Oxalis oreophila Salter MO97 [Angustatae-Lineares] L. Dreyer 666 (Western Cape, South Africa): EU436975; AJ582341; EU437261. Oxalis orthopoda Salter MO351 [Oppositae-Bifurcatae] K. Oberlander 0068 (Western Cape, South Africa): EU436956; EU437093; EU437242. Oxalis cf. pachyrrhiza Weddell MO460 [Carnosae Reiche] L. Dreyer 809 (Ex Hort.): EU436871; EU437037; EU437160. Oxalis pallens E. & Z. MO85 [Angustatae-Lineares] L. Dreyer 654 (Western Cape, South Africa): EU436997; AJ582342; EU437283. Oxalis palmifrons Salter MO403 [Angustatae-Multifoliolatae] J. Holmes 0019 (Ex Hort.): EU437021; EU437124; EU437307. Oxalis pendulifolia Salter MO350 [Oppositae-Subintegrae] K. Oberlander 0067 (Western Cape, South Africa): EU437019; EU437122; EU437305. Oxalis perdicaria (Molina) Bertero MO647 [Ionoxalis] Till Botrand 0002 (Ex Hort.): EU436878; EU437043; EU437167. Oxalis pes-caprae L. MO93 [Cernuae-Eu-Cernuae] L. Dreyer 662 (Western Cape, South Africa): EU436882; AJ582343; EU437171. Oxalis phloxidiflora Schltr. MO109 [Angustatae-Lineares] L. Dreyer 678 (Western Cape, South Africa): EU436984; EU437106; EU437270. Oxalis pillansiana Salter & Exell MO549 [Angustatae-Xanthotrichae] K. Oberlander 0135 (Western Cape, South Africa): EU436938; EU437082; EU437225. Oxalis pocockiae L. Bolus MO37 [Foveolatae] L. Dreyer 626 (Western Cape, South Africa): EU436911; AJ582344; EU437199. Oxalis polyphylla Jacq. MO47 [Angustatae-Lineares] C. Cilliers 0002 (Western Cape, South Africa): EU437010; AJ582345; EU437296. Oxalis porphyriosiphon Salter MO715 [Angustatae-Sessilifoliatae] K. Oberlander 0167 (Western Cape, South Africa): EU436987; EU437107; EU437273. Oxalis primuloides R. Knuth MO142 [Angustatae-Lineares] L. Dreyer 711 (Northern Cape, South Africa): EU436966; AJ582346; EU437252. Oxalis pseudo-cernua R. Knuth MO614 [Cernuae-Costatae R. Knuth] P. Craven 4781 (Karas Region, Namibia): EU436880; EU437045; EU437169. Oxalis psilopoda Turcz. MO347 [Oppositae-Subintegrae] K. Oberlander 0064 (Western Cape, South Africa): EU436891; EU437052; EU437179. Oxalis pulchella Jacq. MO23 [Foveolatae] L. Dreyer 612 (Western Cape, South Africa): EU436907; AJ582347; EU437195. Oxalis cf. pulvinata Salter MO576 [Crassulae] (Ex Hort.): EU436926; EU437076; EU437213. Oxalis punctata L. f. MO50 [Foveolatae] C. Cilliers 0005 (Western Cape, South Africa): EU436917; AJ582367; EU437205. Oxalis purpurascens Salter MO51 [Cernuae-Costatae] L. Dreyer 634 (Hardap Region, Namibia): EU436881; AJ582348; EU437170. Oxalis purpurea L. MO255 [Stictophyllae] K. Oberlander 0024 (Western Cape, South Africa): EU436902; AJ582349; EU437190. Oxalis pusilla Jacq. MO182 [Angustatae-Lineares] L. Dreyer 746 (Western Cape, South Africa): EU437004; AJ582350; EU437290. Oxalis reclinata Jacq. MO149 [Angustatae-Lineares] L. Dreyer 718 (Northern Cape, South Africa): EU436968; AJ582351; EU437254. Oxalis recticaulis Sond. MO792 [Angustatae-Sessilifoliatae] K. Oberlander 0170 (Western Cape, South Africa): EU437003; EU437115; EU437289. Oxalis salteri L. Bolus MO280 [Crassulae] L. Dreyer 784 (Western Cape, South Africa): EU436932; AJ582352; EU437219. Oxalis setosa E. Mey. ex Sond. MO325 [Foveolatae] DUB 770 (Eastern Cape, South Africa): EU436916; EU437070; EU437204. Oxalis smithiana E. & Z. MO322 [Oppositae-Bifurcatae] DUB 699 (Eastern Cape, South Africa): EU436964; EU437099; EU437250. Oxalis sonderiana (O. Kuntze) Salter MO125 [Latifoliolatae] L. Dreyer 694 (Northern Cape, South Africa): EU436945; EU437087; EU437231. Oxalis stellata E. & Z. MO465 [Cernuae-Stellatae] K. Oberlander 0086 (Western Cape, South Africa): EU436894; EU437054; EU437182. Oxalis stenopetala Salter MO106 [Angustatae-Lineares] L. Dreyer 675 (Western Cape, South Africa): EU436983; AJ582353; EU437269. Oxalis stenoptera Turcz. MO553 [Latifoliolatae] K. Oberlander 0138 (Western Cape, South Africa): EU436934; EU437079; EU437221. Oxalis stictocheila Salter MO185 [Angustatae-Lineares] L. Dreyer 749 (Western Cape, South Africa): EU436996; AJ582355; EU437282. Oxalis stricta L. MO669 [Corniculatae] J. Walter 6518 (Niederösterreich, Austria): EU436869; EU437035; EU437158. Oxalis strigosa Salter MO473 [Oppositae-Subintegrae] K. Oberlander 0093 (Western Cape, South Africa): EU436959; EU437096; EU437245. Oxalis suavis R. Knuth MO385 [Latifoliolatae] K. Oberlander 0075 (Western Cape, South Africa): EU436937; EU437081; EU437224. Oxalis suteroides Salter MO527 [AngustataeGlandulosae] K. Oberlander 0115 (Western Cape, South Africa): EU436963; EU437098; EU437249. Oxalis tenella Jacq. MO70 [Latifoliolatae] L. Dreyer 639 (Western Cape, South Africa): EU436978; AJ582356; EU437264. Oxalis tenuifolia Jacq. MO258 [Angustatae-Sessilifoliatae] K. Oberlander 0027 (Western Cape, South Africa): EU436995; EU437112; EU437281. Oxalis tenuipes Salter MO296 [Angustatae-Glandulosae] L. Dreyer 800: (Western Cape, South Africa) EU436986; AJ582358; EU437271. Oxalis tenuis Salter MO289 [Angustatae-Lineares] L. Dreyer 793 (Western Cape, South Africa): EU436989; AJ582359; EU437275. Oxalis tetraphylla Cav. MO392 [Ionoxalis] (Ex Hort.): EU436877; EU437042; EU437166. Oxalis tomentosa L. f. MO62 [Angustatae-Multifoliolatae] Te Roller 0006 (Western Cape, South Africa): EU437022; AJ582360; EU437308. Oxalis cf. tragopoda Salter MO490 [Cernuae-Stellatae] (Ex Hort.): EU436897; EU437057; EU437185. Oxalis truncatula Jacq. MO16 [Oppositae-Subintegrae] L. Dreyer 605 (Western Cape, South Africa): EU436960; AJ582361; EU437246. Oxalis uliginosa Schltr. MO394 [Campanulatae] (Ex Hort.): EU436933; EU437078; EU437220. Oxalis urbaniana Schltr. MO229 [Angustatae-Sessilifoliatae] Bytebier 2015 (Western Cape, South Africa): EU436999; AJ582362; EU437285. Oxalis valdiviensis Barnéoud MO646 [Alpinae] Till Botrand 0001 (Ex Hort.): EU436873; EU437039; EU437162. Oxalis versicolor L. MO307 [Angustatae-Lineares] K. Oberlander 0034 (Western Cape, South Africa): EU437006; AJ582363; EU437292. Oxalis virginea Jacq. MO580 [Oppositae-Subintegrae] (Ex Hort.): EU436955; EU437092; EU437241. Oxalis viscosa E. Mey. ex Sond. MO73 [Angustatae-Sessilifoliatae] L. Dreyer 642 (Western Cape, South Africa): EU436949; AJ582364; EU437235. Oxalis vulcanicola (Donn.-Sm.) Lourt. MO702 [Lotoideae A. Lourteig] (Ex Hort.): EU436872; EU437038; EU437161. Oxalis xantha Salter MO102 [Angustatae-Lineares] L. Dreyer 671 (Western Cape, South Africa): EU436980; AJ582365; EU437266. Oxalis zeekoevleyensis R. Knuth MO197 [OppositaeSubintegrae] K. Oberlander 0012 (Western Cape, South Africa): EU436900; EU437059; EU437188. Oxalis zeyheri Sond. MO590 [Angustatae-Multifoliolatae] (Ex Hort.): EU436962; EU437097; EU437248. Sarcotheca laxa Knuth SARLAX [–] (Ex Hort.): EU436859; AJ582366; EU437148. Sarcotheca monophylla (Planch. ex Hk. f.) Hallier. f. SARMON [–] (Ex Hort.): EU436860; EU437030; EU437149. 1677