Plant Syst Evol DOI 10.1007/s00606-014-1063-3 ORIGINAL ARTICLE Sibbaldia: a molecular phylogenetic study of a remarkably polyphyletic genus in Rosaceae Torsten Eriksson • Magnus Lundberg • Mats Töpel • Pia Östensson • Jenny E. E. Smedmark Received: 6 January 2014 / Accepted: 14 April 2014  Springer-Verlag Wien 2014 Abstract Using DNA sequence data from nuclear ribosomal ITS in combination with plastid trnLF spacer and trnL intron data, we show that Sibbaldia is a polyphyletic assemblage. It falls into five separate clades of Potentilleae, three within Fragariinae and two within Potentilla (Potentillinae sensu Soják). To a large extent, our results are congruent with Soják’s findings based on morphological characters such as anther structure. Four of the Sibbaldia species included in this study remain in Sibbaldia, while S. adpressa is classified in Sibbaldianthe, S. perpusilloides is considered to represent a new genus in Fragariinae, Chamaecallis Smedmark, S. micropetala is nested within the Potentilla anserina clade, and four species belong to a basal clade of Potentilla. The phylogenetic affinity of Sibbaldiopsis is still unsettled, but one of the three species that have been classified in the genus is found to belong inside Sibbaldia, and it should be named Sibbaldia retusa (O.F. T. Eriksson Bergius Foundation, Royal Swedish Academy of Sciences, Box 50017, 10405 Stockholm, Sweden Present Address: T. Eriksson (&) The Natural History Collections, University of Bergen, P.O. Box 7800, 5020 Bergen, Norway e-mail: torsten.eriksson@adm.uib.no M. Lundberg
P. Östensson Department of Botany, Stockholm University, 10691 Stockholm, Sweden M. Töpel Department of Biological and Environmental Sciences, Gothenburg University, Box 461, 40530 Göteborg, Sweden J. E. E. Smedmark University Museum of Bergen, The Natural History Collections, University of Bergen, P.O. Box 7800, 5020 Bergen, Norway Müller) T. Erikss. Further study will show whether the remaining two species, Potentilla cuneifolia and P. miyabei, are more closely related to Sibbaldia, Sibbaldianthe, or if they make up a distinct lineage separate from either of these. Keywords Bayesian inference
Phylogeny
Polyphyly
Potentilla
Sibbaldia
Taxonomy Introduction Sibbaldia constitutes a rather small group of species in the rose family (Rosaceae), found mostly in mountain regions of the Old World, particularly in Asia. The most widely known and widely distributed species, Sibbaldia procumbens, is circumboreal. Although more than 40 names at the species level have been published (see ‘‘List of taxa’’), recent treatments list ca. 10–15 species (Dixit and Panigrahi 1981; Rajput et al. 1997; Dikshit and Panigrahi 1998; Li et al. 2003), or even as little as two (Soják 2008). Sibbaldia has been classified in Potentilleae within Rosoideae, close to Potentilla, and recent phylogenetic analyses using DNA sequence data have confirmed this position (Eriksson et al. 1998, 2003; Potter et al. 2007; Lundberg et al. 2009; Dobeš and Paule 2010). The species that have been referred to Sibbaldia are mostly small-flowered perennials with yellow, white or purple flowers, including cushion plants of high altitudes in the Himalayas. The leaves are typically trifoliate or pinnate, although there is considerable variation among the species. The consistent diagnosing characteristic is ‘‘fewer’’ stamens than Potentilla (i.e. ten or less). Linnaeus described the first species of Sibbaldia and based his system mainly on pistil and stamen number (and position). There was no problem at the time in distinguishing his two species of Sibbaldia from Potentilla by their fewer stamens (Linnaeus 1753, 1754). 123 T. Eriksson et al. However, as additional species referred to these and other close genera were discovered and described, this simple and technical distinction was blurred. Still, the distinction between these genera was commonly maintained based on the artificial ‘‘fewer stamens’’ character (e.g. Nestler 1816; Seringe 1825; Bunge 1829; Endlicher 1840; Focke 1888; Rydberg 1898, 1908; Wolf 1908; Muravjova 1936; Chatterjee 1938; Juzepchuk 1941; Hutchinson 1964; Dixit and Panigrahi 1981; Rajput et al. 1997; Dikshit and Panigrahi 1998). Some authors raised doubts as to the rationale for keeping Sibbaldia separate (Bentham 1858; Wolf 1908), and in a few cases, Sibbaldia has been joined with Potentilla (e.g. Spenner 1829; Hooker 1865, 1878). When species referred to Sibbaldia in recent revisions and floras (Dixit and Panigrahi 1981; Rajput et al. 1997; Dikshit and Panigrahi 1998; Li et al. 2003) are compared even superficially, one is immediately struck by the variation in general characteristics such as leaf form and division, stamen number, stamen placement, anther form, presence or absence of a floral disc and flower colour. In addition, there are indications that the lower stamen numbers in Sibbaldia species may have been reduced in parallel, mainly because the position of the stamens in relation to petals and sepals is inconsistent. A few species traditionally classified in Sibbaldia have from time to time been moved into separate genera (Endlicher 1840; Juzepchuk 1941), but despite the morphological heterogeneity of the genus, it is not until recently that its circumscription has been seriously questioned (Soják 2008). There have been indications based on DNA sequence data that the genus may not be monophyletic (Lundberg et al. 2009; Dobeš and Paule 2010), but as previous phylogenetic analyses have only included one or a few Sibbaldia species, we wanted to test the monophyly of the genus in the light of its morphological variation. In a recent study, Soják (2008) investigated two morphological characters for all of Potentilleae. One of his conclusions was that these two characters, the number and structure of thecae in anthers and the position of styles on achenes, divided the Sibbaldia species, and he suggested that some of them should be included in Potentilla, but not all. A secondary goal for this study then was to test the classification suggested by Soják (2008) using a molecular phylogeny. While discussing the species in this paper, we will initially use the names in the most recent floras and revisions, i.e. as classified in Sibbaldia. Materials and methods Taxon sampling The intention was to sample as many species of Sibbaldia in a wide sense as possible, along with representatives from 123 all genera in Potentilleae. For rooting purposes, three outgroup taxa, Rosa majalis, Sanguisorba officinalis and Aremonia agrimonioides, were selected to represent the two clades most close to Potentilleae (Eriksson et al. 2003). All sampled taxa are listed in Table 1. For the in-group, we selected two or more species of each genus and attempted to use type species. When possible, we selected species that have been resolved in early splits in their respective ‘‘genus clade’’ in previous analyses. We also favoured diploids over polyploids where possible. From Potentilla, we sampled six species from subgenus Fragariastrum (Heister ex Fabr.) Reichenb. because their leaves and flower colours were reminiscent of some Sibbaldia species and an additional seven species representing early splits in the Potentilla clade (Töpel et al. 2011). In total, we selected 43 in-group species plus three outgroup species. Two Sibbaldia species were only available to us as type collections, and for this reason, we were unable to extract DNA from S. trullifolia and S. tenuis. Sequencing Two DNA regions were sequenced. Their utility has been assessed for Potentilleae in previous studies (Eriksson et al. 1998, 2003; Gehrke et al. 2008; Lundberg et al. 2009), and they were well known to be relatively easy to amplify, even from old herbarium collections. We used nuclear ribosomal internal transcribed spacers (ITS) and one region of plastid DNA comprising the spacer between the trnL and trnF genes, as well as an intron in the trnL gene (trnLF). The ribosomal DNA is known to be affected by concerted evolution, but the level of homogenisation differs between groups. We did not perform any cloning to confirm that only a single repeat type was present. However, we sequenced different PCR products in tests, both using the same DNA extraction and different DNA extractions. In rare cases, we noted double signal for single bases, but low-quality sequences caused by low-quality DNA after extracting old herbarium specimens were more of a problem. For this reason, we had to try different accessions in several species before getting good enough sequences. When comparing such partial reads, we found no evidence of different repeat types and the different accessions gave the same sequence for the parts we could read. We used ca. 0.05 g of silica-dried or herbarium plant material that was homogenised in CTAB using a minibeadbeater and silica beads. Extraction of total genomic DNA followed a downscaled version of the protocol described by Doyle and Doyle (1990). Polymerase chain reactions (PCR) using Taq DNA polymerase (Roche Applied Science, Germany) and amplification conditions were as described in Lundberg et al. (2009). We used the ITS-I (Urbatsch et al. 2000) and ITS4 (White et al. 1990) Sibbaldia: a molecular phylogenetic study Table 1 List of species and sequences used, along with voucher specimens, sequence database accession numbers and literature references to original publication of sequence Species Vouchers ITS accessions trnLF accessions References Alchemilla alpina R. Eriksson s.n. (GH, S) U90816, U90817 AJ512217 Eriksson et al. (1998, 2003) Alchemilla cryptantha Eriksson 914 (S) FJ356153 FJ422306 Lundberg et al. (2009) Aremonia agrimonioides Karlsson 94076 (LD) U90799 AJ512230, AJ512231 Eriksson et al. (1998, 2003) Chamaerhodos erecta Lackschewitz 11453 (GH), Norlindh and Ahti 10161A (S) U90794 AJ512219 Eriksson et al. (1998, 2003) Chamaerhodos mongolica E. Rosenius 1028 (S) FJ356155 FJ422285 Lundberg et al. (2009) Comarum palustre Comarum salesovianum Eriksson 659 (GH, S) Eriksson and Vretblad TE751 (S) AJ511777 AJ511779 AJ512237 AJ512228 Eriksson et al. (2003) Eriksson et al. (2003) Dasiphora fruticosa Karlsson 94074 (LD), Eriksson 806 (S) U90808, U90809 AJ512233 Eriksson et al. (1998, 2003) Drymocallis agrimonioides Laferrière 2357 (A), Lundberg 15 (S) U90787 FJ422289 Eriksson et al. (1998); Lundberg et al. (2009) Drymocallis rupestris Lundberg 6 (S) FJ356163 FJ422321 Lundberg et al. (2009) Fragaria vesca Eriksson and Smedmark 43 (S) AJ511771 AJ512232 Eriksson et al. (2003) Fragaria viridis Lundberg 16 (S) FJ356166 FJ422295 Lundberg et al. (2009) Potaninia mongolica Potentilla alba Norlindh and Ahti 10384 (S) Töpel MA122 (GB) AM286742 FN430774 AM286743 FN556397 Potter et al. (2007) Töpel et al. (2011) Potentilla alchemilloides Anderberg and Anderberg 26 (S) FJ356168 FJ422297 Lundberg et al. (2009) Potentilla anserina Eriksson and Smedmark 44 (S) AJ511773 AJ512238 Eriksson et al. (2003) Potentilla biflora Viereck 5042 (S) KJ396292/TE KJ396304/TE This study Potentilla cuneifolia Lundberg 39 (S) FJ356169 FJ422298 Lundberg et al. (2009) Potentilla curviseta Stewart Aug. 16, 1922 (S) KJ396293/ML KJ396305/ML This study Potentilla dickinsii Sun et al. s.n. (A), Crompton et al. 139 (E) U90785 AJ512243 Eriksson et al. (1998, 2003) Potentilla divina Soják 17 VII 1984 (S) KJ396294/ML KJ396306/ML This study Potentilla fragarioides Eriksson s.n. (S) FN555610 AJ512226 Töpel et al. (2011); Eriksson et al. (2003) Potentilla lignosa Töpel MA132 (GB) FJ356171 FJ422299 Lundberg et al. (2009) Potentilla micrantha Karlsson 94075 (LD), Eriksson and Smedmark 42 U90812, U90813 AJ512227 Eriksson et al. (1998, 2003) Potentilla nitida GBG Bot Gard. no voucher KJ396295/TE KJ396307/TE This study Potentilla peduncularis Eriksson and Vretblad TE758 (S) AJ511778 AJ512239 Eriksson et al. (2003) Potentilla recta No voucher FN430784 FN556419 Töpel et al. (2011) Potentilla reptans Eriksson 650 (GH, S), Eriksson 822 (S) U90784 AJ512241 Eriksson et al. (1998, 2003) Potentilla stenophylla Eriksson and Vretblad TE763 (S) AJ511780 AJ512240 Eriksson et al. (2003) Potentilla stipularis Petrovsky 28.VI.1975 (S) KJ396296/ML KJ396308/ML This study Potentilla valderia Rosa majalis Lundberg 32 (S) Eriksson 641 (GH, S) KJ396297/ML U90801 KJ396309/ML AJ512229 This study Eriksson et al. (1998, 2003) Sanguisorba officinalis Helfgott Ben-5 (TEX) AF183533, AF183556 AJ416465 Helfgott et al. (2000) Sibbaldia adpressa (Sibbaldianthe adpressa) Gusev 391 (S) FJ356176 FJ422304 Lundberg et al. (2009) Sibbaldia cuneata Binns 5 (E) FJ356173 FJ422301 Lundberg et al. (2009) Sibbaldia micropetala (Potentilla micropetala) Ludlow et al. 16237 (S) KJ396298/PÖ KJ396310/PÖ/ ML This study Sibbaldia parviflora Lundberg 4 (S) FJ356174 FJ422302 Lundberg et al. (2009) Sibbaldia pentaphylla (Potentilla clandestina) Smith 11205 (S) KJ396299/PÖ KJ396311/PÖ This study 123 T. Eriksson et al. Table 1 continued Species Vouchers ITS accessions trnLF accessions References Sibbaldia perpusilloides Forrest 16456 (S), Forrest 26836 (S) KJ396300/ML KJ396312/ML This study Sibbaldia procumbens Aronsson s.n. (S), Eriksson 698 (S) U90820, U90821 AJ512235 Eriksson et al. (1998, 2003) Sibbaldia purpurea (Potentilla purpurea) Ludlow et al. 20749 (S) KJ396301/PÖ KJ396313/PÖ This study Sibbaldia semiglabra Klackenberg 820621-11 (S) FJ356175 FJ422303 Lundberg et al. (2009) Sibbaldia sikkimensis (Potentilla sikkimensis) Rock 16987 (S) KJ396302/PÖ KJ396314/PÖ This study Sibbaldia tetrandra (Potentilla tetrandra) Ahlgren s.n. (S) KJ396303/PÖ/ ML KJ396315/PÖ This study Sibbaldianthe bifurca Karis 412 (S), Eriksson 811 (S) U90786 AJ512224 Eriksson et al. (1998, 2003) Sibbaldiopsis tridentata (Sibbaldia retusa) Hill 17146 (A), Eriksson and Smedmark 40 (S) U90791 AJ512236 Eriksson et al. (1998, 2003) (Chamaecallis perpusilloides) New species names are in parentheses. In cases where two voucher specimens or references are listed, the first concerns ITS and the second trnLF. Herbarium abbreviations follow Holmgren et al. (1990) primers to amplify the ITS region. Some problematical taxa were necessary to amplify in two parts using also the ‘‘internal’’ primers ITS2 (White et al. 1990) and ITS3B (Baum et al. 1994). For amplification of the trnLF region of plastid DNA, the trnL(c) and trnL(f) primers were used (Taberlet et al. 1991). Problematic taxa were amplified in two parts with the additional primers trnL(d) and trnL(e). Primers used were obtained from MWG Biotech AG, Germany. For taxa that did not amplify with Taq polymerase, the high-fidelity and proofreading Phusion DNA polymerase (Finnzymes, Finland) was used. PCR set-up and amplification conditions using Phusion were the same as described by Lundberg et al. (2009). Amplification products were cleaned using MultiScreen PCR plates for DNA clean-up (Millipore, Ireland). Sequencing reactions, conditions for producing extensions, precipitation and visualisation were also the same as described in Lundberg et al. (2009). BigDye version 3.1 (Applied Biosystems) was used for sequencing reactions, and the primers used for amplification were used for sequencing as well. Sequences were assembled and proofread using the Staden package (Staden 1996). All DNA sequences have been deposited in EMBL. Alignment of DNA sequences We used Prank (Löytynoja and Goldman 2005, 2008) with the -F option for sequence alignment of the two data sets, with subsequent manual adjustments in SeAl (Rambaut 2002). Manual adjustments were particularly important for the plastid data set. Data matrices are available from the first author. 123 Phylogenetic estimation We tested molecular models for each of the two data sets by using mrAIC (Nylander 2004) which uses PHYML (Guindon and Gascuel 2003) as a maximum-likelihood back end. Phylogenies were estimated using Bayesian inference (Yang and Rannala 1997) in MrBayes 3.1.2 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). Each of the data sets as well as the combined was run for 10 million generations, sampling every 100 generations. Settings were nruns = 2 and nchains = 4, and we used the models suggested by the AICc criterion (Akaike Information Criterion with small sample correction; Sugiura 1978; Hurvich and Tsai 1989), namely the general time-reversible model (GTR; Tavaré 1986) with gamma distribution of rates (GTR ? G; Yang 1993). In the combined analysis, each data set was allowed to estimate its own model parameters, i.e. the model parameters (except the topology) were unlinked. The analyses started with default priors and a randomly generated default topology. The log files from the MrBayes runs and the output from the sump command were studied for information on analysis convergence and if the chains could be assumed to have reached stationarity. We also checked chain stationarity using the plot output from the web interface of AWTY (Nylander et al. 2008). Fifty percent of burnin (50,000 trees) was removed from each analysis prior to computing majority rule consensus trees, by using the sumt command. Hence, the tree in Fig. 1 is based on 100,002 trees. Sibbaldia: a molecular phylogenetic study Results In addition to data that have been published elsewhere (Eriksson et al. 2003; Lundberg et al. 2009; Töpel et al. 2011), this study includes new ITS and trnLF sequences from six additional Sibbaldia species and seven additional Potentilla species. The matrices consisted of 46 taxa, with 675 and 1,425 aligned characters in the ITS and trnLF data sets, respectively. The Prank alignment program performed reasonably well on the ITS data, but not on the trnLF data, and that alignment had to be extensively corrected, in particular in regions with large indels. The GTR ? G model was preferred for both data partitions. For the combined data set, the amount of indels and unsequenced parts was 28 % of a total of 96,600 aligned positions; the scored uncertainties (bad signal, high noise level or double peaks) were 0.3 %; the G ? C content is 43 %. The phylogenetic analyses in MrBayes performed well for all data sets. The chain swap acceptance rates were 22–36 % (zero for one of the chains in the separate trnLF analysis), the model parameter acceptance rates were 11–47 %, and the average standard deviation of split frequencies did not exceed 0.003 at the end of the runs. The potential scale reduction factors of the parameters were 1.000 (1.001 for four parameters in the combined analysis). The AWTY convergence plots indicated that the chains had reached stationarity. The separate phylogenies (not shown) both divide the Potentilleae into two clades: the Fragariinae and Potentilla (Potentillinae sensu Soják 2008). The two markers show a different resolving power. ITS is less resolved than trnLF, in particular at the base of the Potentilla clade. Comparing the analyses of the separate markers reveals four more or less well-supported differences, two in Fragariinae and two in the Potentilla clade. In Fragariinae, the conflicts are the same as two conflicts noted by Lundberg et al. (2009), namely Potentilla cuneifolia shifting sister group relationships and in the other case, Drymocallis, Dasiphora and Chamaerhodos showing different interrelationships. In the Potentilla clade, one conflict results in Potentilla dickinsii and P. reptans exchanging sister group relations, and the other concerns the entire Anserina clade (Potentilla sect. Anserina sensu Soják 2007, including P. lignosa). The latter clade shifts between being sister to Potentilla (trnLF) and being sister to Fragariinae (ITS, with slightly lower support). Except for these conflicts, the trees are congruent, and since none of the discrepancies affect the position of any Sibbaldia species or the monophyly of any clade classified to genus (except for the Anserina clade which might be considered a genus of it own), we will mainly report results from the combined analysis here. The combined analysis (Fig. 1) showed the same main tree structure as previous analyses comprising Potentilleae (Eriksson et al. 2003; Potter et al. 2007; Lundberg et al. 2009). In the Fragariinae, a basal split divides the clade into one subclade comprising Fragaria as sister to a clade of Dasiphora, Potaninia, Drymocallis and Chamaerhodos. The other Fragariinae clade consists of Alchemilla in the wide sense (cf. Gehrke et al. 2008) with two species of Comarum (or Comarum and Farinopsis) as unresolved sisters, and the Sibbaldianthe clade weakly supported as sister to one Sibbaldia clade. (Details discussed below.) For Potentilla (sister clade to Fragariinae), the Anserina clade plus Potentilla (Tylosperma) lignosa is sister to the rest of Potentilla. The latter clade would contain almost all species of Potentilla, but the sample here is focused on a few clades of Potentilla that represent early splits (‘‘basal’’ clades). We can identify three well-supported clades here, but the support for relationships among them is weak: the Potentilla alba clade (subgen. Fragariastrum with two yellow-flowered species as sister clade), one Sibbaldia clade and a clade containing most Potentilla species including the type species P. reptans. Three of the discrepancies between the plastid and nuclear trees mentioned above were resolved in favour of the plastid tree resolution in the combined analysis, all with the same or slightly increased level of support. The clade containing Dasiphora and others (C in Fig. 1) was well supported (pp 1.0) as sister to Fragaria like in both of the separate trees, and the relationships within clade C were resolved as in the ITS phylogeny with slightly increased levels of support in the combined analysis. The analysed Sibbaldia species did not form a clade. Rather, they fell into five separate clades, two of which are relatively larger, namely the Sibbaldia sensu stricto clade in Fragariinae (A in Fig. 1, four accessions), and one clade of Himalayan species which is found within the Potentilla clade (D in Fig. 1, four accessions). The Himalayan clade is strongly supported to be within Potentilla, but very weakly supported (pp 0.59) as sister to a clade containing Potentilla subgen. Fragariastrum. In addition, three species in our Sibbaldia sample are resolved in three separate clades. The first two in Fragariinae and the third in Potentilla: Sibbaldia adpressa is sister to Sibbaldianthe bifurca (clade B in Fig. 1), Sibbaldia perpusilloides is for all practical purposes unresolved in a trichotomy at the base of the Dasiphora clade (C in Fig. 1), and Sibbaldia micropetala is found within Potentilla sect. Anserina (E in Fig. 1). The position of all five Sibbaldia instances is very well supported and not affected by any incongruencies between plastid and nuclear ribosomal data. Discussion Our results agree with previously published phylogenies that included more than just a few species from Rosoideae 123 T. Eriksson et al. 123 Sibbaldia: a molecular phylogenetic study b Fig. 1 Majority-rule consensus tree (50 %) based on 100,002 trees sampled after burnin in the Bayesian inference analysis in MrBayes 3.1.2. Estimated posterior probabilities for clades are noted at each node. Five clades including species that are or have been classified in Sibbaldia are indicated with grey and the capital letters A–E. Clade A corresponds to Sibbaldia in the strict sense. Species names in bold are current names. The circumscription of the Fragariinae and Potentilla (Potentillinae sensu Soják 2008) clades is indicated to the right. Branches are drawn proportional to the mean estimated branch lengths, with a scale bar at the base (Eriksson et al. 1998, 2003; Potter et al. 2007; Lundberg et al. 2009; Dobeš and Paule 2010). The Potentilleae clade is split into two main clades (Fig. 1). One is Fragariinae, containing a number of small or large genera: Alchemilla (incl. Aphanes and Lachemilla), Chamaerhodos, Comarum (possibly including Farinopsis), Dasiphora, Drymocallis, Fragaria, Potaninia, Sibbaldia, Sibbaldianthe, Sibbaldiopsis, as well as a few former Potentilla species which have only recently been classified in other genera (Lundberg et al. 2009; Paule and Soják 2009; Soják 2008). The other clade is Potentilla, which contains some subclades that often have been treated as genera, such as Duchesnea and Horkelia. Potentilla has also been classified as Potentillinae (Soják 2008). There is also morphological support for this division of Potentilleae into two clades. Soják (2008) made a detailed description of characters that support this split (primarily anther structure). He had previously published a brief account of his finding, but unfortunately in a publication which was not widespread enough that the anther structure character was acknowledged (Soják 1989). Based on the results presented in this study, anthers with thecae that are confluent at the apex (and in some cases also at the base) are a synapomorphy for Fragariinae, and this character is also congruent with the results from previous phylogenetic studies based on plastid spacer and intron data, as well as a combination of seven DNA markers from both the plastid and nuclear genomes (Potter et al. 2007; Lundberg et al. 2009; Dobeš and Paule 2010). All species in Potentilleae are characterised by having styles that are not apically placed on the achene. In the phylogenetic hypothesis presented here (Fig. 1), the Anserina clade, which has styles with a lateral insertion, is the sister group of Potentilla s.str. with subterminal styles, rather than with Fragariinae with lateral-to-basal styles. The results from the molecular analyses are, however, not unambiguous. Some data (notably nuclear ribosomal ITS) favour that the Anserina clade instead is sister to Fragariinae (Töpel et al. 2011). The style insertion character may be regarded as supporting the latter position, as well as for classifying the species in the Anserina clade in genera separate from Potentilla, as has been suggested previously (Hill 1756; Rydberg 1898; Soják 2010). Should this latter topology stabilise as more data are added, this would imply that lateral-to-basal styles are a synapomorphy for Fragariinae plus the Anserina clade. Else, a lateral-to-basal insertion of the style is the plesiomorphic condition in Potentilleae, and subterminal styles have evolved in Potentilla s.str. At present, in the light of conflicting data, we consider it premature to follow Soják’s (2010) classification and prefer to leave the species in the Anserina clade in Potentilla for the time being. Apart from the conflicting topologies based on different genetic markers, the relationships within the Anserina clade have not been properly addressed. For example, Potentilla micropetala is nested inside the Anserina clade in our tree (Fig. 1), but it is classified separately in Piletophyllym by Soják (2010). Although the morphological characters discussed by Soják (2008) are congruent with the estimated phylogeny, his preferred subtribal classification was not. He divided Potentilleae into four subtribes: Potentillinae, Fragariinae, Chamaerhodotinae and Alchemillinae. The first two were based on his anther structure and style position characters, but the latter two were based on special cases of the anther character, Chamaerhodos with a single lateral theca and Alchemilla with a single ‘‘ventral’’ theca. The molecular phylogeny shows Chamaerhodotinae and Alchemillinae to be in-groups in Fragariinae, which is not unexpected (Eriksson et al. 1998, 2003; Dobeš and Paule 2010) since they both have lateral-to-basal styles. If Alchemillinae and Chamaerhodotinae would just be treated as clade names with the understanding that they are subclades of Fragariinae, we think that the names might not necessarily cause confusion, but in a more formal (ICBN) sense, the use of Chamaerhodotinae and Alchemillinae as subtribes alongside Potentillinae and Fragariinae would be unfortunate. It would result, for example, in the closest relatives of Alchemillinae being nested well within Fragariinae, and some (but not all) members of Fragariinae would have their closest relatives in Alchemillinae. We suggest that it might be best not to use the subtribes Alchemillinae and Chamaerhodotinae in the sense of Soják (2008). The phylogeny presented here is based on two molecular DNA sequence regions, the nuclear ribosomal internal transcribed spacers (ITS) and a plastid DNA region that includes the spacer between the trnL and trnF genes, and an intron in the trnL gene. Our results are fully congruent with to those of Lundberg et al. (2009) although in that study, two additional regions were added (a part of nuclear ribosomal ETS and the plastid trnS–G spacer region). Much of the Sibbaldia collections are fairly old, and in many cases, we had access only to herbarium material. For this reason, we decided to only use two regions even though some species were available also for the two extra regions. For much of the important species, it was simply too difficult to get sequences for any but the most easily 123 T. Eriksson et al. amplified regions. When considering the resolution and support in the combined tree, it is clear that for the purpose of studying the monophyly of Sibbaldia, it was sufficient to use only these two regions. Lundberg et al. (2009) specifically studied incongruencies between nuclear and plastid phylogenies in this group in order to find potential occurrences of reticulate evolution, such as allopolyploidy. They found a number of such cases of incongruence, and some of those were also seen in our trees when data were analysed and compared separately (not shown). However, species classified as Sibbaldia were not involved in these incongruencies, and we chose to combine the data sets into a joint analysis, keeping incongruencies from the separate tree in mind when studying the results. We believe that for the purpose of studying the monophyly (or non-monophyly) of Sibbaldia sensu lato, the combined tree may be trusted because it is not in conflict with the separate trees. Below, we discuss the details of the five Sibbaldia clades in our combined tree (Fig. 1), along with the taxonomic and nomenclatural implications. Clade A in Fragariinae: Sibbaldia sensu stricto The Fragariinae clade A includes Sibbaldia procumbens, which may be considered the type species of Sibbaldia. Smith (1811) published a previously unpublished account of Linnaeus’ trip to Lapland. It contains a brief description of Sibbaldia procumbens as ‘‘Jussiea’’, and Smith comments in a footnote ‘‘… in this and many following instances, the original names in the manuscript are here retained, as a matter of curiosity to the learned botanist, who will be interested in seeing to whom Linnaeus extemporaneously dedicated his new genera as they occurred…’’. Although Farr et al. (1979) consider this a split of Sibbaldia, leaving Sibbaldia erecta as type for Sibbaldia, it is clear that neither Linnaeus (since he had clearly reconsidered the name after writing this manuscript and published the genus as ‘‘Sibbaldia’’ including the species mentioned) nor Smith (since he explicitly considered the name preliminary and published only as a curiosity) accept the genus Jussiea when it was published. Under article 34 of the International Code of Botanical Nomenclature, Jussiea was therefore not validly published. When describing Chamaerhodos, Bunge (1829) included three species from Sibbaldia, namely C. erecta, C. altaica and C. grandiflora. Of the original species described in Sibbaldia by Linnaeus (1753), only S. procumbens remained then, and Rydberg (1908) explicitly noted S. procumbens as type species. The clade A therefore is the clade that should be used to circumscribe the genus Sibbaldia, and for this reason, we call this clade ‘‘Sibbaldia sensu stricto’’. 123 Soják (2008) was of the opinion that Sibbaldia only comprised two species: S. procumbens in a wide sense (including S. cuneata, S. parviflora and S. semiglabra) and S. trullifolia. Unfortunately, we were unable to sample S. trullifolia for the present study. Our phylogeny shows some supported structure within this clade, and furthermore, Sibbaldiopsis tridentata is nested inside it as sister to procumbens plus semiglabra. It seems at least that lumping all of these four species (S. cuneata, S. parviflora, S. procumbens and S. semiglabra) into a single one may be overly drastic, especially in the light of the position of Sibbaldiopsis. In order to solve the question of species delimitation in this circumboreal clade, it may be necessary to sample a wider range of populations. Rydberg (1908) cited ‘‘?Potentilla retusa O. F. Müller’’ as a questionable synonym for Sibbaldiopsis tridentata. Potentilla retusa was described in Flora Danica and based on a collection from Greenland, and for some reason, the flower colour on the plate is yellow. It is clear from the type material (in herbarium C), however, that Potentilla retusa is indeed white-flowered and conspecific with Sibbaldiopsis tridentata, and hence, the correct epithet at specific level should be ‘‘retusa’’ (cf. Taxonomy section). Sibbaldiopsis tridentata is enclosed within the Sibbaldia sensu stricto clade in our tree, as well as in the combined tree of Lundberg et al. (2009) which was based on four markers. The signal for this result seems to come mainly from the plastid data because in the nuclear ribosomal tree, Sibbaldiopsis tridentata is unresolved in a trichotomy at the base of clade A in Fig. 1 (cf. Fig. 1a in Lundberg et al. 2009). We choose to follow Paule and Soják (2009) and classify the species in Sibbaldia, in order for Sibbaldia to be monophyletic. Thus, the name of the species should be Sibbaldia retusa. However, even though Sibbaldia retusa is consistently resolved in the Sibbaldia sensu stricto clade, some uncertainties remain. This species differs in some characters from the other species in the Sibbaldia sensu stricto clade, and these characters have a long-standing use in distinguishing genera in the Potentilleae. The flowers are white and have fairly long petals rather than small yellow petals. The number of stamens is 20 instead of five, and the achenes are hairy rather than glabrous. The number of stamens and the hairy achenes can safely be considered plesiomorphies and the white flowers possibly an autapomorphy in this clade, so these characteristics can not support any group. However, as Sibbaldia retusa joins two of the Sibbaldia sensu stricto species, homoplasy is implied in the characters. Furthermore, Sibbaldia retusa is a tetraploid (Dalgaard 1989), and although our results so far do not show that it is an allopolyploid, it is still a possibility. In that case, part of its ancestry may be outside Sibbaldia, and under such a scenario, it might have been preferable to keep it in Sibbaldiopsis. Sibbaldia: a molecular phylogenetic study Wolf (1908) classified Sibbaldia retusa (as Potentilla tridentata) in a group (‘‘grex’’) along with P. cuneifolia (as P. ambigua) and P. miyabei (the latter included in Lundberg et al. 2009, but not here). The classification of Wolf may have prompted the idea that these three species were close relatives (Robertson 1974), and subsequently, the two latter species were first recombined into Sibbaldiopsis (Soják 2004, 2008) and later into Sibbaldia (Paule and Soják 2009). There is no indication that these three species form a clade, and to join them in Sibbaldiopsis creates a non-monophyletic genus. Indeed, although P. cuneifolia and P. miyabei are similar to each other, neither is similar to Sibbaldia retusa, and it may be more reasonable from a morphological standpoint to classify them in different genera. Potentilla cuneifolia and P. miyabei display a complex relationship to each other and to the Sibbaldia sensu stricto clade on the one hand and to the Sibbaldianthe clade on the other hand (Lundberg et al., 2009). Potentilla cuneifolia, at least, is a polyploid (Ikeda 1989), and due to their potentially allopolyploid origin, it is currently not clear what the exact relationships of these two species are. Therefore, we think that classifying P. cuneifolia and P. miyabei in Sibbaldia (Paule and Soják 2009) is undesirable, unless the name Sibbaldia is applied to a much more inclusive, still unresolved, clade. Thus, we choose to retain their Potentilla names here. Clade B in Fragariinae: the Sibbaldianthe clade In all of our analyses as well as in preliminary studies (Kurtto and Eriksson 2003; Lundberg et al. 2009), Sibbaldia adpressa is well supported as sister to Potentilla bifurca (sometimes classified as Schistophyllidium or Sibbaldianthe), and this clade (B in Fig. 1) is sister to the Sibbaldia sensu stricto clade, plus Potentilla cuneifolia, with low support. Potentilla bifurca is similar to S. adpressa, although somewhat larger. The leaves of S. adpressa are similar to those of P. bifurca but smaller (cf. Soják 2004, Fig. 3:3 and Fig. 3:13–15), both have relatively small flowers, although the flowers of S. adpressa are smaller. The number of stamens in P. bifurca is 20, which is the plesiomorphic number for all of Potentilleae, while S. adpressa has ten. The genus Sibbaldianthe was described for Sibbaldia adpressa (Juzepchuk 1941), and with our sample, retaining the name Schistophyllidium for bifurca would not violate a criterion of strict monophyly for genera. However, in the light of the general similarities between these species, there seems to be little gained in keeping them apart in separate genera. Hence, contrary to the view of Soják (2004) but in agreement with Paule and Soják (2009), we suggest that it is reasonable to classify both species in this clade in Sibbaldianthe (Kurtto and Eriksson 2003). Clade C in Fragariinae: the Dasiphora clade Sibbaldia perpusilloides is a tiny and prostrate Himalayan species with white flowers, ten stamens and very small trifoliate leaves. In collections, it is often found to be growing more or less intertwined in moss cushions (cf. Rajput et al. 1997, Fig. 7). Trifoliate leaves in Sibbaldia species are generally toothed only at the apex, but S. perpusilloides is toothed also along the sides of its tiny leaflets. Soják (2008) considered this species to be isolated taxonomically as compared to the other species classified in Sibbaldia, and he remarked that the anthers were intermediate between the Potentilla type and the Fragariinae type, but we have been unable to distinguish them from the type seen in other Fragariinae. As mentioned by Soják (2008), the position of the style is lateral like in the Fragariinae, and the species is here well supported within a Fragariinae clade (Fig. 1, clade C) along with Dasiphora, Drymocallis, Potaninia and Chamaerhodos. Because Sibbaldia perpusilloides does not resemble any species in any of these four genera, there are good reasons to classify it in its own new genus, Chamaecallis. See ‘‘Taxonomy’’ section below. Clade D in Potentilla: the Himalayan clade Four Sibbaldia species from our sample fall within Potentilla, and this is in full agreement with the assessment by Soják (2008) based on anther structure and position of the style on the achenes. These species (S. pentaphylla, S. purpurea, S. sikkimensis, S. tenuis and S. tetrandra), which are all found at more or less high elevation in the Himalayas, share the possession of a relatively wide disc inside the stamens (adaxially), and all have pink or purple petals except S. tetrandra, which has pale yellow petals. Sibbaldia tenuis and S. macropetala also belong here, although they were not included in this study, by virtue of the presence of the flower disc and their purple flower colour (Muravjova 1936, Fig. 6; Handel-Mazzetti 1939). Two species in this clade have pinnate leaves (S. sikkimensis and S. tenuis), and the leaflets of the latter are intermediate between those of S. sikkimensis and the rest of the species in the clade, which all have palmate leaves with three to five apically toothed leaflets. Species with pink or purple flowers are uncommon in Potentilleae and occur in this clade, in the Potentilla alba clade (subgen. Fragariastrum) and in a few other cases (such as Potentilla nepalensis and P. atrosanguinea). In relation to this, it is noteworthy that clade D is resolved as sister to the Potentilla alba clade (plus a few yellow-flowered species) in our combined tree, albeit with very low support. A very similar resolution is reported by Dobeš and Paule (2010). 123 T. Eriksson et al. Two species that have been named Sibbaldia probably belong to the Potentilla alba clade, namely Sibbaldia argentea (Potentilla ghalgana) and Sibbaldia omeiensis (Potentilla omeiensis) as they are white-flowered and similar to other species of that clade. Soják (2007) writes that the type specimen of Sibbaldia omeiensis has dark purple flowers. However, an image of the specimen is available in digital form from herbarium PE. The specimen on the sheet can be clearly seen to have been the basis for the original illustration (Yü and Li 1981), and its flowers are white rather than purple. Clade E in Potentilla: the Anserina clade A number of species that have been described in Sibbaldia have interruptedly pinnate leaves that are very similar to those of the Anserina clade (sect. Anserina sensu Soják 2007), namely Sibbaldia axilliflora, S. byssitecta, S. emodi, S. micropetala, S. phanerophlebia and S. potentilloides. Type material, descriptions and published photographs indicate that these species are very closely related, and revisional work may be needed to solve the problem of what names should be treated as synonyms or recombined in Potentilla. The accession we have included of Sibbaldia micropetala is consistently resolved well supported inside the Anserina clade as was indicated by its leaves, and there is no doubt that it belongs there. Hence, following these phylogenetic results, Soják’s (2008) genus Piletophyllum, in which he classified Sibbaldia emodi and Potentilla micropetala (as Piletophyllum emodi and Piletophyllum micropetalum, respectively) becomes either a synonym of Potentilla or a synonym of Argentina, depending on where genus rank is applied in the tree. Another notable result in relation to the Anserina clade (E in Fig. 1) is that Potentilla lignosa is strongly supported as sister group to the rest of the sampled species from that clade (see also Dobeš and Paule 2010). Potentilla lignosa is a white-flowered prostrate shrub with pinnate leaves found in mountain areas of western Asia, with a close relative (P. sericophylla) in north-western India. Wolf (1908) classified P. lignosa in his Potentilla ‘‘grex’’ Xylorrhizae. Recently, Soják (2004, 2008) resurrected Tylosperma for these two species. Treating these species as a separate genus is not a problem if the Anserina clade is classified as the genus Argentina separate from Potentilla because they would just become sister genera. However, if the Anserina clade is retained in Potentilla, accepting Tylosperma would cause Potentilla to become paraphyletic. Conclusions Sibbaldia in the wide sense is shown to be a non-monophyletic assemblage of at least five lineages. Some 123 additional species sometimes referred to Sibbaldia, but clearly not closely related, are also included in the list of taxa below. The five Sibbaldia clades discussed here are related to different parts of Potentilleae, three in Fragariinae and two in Potentilla (Potentillinae sensu Soják 2008). The taxonomic conclusions that we present here are based on molecular phylogenetic results that are well supported and not affected by conflicts between the ITS and trnLF regions. Sibbaldia in the strict sense is a small group centred on Sibbaldia procumbens and consists of three to six species (S. cuneata, S. parviflora, S. procumbens, S. semiglabra, S. retusa and possibly S. trullifolia), depending on species delimitation. The well-supported resolution that we see within the clade seems to favour the larger number of species. For the other two lineages of former Sibbaldia species found in Fragariinae, two additional genera are necessary: Sibbaldianthe for Sibbaldia adpressa and relatives and a new genus for Sibbaldia perpusilloides. All other species (of these five groups) should be classified in Potentilla. We suggest using the following genera for the species that used to be classified as Sibbaldia: Sibbaldia, Sibbaldianthe, the new genus Chamaecallis and Potentilla. In addition, we suggest that S. tridentata should be in Sibbaldia, but the species epithet should be retusa. See the ‘‘Taxonomy’’ section and ‘‘List of taxa’’ below for details. The molecular phylogeny is consistent with Soják’s (2008) anther character but not with his classification of Potentilleae, in which Fragariinae is rendered non-monophyletic. The position of the style does not seem to be a good character to use as a basis for classification, other than to define Potentilla s.str., since chloroplast data and combined nuclear and cpDNA data indicate that lateral-tobasal styles may be plesiomorphic in Potentilleae. However, the ITS tree which shows the Anserina clade as sister to Fragariinae is fully congruent with that character. Taxonomy Chamaecallis Smedmark, gen. nov. Type species: Chamaecallis perpusillodes (W.W.Sm.) Smedmark, comb. nov. : Potentilla perpusilloides W.W.Sm., Records of the Botanical Survey of India 4(5):188–189. 1911. Type specimen: Sikkim, Zemmu Valley, 4,270 m, July 13, 1909, Smith, Cave 1383a (CAL). : Sibbaldia perpusilloides (W.W.Sm.) Hand.Mazz. Symbolae Sinicae, Part 7:520. 1933. = Potentilla brachystemon Hand.-Mazz. Description: herbs perennial, glabrous or slightly tomentose, prostrate, tufted. Stems 1–1.5 cm. Leaves trifoliolate, margin of leaflets 2–5 serrate. Flowers solitary, Sibbaldia: a molecular phylogenetic study terminal, 5-merous. Petals white-to-cream-coloured. Stamens 10, antisepalous, 2 between each petal, filaments short, anthers elliptical, thecae confluent at apex. Disc present. Carpels glabrous, style flattened, lateral. Diagnosis: Chamaecallis can be distinguished from Potentilla and Sibbaldia, where it has been classified previously, as well as from Drymocallis, Chamaerhodos, Dasiphora and Potaninia, to which it is closely related, by its almost glabrous nature and by the combination of its prostrate, cushion-like habit, 5-merous flowers and trifoliolate leaves with serrate leaflet margins. Known distribution: Afghanistan, Myanmar, Bhutan, Nepal, China and Sikkim. Etymology: Chamaecallis is formed by combining the Greek chamae, meaning ground, low or creeping, and calli, meaning beautiful. The name is also a combination the first and last parts of the names of two of the most closely related genera, Chamaerhodos and Drymocallis. Sibbaldia retusa (O. F. Müller) T. Erikss. comb. nov. : Potentilla retusa O. F. Müller, Flora Danica 5(14), tab. 799, 1780. Type: Greenland, Godthaab [‘‘E Grönlandia misit reverendiss. Balvig. Missionarius’’.], specimen without collector and number (C!). Note. The sheet bears five specimens marked 1, 1, 2, 3 and 4, respectively, and different localities for each of the numbers. Only one, i.e. nr. 2 collected at ‘‘Godthaab’’, has hairy leaves that conforms to Müller’s description and is interpreted as holotype. Böcher et al. (1966) note that Potentilla tridentata f. hirsutifolia Paese is known to occur rarely in Greenland. = Potentilla tridentata Solander in Aiton, W. Hortus Kewensis 2: 216, tab. 9, 1789. : Sibbadiopsis tridentata (Solander in Aiton) Rydb. Rydberg, P. A. Mem. Columb. Univ. II: 187, 1898. : Sibbaldia tridentata (Solander in Aiton) Paule and Soják. Journal of the National Museum (Prague), Natural History Series, 178:15–16, 2009. Sibbaldia cuneata Hornem. Sibbaldia maxima Kesselr. ex Murav. (not seen) Sibbaldia olgae Juz. and Ovcz. (fide Soják 2004) Sibbaldia taiwanensis H.L.Li (fide Li et al. 2003) Potentilla cuneata Lehm. (K!) Lehmann cites Wallich’s catalogue number 1015 and the locality ‘‘Gossain Than’’. There are two sheets in the Wallich herbarium with that number, and they bear at least three different species: Sibbaldia cuneata, Potentilla cuneifolia and some unidentified specimen. The Potentilla specimens are marked 1015.1 ‘‘ex alpibus Kumaon’’, and the Sibbaldia specimens are marked 1015.2 ‘‘Gossain Than’’. Since Lehmann cites the latter locality, and the plant described clearly is a Sibbaldia rather than Potentilla cuneifolia, we agree with Soják (2008 and elsewhere) that Potentilla cuneata is synonymous with Sibbaldia cuneata. Sibbaldia parviflora Willd. Sibbaldia unguiculata Rajput and Tahir. Described by Rajput and Tahir (2008) (not Rajput et al. 1997), based on a single collection from northern India (BM!). We were allowed to sample a leaflet from the type specimen, and the ITS sequence is identical to that of Sibbaldia parviflora. Sibbaldia procumbens L. Sibbaldia aphanopetala Hand.-Mazz. (fide Li et al. 2003) Sibbaldia macrophylla Turcz. ex Murav. (fide Soják 2004) Sibbaldia octopetala Mill. (not seen) Sibbaldia perpusilla (Hook. f.) Chatterjee (K!) Potentilla sibbaldii Hall. f. Sibbaldia retusa (O.F. Müll.) T. Erikss. List of taxa Below are listed species that have at some point been referred to Sibbaldia and their tentative classification. Note that the synonymy only covers Sibbaldia species names and that it is not a full taxonomic revision. It is based on molecular phylogenetic results from this and other studies, in addition to the study of herbarium specimens, previously published pictures, descriptions and views of other authors. It is included here in the hope that it may be useful for further studies. Sibbaldia L. Coelas Dulac (Jussiea L. ex Sm. Name not validly published.) Potentilla retusa O.F. Müll. (C!) Potentilla tridentata Aiton Sibbadiopsis tridentata (Aiton) Rydb. Sibbaldia tridentata (Aiton) Paule and Soják Sibbaldia semiglabra C.A.Mey. Sibbaldia trullifolia (Hook.f.) Chatterjee (fide Soják 2008; K!) Sibbaldianthe Juz. Sibbaldianthe adpressa (Bunge) Juz. Sibbaldia adpressa Bunge Sibbaldia fragariastrum Tucz. ex Ledeb. Sibbaldia minutissima Kitam. (fide Ikeda et al. 2004) Potentilla lindenbergii Lehm. (fide Soják 1970) 123 T. Eriksson et al. Sibbaldianthe sericea Grubov Sibbaldia sericea (Grubov) Soják Sibbaldianthe bifurca (L.) Kurtto and T. Erikss. Potentilla bifurca L. Schistophyllidium bifurcum (L.) Ikonn. Former Potentilla Potentilla cuneifolia Bertol. Potentilla ambigua Cambess. (K!) This name is not valid since Potentilla ambigua Gaud. is an older name. Sibbaldiopsis cuneifolia (Bertol.) Soják Sibbaldia cuneifolia (Bertol.) Paule and Soják See Sibbaldia cuneata for a discussion on Potentilla cuneata Lehm., a name we have previously used for this taxon. Potentilla miyabei Makino Sibbaldiopsis miyabei (Makino) Soják Sibbaldia miyabei (Makino) Paule and Soják Anserina clade; Potentilla sect. Anserina [in part] Potentilla micropetala D. Don Potentilla axilliflora Hook. f. (K!) Sibbaldia axilliflora (Hook. f.) Chatterjee Sibbaldia byssitecta Soják (K!) Sibbaldia potentilloides Cambess. Sibbaldia emodi H. Ikeda and H. Ohba (close or conspecific with P. micropetala). Yet to be recombined in Potentilla. Sibbaldia phanerophlebia T.T.Yü and C.L.Li (close to P. micropetala). Yet to be recombined in Potentilla. Potentilla glabriuscula (T.T.Yü and C.L.Li) Soják Sibbaldia glabriuscula T.T.Yü and C.L.Li (Fide Soják 1988; Ikeda and Ohba 1999) Potentilla microphylla D.Don Sibbaldia pusilla (W.W.Sm.) Dikshit and Panigrahi Potentilla sensu stricto clade [in part] Potentilla sericea L. Sibbaldia compacta (W.W.Sm. and Cave) Dixit and Panigrahi (K!) Horkelia californica Cham. and Schltdl. Sibbaldia californica Spreng. (fide Keck 1938). Horkelia congesta (fide Keck 1938). Sibbaldia congesta D.Dietr. 123 Potentilla pulvinata (T.T.Yü and C.L.Li) Soják Sibbaldia pulvinata T.T.Yü and C.L.Li (fide Soják 1994) Potentilla alba clade; subg. Fragariastrum Potentilla ghalghana Juz. Sibbaldia argentea Owerin ex Juz. (fide Juzepczuk 1941) Potentilla omeiensis (T.T.Yü and C.L.Li) Soják Sibbaldia omeiensis T.T.Yü and C.L.Li Potentilla sikkimensis Prain Sibbaldia sikkimensis (Prain) Chatterjee Sibbaldia melinotricha Hand.-Mazz. Potentilla tenuis (Hand.-Mazz.) Soják Sibbaldia tenuis Hand.-Mazz. Potentilla purpurea (Royle) Hook. f. Sibbaldia purpurea Royle Potentilla suavis Soják Sibbaldia macropetala Murav. Potentilla clandestina Soják Sibbaldia pentaphylla J.Krause Potentilla tetrandra (Hook. f.) Bunge Sibbaldia tetrandra Bunge Dryadanthe tetrandra (Bunge) Juz. Chamaerhodos Bunge [in part] Chamaerhodos erecta (L.) Bunge Sibbaldia erecta L. Sibbaldia polygyna Willd. ex Schult. Chamaerhodos altaica (Laxm.) Bunge Sibbaldia altaica Laxm. Chamaerhodos grandiflora Pall. ex Schult. Sibbaldia grandiflora Pall. Chamaerhodos sabulosa Bunge Sibbaldia sabulosa (Bunge) Steud. Chamaecallis Smedmark Chamaecallis perpusillodes (W.W.Sm.) Smedmark Potentilla perpusilloides W.W.Sm. Sibbaldia perpusilloides (W.W.Sm.) Hand.-Mazz. Potentilla brachystemon Hand.-Mazz. Sibbaldia: a molecular phylogenetic study Acknowledgments We are grateful for the access to herbarium material from the herbaria in Stockholm (S), Copenhagen (C) and London (BM, K). We are also grateful for being able to cultivate certain species in the Bergius Botanic Garden, Stockholm. We thank A. Yoder (Duke University) and E. Conti (Universität Zürich) for help with sequencing during lean times and B. Ertter (University of California, Berkeley) for information on Horkelia. References Baum DA, Sytsma KJ, Hoch PC (1994) A phylogenetic analysis of Epilobium (Onagraceae) based on nuclear ribosomal DNA sequences. Syst Bot 19:363–388 Bentham G (1858) Handbook of the British flora. L Reeve Co, London Böcher TW, Holmen K, Jakobsen K (1966) Grønlands flora, 2nd edn. Haase sons, København Bunge AA (1829) Pentgynia. In: Ledebour CF (ed) Flora altaica, vol 1. G Reimer, Berlin, pp 428–433 Chatterjee D (1938) A note on the Indian and Chinese species of the genus Sibbaldia. Linn Notes RBG Edinb 19:325–327 Dalgaard V (1989) Additional chromosome numbers in vascular plants from the Disko Bugt area (west Greenland). Willd 19:199–213 Dikshit BK, Panigrahi G (1998) The family Rosaceae in India (revisionary studies on Potentilla L., Sibbaldia L., Brachycaulos Dikshit, Panigr.). Bishen Singh Mahendra Pal Singh, Dehra Dun Dixit BK, Panigrahi G (1981) Revision of the genus Sibbaldia L. (Rosaceae). India Proc Indian Acad Sci (Plant Sci) 90:253–272 Dobeš C, Paule J (2010) A comprehensive chloroplast DNA-based phylogeny of the genus Potentilla (Rosaceae): implications for its geographic origin, phylogeography and generic circumscription. Mol Phylogenet Evol 56:156–175 Doyle JJ, Doyle JL (1990) Isolation of plant DNA from fresh tissue. Focus 12:13–15 Endlicher S (1840) Genera plantarum, vol 2. Fr Beck, Wien Eriksson T, Donoghue MJ, Hibbs MS (1998) Phylogenetic analysis of Potentilla using DNA sequences of nuclear ribosomal internal transcribed spacers (ITS), and implications for the classification of Rosoideae (Rosaceae). Pl Syst Evol 211:155–179 Eriksson T, Hibbs MS, Yoder AD, Delwiche CF, Donoghue MJ (2003) The phylogeny of Rosoideae (Rosaceae) based on sequences of the internal transcribed spacers (ITS) of nuclear ribosomal DNA and the trnL/F region of chloroplast DNA. Int J Pl Sci 164:197–211 Farr ER, Leussink JA, Stafleu FA (1979) Index nominum genericorum (Plantarum), vol 3. Regnum vegetabile 102. Bohn, Scheltema and Halkema, Utrecht Focke WO (1888) Potentilleae. In: Engler A, Prantl K (eds) Die naturlichen Pflanzenfamilien III(3). Engelmann, Leipzig, pp 32–36 Gehrke B, Bräuchler C, Romoleroux K, Lundberg M, Heubl G, Eriksson T (2008) Molecular phylogenetics of Alchemilla, Aphanes and Lachemilla (Rosaceae) inferred from plastid and nuclear intron and spacer DNA sequences, with comments on generic classification. Mol Phylogenet Evol 47:1030–1044 Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52:696–704 Handel-Mazzetti H (1939) Plantae Sinensis. Rosaceae. 1: Potentillineae. Acta Horti Gothob 13:289–334 Helfgott DM, Francisco-Ortega J, Santos-Guerra A, Jansen RK, Simpson BB (2000) Biogeography and breeding system evolution of the woody Bencomia alliance (Rosaceae) in Macaronesia based on ITS sequence data. Syst Bot 25:82–97 Hill J (1756) The British herbal: an history of plants and trees, natives of Britain cultivated for use or raised for beauty. T. Osborne, London Holmgren PK, Holmgren NH, Barnett LC (eds) (1990) Index herbariorum, part I: the herbaria of the world, 8th edn. IAPT and New York Botanical garden, New York Hooker JD (1865) Rosaceae. In: Bentham G, Hooker JD (eds) Genera plantarum, vol 1. Reeve and co., London, pp 600–629 Hooker JD (1878) Rosaceae (Potentilla). The flora of British India, vol II. L. Reeve Co., London, pp 345–360 Huelsenbeck JP, Ronquist FR (2001) MRBAYES: Bayesian inference of phylogeny. Bioinformatics 17:754–755 Hurvich CM, Tsai C-L (1989) Regression and time series model selection in small samples. Biometrika 76:297–307 Hutchinson J (1964) The genera of flowering plants, vol 1, dicotyledons. Clarendon Press, Oxford Ikeda H (1989) Chromosome numbers of Himalayan Potentilla (Rosaceae). J Jap Bot 64:361–367 Ikeda H, Ohba H (1999) A systematic revision of Potentilla L. section Leptostylae (Rosaceae) in the Himalaya and adjacent regions. Bull Univ Mus Univ Tokyo 39:31–117 Ikeda H, Ohba H, Subedi MN (2004) Sibbaldia minutissima Kitam. (Rosaceae) is conspecific to S. adpressa. Studies of the Flora of Mustang, Central Nepal, 2. J Jap Bot 79:91–95 Juzepchuk SV (1941) Rosoideae. In: Flora of the USSR, vol X. Izdatelstvo Akademii Nauk SSSR, Moskva-Leningrad, pp 3–380 Keck DD (1938) Revision of Horkelia and Ivesia. Lloydia 1:75–142 Kurtto A, Eriksson T (2003) Atlas Florae Europaeae notes. 15. Generic delimitation and nomenclatural adjustments in Potentilleae (Rosaceae). Ann Bot Fenn 40:135–141 Li C, Ikeda H, Ohba H (2003) Sibbaldia. Flora of China. 9. Science Press, Beijing, pp 329–333 Linnaeus C (1753) Species plantarum 1. Salvius, Stockholm Linnaeus C (1754) Genera plantarum. Salvius, Stockholm Löytynoja A, Goldman N (2005) An algorithm for progressive multiple alignment of sequences with insertions. Proc Natl Acad Sci USA 102:10557–10562 Löytynoja A, Goldman N (2008) Phylogeny-aware gap placement prevents errors in sequence alignment and evolutionary analysis. Science 320:1632–1635 Lundberg M, Töpel M, Eriksen B, Nylander JAA, Eriksson T (2009) Allopolyploidy in Fragariinae (Rosaceae): comparing four DNA sequence regions, with comments on classification. Mol Phylogenet Evol 51:269–280 Muravjova OA (1936) The genus Sibbaldia L. and its species. Acta Inst Bot Acad Scient URSS Sér 1 Fasc 2:217–241 Nestler CG (1816) Monographia de Potentilla. Treuttel, Würtz, Paris Nylander JAA (2004) MrAIC.pl. https://github.com/nylander/MrAIC Nylander JAA, Wilgenbusch JC, Warren DL, Swofford DL (2008) AWTY (are we there yet?): a system for graphical exploration of MCMC convergence in Bayesian phylogenetics. Bioinformatics 24:581–583 Paule J, Soják J (2009) Taxonomic comments on the genus Sibbaldiopsis Rydb. (Rosaceae). J Natl Mus (Prague), Nat Hist Ser 178:15–16 Potter D, Eriksson T, Evans RC, Oh S, Smedmark JEE, Morgan DR, Kerr M, Robertso KR, Arsenault M, Dickinson TA, Campbell CS (2007) Phylogeny and classification of Rosaceae. Pl Syst Evol 266:5–43 Rajput MTM, Tahir SS (2008) A new species of the genus Sibbaldia (Rosaceae) from south east Asia. Pak J Bot 40:2255–2258 Rajput MTM, Tahi SS, Hussain SZ, Spongberg SA (1997) The genus Sibbaldia (Rosaceae). Pak J Bot 29:1–38 Rambaut A (2002) Se-Al. Computer software distributed by the author. Institute of Evolutionary Biology, University of Edinburgh. Version 2.0. http://tree.bio.ed.ac.uk/software/seal/ 123 T. Eriksson et al. Robertson K (1974) The genera of Rosaceae in the Southeastern United States. J Arnold Arbor 55:303–401 Ronquist F, Huelsenbeck JP (2003) MRBAYES 3: bayesian phylogenetic inference under mixed models. Bioinformatics 19:1572–1574 Rydberg PA (1898) A monograph of the North American Potentilleae. In: Memoirs from the Department of Botany of Columbia University, Columbia University, New York, pp 1–223 Rydberg PA (1908) Rosaceae. In: North American flora, vol 22. The New York Botanical Garden, New York Seringe NC (1825) Potentilla. In: De Candolle AP (ed) Prodromus systematis naturalis regni vegetabilis. Treuttel, Würtz, Paris, pp 571–586 Smith JE (1811) Lachesis Lapponica or a tour in Lapland by Linnaeus, vol I. Taylor Co, London Soják J (1970) Taxonomische und phytogeographische Anmerkungen zur Gattung Potentilla. Preslia 42:70–81 Soják J (1988) Notes on Potentilla (Rosaceae) VII. Some Himalayan taxa. Candollea 43:437–453 Soják J (1989) Generická problematika Potentilla s.l. Cas Nár Muz Rada Prir 154:117–118 Soják J (2004) Potentilla L. (Rosaceae) and related genera in the former USSR (identification key, checklist and figures). Notes on Potentilla XVI. Bot Jahrb Syst 125:253–340 Soják J (2007) Potentilla (Rosaceae) in China. Notes on Potentilla XIX. Harv Pap Bot 12:285–324 Soják J (2008) Notes on Potentilla XXI. A new division of the tribe Potentilleae (Rosaceae) and notes on generic delimitations. Bot Jahrb Syst 127:349–358 Soják J (2010) Argentina Hill, a genus distinct from Potentilla (Rosaceae). Thaiszia: J Bot 20:91–97 Spenner FCL (1829) Flora friburgensis. Brisgoviae, Friburg, p 1084 123 Staden R (1996) The Staden sequence analysis package. Mol Biotechnol 5:233–241 Sugiura N (1978) Further analysis of data by Akaike’s information criterion and finite corrections. Comm St A 7:13–26 Taberlet P, Gielly L, Pautou G, Bouvet J (1991) Universal primers for amplification of three non-coding regions of chloroplast DNA. Pl Mol Biol 17:1105–1109 Tavaré S (1986) Some probabilistic and statistical problems in the analysis of DNA sequences. Lect Math Life Sci 17:57–86 Töpel M, Lundberg M, Eriksson T, Eriksen B (2011) Molecular data and ploidal levels indicate several putative allopolyploidization events in the genus Potentilla (Rosaceae). PloS Curr. doi:10. 1371/currents.RRN1237 Urbatsch LE, Baldwin B, Donoghue MJ (2000) Phylogeny of coneflowers and relatives (Heliantheae: Asteraceae) based on nuclear rDNA internal transcribed spacer (ITS) sequences and chloroplast DNA restriction site data. Syst Bot 25:539–565 White TJ, Bruns T, Lee S, Taylor J (1990) Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ (eds) PCR protocols: a guide to methods and applications. Academic Press, London, pp 315–322 Wolf T (1908) Monographie der Gattung Potentilla. Biblioth Bot 16(Heft 71):1–714 Yang Z (1993) Maximum-likelihood estimation of phylogeny from DNA sequences when substitution rates differ over sites. Molec Biol Evol 10:1396–1401 Yang Z, Rannala B (1997) Bayesian phylogenetic inference using DNA sequences: a Markov chain monte carlo method. Molec Biol Evol 14:717–724 Yü T-T, Li C-L (1981) New species of Sibbaldia from China. Acta Phytotax Sin 19:515–518