A phylogenetic analysis of the genus Paspalum (Poaceae) based on cpDNA and morphology

Magdalena Vaio

The present work intends to clarify the phylogenetic relationships among the species of Paspalum L. belonging to the informal groups Notata/Linearia and Dilatata, and to raise some preliminary hypotheses on the phylogeny of the genus as a whole. A combined dataset including morphological and molecular characters was used to analyze 28 species of Paspalum plus some representatives of related genera of the tribe Paniceae. Analyses were performed using both parsimony and maximum likelihood. The monophyly of Paspalum is not supported nor contradicted. The circumscription of informal groups of Paspalum is discussed, as well as the cladistic treatment of allopolyploid taxa, especially those comprising the Dilatata group. The relationships of members of the Dilatata with their putative progenitors is confirmed, but the monophyly of the group as a whole is not. A close relationship between P. dilatatum Poir. and P. lividum Trin. ex Schltdl. is shown. Our analysis is consistent with the monophyly of a group comprising Notata+Linearia, with a monophyletic Notata group nested within it. The delimitation of the core Notata is proposed by including P. conduplicatum Canto-Dorow, Valls and Longhi-Wagner, P. notatum Flüggé, P. minus E. Fourn., P. pumilum Nees and P. subciliatum Chase.

Genus Paspalum L. comprises approximately 400 species worldwide and about 220 in Brazil. Paspalum is ecologically and economically important, and has been very useful as pasture. Traditionally, taxonomists use informal groups, most of which described by Agnes Chase in 1929. Some groups present a problematic circumscription, this is the case of the Linearia and Notata groups. This work uses a phylogenetic approach to study these groups and related species. DNA sequences from ITS of nuclear rRNA, from chloroplast intergenic spacer psbA-trnH and chloroplast trnL intron were used to perform the analyses. The informal groups studied were considered highly artificial, being the representatives from several informal groups splitted throughout the trees. Only a small core of species from Notata group could be accepted as a good formal clade.

Paspalum is one of the most important genera of the Poaceae family due to its large number of species and diversity. The subgenus Anachyris comprises six species mainly from South America grouped together by sharing rare spikelet characteristics. A genetic analysis using ISSR markers, compared with the morphological and phenotypic variation observed in each one species, was used to establish genetic relationships among 40 accessions with several ploidy levels, belonging to 5 species of the subgenus Anachyris. Fourteen accessions of Paspalum malacophyllum (2x and 4x), 12 of P. simplex (2x, 3x, 4x and 6x), 4 of P. procurrens (2x and 4x), 4 of P. usterii (4x) and 6 of P. volcanensis (4x) were analysed. A total of 227 ISSR loci (98.7% polymorphic) were detected among all accessions, with variable loci number and percentages of poly-morphism according to species delimitations. Six main groups were identified by cluster analysis based on Jaccard’s genetic distance and UPGMA, four of which matched all the respective accessions of P. simplex, P. procurrens, P. usterii and P. volcanensis, while the other two were consistent with two different groups of accessions of P. malacophyllum, one involving most tetraploid accessions, and the other one grouping together a tetraploid and two diploid accessions. The distinctive morphological characteristics and the separate clustering of these tetraploid and diploid cytotypes suggest to consider a new multiploidspecies complex inside the subgenus Anachyris. Both cytotypes of P. procurrens, and the four co-specific cytotypes of P. simplex consistently clustered together forming two specific groups for the two multiploid taxa. This is in agreement with the existence of high phenotypic similarities between diploid andtetraploid cytotypes of P. procurrens, and among diploid, triploid, tetraploid and hexaploid cytotypes of P. simplex. Since the polyploid cytotypes of these species are reproduced by apomixis, the specific genetic clustering by ISSR markers and morphological and cytological results support the hypothesis that the two multiploid species were originated by autopolyploidy. Our results confirm previous studies suggesting a monophyletic origin for the subgenus Anachyris and are concordant with previous data regarding genomic homologies and phylogenetic analyses in the genus.

ISSN 0378-2697, Volume 288, Combined 3-4 This article was published in the above mentioned Springer issue. The material, including all portions thereof, is protected by copyright; all rights are held exclusively by Springer Science + Business Media. The material is for personal use only; commercial use is not permitted. Unauthorized reproduction, transfer and/or use may be a violation of criminal as well as civil law. Author's personal copy Plant Syst Evol (2010) 288:227–243 DOI 10.1007/s00606-010-0327-9 ORIGINAL ARTICLE A phylogenetic analysis of the genus Paspalum (Poaceae) based on cpDNA and morphology Gabriel H. Rua • Pablo R. Speranza Magdalena Vaio • Mónica Arakaki • Received: 18 November 2009 / Accepted: 1 July 2010 / Published online: 3 August 2010 Springer-Verlag 2010 Abstract With about 350 species, Paspalum is one of the richest genera within the Poaceae. Its species inhabit ecologically diverse areas along the Americas and they are largely responsible for the biodiversity of grassland ecosystems in South America. Despite its size and relevance, no phylogeny of the genus as a whole is currently available and infrageneric relationships remain uncertain. Many Paspalum species consist of sexual-diploid and apomicticpolyploid cytotypes, and several have arisen through hybridization. In this paper we explore the phylogenetic structure of Paspalum using sequence data of four noncoding cpDNA fragments from a wide array of species which were combined with morphological data for a subset of diploid taxa. Our results confirmed the general monophyly of Paspalum if P. inaequivalve is excluded and the small genus Thrasyopsis is included. Only one of the four currently recognized subgenera was monophyletic but nested within the remainder of the genus. Some informal morphological groups were found to be polyphyletic. The G. H. Rua and P. R. Speranza contributed equally to this paper. G. H. Rua (&) Cátedra de Botánica Agrı́cola, Facultad de Agronomı́a, Universidad de Buenos Aires, Avenida San Martı́n 4453, C1417DSE Buenos Aires, Argentina e-mail: ruagabri@agro.uba.ar P. R. Speranza M. Vaio Depto. de Biologı́a Vegetal, Facultad de Agronomı́a, Universidad de la República, Garzón 780, 12900 Montevideo, Uruguay M. Arakaki Department of Ecology and Evolutionary Biology, Brown University, 80 Waterman St., Box G-W, Providence, RI 02912, USA placement of known allopolyploid groups is generally congruent with previously stated hypotheses although some species with shared genomic formulae formed paraphyletic arrangements. Other species formed a basal grade including mostly umbrophilous or hygrophilous species. It is hypothesized that the genus may have diversified as a consequence of the expansion of C4 grass-dominated grasslands in South America. Keywords Paniceae Paspalum Polyploidy cpDNA Phylogeny Introduction With about 350 species (Denham 2005; Zuloaga and Morrone 2005), Paspalum L. is one of the richest genera within the Poaceae. Its species inhabit ecologically diverse areas along North, Meso, and South America, and centers of highest diversity have been recognized in the Brazilian Cerrados and the Campos of Argentina, Uruguay, and southern Brazil. A few species are found in Africa, Asia, and Oceania, and three or four can be regarded as cosmopolitan, but the genus is thought to have originated in tropical South America (Chase 1929; Nicora and Rúgolo de Agrasar 1987; Judziewicz 1990). One species, P. scrobiculatum L. (Kodo millet), is cultivated as a cereal in Asia, and several others including P. notatum Flüggé (bahiagrass) and P. dilatatum Poir. (dallisgrass) are regarded as valuable forage grasses (Bennett 1962; Allem and Valls 1987; Filgueiras 1992). These species have been favored by research programs; however, living collections and data on genetics and reproduction for most species are far scarcer. Species of Paspalum are largely responsible for the biodiversity of 123 228 Author's personal copy grassland ecosystems in South America, which are severely threatened by the expansion of agriculture. Greater insight into the evolution of the genus as a whole can help understanding the evolution of these rich ecosystems, and consequently help us design better strategies to manage their biodiversity and to facilitate efforts to domesticate those species identified as potential crops. Morphologically, species of Paspalum are characterized by their plano-convex spikelets, probably the only morphological synapomorphy for the genus. They are also recognized by their dorsiventral, raceme-like partial inflorescences and, with few exceptions, by the lack of a lower glume. Molecular analyses of the tribe Paniceae and the subfamily Panicoideae (Gómez-Martı́nez and Culham 2000; Duvall et al. 2001; Giussani et al. 2001) show that Paspalum belongs in a clade whose members share a base chromosome number of x = 10 and is related to other genera that have a NADP-ME photosynthetic pathway. The taxa most closely related to Paspalum are Anthaenantiopsis Mez ex Pilg., the monotypic Hopia Zuloaga and Morrone (Zuloaga et al. 2007), and two species currently included in Panicum L. s.l.: P. validum Mez and P. tuerckheimii Hack. (Gómez-Martı́nez and Culham 2000; Giussani et al. 2001; Aliscioni et al. 2003; Denham and Zuloaga 2007). As currently circumscribed, Paspalum includes the former genus Thrasya Kunth, which intergrades with the informal group ‘‘Decumbentes’’ of Paspalum (Burman1985; Denham 2005; Denham and Zuloaga 2007). Additionally, recent analyses have suggested that Thrasyopsis Parodi and Reimarochloa Hitchc. are also closely related to Paspalum and they should be probably subsumed within it (Rua et al. 2007; Scataglini et al. 2007). Considerable taxonomic efforts have been devoted to the genus Paspalum, especially regarding infrageneric classification (Chase 1929, and unpubl. manuscript; Pilger 1941). The informal grouping originally proposed by Chase (1929), which is based on morphological similarities, is widely accepted. Despite the publication of a number of taxonomic revisions of particular groups in the last 10 years (Oliveira and Valls 2002; Oliveira 2004; Zuloaga and Morrone 2001, 2005; Zuloaga et al. 2004; Denham 2005; and references therein), phylogenetic hypotheses are still fragmentary (Rua and Aliscioni 2002; Denham et al. 2002; Souza-Chies et al. 2006; Denham and Zuloaga 2007; Rua et al. 2007; Scataglini et al. 2007; Giussani et al. 2009). Consequently, no phylogeny of the genus as a whole is currently available and infrageneric relationships remain uncertain and require further studies. Many Paspalum species consist of sexual-diploid and apomictic-polyploid cytotypes, and several have been shown to have arisen through hybridization (Quarin and Norrmann 1990). Furthermore, interspecific hybridization 123 G. H. Rua et al. and allopolyploidy may not be morphologically evident in some species complexes in Paspalum (Vaio et al. 2005). This fact poses further challenges to the interpretation of molecular and, particularly, morphological data and thus reticulate phylogenetic histories cannot be reconstructed by methods designed to uncover hierarchical relationships (Bachmann 2000; Linder and Riesberg 2004). Moreover, the inclusion of hybrid taxa in a phylogenetic analysis may severely disturb the inference of the relationships among other taxa included in the matrix (McDade 1992; Vriesendorp and Bakker 2005). Uniparentally transmitted sequences are always expected to evolve hierarchically and their phylogenies can be directly inferred by conventional methods. Because of this, chloroplast DNA sequences from species of different ploidy levels can be included in a phylogenetic study even in the absence of complete information about the allo or autopolyploid nature of the species involved; moreover, as taxon sampling can be increased by using all available material at this stage, the overall phylogenetic hypothesis may even be strengthened. For this reason, although phylogenetic relationships among plastid sequences may have different degrees of incongruence with organismal phylogenies (Rieseberg and Soltis 1991), the use of chloroplast DNA sequences is more convenient and cost effective in exploratory analyses. Putatively more informative sequences such as single-copy nuclear genes (Sang 2002) or ribosomal ITS sequences (Álvarez and Wendel 2003) may involve further technical complications and interpretation, particularly if sequence polymorphism may be expected as in the case of hybrid taxa. In the case of known diploid species, on the other hand, as less incongruence is expected a priori between plastid and nuclear coded morphology, the simultaneous use of both sources of information may be attempted. Preliminary attempts to establish major evolutionary lineages within Paspalum have combined different genomic sources of molecular data and morphology without previously discriminating diploid and polyploid taxa; moreover, partial taxon sampling has emphasized only specific morphological groups (Vaio et al. 2005; SouzaChies et al. 2006; Giussani et al. 2009). These analyses have shown that sequences with different inheritance patterns clearly affect the placement of known hybrid taxa and that a larger number of species representing the overall diversity of the genus must be sampled to assess the phylogenetic relationships of the genus as a whole (SouzaChies et al. 2006). In this paper we will explore the phylogenetic structure of the genus Paspalum using sequence data of four non-coding cpDNA fragments from a wide array of species and morphological data for a subset of diploid taxa. Author's personal copy A phylogenetic analysis of the genus Paspalum (Poaceae) Materials and methods Taxon sampling Seventy-one species of Paspalum were included in our analysis (Table 1), including one species formerly placed in the genus Thrasya. Although all members of the Dilatata-complex were sequenced, the assemblage was represented in our data matrix by a single sequence, because all sequences were identical (Speranza and Malosetti 2007). The two known species of Thrasyopsis and a representative of Anthaenantiopsis were also included. Two species of Axonopus P. Beauv. were used for the purpose of rooting the cladograms. Voucher specimens are listed in Table 1, including origin, chromosome counts (when available), informal grouping so far suggested in the literature, and GenBank sequence identifiers. DNA isolation, sequencing, and alignment For all plant materials, DNA was isolated from fresh leaves or silica-gel-dried leaves using the Sigma GeneluteTM kit (Sigma–Aldrich, St Louis, MO, USA) according to the manufacturer’s instructions. Four cpDNA regions were amplified and sequenced: the trnL(UAA) intron, the trnL(UAA)-trnF(GAA) spacer, the atpB-rbcL spacer, and the trnG(UCC) intron. Amplification and sequencing conditions, and primer information are reported in Vaio et al. (2005). All regions were sequenced in both directions on a CEQ 8000 capillary sequencer (Beckman-Coulter, Fullerton, CA, USA). The sequences were edited manually using SequencherTM (V4.1.4, Genecodes, AnnArbor, MI, USA) and all ambiguous end regions removed. The resulting partial sequences were prealigned with the Clustal-W (Thompson et al. 1994) algorithm implemented in BioEdit (ver. 7.0.9.0, Hall 1999) and the resulting alignments were manually adjusted. Morphological characters Morphological characters were scored mainly from herbarium material deposited at BAA, CEN, CTES, MVFA, and the Herbarium of the Universidad Nacional de Misiones, Posadas, Argentina (MNES, not indexed in Holmgren and Holmgren (1998 onwards)). A set of 115 characters was defined for use in the analyses (Appendix), including 63 characters of spikelets, 20 of inflorescences, 31 of vegetative growth, and one of reproduction. Anatomic features of rachises were scored according to Aliscioni and Denham (2008). Obvious autapomorphies were not included. All characters were treated as non-additive. Polymorphic characters were scored as such, as recommended when the polarity of the characters is unknown from 229 previous analyses (Kornet and Turner 1999). Missing data (including unavailable and inapplicable data) represent 7.6% of the entries in the morphological data matrix. Phylogenetic analysis A single matrix containing all four cpDNA sequences was assembled for the entire set of taxa (hereafter referred to as ‘‘Matrix A’’). This matrix consisted of 2,223 positions 145 of which were phylogenetically informative. Indels that could not be unambiguously aligned were rendered uninformative by introducing gaps in the rest of the taxa. Indels were coded as present/absent only when the alignment of the flanking sequences was unambiguous. A total of 15 indels between 1 and 17 bp were coded and analyzed. Parsimony analyses were performed using TNT ver. 1.1 (Goloboff et al. 2003b, 2008). Matrix A was analyzed under equal character weights. A heuristic search strategy was adopted, consisting of 10,000 random addition sequences followed by TBR swapping, using Wagner trees as starting trees and holding a maximum of two trees each time. The trees obtained were submitted to an additional round of TBR swapping. In order to detect possible islands, the trees obtained from the analysis described above were further submitted to 20,000 iterations of Parsimony Ratchet (Nixon 1999) followed by an additional round of TBR. Branches with ambiguous support (min. length = 0) were collapsed. Group support was quantified through two measures: 1 the decay index of Bremer (BS, Bremer 1994), and 2 the symmetric jackknife frequency (SJF, Goloboff et al. 2003a). For a model-based approach, Bayesian inference was conducted upon Matrix A using Mr. Bayes 3 (Huelsenbeck and Ronquist 2001). For this analysis, the unequal-frequency General Time Reversible plus Gamma (GTR ? I ? G) model of evolution was selected with MrModeltest 2.3 (Nylander 2004). Indels were coded as restriction site data following the recommendations in the software documentation. The analysis was performed for two million generations and the chain was sampled every 100 generations. The first 5,000 trees in each run were discarded as burn-in. A second matrix (hereafter ‘‘Matrix B’’) was assembled combining the cpDNA and morphological data of diploid species. In the case of species where diploids were not available but polyploids had been reported to be morphologically undistinguishable from the corresponding diploids, available polyploid cytotypes were used to represent the species’ morphologies. Species known to be allopolyploid, and species for which diploid cytotypes have not been reported (with exception of the two Thrasyopsis species, for which no cytogenetic data were available) were 123 Species Voucher Country Anthaenantiopsis rojasiana Parodi GG Roitman et al. s.n. (Herb. GH Rua 609) (BAA) Argentina Axonopus furcatus (Flüggé) Hitchc. PR Speranza s.n. (CEN) USA EU627202 EU627280 EU627358 EU627436 Axonopus rosengurttii G.A. Black PR Speranza s.n. (MVFA) Uruguay EU627206 EU627284 EU627362 EU627440 Paspalum acuminatum Raddi A Asenjo s.n. (Herb. GH Rua 495) (BAA) Argentina Paspalum alcalinum Mez GH Rua et al. 303 (BAA) Paraguay 60 Livida EU627203 EU627281 EU627359 EU627437 Paspalum almum Chase GH Rua and J Fernández 582 (BAA) Argentina 12 Alma EU627204 EU627282 EU627360 EU627438 Paspalum arundinaceum Poir. GH Rua 926 (BAA) Argentina Virgata? EU627207 EU627285 EU627363 EU627441 Paspalum atratum Swallen GH Rua et al. 261 (BAA) Brazil 40 Plicatula EU627208 EU627286 EU627364 EU627442 Paspalum bertonii Hack. Paspalum bicilium Mez GH Rua et al. 482 (BAA) RC Oliveira and GH Rua 1505 (UB, BAA) Argentina 20 Bertoniana Brazil Subg. Ceresia EU627209 EU627287 EU627365 EU627443 EU627210 EU627288 EU627366 EU627444 Paspalum ceresia (Kuntze) Chase GH Rua and L Aagesen 327 (BAA) Bolivia EU627211 EU627289 EU627367 EU627445 Paspalum chacoense Parodi PI 404691 (BAA) Paraguay 20 Caespitosa EU627212 EU627290 EU627368 EU627446 Paspalum chaseanum Parodi Saravia Toledo 13894 (CTES) Bolivia 40 Plicatula EU627213 EU627291 EU627369 EU627447 Paspalum commune Lillo GH Rua and L Aagesen 325 (BAA) Bolivia 40 Virgata EU627214 EU627292 EU627370 EU627448 Paspalum compressifolium Swallen Krapovickas 40758 (CTES) Brazil 40 Plicatula EU627215 EU627293 EU627371 EU627449 Paspalum conjugatum P.J. Bergius GH Rua, MC Gróttola and LG Frank 178 (BAA) Argentina 40 Conjugata EU627216 EU627294 EU627372 EU627450 Paspalum conspersum Schrad. CL Quarin 2319 (CTES) Argentina 60 Virgata EU627217 EU627295 EU627373 EU627451 Paspalum cromyorhizon Trin. ex Döll GG Roitman and al. (Herb. GH Rua 462) (BAA) Argentina Notata EU627219 EU627297 EU627375 EU627453 Paspalum denticulatum Trin. GH Rua 466 (BAA) Argentina 40 Livida EU627220 EU627298 EU627376 EU627454 Paspalum gr. Dilatata MVFA26505 Uruguay Dilatata EU627221 EU627299 EU627377 EU627455 Paspalum distichum L. GH Rua and L Aagesen 323 (BAA) Bolivia 60 Disticha EU627222 EU627300 EU627378 EU627456 Paspalum durifolium Mez GH Rua et al. 492 (BAA) Argentina Argentina GH Rua et al. 472 (BAA) GH Rua et al. 486 (BAA) Argentina RC Oliveira and GH Rua s.n. (Herb. GH Rua 613) Brazil (BAA) Paspalum exaltatum J. Presl Paspalum falcatum Nees ex Steud. atpB–rbcL trnG intron trnL intron trnL–trnF EU627205 EU627283 EU627361 EU627439 Dissecta 60 Subg. Ceresia Ungrouped EU627200 EU627278 EU627356 EU627434 Author's personal copy Paspalum ellipticum Döll Paspalum equitans Mez Paspalum erianthum Nees ex Trin. 2n Group 230 123 Table 1 List of the species included in the analysis, voucher specimens, country of origin, chromosome number, informal grouping (sensu Zuloaga and Morrone 2005), and GenBank accession codes EU627223 EU627301 EU627379 EU627457 EU627224 EU627302 EU627380 EU627458 EU627225 EU627303 EU627381 EU627459 EU627226 EU627304 EU627382 EU627460 P Laterra s/n (Herb. GH Rua 915) (BAA) Argentina 20 Quadrifaria EU627227 EU627305 EU627383 EU627461 C Quarin 4052 (CTES) Brazil 20 Falcata EU627228 EU627306 EU627384 EU627462 Paspalum fasciculatum Willd. ex Flüggé C Quarin 3934 (CTES) Brazil 20 Fasciculata EU627229 EU627307 EU627385 EU627463 Paspalum flavum J. Presl M Arakaki and Y Ramı́rez 1590 (CEN) Perú Paspalum foliiforme S. Denham RC Oliveira and GH Rua s.n. (Herb. GH Rua 615) Brazil (BAA) Paspalum glabrinode (Hack.) Morrone and Zuloaga GH Rua et al. 177 (BAA) Argentina 20 Ungrouped EU627232 EU627310 EU627388 EU627466 Paspalum guenoarum Arechav. GH Rua et al. 567 (BAA) Paraguay 40 Plicatula EU627233 EU627311 EU627389 EU627467 Racemosa EU627230 EU627308 EU627386 EU627464 Subg. Harpostachys EU627231 EU627309 EU627387 EU627465 G. H. Rua et al. Notata Ungrouped Eriantha Species Voucher Country 2n Group Paspalum haumanii Parodi C Quarin 3860 (CTES) Argentina 20 Quadrifaria EU627234 EU627312 EU627390 EU627468 Paspalum humboldtianum Flüggé GH Rua and L Aagesen 356 (BAA) Bolivia EU627235 EU627313 EU627391 EU627469 20 Subg. Ceresia atpB–rbcL trnG intron trnL intron trnL–trnF GH Rua et al. 209 (BAA) Paraguay 60 Inaequivalvia EU627236 EU627314 EU627392 EU627470 Paspalum inconstans Chase GH Rua and L Aagesen 353 (BAA) Bolivia EU627237 EU627315 EU627393 EU627471 Paspalum indecorum Mez GH Rua et al. 490 (BAA) Argentina 20 Caespitosa EU627238 EU627316 EU627394 EU627472 Paspalum intermedium Munro ex Morong and Britton Paspalum ionanthum Chase GH Rua et al. 35 (BAA) Paraguay 20 Quadrifaria EU627239 EU627317 EU627395 EU627473 GH Rua et al. 309 (BAA) Argentina 40 Notata EU627240 EU627318 EU627396 EU627474 Paspalum juergensii Hack. GH Rua and L Aagesen 354 (BAA) Bolivia EU627241 EU627319 EU627397 EU627475 Paspalum aff. jujuyense Zuloaga GH Rua et al. 507 (BAA) Paraguay 40 Livida Paspalum lenticulare Kunth GH Rua et al. 45 (BAA) Paraguay 40 Plicatula EU627244 EU627322 EU627400 EU627478 Paspalum lepton Schult. GH Rua 920 (BAA) Argentina EU627242 EU627320 EU627398 EU627476 Subg. Harpostachys Paniculata Plicatula EU627201 EU627279 EU627357 EU627435 Paspalum lilloi Hack. GH Rua et al. 127 (BAA) Argentina 20 Bertoniana EU627243 EU627321 EU627399 EU627477 Paspalum lineare Trin. T Killeen 2218 (CTES) Bolivia EU627245 EU627323 EU627401 EU627479 Paspalum maculosum Trin. GH Rua et al. 487 (BAA) Argentina 20 Notata EU627246 EU627324 EU627402 EU627480 Paspalum malacophyllum Trin. JFM Valls et al. 14855 (CEN) Brazil EU627247 EU627325 EU627403 EU627481 Paspalum mandiocanum var. subaequiglume I.L. Barreto GH Rua et al. 301 (BAA) Paraguay 50 Corcovadensia EU627248 EU627326 EU627404 EU627482 Paspalum modestum Mez GH Rua et al. 146 (BAA) Argentina 20 Plicatula EU627249 EU627327 EU627405 EU627483 Paspalum notatum Flüggé Paspalum orbiculatum Poir. GH Rua et al. 296 (BAA) GH Rua et al. 590 (BAA) Brazil 20 Notata Paraguay 20 Orbiculata EU627250 EU627328 EU627406 EU627484 EU627251 EU627329 EU627407 EU627485 80 Notata 20 Subg. Anachyris Paspalum ovale Nees ex Steud. GH Rua et al. 476 (BAA) Argentina 80 Ovalia EU627252 EU627330 EU627408 EU627486 Paspalum palustre Mez C Quarin 3648 (CTES) Argentina 20 Plicatula EU627253 EU627331 EU627409 EU627487 GH Rua 587 (BAA) Argentina Paspalum paucifolium Swallen GH Rua et al. 313 (BAA) Argentina 40 Eriantha Paniculata EU627255 EU627333 EU627411 EU627489 Paspalum pilosum Lam. GH Rua 528 (BAA) Argentina EU627256 EU627334 EU627412 EU627490 Paspalum plicatulum Michx. AI Honfi 14 (CTES, MNES) Argentina 20 Plicatula EU627257 EU627335 EU627413 EU627491 Paspalum polyphyllum Nees ex Trin. D Hojsgaard 264 (MNES) Argentina 40 Subg. Ceresia EU627258 EU627336 EU627414 EU627492 Paspalum pumilum Nees GH Rua et al. 592 (BAA) Argentina Notata EU627259 EU627337 EU627415 EU627493 Paspalum quadrifarium Lam. PR Speranza 27 (MVFA) Uruguay 20 Quadrifaria EU627260 EU627338 EU627416 EU627494 Paspalum quarinii Morrone and Zuloaga Paspalum remotum J. Rémy JFM Valls 11268 (CEN) GH Rua 543 (BAA) Brazil Bolivia 40 Quadrifaria 80 Livida EU627261 EU627339 EU627417 EU627495 EU627262 EU627340 EU627418 EU627496 Paspalum repens P.J. Bergius GH Rua et al. (BAA) Paraguay Dissecta EU627263 EU627341 EU627419 EU627497 Paspalum rufum Nees ex Steud. GH Rua and IB Boccaloni 156 (BAA) Argentina Ungrouped EU627264 EU627342 EU627420 EU627498 Subg. Harpostachys EU627254 EU627332 EU627410 EU627488 231 123 Paspalum paniculatum L. Author's personal copy Paspalum inaequivalve Raddi A phylogenetic analysis of the genus Paspalum (Poaceae) Table 1 continued EU627271 EU627349 EU627427 EU627505 Brazil GH Rua s.n. (CEN) Thrasyopsis juergensii (Hack.) Soderstr. and A.G. Burm. Thrasyopsis repanda (Nees) Parodi EU627277 EU627355 EU627433 EU627511 Argentina 40 Plicatula Brazil C Quarin 4158 (CTES) GH Rua and JL Rosa 728 (CEN) Paspalum wrightii Hitchc. and Chase EU627276 EU627354 EU627432 EU627510 40 Virgata French Guiana M Tourn and F Perret s.n. (BAA) Paspalum virgatum L. EU627275 EU627353 EU627431 EU627509 EU627274 EU627352 EU627430 EU627508 Disticha Argentina 40 Subg. Anachyris Paraguay AI Honfi 1175 (MNES) GH Rua et al. 559 (BAA) Paspalum usteri Hack. Paspalum vaginatum Sw. EU627273 EU627351 EU627429 EU627507 Argentina 40 Subg. Harpostachys I Boccaloni (Herb. GH Rua 506) (BAA) Paspalum unispicatum (Scribn. and Merr.) Nash EU627269 EU627347 EU627425 EU627503 EU627272 EU627350 EU627428 EU627506 Caespitosa Brazil RC Oliveira and GH Rua 1502 (UB, BAA) Paspalum trichostomum Hack. Subg. Ceresia Argentina GH Rua et al. 192 (BAA) Paspalum stellatum Humb. and Bonpl. ex Flüggé EU627267 EU627345 EU627423 EU627501 Argentina 40 Subg. Anachyris GH Rua et al. 308 (BAA) Paspalum simplex Morong Plicatula 20 Setacea EU627268 EU627346 EU627424 EU627502 USA PR Speranza s.n. (CEN) Paspalum setaceum Michx. EU627265 EU627343 EU627421 EU627499 Argentina GH Rua 520 (BAA) Paspalum scrobiculatum L. 2n Group atpB–rbcL trnG intron trnL intron trnL–trnF Country Voucher Species Table 1 continued 123 EU627270 EU627348 EU627426 EU627504 Author's personal copy 232 G. H. Rua et al. excluded from this matrix. The resulting Matrix B contained 3.5% of missing entries, including both unavailable and inapplicable data. Because of the high levels of homoplasy found in preliminary assessments of the morphological data set, Matrix B was analyzed using implied weights (Goloboff 1993). When using implied weights, TNT downweights homoplastic characters in proportion to their amount of extra steps (homoplasy), and saves trees that minimize ‘‘distortion’’ (D), which is an increasing function of homoplasy (Goloboff et al. 2003b). D ¼ e=ðe þ kÞ; where e = extra steps and k = constant of concavity. The strength with which a homoplastic character is downweighted depends on the concavity value (k) of the weighting function: the lower the k value the stronger the weighting function. To explore the stability of the results, analyses were performed under 33 different k values. Because distortion is not a linear function of concavity, k values were selected in such a way that they produce regular distortion increments of 1.25%, within a range of 50–90% related to an average non-homoplastic character (Mirande 2009). To test tree stability related to variations of k, comparisons between pairs of contiguous trees (i.e. between trees obtained using kn and kn-1) were performed, by calculation of 1 SPR-difference, i.e. the number of SPR-swaps required to convert tree n into tree n - 1; 2 number of shared taxa (=nodes in agreement subtree); and 3 number of shared groups (=nodes in strict consensus tree). Calculation of k values, tree searches, and calculation of stability measures were all performed using a TNT script written by J. Marcos Mirande (unpublished), who kindly made it available to us. Because support measures are not comparable when using different weighting functions, BS and SJF values were independently calculated for each concavity. The searching routine was otherwise identical with that described for Matrix A. Congruence of both data sets was tested through ILD test (Farris et al. 1995) under k values ranging between 3 and 20, using a custom TNT script (Goloboff et al. 2003b; Ramı́rez 2006). Results Matrix A As a result of the parsimony analysis of Matrix A, 31,080 equally parsimonious trees were found, each 426 steps long Author's personal copy A phylogenetic analysis of the genus Paspalum (Poaceae) (CI = 0.75, RI = 0.82). The strict consensus obtained with the parsimony analysis is shown in Fig. 1. The trees obtained with parsimony and Bayesian approaches were fully congruent; furthermore, the tree obtained with Mr. Bayes was nearly identical with the majority rule consensus of the trees obtained with parsimony (not shown). Almost all the nodes present in the parsimony consensus tree received 100% posterior probabilities in the Bayesian analysis, the nodes receiving lower posterior probabilities are indicated in Fig. 1. Only one node not present in the parsimony consensus received 100% posterior probability in the Bayesian analysis. This node was included in Fig. 1 and shown as a dashed line. In all trees (Fig. 1) [Paspalum ? Anthaenantiopsis] formed a strongly supported clade (node 1), and a clade comprising [Anthaenantiopsis rojasiana ? P. inaequivalve] (node 2) was sister to the remaining species of Paspalum (node 3). The basal region of the tree was poorly resolved and mostly included umbrophilous species with more or less plagiotropic culms (P. orbiculatum, P. flavum, P. conjugatum, P. setaceum, P. mandiocanum, and P. inconstans) and floating grasses (P. repens and P. acuminatum). This poorly resolved portion of the tree also included the rhizomatose-stoloniferous, heliophilous species P. distichum and P. vaginatum, and P. glabrinode, a morphologically very distinct species. The pairs [P. distichum ? P. vaginatum] and [P. mandiocanum ? P. inconstans] were highly supported, whereas [P. conjugatum ? P. repens] formed a moderately supported clade (node 4). A weakly supported clade (node 5) although with a 100% posterior probability comprised all the remaining species, distributed in two major clades (nodes 6 and 11). These clades appeared consistently, although weakly supported in all analyses. An unresolved region in this clade includes P. almum, P. ceresia, and three species belonging to subg. Harpostachys (Denham 2005). One of the two major clades (node 6, hereafter referred to as the ‘‘CQPA’’ clade) including mostly species with paired spikelets, showed a basal polytomy from which P. durifolium and three clades emerge. One of these (node 7) was weakly supported and roughly corresponds to the informal group ‘‘Caespitosa’’ ? P. fasciculatum. The second clade (node 8) was, in contrast, highly supported, and included P. intermedium and allied species, currently members of the ‘‘Quadrifaria’’ group, plus P. ovale (‘‘Ovalia’’ group) and P. denticulatum ? P. aff. jujuyense (‘‘Livida’’ group). Finally, a third clade (node 9) included the remaining species of the group ‘‘Quadrifaria’’ (P. quadrifarium and P. quarinii), the group ‘‘Paniculata’’, P. remotum (‘‘Livida’’ group), and a highly supported clade (node 10) containing the group ‘‘Humboldtiana’’ (sensu Parodi and Nicora, 233 unpubl. manuscript, also including P. paucifolium), P. falcatum, and subg. Anachyris. The other major clade (node 11, hereafter the ‘‘NPBT’’ clade) contained species belonging to groups ‘‘Notata’’ sensu lato (including the former group ‘‘Linearia’’, Zuloaga et al. 2004; Souza-Chies et al. 2006), ‘‘Plicatula’’ (Oliveira 2004; Oliveira and Valls 2008), and ‘‘Bertoniana’’ (Zuloaga and Morrone 2005). It also includes P. stellatum (subgen. Harpostachys), P. alcalinum (group ‘‘Livida’’), P. rufum (ungrouped according to Zuloaga and Morrone 2005), and a highly supported clade comprising the two species of Thrasyopsis (node 12). Also well supported were the clades corresponding to group ‘‘Bertoniana’’ (node 13) and to two groups of species belonging to the ‘‘Plicatula’’, whereas the group ‘‘Plicatula’’ itself remained unresolved. Matrix B Although morphological and molecular data sets were significantly incongruent at 99% confidence level under all k values tried, combined analysis were performed in order to maximize the explanatory power of the available evidence (Nixon and Carpenter 1996). Nevertheless, both data sets were also analyzed separately for comparison. Analysis of Matrix B yielded eight different trees when analyzed under 33 values of k ranging between 2.3 and 20.5 (with a single most parsimonious tree under each k value).The most stable topology (Fig. 2) was obtained under k ranging between 3.1 and 7.8. When molecular data were analyzed separately, 36 most parsimonious trees were obtained under all concavity values tried (results not shown). Combined analysis of Matrix B yielded a weakly supported monophyletic genus Paspalum, with Anthenantiopsis as its sister group. Moreover, Anthenantiopsis was nested within Paspalum when searches were tried using equal weights (trees not shown). The clade comprising [Paspalum ? Anthenantiopsis] was very strongly supported under all search strategies. The position of P. inaequivalve could not be tested further because it is a hexaploid species and polyploids were explicitly excluded from Matrix B. Most groups were consistent with those obtained from Matrix A; however, the species forming the basal grade on analysis of Matrix A were grouped differently, although with no jackknife support (Fig. 2). For the two main internal clades, further resolution was achieved as a consequence of the addition of the morphological signal, especially within the NPBT-clade, where monophyletic groups ‘‘Plicatula’’ (node 14) and ‘‘Notata’’ sensu stricto (node 15) were recovered (Fig. 2). Nevertheless, support for branches was not substantially improved in general and all additional nodes were weakly supported. 123 Author's personal copy G. H. Rua et al. 1 53 11 P. atratum - PLICATULA P. chaseanum - PLICATULA P. ellipticum - NOTATA p. erianthum - ERIANTHA P. guenoarum - PLICATULA P. lineare - NOTATA P. maculosum - NOTATA P. scrobiculatum - PLICATULA P. plicatulum - PLICATULA P. pumilum - NOTATA P. notatum - NOTATA 2 P. alcalinum - LIVIDA 78 P. rufum - ungrouped 5 Thrasyopsis juergensii 99 Thrasyopsis repanda 12 4 98 P. bertonii - BERTONIANA P. lilloi - BERTONIANA NPBT-clade Axonopus furcatus Axonopus rosengurttii 6 Anthaenantiopsis rojasiana 99 P. inaequivalve - INAEQUIVALVIA P. flavum - RACEMOSA P. orbiculatum - ORBICULATA >10 2 2 P. conjugatum - CONJUGATA 100 59/96 P. repens - DISSECTA P. setaceum - SETACEA 4 5 1 1 P. acuminatum - DISSECTA 94 51/97 P. glabrinode - ungrouped 5 P. distichum - DISTICHA 3 99 P. vaginatum - DISTICHA 6 P. inconstans - subg. Harpostachys 1 99 P. mandiocanum - CORCOVADENSIA P. almum - ALMA P. ceresia - subg. Ceresia P. foliiforme - subg. Harpostachys P. pilosum - subg. Harpostachys P. unispicatum - subg. Harpostachys P. durifolium - ungrouped P. arundinaceum - VIRGATA ? 1 P. chacoense - CAESPITOSA 1 P. trichostomum - CAESPITOSA -/97 P. fasciculatum - FASCICULATA 1 P. indecorum - CAESPITOSA -/66 7 P. exaltatum - QUADRIFARIA 3 P. haumanii - QUADRIFARIA 94 1 P. ovale - OVALIA 7 -/82 P. virgatum* - VIRGATA 99 2 P. aff. jujuyense - LIVIDA 3 6 89 P. denticulatum - LIVIDA 93 8 2 P. conspersum* - VIRGATA 83 P. intermedium - QUADRIFARIA P. quadrifarium - QUADRIFARIA DILATATA -complex* 2 P. quarinii - QUADRIFARIA 1 P. remotum - LIVIDA 69/99 P. commune* - VIRGATA 1 2 P. juergensii - PANICULATA 59 9 83 5 P. paniculatum - PANICULATA P. humboldtianum - subg. Ceresia 3 P. bicilium - subg. Ceresia 89 5 P. polyphyllum - subg. Ceresia 1 P . falcatum - FALCATA 97 51/99 P. paucifolium - ERIANTHA 1 P. simplex - subg. Anachyris 10 1 64 P. malacophyllum - subg. Anachyris -/96 P. usterii - subg. Anachyris CQPA-clade 234 13 compressifolium - PLICATULA lepton - PLICATULA lenticulare - PLICATULA wrightii - PLICATULA 2 1 P. modestum - PLICATULA 85 56/99 P. palustre - PLICATULA p. equitans - ungrouped 1 P . stellatum - subg. Ceresia 1 53 3 P. cromyorhizon - NOTATA 64/98 90 P. ionanthum - NOTATA 3 92 P. P. P. P. Fig. 1 Strict consensus of 30,080 most parsimonious trees obtained from analysis of Matrix A (all terminals, cpDNA data only) using equal character weighting. Values above branches, Bremer supports; values below branches: left, symmetric jackknife frequency; right, posterior 123 probabilities in Bayesian analysis, only values below 100% are shown; all other nodes received 100% posterior probabilities. The branch shown in dashed lines had posterior probability = 100% but did not appear in the MP consensus. Numbered nodes are referred to in the main text Author's personal copy A phylogenetic analysis of the genus Paspalum (Poaceae) 235 supported by molecular data. Finally, 14 groups were diagnosed exclusively by morphological synapomorphies, of which only two groups [P. humboldtianum ? P. paucifolium] and [P. cromyorhizon ? P. equitans] were reasonably well supported. Most groups recovered in the combined analysis can be recognized by morphological synapomorphies (Fig. 3). Morphological ‘‘hard’’ synapomorphies (i.e. such characters states that are synapomorphic for a group and not represent homoplasies in another part of the tree) were rare and included lack of upper glumes [character 3] and plurinerved upper lemmas [ch. 58] in [P. malacophyllum ? P. simplex] (representatives of subgen. Anachyris), multinerved upper glumes [ch. 12] in the Thrasyopsis clade, lack of sclerenchyma rings in rachises [ch. 73] in [P. vaginatum ? P. conjugatum ? P. repens], and air lacunae in midveins [ch. 113] in [P. modestum ? P. palustre]. ‘‘Hard’’ synapomorphies were otherwise molecular. The most stable characters supporting the main crown clade (node 5, Fig. 1; node 50 , Fig. 2) were outer upper glume veins approximate to margins [ch. 16], palea margins with overlapping wings [ch. 54], and the lack of enrichment branches in the inflorescence [ch. 84]. The CPQA clade was partially supported by single lateral lower lemma veins [ch. 33] and slightly indurate upper florets [ch. 42], and homogeneously or distally paired spikelets [ch. 79]. On the other hand, the NPBT clade did not show any morphological synapomorphy and it is entirely Discussion Monophyly of Paspalum and outgroup relationships Our results confirmed the general monophyly of Paspalum as currently circumscribed, i.e. including the species formerly placed in Thrasya (Giussani et al. 2001; Denham 2005; Denham and Zuloaga 2007). Nevertheless, the monophyly of Paspalum as sampled in our dataset requires either the exclusion of P. inaequivalve or the inclusion of Anthaenantiopsis rojasiana. Paspalum inaequivalve is a rather atypical species which is not obviously related to any other species of Paspalum except, perhaps, to P. microstachyum J. Presl (Aliscioni and Denham 2009), not included in this analysis. Further sampling of the taxa closely related to Paspalum is needed to assess the correct Axonopus furcatus Axonopus rosengurttii Anthaenantiopsis rojasiana P. orbiculatum 0.07 P. vaginatum 0.07 0.33 P. conjugatum P. repens 93 4 0.14 0.05 P. setaceum P. foliiforme P. glabrinode P. almum 0.05 3 0.10 50 0.08 0.12 55 7 0.05 0.04 6 5’ 0.02 0.03 0.05 11 0.03 P. fasciculatum P. chacoense P. indecorum 0.26 P. exaltatum 1.00 92 P. haumanii 0.25 P. intermedium 99 83 P. aff. jujuyense 8 P. quadrifarium P. quarinii 0.01 0.33 0.19 P. juergensii 50 90 74 P. paniculatum 0.34 0.12 P. humboldtianum 83 9 84 0.70 P. paucifolium P. falcatum 99 0.04 12 0.48 P. malacophyllum 57 10 P. simplex 99 0.83 Thrasyopsis juergensii Thrasyopsis repanda 99 P. ellipticum 0.04 P. rufum 0.09 0.50 P. bertonii P. lilloi 95 13 P. plicatulum 0.12 P. compressifolium 0.69 51 0.05 P. lenticulare 99 P . wrightii 0.17 P. modestum 0.17 14 68 P. palustre 71 P. stellatum 0.04 0.07 P. cromyorhizon 0.04 P. equitans 80 P. maculosum 0.04 0.10 P. notatum 57 P. pumilum 64 0.04 CQPA-clade 59 NPBT-clade 1 15 Fig. 2 Most parsimonious tree obtained from analysis of Matrix B (only terminals with known diploid cytotypes, cpDNA and morphology), using implied weights under concavity k = 6. Values above branches, Bremer supports; values below branches, symmetric Jackknife frequency. Nodes 1–13 as in Fig. 1, except for node 50 which differs from node 5 in Fig. 1 for not including P. foliiforme; nodes 14 and 15 do not appear in Fig. 1, and they are referred to in the main text 123 Author's personal copy 236 G. H. Rua et al. Anthaenantiopsis rojasiana P. orbiculatum P. vaginatum P. conjugatum P. repens P. setaceum P. foliiforme P. glabrinode P. almum P. fasciculatum P. chacoense P. indecorum P. intermedium P. aff. jujuyense P. exaltatum P. haumanii P. quadrifarium P. quarinii P. juergensii P. paniculatum P. humboldtianum P. paucifolium P. falcatum P. malacophyllum P. simplex Thrasyopsis juergensii Thrasyopsis repanda P. ellipticum P. rufum P. bertonii P. lilloi P. maculosum P. notatum P. pumilum P. stellatum P. cromyorhizon P. equitans P. plicatulum P. compressifolium P. lenticulare P. wrightii P. modestum P. palustre Fig. 3 Synapomorphies mapped onto the tree of Fig. 2. Shapes on branches represent synapomorphic changes. Autapomorphic changes are not shown. Ambiguous character reconstructions are not depicted. Squares, morphological characters; circles, cpDNA characters; black symbols, ‘‘hard’’ synapomorphies; white symbols, ‘‘soft’’ synapomorphies placement of these species. Despite this, our data support the view that with few exceptions, the great majority of the species presently included in Paspalum form a well supported monophyletic assemblage. The genus Thrasyopsis has been suggested to be phylogenetically related to Paspalum (Denham and Zuloaga 2007), and morphologically related to the informal group ‘‘Crassa’’ (Chase, unpubl. manuscript). In fact, in this analysis the two species of Thrasyopsis were deeply nested within Paspalum and must certainly be transferred to it (new combinations will be formally published in an ongoing paper). Unfortunately, no species of the ‘‘Crassa’’ group were available to assess their phylogenetic affinities with the species of Thrasyopsis. 123 Author's personal copy A phylogenetic analysis of the genus Paspalum (Poaceae) Infrageneric classification Four subgeneric entities are currently recognized within Paspalum: subg. Ceresia, subg. Anachyris, subg. Harpostachys, and subg. Paspalum (Denham 2005; Zuloaga and Morrone 2005). Such a subgeneric classification is not supported by our results because it does not reflect natural phylogenetic relationships. Only the species included in subgenus Anachyris formed a clearly monophyletic and well supported clade in this study. This assemblage consists of a morphologically distinct alliance of species characterized by having navicular, concavo-convex spikelets with an upper glume reduced or lacking and an upper flower sharply nerved (Morrone et al. 2000). This group is, however, deeply embedded within the genus, so its recognition as a subgenus would render the subgenus Paspalum paraphyletic. A different situation is presented by the species included in subg. Ceresia. Our data suggest no close relationships between P. stellatum and P. ceresia; or between P. stellatum and P. humboldtianum, P. polyphyllum, and P. bicilium, a fact already suggested on the basis of morphological evidence (Rua and Aliscioni 2002). The grouping of the last three species together with subg. Anachyris, P. paucifolium and P. falcatum is relatively well supported, whereas P. ceresia and P. stellatum appear in very distant positions. Relationships among species of subg. Harpostachys were very poorly resolved. These species appear as part of a grade that is basal to the two core clades of the genus. The close relationship between P. inconstans (subg. Harpostachys) and P. mandiocanum (subg. Paspalum, ‘‘Corcovadensia’’ group) recovered suggests that, as currently circumscribed (Denham 2005), subg. Harpostachys is also not monophyletic. The taxonomy of Paspalum has always been a difficult issue. Delimitation of putative taxa within Paspalum has been obscured by frequent overlapping of character distributions probably caused by both homoplasy and hybridization. Because of this, a formal classification has long been eschewed. Indeed, the informal grouping proposed by Chase (1929) on the basis of morphological similarity has taken the place of a formal taxonomy below subgeneric rank (Zuloaga and Morrone 2005), despite the fact that Chase (1929) recognized that only ‘‘some groups … are natural aggregations of closely related species’’ whereas ‘‘the constituents of other groups are less obviously allied’’ (Chase 1929, p. 7). Our molecular and morphological analysis shows that several currently accepted informal groups are clearly not monophyletic, whereas others probably are. It would be sound to dismiss non-monophyletic groups (see below), or perhaps to re-define them. The convenience of provisionally maintaining informal groups may be questioned, but a 237 new, phylogeny-based circumscription for infrageneric groups is not yet possible. At the current stage of our knowledge about the phylogeny of Paspalum, favoring any particular taxonomic grouping seems at least premature. Polyphyletic morphological groups Among the morphologically identified groups that were not recovered in our analysis, ‘‘Eriantha’’ and ‘‘Livida’’ appear as particularly polyphyletic. Sampled species belonging to these two groups appear scattered over the entire topology, suggesting that they are highly artificial assemblages. The ‘‘Eriantha’’ group (Morrone et al. 2004) was represented in our analyses by P. erianthum and P. paucifolium only. Despite the apparent similarity of their spikelets, the former species was included in the NPBT clade, whereas the latter grouped together with the P. humboldtianum alliance, deeply nested within the CQPA clade. P. erianthum groups together with species belonging to the CQPA clade in the analysis performed by Giussani et al. (2009). High ploidy levels only have been reported for this species (Norrmann et al. 1994) and consequently a complex cytogenetic architecture cannot be ruled out. Resolving the phylogenetic affinities of this kind of species may require more detailed genetic characterization. Four species currently classified in the ‘‘Livida’’ group (Zuloaga and Morrone 2005; Denham et al. 2010) were included in our analysis. Paspalum denticulatum, P. aff. jujuyense, and P. remotum were placed in the CQPA clade. Nevertheless, P. denticulatum and P. aff. jujuyense were nested within the P. intermedium alliance, whereas the Andean species P. remotum was rather related to the ‘‘Quadrifaria–Paniculata–Anachyris’’ alliance. On the other hand, P. alcalinum grouped together with P. rufum, within the NPBT clade. Despite putatively diagnostic morphological differences, P. jujuyense has been synonymized with P. denticulatum because intermediate morphologies were found which obscure the limits between both species (Zuloaga and Morrone 2005; Denham et al. 2010). Because available data are currently insufficient, both species were tentatively maintained as separate in our analysis. Nevertheless, the almost identical cpDNA sequences (only one position differs) suggest at least very close affinity between them. The NPBT clade This clade supported by molecular characters comprises all the species belonging in the informal groups ‘‘Notata’’, ‘‘Plicatula’’, and ‘‘Bertoniana’’, plus the two species currently included in the genus Thrasyopsis, and P. erianthum and P. alcalinum, which are currently included in groups ‘‘Eriantha’’ and ‘‘Livida’’, respectively (Zuloaga and Morrone 2005). 123 238 Author's personal copy Monophyly of the group ‘‘Bertoniana’’ was highly supported by both molecular and morphological data. This group comprises only two species, P. bertonii and P. lilloi, characterized by having pilose spikelets, the upper floret open at the top, and sharply costate leaf blades (Zuloaga and Morrone 2005). Both species share a highly restricted habitat in river margins in the high Paraná basin. Relationships of the group ‘‘Bertoniana’’ within the NPBTclade were ambiguously resolved. In this analysis, the species of the ‘‘Notata’’ group sensu lato did not group together; however, a ‘‘core-Notata’’ clade including P. notatum, P. pumilum, and P. maculosum was recovered when morphological and sequence information were both taken into account (Matrix B, Figs. 2 and 3). The ‘‘Notata’’ group comprises species with solitary, mostly glabrous spikelets, and inflorescences composed of two conjugate racemes, sometimes accompanied by one or two additional, more distant ones (Chase 1929; Barreto 1957; Canto-Dorow et al. 1996; Parodi and Nicora, unpubl. manuscript). Recently (Zuloaga et al. 2004), the circumscription of the ‘‘Notata’’ group has been extended to include the species formerly placed in the ‘‘Linearia’’ group (Chase 1929; Parodi and Nicora, unpubl. manuscript; Oliveira and Valls 2002). A preliminary phylogeny based on ITS data and morphology (Souza-Chies et al. 2006) was consistent with merging the two groups, and recovered a monophyletic ‘‘core-Notata’’ group within a clade comprising species of both ‘‘Notata’’ and ‘‘Linearia’’ groups. P. cromyorhizon was related to P. equitans, a species alternatively excluded (Zuloaga et al. 2004) or included (Souza-Chies et al. 2006) within the ‘‘Notata’’ group. Despite having inflorescences with 4–7 primary branches, P. equitans shows clear morphological similarities with species of the ‘‘Notata’’ group, for example P. ionanthum and P. ramboi (Parodi and Nicora, unpubl. manuscript; Barreto 1983). On the other hand, P. ellipticum is currently placed in the ‘‘Notata’’ group sensu lato (Oliveira and Valls 2002; Zuloaga et al. 2004) despite its pilose rather than glabrous spikelets. In our analysis, P. ellipticum was undoubtedly placed in the NPBT clade, but its sister relationships were not clearly resolved. The ‘‘Plicatula’’ group includes species characterized by a sharply plano-convex spikelet and a dark brown, shiny upper floret (Chase 1929; Oliveira 2004). Analysis of Matrix B yielded a monophyletic ‘‘Plicatula’’ group with a clade containing P. wrightii, P. modestum, and P. palustre nested within it. These three species are hydrophytic grasses sharing spongy leaf sheaths, a whitish midvein in the leaf blades, an upper floret neither so dark nor so bowed as those of typical ‘‘Plicatula’’ species (Parodi and Nicora, unpubl. manuscript; Barreto 1974; Oliveira 2004), and chromosome pairing affinity (Martı́nez and Quarin 1999). They were segregated as a different group by Barreto (1974), together with the presumably annual P. boscianum. 123 G. H. Rua et al. Our analysis thus supports the inclusion of these species within the ‘‘Plicatula’’ group, as currently accepted by most authors (Oliveira 2004; Zuloaga and Morrone 2005; Oliveira and Valls 2008). Our analysis placed Paspalum rufum in the NPBT clade, but its relationships with other taxa within the clade were not clear. Some accessions of this species were also allied to the ‘‘Plicatula’’ group in other molecular analysis (Giussani et al. 2009). This species has morphological similarities with species of the groups ‘‘Eriantha’’ and ‘‘Virgata’’ (Barreto 1954; Zuloaga and Morrone 2005) and a degree of chromosome pairing with species of the ‘‘Quadrifaria’’ group (Quarin and Norrmann 1990); it also shares with the ‘‘Plicatula’’ group the occurrence of a dark brown upper floret. As reconstructed on our phylogenetic hypothesis, chromosome pairing affinity with the ‘‘Quadrifaria’’ group seems to be the result of synplesiomorphy. The placement of P. stellatum near P. cromyorhizon, and P. equitans was challenging. P. stellatum is a very distinct species, morphologically related to P. ceresia and other species with a dilated, membranaceous rachis currently placed in the subgenus Ceresia (Denham et al. 2002). Known polyploids of P. stellatum have unusual chromosome numbers (2n = 32 and 52; Honfi et al. 1990) suggesting an allopolyploid origin including at least one genome with a base chromosome number not equal to 10. Because no confirmed diploid material of this species was available, the hypothesis of allopolyploidy involving a maternal chromosome donor belonging to the NPBT clade is to be further explored. On the other hand, inclusion of diploid material of this species would be necessary to analyze its phylogenetic relationships. The CQPA clade The CQPA clade includes, distributed in three subclades, representatives of informal groups ‘‘Caespitosa’’, ‘‘Quadrifaria’’, and ‘‘Paniculata’’, and of subgenus Anachyris, along with several species currently placed in different groups. One of these clades corresponds to the ‘‘Caespitosa’’ group and it also includes P. arundinaceum, a species currently assigned to the ‘‘Virgata’’ group (Chase 1929; Judziewicz 1990) or at least related to it (Giussani et al. 2009). This clade further includes P. fasciculatum, a morphologically distinct, creeping, paludose grass with flabellate inflorescences, which is currently ascribed to the monotypic ‘‘Fasciculata’’ group (Zuloaga and Morrone 2005). However, P. fasciculatum shares with the other species in this clade the occurrence of solid culms and basally constricted upper florets and the occasional occurrence of a lower glume. The other two clades were both highly supported. Species of the informal group ‘‘Quadrifaria’’, comprising Author's personal copy A phylogenetic analysis of the genus Paspalum (Poaceae) robust tussock grasses with multi-racemed inflorescences (Barreto 1966, 1974; Gomes and Monteiro 1996; Zuloaga and Morrone 2005), appeared partitioned between these two groups, a result also found in a cpDNA-based analysis of the groups ‘‘Quadrifaria’’ and ‘‘Virgata’’ (Giussani et al. 2009), and supported by fluorescent in situ hybridization (FISH) data of the 45S rDNA (Vaio et al. 2005). Indeed, an alliance of species related to P. intermedium was placed in one clade, whereas P. quadrifarium and P. quarinii were placed in the other. Interestingly, known allopolyploid species of groups ‘‘Virgata’’ and ‘‘Dilatata’’ were also scattered in both clades, suggesting different maternal affinities (Vaio et al. 2005, see below). The clade of P. quadrifarium and allied species is characterized by having upper florets with a bowed lemma and a more or less concave palea, an upper glume without folded margins, rhizomes with cataphylls, and dorsally rounded leaf sheaths, although all characters show further reversions and no one is exclusive. This clade includes the ‘‘Paniculata’’ group, represented by P. paniculatum and P. juergensii, and, unexpectedly, a clade comprising a group of species morphologically related to P. humboldtianum which includes P. paucifolium (currently in the ‘‘Eriantha’’ group), the odd P. falcatum, and the species included in the subgenus Anachyris. The morphological relationship among P. humboldtianum, P. polyphyllum, and P. paucifolium was already shown by Parodi and Nicora (unpubl. manuscript) who included all these species in the informal group ‘‘Humboldtiana’’, no longer recognized by later authors. All these species are rhizomatous grasses and have an upper glume with pilose to ciliate margins. Additionally, the sampled cpDNA fragments of P. polyphyllum and P. bicilium differ in eight positions (two positions autamorphic of P. bicilium and six autapomorphic of P. polyphyllum) and a 19 bp insertion in the trnL-trnF spacer, suggesting these entities are not conspecific as currently recognized (Zuloaga and Morrone 2005). Nevertheless, this fact needs corroboration by further sampling at both the species and population levels. Paspalum falcatum is a very distinct species with no obvious morphological affinities, so that it has been alternatively placed in groups ‘‘Lachnea’’ (Chase, unpubl. manuscript), ‘‘Stellata’’ (Barreto 1974), or the monotypic ‘‘Falcata’’ (Zuloaga and Morrone 2005; Parodi and Nicora, unpubl. manuscript). The phylogenetic relationships resulting from our analysis were, however, not completely unpredictable, because P. falcatum shares with P. polyphyllum its growth habit and vegetative morphology, and with P. malacophyllum (subgenus Anachyris) its concavoconvex spikelets and ciliate rachis. Finally P. durifolium, a species included in the group Quadrifaria by Barreto (1966) and Gomes and Monteiro (1996), and left ungrouped by Zuloaga and Morrone 239 (2005), remained ungrouped within the CQPA clade. The situation of this known allopolyploid species is discussed below. Allopolyploid groups As stated above, the cladistic treatment of allopolyploid taxa implies several difficulties. The informal groups ‘‘Dilatata’’ and ‘‘Virgata’’ have been extensively studied and they have been shown to be entirely composed of allopolyploid species, on the basis of cytogenetic evidence (Fernandes et al. 1968; Burson et al. 1973; Burson 1978, 1979, 1983, 1991, 1995; Burson and Bennett 1976; Burson and Quarin 1982; Caponio and Quarin 1990; Honfi et al. 1990; Quarin and Caponio 1995). Chromosome-pairing data have shown that members of the groups ‘‘Dilatata’’ and ‘‘Virgata’’ have partially homologous genomes. The tetraploid members of the Dilatata group have been assigned the IIJJ genomic formula (reviewed in Speranza 2009) whereas the species included in the ‘‘Virgata’’ group, P. virgatum and P. conspersum have been assigned the genomic formulae I2I2JJ and IIJ2J2, respectively (Burson 1978; Burson and Quarin 1982). The I and J genomes were originally named after P. intermedium and P. juergensii belonging in the groups ‘‘Quadrifaria’’ and ‘‘Paniculata’’, respectively. On the basis of the assigned genomic denominations it is clear that the genome donors for the Dilatata and Virgata groups may have been different. The apomictic members of the Dilatata group include a third genome designated X for which no sound phylogenetic hypotheses are available. This genome is thought to have been contributed to the group by a paternal progenitor and consequently no information about its origin is expected to derive from the analysis of cpDNA data (Speranza 2009). Interestingly, in our analysis, these two allopolyploid groups with partially homologous genomes were placed differently within the CQPA clade. Whereas the two well characterized species belonging to the ‘‘Virgata’’ group, P. conspersum and P. virgatum, were nested within the P. intermedium alliance, which is thought to contribute I-type genomes, P. commune and the entire ‘‘Dilatata’’ complex were grouped within the clade containing P. quadrifarium nearer the ‘‘Paniculata’’ group, a proposed source of J-type genomes. While cpDNA sequence identity within the Dilatata group suggests a narrow genetic base for the origin of all of its member species, the members of the Virgata group seem to have originated independently with maternal contributions of different members of the P. intermedium alliance. Another species in which cytogenetic affinities with the IIJJ genomic combination have been found is P. durifolium (IIJJXX) (Burson 1985). The position of the maternal genome donor for this species remained unresolved. 123 240 Author's personal copy Whereas the J genomes have been reported for species of the ‘‘Paniculata’’ group only, the I-type genomes occur in species of at least the two clades into which the group ‘‘Quadrifaria’’ was split, and also in P. rufum, a species of the NPBT clade. Thus, it becomes clear from the phylogenetic relationships obtained from our data that the definitive identification of the genomic sources for the species of the ‘‘Dilatata’’ and ‘‘Virgata’’ groups is far from being achieved. The currently proposed sources of I genomes form a paraphyletic/polyphyletic assemblage, whereas the proposed donor of the J genome (presumably P. juergensii or P. paniculatum) is nested within a part of this assemblage. According to this, the ability to pair with the I-type genomes seems to be a plesiomorphic condition at least within the crown Paspalum species. The current assignment of the I and J genomes may have been biased by the diploid materials and knowledge available when the original crosses were made. A general understanding of the relationships among the main clades within the genus will provide the basis for a more systematic analysis of genomic relationships between the polyploid groups and diploid species in the future. Diversification of the genus and origin of the major clades Among the species diverging early in our phylogeny, Paspalum inaequivalve, P. flavum, P. orbiculatum, and P. conjugatum are creeping species inhabiting shaded forest areas or forest margins, and P. repens is a freely floating grass. Hence, most basally diverging lineages within Paspalum include hygrophytic grasses with plagiotropic culms. These species are sister to a poorly resolved backbone while branch lengths are relatively longer within the rest of the tree and autapomorphies are relatively abundant. This fact suggests an understory origin of the genus and subsequent relatively rapid radiation into open grassland habitats that gave rise to the groups containing most of the species of the genus. Interestingly, if such a scenario were correct, the evolutionary history of the genus Paspalum would have recapitulated the history of the family Poaceae as a whole (GPWG 2001). Major diversification and expansion of C4 grass-dominated ecosystems both globally and in South America is thought to have taken place during the Miocene–Pliocene boundary (Cerling et al. 1997; Jacobs et al. 1999). It can be hypothesized that the radiation that gave rise to the major clades of Paspalum may be chronologically linked to the same climatic repatterning and biological events that promoted the expansion of C4 grasses in South America and globally. The divergence of shade-tolerant species appearing in the basal grade may have taken place before the widespread availability of open grassland habitats. 123 G. H. Rua et al. Such a diversification timing is generally congruent with the chronological framework presented by Vicentini et al. (2008) where the phylogenetic origin of C4 grass clades precedes their expansion and dominance of grassland communities. Because of the abundance, species richness, and widespread distribution of Paspalum in South America, further analysis of ecological and biogeographical patterns within the genus on a sound phylogenetic context may potentially provide invaluable information of the recent biogeographical history of the continent. Acknowledgments We are indebted to Pamela Soltis for laboratory facilities at the University of Florida, USA, to the Willi Hennig Society for making available a sponsored version of TNT, and to J. Marcos Mirande for kindly making available his unpublished TNT script. The following persons collaborated with us during field collection work and/or made available plant material from their own collections: Lone Aagesen, Alejandro Asenjo, Isabel B. Boccaloni, M. Eugenia Buela, Irene Caponio, Ana M. Carrión, Julio Daviña, Juan R. Fernández, Luis Frank, M. Cecilia Gróttola, Diego H. Hojsgaard, Ana I. Honfi, Graciela Lavia, Regina C. Oliveira, Florencia Perret, Camilo L. Quarin, Yamil Ramı́rez, Germán G. Roitman, José L. Rosa, G. Mónica Tourn, Nicolás Trillo, José F. M. Valls. This work received financial support through grants PIP 6568 (CONICET, Argentina), UBACyT JG21, and G806 (University of Buenos Aires, Argentina), and to GHR from the ‘‘Conselho Nacional de Desenvolvimento Cientı́fico e Tecnológico (CNPq)’’, Brazil. GHR is a member of the ‘‘Carrera del Investigador’’ of the ‘‘Consejo Nacional de Investigaciones Cientı́ficas y Técnicas (CONICET)’’, Argentina. Appendix Morphological characters used for phylogenetic analysis. 1. Lower glume, presence: lacking [0], present [1]. 2. Lower glume, position: on the medial plane of the spikelet [0], turned to one side of the spikelet [1]. 3. Upper glume, presence: lacking [0], present [1]. 4. Upper glume, length relative to the lower lemma: nearly equal [0], conspicuously shorter [1]. 5. Upper glume, edge: not edgeforming [0], forming an edge around the lower lemma [1]. 6. Upper glume, distal portion: rounded [0], obtuse [1], acute [2], acuminate [3]. 7. Upper glume, apex: rounded [0], obtuse [1], apiculate [2], acute [3], acuminate [4], truncate [5]. 8. Upper glume, dorsum: flat [0], bowed [1]. 9. Upper glume, a tuft of long hairs at the base: lacking [0], present [1]. 10. Upper glume, consistency: hyaline, tiny [0], membranous [1]. 11. Upper glume, a marginal fringe of tuberculate hairs: lacking [0], present [1]. 12. Upper glume, number of lateral veins on each glume half: none [0], one [1], two [2], three [3], four or more [4]. 13. Upper glume, abaxial vein protrusion: not prominent [0], prominent [1]. 14. Upper glume, distal convergence of inner lateral veins: not convergent [0], convergent [1]. 15. Upper glume, midrib: lacking [0], present [1]. 16. Upper glume, vein distribution: equidistant, no veins approximate towards Author's personal copy A phylogenetic analysis of the genus Paspalum (Poaceae) margins [0], only outer lateral veins approximate towards margins [1], all lateral veins approximate towards margins [2]. 17. Upper glume, outer vein position: marginal [0], not marginal [1]. 18. Upper glume, marginal region: folded [0], flat [1]. 19. Upper glume, between-vein indumentum: glabrous to scabrous [0], pubescent at the base [1], fully pubescent [2], pubescent towards the apex [3]. 20. Upper glume, marginal region (beyond outer veins) indumentum: glabrous to scabrous [0], scarcely pubescent at base [1], pubescent [2]. 21. Upper glume, hair base: simple [0], with [tinged] cushions [1], both [2]. 22. Upper glume, apex: navicular or cucullate [0], flat [1]. 23. Upper glume, surface: smooth [0], transversely crumpled or wrinkled [1]. 24. Upper glume, whether flabby or not: tight [0], flabby [1]. 25. Upper glume, symmetry: symmetrical [0], asymmetrical [1]. 26. Lower lemma, distal portion: rounded [0], obtuse [1], acute [2], acuminate [3]. 27. Lower lemma, apex: rounded [0], obtuse [1], apiculate [2], acute [3], acuminate [4], truncate [5], emarginate [6]. 28. Lower lemma, dorsum: flat [0], bowed [1], concave [2], sulcate [3]. 29. Lower lemma, consistency: hyaline, tiny [0], membranous [1], laterally indurate [2]. 30. Lower lemma, a basal tuft of hairs: wanting [0], present [1]. 31. Lower lemma, a marginal fringe of tuberculate hairs: wanting [0], present [1]. 32. Lower lemma, between-vein indumentum: glabrous [0], distally pubescent [1], basally pubescent [2], fully pubescent [3]. 33. Lower lemma, number of lateral veins on each side: none [0], one [1], two [2], three [3]. 34. Lower lemma, midvein: lacking [0], present [1]. 35. Lower lemma, distal convergence of inner lateral veins: not convergent [0], convergent [1]. 36. Lower lemma, marginal region: folded [0], flat [1]. 37. Lower lemma, marginal region (beyond outer veins) indumentum: glabrous [0], pubescent [1]. 38. Lower lemma, apex: navicular to cucullate [0], flat [1], folded [2]. 39. Lower lemma, internerval space: smooth [0], wrinkled [1]. 40. Lower lemma, axillary flower: wanting [0], reduced to a palea [1], well developed [2]. 41. Upper floret, pigmentation: pale to stramineous [0], brown [1], dark brown, shining [2], purplish [3]. 42. Upper floret, induration: not indurate [0], slightly indurate [1], strongly indurate [2]. 43. Upper floret, length relative to the lower lemma: nearly equal [0], conspicuously shorter [1]. 44. Upper floret, shape: elliptical [0], ovate [1], obovate [2], rhomboid [3], orbicular [4]. 45. Upper floret, basal constriction: lacking [0], present [1]. 46. Upper floret, callus: glabrous [0], laterally tufted [1], tufted around [2]. 47. Upper floret, lemma dorsum: more or less flattened [0], bowed towards the base [1], sharply bowed [2], gibbose [3]. 48. Upper floret, palea dorsum: convex [0], flat [1], concave [2]. 49. Upper floret, lemma apex: acute, pointed [0], acute but blunt at the very apex [1], rounded [2], acuminate [3], obtuse [4]. 50. Upper floret, lemma nerves: not prominent [0], sharply prominent 241 [1]. 51. Upper floret, epidermal papillae on the lemma: wanting [0], present [1]. 52. Upper floret, abscission at maturity: none [0], occurring [1]. 53. Upper floret, apex indumentum: wanting [0], present [1]. 54. Upper floret, palea margins: not winged [0], with non-overlapping wings [1], with overlapping wings [2]. 55. Upper floret, palea adaxial surface: smooth [0], papillose [1]. 56. Upper floret, lemma apex: open [0], cucullate [1]. 57. Upper floret, lemma margins: not thickened [0], slightly thickened [1], sharply thickened [2]. 58. Upper floret, number of lateral veins on each lemma half: none [0], one [1], two [2], three or more [3]. 59. Upper floret, midvein: lacking [0], present [1]. 60. Upper floret, lemma [inner]-nerves: distant from margins [0], submarginal [1]. 61. Upper floret, anther pigmentation: yellow [0], purple-tinged [1], deep purple [2]. 62. Pigmentation of stigmas: whitish [0], lila [1], purple [2], yellow [3]. 63. Upper floret, caryopsis hilum: punctiform [0], elliptical [1], linear [2]. 64. Inflorescence, terminal spikelet: wanting [0], present [1]. 65. Inflorescence, arrangement of primary branches: several branches along an axis with conspicuous internodia [0], several branches, the two distal ones conjugate [1], only two conjugate primary branches [2], one branch alone [3]. 66. Inflorescence, arrangement of primary branches II: all alternate [0], verticillate at least in the lower nodes [1]. 67. Maximum number of orthostichies: none [0], one [1], two [2], three [3], four [4]. five [5], six [6]. 68. Inflorescence, main axis cross section: polygonal [0], flattened [1]. 69. Inflorescence, pubescence on pulvinula: glabrous [0], [shortly] pubescent [1]. 70. Inflorescence, long cilia arising from pulvinula: wanting [0], present [1]. 71. Inflorescence, a spikelet ending each primary branch: present [0], lacking, lateral spikelets becoming rudimentary towards apex [1]. 72. Inflorescence, rachis cross section: trigonous, not expanded [0], laterally expanded into wings having chlorenchyma [1], laterally expanded into membranous epidermal wings [2]. 73. Inflorescence, rachis, sclerenchima ring: lacking [0], present [1]. 74. Inflorescence, rachis, medullar lacunae: lacking [0], present [1]. 75. Inflorescence, rachis venation: a midnerve thick and prominent [0], several parallel equal-range nerves [1]. 76. Inflorescence, rachis margin: smooth [0], scabrous [1], with more or less scattered cilia [2], with a dense fringe of tuberculate cilia [3]. 77. Inflorescence, rachis: straight to slightly sinuous [0], sharply sinuous [1]. 78. Inflorescence, rachis surface indumentum: glabrous [0], scabrous [1], pubescent [2]. 79. Inflorescence, arrangement of spikelets: solitary [0], homogeneously paired [1], proximally paniculate, distally paired [2]. 80. Inflorescence, concrescence of sPc branchlets with the rachis: (nearly) free, the spikelets appear pedicellate [0], concrescent, the spikelets appear subsessile [1]. 81. Inflorescence, a crown of hairs at the top of pedicells: lacking [0], present [1]. 82. Inflorescence, pedicells 123 242 Author's personal copy cross section: terete [0], trigonous/tetragonous [1], flattened [2]. 83. References Aliscioni SS, Denham SS (2008) Rachis of the genus Paspalum L. (Poaceae: Panicoideae: Paniceae): anatomy and taxonomic significance of the primary branches of the inflorescences. Flora 203:60–76 Aliscioni SS, Denham SS (2009) Atypical foliar anatomy related to Kranz syndrome in Paspalum inaequivalve and Paspalum microstachyum (Poaceae: Panicoideae: Paniceae). Flora 204:718–729 Aliscioni SS, Giussani LM, Zuloaga FO, Kellogg EA (2003) A molecular phylogeny of Panicum (Poaceae: Paniceae): tests of monophyly and phylogenetic placement within the Panicoideae. Am J Bot 90:796–821 Allem AC, Valls JFM (1987) Recursos forrageiros nativos do Pantanal mato-Grossense. Empresa Brasileira de Pesquisa Agropecuária. Centro Nacional de Recursos Genéticos, Brazil Álvarez I, Wendel JF (2003) Ribosomal ITS sequences and plant phylogenetic inference. Mol Phyl Evol 29:417–434 Bachmann K (2000) Molecules, morphology and maps: new directions in evolutionary genetics. Plant Species Biol 15:197–210 Barreto IL (1954) Las especies afines a Paspalum virgatum en la América del Sur. Rev Arg Agr 21:125–142 Barreto IL (1957) Las especies de Paspalum con dos racimos conjugados en Rı́o Grande del Sur (Brasil). Rev Arg Agr 24:89–117 Barreto IL (1966) Las especies afines a Paspalum quadrifarium (Gramineae) en la América del Sur de clima subtropical y templado. Darwiniana 14:130–160 Barreto IL (1974) O gênero Paspalum (Gramineae) no Rio Grande do Sul. Thesis, ‘Livre docência’ Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brazil Barreto IL (1983) Una nueva especie de Paspalum (Gramineae) del sur de Brasil. Bonplandia 5:291–293 Bennett HW (1962) Dallisgrass, bahiagrass, and vaseygrass. In: Hughes HD, Heath ME, Metcalfe DS (eds) Forages. The science of grassland agriculture, 2nd edn. The Iowa State University Press, Ames, pp 281–285 Bremer K (1994) Branch support and tree stability. Cladistics 10:295–304 Burman AG (1985) The genus Thrasya H.B.K. (Gramineae). Acta Bot Venez 14(4):7–93 Burson BL (1978) Genome relations between Paspalum conspersum and two diploid Paspalum species. Can J Gen Cytol 20:365–372 Burson BL (1979) Cytogenetics of Paspalum urvillei 9 P. intermedium and P. dilatatum 9 P. paniculatum hybrids. Crop Sci 19:534–538 Burson BL (1983) Phylogenetic investigation of Paspalum dilatatum and related species. In: Smith JA, Hays VW (eds) Proceedings of 14th international grassland congress, Lexington, Kentucky, USA, 1981. Westview Press, Boulder, pp 170–173 Burson BL (1985) Cytology of Paspalum chacoense and P. durifolium and their relationship to P. dilatatum. Bot Gaz 146:124–129 Burson BL (1991) Genome relationships between tetraploid and hexaploid biotypes of dallisgrass, Paspalum dilatatum. Bot Gaz 152:219–223 Burson BL (1995) Genome relationship and reproductive behavior of intraspecific Paspalum dilatatum hybrids: yellow-anthered 9 Uruguaiana. Int J Plant Sci 56:326–331 123 G. H. Rua et al. Burson BL, Bennett EC (1976) Cytogenetics of Paspalum conspersum and its genomic relationship with yellow-anthered P. dilatatum and P. malacophyllum. Can J Gen Cytol 18:708–791 Burson BL, Quarin CL (1982) Cytology of Paspalum virgatum and its relationship with P. intermedium and P. jurgensii. Can J Gen Cytol 24:219–226 Burson BL, Lee H, Bennett EC (1973) Genome relations between tetraploid Paspalum dilatatum and four diploid Paspalum species. Crop Sci 13:139–143 Canto-Dorow TS, Longhi-Wagner HM, Valls JFM (1996) Revisão taxonômica das espécies de Paspalum L. grupo Notata (PoaceaePaniceae) do Rio Grande do Sul, Brasil. Iheringia 47:3–44 Caponio I, Quarin CL (1990) Intra- and interspecific hybridization between dallisgrass and vaseygrass. Crop Sci 30:362–364 Cerling TE, Harris JM, MacFadden BJ, Leakey MG, Quade J, Eisenmann V, Ehleringer JR (1997) Global vegetation change through the Miocene/Pliocene boundary. Nature 389:153–158 Chase A (1929) The North American species of Paspalum. Contributions from the United States National Herbarium, vol 28, Part 1 Denham SS (2005) Revisión sistemática del subgénero Harpostachys de Paspalum (Poaceae, Panicoideae, Paniceae). Ann Missouri Bot Gard 92:463–532 Denham SS, Zuloaga FO (2007) Phylogenetic relationships of the ‘Decumbentes’ group of Paspalum, Thrasya and Thrasyopsis (Poaceae: Panicoideae: Paniceae). Aliso 23:545–562 Denham SS, Zuloaga FO, Morrone O (2002) Systematic revision and phylogeny of Paspalum subgenus Ceresia (Poaceae: Panicoideae: Paniceae). Ann Mo Bot Gard 89:337–399 Denham SS, Morrone O, Zuloaga FO (2010) Estudios en el género Paspalum (Poaceae, Panicoideae, Paniceae): Paspalum denticulatum y especies afines. Ann Mo Bot Gard 97:11–33 Duvall MR, Noll JD, Minn AH (2001) Phylogenetics of Paniceae (Poaceae). Am J Bot 88:1988–1992 Farris JS, Källersjö M, Kluge AG, Bult C (1995) Testing significance of incongruence. Cladistics 10:315–319 Fernandes MIBM, Barreto IL, Salzano FM (1968) Cytogenetic, ecologic and morphologic studies in Brazilian forms of Paspalum dilatatum. Can J Genet Cytol 10:131–138 Filgueiras TS (1992) Gramı́neas forrageiras nativas no Distrito Federal. Brasil Pesqui Agropecu Bras 27:1103–1111 Giussani LM, Cota-Sánchez H, Zuloaga FO, Kellogg EA (2001) A molecular phylogeny of the grass subfamily Panicoideae (Poaceae) shows multiple origins of C4 photosynthesis. Am J Bot 88:1993–2012 Giussani LM, Zuloaga FO, Quarin CL, Cota-Sánchez H, Ubayasena K, Morrone O (2009) Phylogenetic relationships in the genus Paspalum (Poaceae: Panicoideae: Paniceae): an assessment of the Quadrifaria and Virgata informal groups. Syst Bot 34:32–43 Goloboff PA (1993) Estimating character weights during tree search. Cladistics 9:83–91 Goloboff PA, Farris JS, Källersjö M, Oxelman B, Ramı́rez MJ, Szumik CA (2003a) Improvements to resampling measures of group support. Cladistics 19:324–332 Goloboff PA, Farris JS, Nixon KC (2003b) TNT: tree analysis using new technology. Program and documentation. http://www. zmuc.dk/public/phylogeny. Accessed 12 Nov 2009 Goloboff PA, Farris JS, Nixon KC (2008) TNT, a free program for phylogenetic analysis. Cladistics 24:774–786 Gomes MJIR, Monteiro R (1996) As espécies de Paspalum L. (Poaceae) dos grupos Virgata e Quadrifaria no Brasil—distribução e ecologia. Arq Biol Tecnol 39:1013–1019 Gómez-Martı́nez R, Culham A (2000) Phylogeny of the subfamily Panicoideae with emphasis on the tribe Paniceae: evidence from the trnL-F cp DNA region. In: Jacobs WL, Everett J (eds) Grass systematics and evolution, vol. 2 of proceedings of the second Author's personal copy A phylogenetic analysis of the genus Paspalum (Poaceae) international conference on the comparative biology of the monocots, Sydney, September 1998. CSIRO, Melbourne, pp 136–140 GPWG (Grass Phylogeny Working Group) (2001) Phylogeny and subfamilial classification of the grasses (Poaceae). Ann Mo Bot Gard 88:373–457 Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser 41:95–98 Holmgren PK, Holmgren NH (1998, continuously updated) Index herbariorum: a global directory of public herbaria and associated staff. New York Botanical Garden’s Virtual Herbarium, New York. http://sweetgum.nybg.org/ih/. Accessed 12 Nov 2009 Honfi AI, Quarin CL, Valls JFM (1990) Estudios cariológicos en gramı́neas sudamericanas. Darwiniana 30:87–94 Huelsenbeck JP, Ronquist F (2001) MR. BAYES: Bayesian inference of phylogeny. Bioinformatics 17:754–755 Jacobs BF, Kingston JD, Jacobs L (1999) The origin of grassdominated ecosystems. Ann Mo Bot Gard 86:590–643 Judziewicz EJ (1990) Poaceae (Gramineae). In: Görts-van Rijn ARA (ed) Flora of the Guianas. Koeltz, Königstein, pp 1–725 Kornet DJ, Turner H (1999) Coding polymorphism for phylogeny reconstruction. Syst Biol 48:365–379 Linder CR, Riesberg LH (2004) Reconstructing patterns of reticulate evolution in plants. Am J Bot 9:1700–1708 Martı́nez EJ, Quarin CL (1999) Citoembriologı́a y comportamiento reproductivo de un citotipo diploide de Paspalum hydrophilum y sus hı́bridos con P. palustre (Poaceae, Paniceae). Darwiniana 37:243–251 McDade L (1992) Hybrids and phylogenetic systematics: II. The impact of hybrids on cladistic analysis. Evolution 46:1329–1346 Mirande JM (2009) Weighted parsimony phylogeny of the family Characidae (Teleostei: Characiformes). Cladistics 25:574–613 Morrone O, Denham SS, Aliscioni SS, Zuloaga FO (2000) Revisión de las especies del género Paspalum (Panicoideae: Paniceae), subgénero Anachyris. Candollea 55:105–135 Morrone O, Denham SS, Zuloaga FO (2004) Revisión taxonómica del género Paspalum grupo Eriantha (Poaceae, Panicoideae, Paniceae). Ann Mo Bot Gard 91:225–246 Nicora EG, Rúgolo de Agrasar ZE (1987) Los géneros de Gramı́neas de América Austral. Hemisferio Sur, Buenos Aires Nixon KC (1999) The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics 15:407–414 Nixon KC, Carpenter JM (1996) On simultaneous analysis. Cladistics 12:221–241 Norrmann GA, Quarin CL, Killeen TJ (1994) Chromosome numbers in Bolivian grasses (Gramineae). Ann Mo Bot Gard 81:768–774 Nylander JAA (2004) MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University Oliveira RC (2004) Estudo taxonômico das espécies de Paspalum L., grupo Plicatula (Poaceae-Panicoideae-Paniceae) no Brasil. PhD thesis, University of Campinas, Campinas, SP, Brazil Oliveira RC, Valls JFM (2002) Taxonomia de Paspalum L., grupo Linearia (Gramineae-Paniceae) do Brasil. Rev Bras Bot 25:371–389 Oliveira RC, Valls JFM (2008) Novos sinônimos e ocorrências em Paspalum L. (Poaceae). Hoehnea 35:289–295 Pilger RKF (1941) Gramineae, Panicoideae. In: Engler A, Prantl KAE (eds) Die natürlichen Pflanzenfamilien, vol 14e, 2nd edn. W. Engelmann, Leipzig, pp 1–208 243 Quarin CL, Caponio I (1995) Cytogenetics and reproduction on P. dasypleurum and its hybrids with P. urvillei and P. dilatatum ssp. flavescens. Int J Plant Sci 156:232–235 Quarin CL, Norrmann GA (1990) Interspecific hybrids between five Paspalum species. Bot Gaz 151:366–369 Ramı́rez M (2006) Further problems with the incongruence length difference test: ‘‘hypercongruence’’ effect and multiple comparisons. Cladistics 22:289–295 Rieseberg LH, Soltis DE (1991) Phylogenetic consequences of cytoplasmic gene flow in plants. Evol Trend Plant 5:65–84 Rua GH, Aliscioni SS (2002) A morphology-based cladistic analysis of Paspalum sect. Pectinata (Poaceae). Syst Bot 27:489–501 Rua GH, Speranza PR, Arakaki M, Vaio M (2007) Filogenia del género Paspalum (Poaceae, Panicoideae): Análisis preliminares de secuencias de 4 fragmentos del cpDNA. Darwiniana 45:120–121 Sang T (2002) Utility of low-copy nuclear gene sequences in plant phylogenies. Crit Rev Biochem Mol Biol 37:121–147 Scataglini MA, Giussani LM, Denham SS, Zuloaga FO, Morrone O (2007) Una aproximación a la filogenia de Paspalum ((Poaceae, Panicoideae, Paniceae) utilizando tres marcadores de ADN de cloroplasto. Darwiniana 45:124–125 Souza-Chies TT, Essi L, Rua GH, Valls JFM, Miz R (2006) A preliminary approach to the phylogeny of the genus Paspalum (Poaceae). Genetica 126:15–32 Speranza P (2009) Evolutionary patterns in the Dilatata group (Paspalum, Poaceae). Plant Syst Evol 282:43–56 Speranza P, Malosetti M (2007) Nuclear and cytoplasmic microsatellite markers for the species of the Dilatata group of Paspalum (Poaceae). Plant Genet Resour 5:14–26 Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22:4673–4680 Vaio M, Speranza P, Valls JFM, Guerra M, Mazzella C (2005) Localization of the 45S and 5S rDNA and cpDNA sequence analysis in species of the Quadrifaria group of Paspalum (Poaceae, Paniceae). Ann Bot 66:191–200 Vicentini A, Barber JC, Aliscioni SS, Giussani LM, Kellogg EA (2008) The age of the grasses and clusters of origins of C4 photosynthesis. Global Change Biol 14:2963–2977 Vriesendorp P, Bakker FT (2005) Reconstructing patterns of reticulate evolution in angiosperms: what can we do? Taxon 54:593–604 Zuloaga FO, Morrone O (2001) Flora fanerogamica Argentina. Fasc. 71. 19. Poaceae, parte 4. Tribu 18. Paniceae, parte B. 21. Paspalum. ProFlora Conicet, Córdoba Zuloaga FO, Morrone O (2005) Revisión de las especies de Paspalum para América del Sur austral (Argentina, Bolivia, sur del Brasil, Chile, Paraguay y Uruguay). Monogr Syst Bot Mo Bot Gard 102:1–297 Zuloaga FO, Pensiero J, Morrone O (2004) Systematics of Paspalum group Notata (Poaceae-Panicoideae-Paniceae). Syst Bot Monogr 71:1–75 Zuloaga FO, Giussani LM, Morrone O (2007) Hopia, a new monotypic genus segregated from Panicum (Poaceae). Taxon 56:145–156 123

RELATED PAPERS

RELATED TOPICS

Log In

A phylogenetic analysis of the genus Paspalum (Poaceae) based on cpDNA and morphology

A phylogenetic analysis of the genus Paspalum (Poaceae) based on cpDNA and morphology

Related Papers

RELATED PAPERS

RELATED TOPICS