Journal of Systematics and Evolution 47 (5): 431–443 (2009) doi: 10.1111/j.1759-6831.2009.00039.x Phylogeny and evolution of Perezia (Asteraceae: Mutisieae: Nassauviinae) 1 Beryl B. SIMPSON∗ 1 2 Mary T. K. ARROYO 1 Sandra SIPE 1 Joshua McDILL 3 Marta DIAS de MORAES (Integrative Biology and Plant Resources Center, The University of Texas, 1 University Station A6700, Austin, TX 78712, USA) 2 (Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago de Chile, Chile) 3 (Universidade Federal do Acre, Campus Floresta, Cruzeiro do Sul, Acre, Brazil) Abstract A molecular phylogenetic analysis of most of the species of Perezia reveals that, as traditionally defined, the genus is not monophyletic with two species more closely related to Nassauvia than to Perezia. In addition, our results show that Burkartia (Perezia) lanigera is related to Acourtia and is the only member of that clade in South America. The remaining species are monophyletic and show a pattern of an early split between a western temperate and an eastern subtropical clade of species. Within the western clade, the phylogeny indicates a pattern of diversification that proceeded from southern, comparatively low-elevation habitats to southern high-elevation habitats, and ultimately into more northern high-elevation habitats. The most derived clades are found in the high central Andes, where significant radiation has occurred. Key words Andes, biogeography, Mutisieae, Nassauviinae, Perezia. Perezia Lag., a genus of 30–35 primarily highelevation species, occurs exclusively in South America, primarily in the central and southern Andes, and thus constitutes a useful model for examining the evolution of high-elevation floras in South America. Considered until 1970 (Vuilleumier, 1970) to be the nominate section of a larger genus that also included Perezia section Acourtia (D. Don) A. Gray, a North American taxon, Perezia is now known to be distinct from Acourtia D. Don and more closely related to South American genera. Species of Perezia occur from sea level in Chile and eastern Argentina to over 4000 m above sea level (a.s.l.) in Bolivia and Peru and from Tierra del Fuego to Colombia. Species range across most of the high Andean habitats, except true páramo, with plant habit often correlated with habitat. The genus has historically included large foliose species that occur in the Nothofagus forests of the southern cone and in the Paraguay– southern Brazil–Uruguay basin, as well as tiny rosettes that grow in the very high-elevation puna of central Bolivia. Perezia is a member of the Nassauviinae, a subtribe of the Mutisieae with 25–27 exclusively America genera (Hind, 2007). Perezia is the fifth largest genus in the subtribe (Crisci, 1980). Vuilleumier (1970) monographed the group (as Perezia sect. Perezia; the other ∗  C Received: 11 March 2009 Accepted: 2 May 2009 Author for correspondence. E-mail: beryl@mail.utexas.edu; Tel.: 512-4714335; Fax: 512-23-29529. 2009 Institute of Botany, Chinese Academy of Sciences section, Acourtia, having been monographed previously by Bacigalupi in 1931). In the 1970 monograph, Vuilleumier recognized 30 species. Since that work was published, four species have been added (or re-added) to the genus based on morphology, namely P. lanigera Hook. and Arn. (but see below), P. eryngioides (Cabrera) Crisci & Marticorena (described as a Trixis but transferred to Perezia by Crisci and Marticorena), P. catharinensis Cabrera, and P. volcanensis Cabrera. Two species placed by Vuilleumier (1970) in synonymy under P. purpurea (i.e. P. atacamensis Phil. and P. burkartii Cabrera) were treated as distinct species by Cabrera (1978). Of these taxa, Perezia lanigera Hook. & Arn. has engendered the most controversy. The entity was originally described as a Perezia but, in her monograph, Vuilleumier (1970) excluded it from the genus because of its aspect and the presence of wooly trichomes in the leaf axils. No Perezia species has this type of trichome. However, Cabrera (1971) considered these trichomes of little importance and consistently treated the species as a Perezia. Crisci (1976), like Vuilleumier (1970), considered the taxon distinct and erected the monotypic genus Burkartia Crisci for it, with the comment that Lophopappus Rusby was its closest relative. In 1970, Vuilleumier suggested a potential relationship between Perezia and Leucheria Lag. based on morphology. Cassini (1825) had earlier placed Perezia with Holocheilus Cass., Leucheria, and Trixis P. Browne in a subgroup of one of his three sections of the Nassauviinae, but Crisci (1974) was the first to address relationships of the genera within the subtribe 432 Journal of Systematics and Evolution Vol. 47 No. 5 in a rigorous way. He scored a wide array of discrete characters, five geographical and 85 morphological (including pollen shape and exine patterning), for 26 taxa that he used in a numerical taxonomic study. In that study, he used various combinations of two different scoring methods and two different measures of similarity to generate three cladograms (Crisci, 1974). The results were, in general, concordant with one another and allowed him to draw several conclusions. One of the conclusions was that the two sections of Perezia were not “close” to one another and should be treated as distinct genera (as suggested previously by Vuilleumier 1970). Second, he commented that Nassauvia and Triptilion were very closely “related”, as well as that Perezia lanigera was more closely “related” to a cluster containing Proustia Lag., Lophopappus Rusby, Acourtia, and Gochnatia glomeriflora A. Gray than to Perezia. Finally, the results suggested that Perezia and Leucheria were each rather isolated within the subtribe, possibly as a “result of the great spectrum of types presented by the two genera” (Crisci, 1974). Shortly after Crisci’s study, Reveal and King (1973) formally re-elevated Acourtia to generic status and provided the necessary new combinations for the caulescent species. A few years later, Turner (1978) moved the North American scapiform Perezia species into Acourtia. Six years after his numerical taxonomic study, Crisci (1980) used the same data set to produce phylogenetic hypotheses of relationships in the Nassauviinae. Character polarity was determined based on several criteria and trees were constructed using a Wagner tree algorithm. Three trees were produced, each with a different outgroup. With a hypothetical outgroup, Perezia was sister to Panphalea DC. and Holocheilus was sister to this pair. This same relationship was produced when Trixis was used as the outgroup. With Dolichlasium Lag. as the outgroup, Perezia was sister to a clade consisting of Panphalea Lag., Moscharia Ruı́z & Pavón, Polyachyrus Lag., Calopappus Meyen, Nassauvia Comm. ex Juss., and Triptilion Ruı́z & Pavón. Within the past 10 years, there have been two published molecular studies that included several members of the subtribe Nassauviinae. Using ndhF sequence data and including representatives of Acourtia, Adenocaulon Hook., Jungia L.f., Leucheria, Nassauvia, Perezia, and Triptilion, Kim et al. (2002) reported that the Nassauviinae had only weak support as a clade but that within the subtribe there were several well-supported subclades, one of which included Perezia as sister to a Nassauvia/Triptilion clade. Acourtia was sister to a clade containing Trixis and Proustia (although this relationship collapsed in the strict consensus tree). Leucheria was in a third clade, sister to Jungia. In a more re- 2009 cent study, Katinas et al. (2008) used internal transcribed spacer (ITS) and trnL-F sequence data to generate a phylogeny that they used to infer the evolution of secondary heads in Nassauviinae. Their phylogeny included 12 of the 25 genera of the tribe: Ameghinoa Speg., Dolichlasium Lag., Holocheilus Cass., Jungia, Leucheria, Moscharia, Nassauvia, Panphalea, Perezia (four species), Polyachyrus Lag., Proustia Lag., and Triptilion. Their results showed Perezia as sister to Panphalea and this clade sister to a clade of Nassauvia plus Triptilion. In a paper circumscribing a segregate genus of Perezia, namely Calorezia Panero, Panero (2007), stated that his unpublished data showed that Perezia nutans Less. (and, by association, P. prenanthoides Less.) was more closely related to Calopappus Meyen, Nassauvia, and Triptilion than to the rest of Perezia and, hence, necessitated a new genus. Panero (2007) considered these three genera along with Panphalea and Perezia to form a “Perezia clade”, called the Perezia group in our discussions. Within Perezia, only Vuilleumier (1970) has made inferences about species relationships. In her 1970 revision, she conducted a numerical taxonomic study of the genus using 24 numerical and 23 non-numerical characters that were measured or scored for over 1200 herbarium specimens. Each species was usually represented by several different “populations” (two or more specimens from the same locality). Each of these populations was designated as a terminal taxonomic unit, with the numerical characters of each unit represented by a mean and variance (calculated from the cluster of specimens measured in the population). These values were used to generate dendrograms based on a Mahalanobis’ generalized distance matrix. The dendrograms were rooted by the a priori placement of P. pungens (H. & B.) Less. as the first taxonomic unit. Perezia nutans and P. prenanthoides and P. multiflora (H. & B.) Less. and similar leafy species tended to be at the base of the dendrogram, but most of the remainder of the terminal units (populations) came out stepwise with terminal taxonomic units of several variable species scattered across the dendrogram. Although there were some cohesive clusters, the dendrograms did not provide easily interpretable relationships among the terminal taxonomic units included. Using the generalized distances as a rough guide and combining them with the morphology underlying them, Vuilleumier (1970) suggested six species groups shown in Fig. 1 with distributions indicated in Fig. 2: A, B. Assessing these groups together with paleoecological data, Vuilleumier (1970) postulated that the genus arose early in the Tertiary in the warm open forests that covered extratropical South America. Specifically, she  C 2009 Institute of Botany, Chinese Academy of Sciences SIMPSON et al.: Phylogeny of the Andean genus Perezia 433 Fig. 1. Species groups modified from Vuilleumier (1970). Note that the prenanthoides and the multiflora groups were considered quite distinct, whereas taxa such as Perezia pilifera and Perezia carduncelloides could not be placed with certainty. Fig. 2. Distribution of the species groups of Perezia (A) as delineated by Vuilleumier (1970) and the number of species in various parts of the generic distribution (B). The species of the former prenanthoides group are excluded, but their distribution is indicated by the brackets that show the northern and southern limits of their distribution in the Nothofagus forests of southern South America. The additional species of the multiflora group added after 1970 are included in both A and B.  C 2009 Institute of Botany, Chinese Academy of Sciences 434 Journal of Systematics and Evolution Vol. 47 No. 5 suggested that the ancestral Perezia was similar in habit to Perezia pungens and inhabited mid-to-low elevation montane habitats of what is now the central Andes. As drying began in the mid-Tertiary, Vuilleumier (1970) postulated a split leading to the ancestor of the multiflora group in southern Brazil–eastern Argentina, the ancestor of the prenanthoides group in the Nothofagus forest, and the ancestor of the remaining species initially at mid altitudes in the central Andean region. She suggested that this last group radiated in the high Andes and southern Patagonia with speciation linked to drying in the late Tertiary, the uplift of the Andes, and Pleistocene climatic fluctuations. Our purpose here is to generate hypotheses of infrageneric species relationships of Perezia using molecular data in order to assess the directions of change in habit and habitat, and to test the patterns of relationships suggested by Vuilleumier (1970). To establish a phylogeny that allows us to examine these patterns, we have included most of the species of Perezia. Based on the studies of Kim et al. (2002), Panero (2007), and Katinas et al. (2008), we have included species of Acourtia, Nassauvia, Panphalea, and Triptilion (all previously linked with Perezia in various ways), as well as Adenocaulon and Lophopappus, two other genera in the Nassauviinae. 1 Material and methods 1.1 Material Material was obtained from herbarium specimens (with permission) or collected in the field. Vouchers and GenBank numbers are listed in Table 1. Included in the ingroup were 28 of the 30–35 species of Perezia, two accessions of Burkartia (Perezia) lanigera, the two former species of Perezia now placed in Calorezia, seven species of Acourtia, five species of Nassauvia, three species of Adenocaulon, and one species each of Leucheria, Panphalea, and Triptilion. Leucheria was designated as the outgroup for the purposes of rooting. 1.2 Methods 1.2.1 DNA Sequencing DNA was isolated from herbarium specimens or silica-dried leaf material using a modified cetyltrimethylammonium bromide (CTAB) protocol (Loockerman & Jansen, 1996). A PCR was used to amplify the ITS region using primers P1 and P4 of Kim and Jansen (1994), with internal primers (P2 and P3) also used for amplification or sequencing when necessary. The chloroplast intergenic spacers rpl32–ndhF and trnL(UAG)–rpl32 were chosen based on their rates of evolution and degree of phylogenetic utility as reported 2009 by Shaw et al. (2007) and Timme et al. (2007). Pereziaspecific internal primers were designed for the present study to assist with amplification and sequencing when necessary. The sequences of the primers used for these regions are given in Table 2. The PCR reactions contained 2.5 µL of 10× PCR buffer; 2 µL of a 10 mmol/L stock solution of combined dNTPs, 2–4 µL of 25 mmol/L MgCl2 , 0.25 µL of a 25 µmol/L stock solution of each forward and reverse primer, 2 µL of 3.3% (w/v) bovine serum albumin (BSA), 1 unit Taq polymerase, 2–8 µL of 1:10 diluted template DNA extract, and water to a final volume of 25 µL. For amplification of ITS, 1.25 µL dimethylsulfoxide (DMSO) was added to the reaction. Annealing temperatures ranged between 45 ◦ C and 52 ◦ C depending on primers and template. The PCR reaction products were purified using exonuclease I and shrimp alkaline phosphatase to degrade unincorporated primers and dNTPs (Werle et al., 1994), and then sequenced via BigDye (v. 3.1) Terminator Cycle Sequencing (Applied Biosystems, Foster City, CA, USA) at the Institute for Cell and Molecular Biology Core Facility at the University of Texas (Austin, TX, USA). Forward and reverse sequence reads were assembled into contigs and edited in Sequencher 4.5 (GeneCodes Corp., 2005), whereas the chloroplast sequences were aligned manually in MacClade 4.08 (Maddison & Maddison, 2005), and ITS sequences were aligned using default settings in MUSCLE (Edgar, 2004), followed by manual adjustments in MacClade. 1.2.2 Phylogenetic analyses The three data matrices were analyzed separately and in combination using maximum parsimony (MP) with PAUP∗ 4.0b10 (Swofford, 2002), Bayesian inference (BI) using MrBayes 3.1.2 (Huelsenbeck & Ronquist, 2001), and maximum likelihood (ML) with GARLI (Zwickl, 2006). Areas of uncertain alignment were excluded from all analyses, as were autapomorphic insertions. The sequenced portions of the 18S and 25S ribosomal subunits were also excluded from the ITS matrix. In PAUP, all characters were treated as equally weighted and unordered, with gaps treated as missing data. The MP tree searches were conducted using the heuristic search option with 1000 randomaddition replicates and tree bisection–reconnection (TBR) branch swapping, with no limitations placed on the number of trees swapped to completion, and all other settings at defaults. For MP bootstrapping, 500 bootstrap replicates were performed using the heuristic search option with one random addition replicate and swapping to completion on a maximum of 10 000 trees. The incongruence length difference (ILD) test (Farris et al., 1995), implemented in PAUP as the  C 2009 Institute of Botany, Chinese Academy of Sciences SIMPSON et al.: Phylogeny of the Andean genus Perezia Table 1 Sources of material Taxon Acourtia coulteri (A. Gray) Reveal & R.M. King A. microcephala DC. A. nana (A. Gray) Reveal & R.M. King A. purpusii (Brandegee) Reveal & R.M. King A. scapiformis (Bacigalupi) Reveal & R.M. King A. runcinata (D. Don) B. L. Turner A. wrightii (A. Gray) Reveal & R.M. King Adenocaulon bicolor Hook. A. chilense Less. A. lyratum S.F. Blake Calopappus acerosus Meyen Leucheria bridgesii Hook. & Arn. Lophopappus foliosus Rusby Nassauvia aculeata Poepp. & Endl. N. digitata Wedd. N. heterophylla (Phil.) Reiche N. lagascae F. Meigen N. pinnigera D. Don Panphalea cardaminifolia Less. Perezia atacamensis (Phil.) Reiche∗ P. calophylla (Phil.) Reiche P. carduncelloides Griseb. P. carthamoides (D. Don) Hook. & Arn. P. ciliaris Hook. & Arn. P. ciliosa (Phil.) Reiche P. cirsiifolia Wedd.∗ P. coerulescens Wedd. P. fonkii (Phil.) Reiche P. integrifolia Wedd.∗ P. kingii Baker P. lactucoides lactucoides (Vahl) Less. P. lactucoides ssp. palustris (Phil.) Vuill. P. lanigera Hook. & Arn.∗ ∗ = Burkartia lanigera (Hook. & Arn.) Crisci P. linearis Less. P. lyrata (Remy) Wedd. P. magellanica (L.f.) Less. P. mandonii Rusby P. megalantha Speg. P. multiflora (H. & B.) Less. P. (Calorezia) nutans Less.∗ P. pedicularidifolia Less. P. pilifera (D. Don) Hook. & Arn. P. pinnatifida (Humb. & Bonpl.) Wedd. P. pungens (Humb. & Bonpl.) Less. P. (Calorezia) prenanthoides Less.∗ ∗ P. purpurata Wedd.  C 435 Voucher Origin Herbarium GenBank No. ITS trnL–rpl32 rpl32–ndhF H.H. Iltis 30784 Tamaulipas, Mexico TEX FJ979680 FJ979729 FJ979781 T. Ross 6616 B.L. Turner 7-22-07 California, USA Texas, USA TEX TEX FJ979679 FJ979682 N/A FJ979732 FJ979780 FJ979784 G.B. Hinton 23607 Nuevo Leon, Mexico TEX FJ979681 FJ979730 FJ979782 J. Calzada 21592 Oaxaca, Mexico TEX FJ979683 FJ979733 FJ979785 W. M. Turner 76 Texas, USA TEX FJ979684 FJ979734 FJ979786 E.J. Lott 4829 Texas, USA TEX N/A FJ979731 FJ979783 G. Helmkamp K-17 Ricardi & Marticorena 1923 D. E. Breedlove 13424 J. Panero & B. Crozier 8457 W. D. Clark 1350 Idaho, USA Malleco, Chile Chiapas, Mexico Los Andes, Chile Santiago, Chile TEX F F TEX TEX FJ979672 FJ979674 FJ979673 FJ979685 FJ979675 FJ979722 FJ979724 FJ979723 FJ979735 FJ979725 FJ979773 FJ979775 FJ979774 FJ979787 FJ979776 Sanders et al. 3325 Marticorena et al. 74 Jujuy, Argentina Curico, Chile TEX F FJ979676 FJ979688 FJ979726 FJ979738 FJ979777 FJ979790 K. Gengler et al. 11 C.P. Cowan 4250 C.P. Cowan 4238 C.P. Cowan 4239 M. Dias de Moraes 832 M.K. Arroyo et al. 94014 Reg. XIII, Chillan, Chile Farellones, Chile Farellones, Chile Farellones, Chile Santa Catarina, Brazil Reg. II Atacama, Chile TEX TEX TEX TEX TEX CONC FJ979690 FJ979687 FJ979686 FJ979691 FJ979669 FJ979657 FJ979740 FJ979737 FJ979736 FJ979741 FJ979719 FJ979707 FJ979792 FJ979789 FJ979788 FJ979793 FJ979770 FJ979758 B. S. Vuilleumier 189 B. B. Simpson 6-II-00-1 E. Wall s.n Rio Negro, Argentina Tucuman, Argentina Mendoza, Argentina GH TEX GH N/A FJ979655 FJ979641 FJ979700 FJ979705 FJ979692 FJ979751 FJ979756 FJ979742 St. Beck et al. 18088 St. Beck 26317 I. Henson 830 E. Garcia 886 Weigend et al. 6824 X. Menhofer X-1900 Rosengurth PE5334 O. Dollenz 648 Cochabamba, Bolivia Arequipa, Peru Cochabamba, Bolivia La Paz, Bolivia Rio Negro, Argentina La Paz, Bolivia Florida, Uruguay Magellanes, Argentina LPB LPB LPB LPB NY LPB GH GH FJ979644 FJ979645 FJ979646 FJ979649 FJ979660 FJ979654 FJ979667 FJ979659 FJ979694 FJ979695 FJ979696 FJ979699 FJ979710 FJ979704 FJ979717 FJ979709 FJ979745 FJ979746 FJ979747 FJ979750 FJ979761 FJ979755 FJ979768 FJ979760 B. Vuilleumier 204 Rio Negro, Argentina GH FJ979642 N/A FJ979743 S. Albert 8-XI-2006-1 S. Albert 8-XI-2006-2 Santa Cruz, Argentina Santa Cruz, Argentina TEX TEX FJ979677 FJ979678 FJ979727 FJ979728 FJ979778 FJ979779 Pirion 3499 Marticorena et al. 194 O. Dollenz 708 I. Henson 1505 E. Pisano V. 5602 M. Madison 1044 J. Wen 7472 F. Jaffuel 3795 M.T.K. Arroyo 20680 Aisen, Chile Reg. VII Talca, Chile Isla Wollaston, Argentina Cochabamba, Bolivia Cerro Corona, Argentina Cuzco, Peru Chile Chillan, Chile Yerba Loca, Chile GH CONC GH LPB GH GH F GH CONC FJ979664 FJ979666 FJ979661 FJ979647 FJ979651 FJ979652 FJ979671 FJ979662 FJ979658 FJ979714 FJ979716 FJ979711 FJ979697 FJ979702 N/A FJ979721 FJ979712 FJ979708 FJ979765 FJ979767 FJ979762 FJ979748 FJ979753 N/A FJ979772 FJ979763 FJ979759 Hutchison 4250 Lima, Peru GH FJ979650 FJ979701 FJ979752 S. King et al. 285 Urubamba, Peru GH FJ979653 FJ979703 FJ979754 G. Seijo 1671 Neuquen, Argentina NY FJ979670 FJ979720 FJ979771 St. Beck 31111 Oruro, Bolivia LPB FJ979643 FJ979693 FJ979744 2009 Institute of Botany, Chinese Academy of Sciences 436 Table 1 Journal of Systematics and Evolution Vol. 47 No. 5 2009 Continued Taxon Voucher Origin Herbarium GenBank No. ITS trnL–rpl32 rpl32–ndhF P. recurvata (Vahl) Less. E. Pisano V. 4045 Patagonia, Argentina GH FJ979663 FJ979713 FJ979713 P. squarrosa ssp. cubaetensis O.S. Ribas et al. 2152 Parana, Brazil TEX FJ979668 FJ979718 FJ979769 (Less.) Vuill. P. sublyrata Domke J. L. Luteyn & L. Door 13773 La Paz, Bolivia TEX FJ979656 FJ979706 FJ979757 P. virens (D. Don) Hook. & Arn. E. Wall 29.XII.1946 Aconcagua, Chile NY FJ979648 FJ979698 FJ979749 P. viscosa Less. G. Montero O. 1304 Cautin, Chile GH FJ979665 FJ979715 FJ979766 Triptilion spinosum Ruiz & Para & Rodriguez 109 Concepcion, Chile F FJ979689 FJ979739 FJ979791 Pavon ∗ Vuilleumier (1970) placed P. atacamensis in P. purpurata and P. cirsiifolia and P. integrifolia in P. coreulescens. Specimens referable to these synonyms are included here because of doubt expressed by Vuilleumier about their placement. ∗∗ These species have been moved to the genera indicated in parentheses. The authorities listed here are for the original description in Perezia. In the monograph of the genus. N/A, not applicable. Table 2 Primers used to amplify and sequence chloroplast intergenic spacers Region ndhF–rpl32 rpl32–trnL Primer Sequence (5′ –3′ ) Reference ndhF 316A 316B 602A 602B 906A 906B rpl32-R rpl32-F 432A 432B 534A 534B 649A 649B trnL (UAG) GAA AGG TAT KAT CCA YGM ATA TT GAG CAA GGA TAA AAA ATT AC GTA ATT TTT TAT CCT TGC TC CRT ATC CTT TAA CAG ATT K MAA TCT GTT AAA GGA TAY G GAG AGA TAA AGA ACG AGA AY RTT CTC GTT CTT TAT CTC TC CCA ATA TCC CTT YYT TTT CCA A CAG TTC CAA AAA AAC GTA CTT C CCC ATC GAC CTT TAC AAT AA TTA TTG TAA AGG TCG ATG GG GAA ATT CAT TGA TTC CAT G CAT GGA ATC AAT GAA TTT C GCY CAA AAC AGA ACT TAA TAG CTA TTA AGT TCT GTT TTG RGC CTG CTT CCT AAG AGC AGC GT Shaw et al., 2007 Present study Present study Present study Present study Present study Present study Shaw et al., 2007 Shaw et al., 2007 Present study Present study Present study Present study Present study Present study Shaw et al., 2007 partition homogeneity test, was used to assess conflict between the chloroplast and nuclear data, with 100 replicates performed using the heuristic search option under the same constraints as used for MP bootstraps. In GARLI, 100 bootstrap replicates were performed for each marker and for the combined matrix, using default settings and the GTR+G+I model. Bootstrap values were determined from a 50% majority rule consensus of the best trees found in each bootstrap replicate. Prior to analysis in MrBayes, the number of substitution types and applicability of gamma rate heterogeneity (G) or invariant sites (I) for each marker were determined with the MrModelTest (Nylander, 2004) using the Akaike Information Criterion (AIC). In MrBayes, each marker was subjected to four million generations of MCMC sampling, with tree topology, estimated model parameters, and likelihood score saved every 100 generations and the automated diagnostic statistic comparing the parameters from two simultaneous runs every ten-thousandth generation (with the first 25% of generations excluded as “burn-in”). Chain heating and priors for model parameters were kept at default values, ex- cept for the nucleotide frequency prior, which was set to a dirichlet distribution. For the combined data analysis in MrBayes, the matrix was partitioned to apply the appropriate substitution model (G) and I to each marker; partition model parameters were unlinked and allowed to vary independently for each partition, except for branch length and topology. All Bayesian analyses were terminated at four million generations as long as the automated diagnostic statistic (average standard deviation of split frequencies) was below 0.01 by that time. Runs were also checked for stationarity in the post-burnin sample by graphical plotting of −ln L scores against generation time. Clade posterior probabilities were determined from a 50% majority rule consensus of the post-burn-in sample of 60 000 trees (30 000 sampled during stationarity from each simultaneous run). 2 Results 2.1 Sequence characteristics Sequences of the ITS could not be obtained from Perezia calophylla and Acourtia wrightii. Perezia  C 2009 Institute of Botany, Chinese Academy of Sciences SIMPSON et al.: Phylogeny of the Andean genus Perezia Table 3 Marker Parameters for the DNA markers used in the present study Aligned length Included bp Variable sites rpl32 1049 829 265 ndhF–rpl32 1229 991 284 Combined CPL 2278 1820 549 ITS 762 599 315 CPL+ITS 3040 2419 864 N/A, not applicable; ITS, internal transcribed spacer; CPL, chloroplast loci. lactucoides subsp. palustris and Acourtia microcephala were lacking sequence from the rpl32–trnL intergenic spacer, and Perezia multiflora was missing sequence from both chloroplast markers. Table 3 provides descriptive statistics from parsimony analyses of the data matrices assembled for the present study and the ML substitution model type selected. Figure 3: A, B shows the majority rule consensus trees for the Bayesian analyses. The parsimony-based ILD test indicated significant conflict between the chloroplast and ITS data partitions (P = 0.01). Topological incongruences between the chloroplast and ITS (Fig. 3: A, B), mostly affecting placement of outgroup taxa, are discussed below. Consequently, we combined our data to generate the phylogeny shown in Fig. 4. 2.2 Phylogenetic results Figure 3 shows a comparison between the phylogenies generated with the combined chloroplast markers and the sequences from ITS 1 and 2. Considering Perezia, several discrepancies should be noted. First, in the chloroplast (cp) DNA tree (Fig. 3: A), P. lactucoides is sister to P. megalantha, whereas the ITS data (Fig. 3: B) show this species in a polytomy with the majority of the species of the genus. However, we were not able to obtain one of the chloroplast sequences from P. lactucoides subsp. palustris, which probably led to this difference. Second, P. virens in the chloroplast tree (Fig. 3: A) is part of a polytomy with members of the high Andean clade; however, in the ITS tree (Fig. 3: B) it branches much lower in the cladogram and is sister to P. linearis. Third, P. ciliosa is in an unresolved clade with P. cirsiifolia and P. atacamensis in the chloroplast tree (Fig. 3: A) but is sister to a clade of P. integrifolia, P. pinnatifida, and P. purpurata in the ITS tree (Fig. 3: B). Fourth, P. viscosa is part of a clade with P. linearis, P. pilifera, and P. recurvata in the chloroplast tree (Fig. 3: A) and in a clade with P. lyrata and P. pedicularidifolia in the ITS tree (Fig. 3: B). The latter placement seems more reasonable in terms of morphology. Finally, P. coerulescens is in a completely unresolved clade of central Andean species in both analyses, but the members of that clade differ in the two trees (Fig. 3: A, B).  C 437 2009 Institute of Botany, Chinese Academy of Sciences Informative sites Mean GC content (%) Substitution model 168 154 322 238 560 24.14 23.69 25.03 56.15 33.17 GTR+G GTR+G N/A GTR+I+G N/A Common to both analyses is the position of the P. multiflora group as sister to or in a basal polytomy with the remaining members of the genus (Fig. 3: A, B). Similarly, in the cp analysis (Fig. 3: A) there is a basal grade of comparatively low elevation, generally humid habit, southern South American species (P. calophylla and P. fonkii) subsequent to the P. multiflora group, whereas in the ITS tree (Fig. 3: B) the multiflora group and several southern South American species (P. fonkii, P. magellanica, and P. megalantha) form a polytomy basal to the remaining species. The prenanthoides group (= Calorezia) is distinct from Perezia, although the chloroplast data (Fig. 3: A) place it as sister to a clade of Calopappus, Nassauvia, and Triptilion and the ITS (Fig. 3: B) places it as sister to the entire Perezia clade. In both analyses, Perezia lanigera (= Burkartia) shows a strong relationship with Acourtia, with the chloroplast data indicating it is sister to Acourtia and the ITS data suggesting that it is embedded within Acourtia. Both show Nassauvia paraphyletic with respect to Triptilion. Although other genera were not sampled thoroughly, our data suggest that Panphalea is always strongly supported as sister to Perezia. The combined tree (Fig. 4) strongly supports the multiflora group as sister to the rest of Perezia (Fig. 4: a; note, the lowercase letters refer to labeled clades in Fig. 4). It also shows a grade (Fig. 4: b, b′ ) of low elevation, humid habitat southern South American species (P. fonkii, P. magellanica + P. megalantha) that is sister to a clade (Fig. 4: c) containing the remaining species of the genus. This large clade (Fig. 4: c) consists of a basal polytomy of southern (south of approximately 40◦ S, except for P. pilifera, which has a distribution that extends from 30◦ to 55◦ S latitude) species (P. lyrata + P. pedicularidifolia, a recurvata clade plus P. viscosa (Fig. 4: d), P. calophylla, P. lactucoides) sister to a clade of central Andean species (Fig. 4: e). Perezia virens and subsequently P. carthamoides, the basal members of this clade, occur relatively further south (∼35◦ S) than the other members of this clade. All members of the most derived clade (Fig. 4: f) occur from northwest Argentina to Ecuador and all occur in high Andean habitats. In the combined analysis, Panphalea is sister to Perezia, Perezia (Burkartia) lanigera is sister to 438 Journal of Systematics and Evolution Vol. 47 No. 5 2009  C 2009 Institute of Botany, Chinese Academy of Sciences Fig. 3. Cladograms from the Bayesian majority rule consensuses constructed using (A) combined chloroplast DNA data and (B) internal transcribed spacer (ITS) data. The Perezia group is the clade containing Acourtia, Burkartia, Calopappus, Calorezia, Nassauvia, Perezia, Panphalea, and Triptilion. Numbers above branches are the Bayesian posterior probabilities. Numbers below the lines are the maximum likelihood/maximum parsimony bootstrap values. SIMPSON et al.: Phylogeny of the Andean genus Perezia 439 Fig. 4. Bayesian majority rule cladogram based on the combined chloroplast and internal transcribed spacer (ITS) data. Numbers above branches are the Bayesian posterior probabilities. Numbers below the lines are the maximum likelihood/maximum parsimony bootstrap values. The lowercase letters refer to clades discussed in the text. The circles show latitudinal limits, with black circles indicating geographical distributions north of 30◦ S and white circles showing distributions south of 30◦ S. Circles that are half black and half white indicate distributions that occur both north and south of 30◦ S. The squares represent elevation, with black squares showing altitudinal distributions above 2500 m and white squares altitudinal distributions below 2500 m. Squares that are half black and half white show that the elevational distribution of a species extends from below 2500 m to higher elevations.  C 2009 Institute of Botany, Chinese Academy of Sciences 440 Journal of Systematics and Evolution Vol. 47 No. 5 Acourtia, the prenanthoides group (Calorezia) is sister to a clade of Calopappus, Nassauvia, and Triptilion, and Triptilion is embedded within Nassauvia. 3 Discussion 3.1 Phylogenetic relationships among species of Perezia Our finding that Perezia prenanthoides/nutans do not cluster with the remainder of Perezia was not surprising given the fact that the two differ in habit and habitat from all other Perezia species and were considered “isolated within the genus” by Vuilleumier (1970). They are large (to 84 cm) branched, soft-leaved, caulescent species that occur in the Nothofagus forest of southern Chile, reaching the subalpine in central Chile. They also have an “Acourtia” type of style branches (Crisci, 1974), unlike the species of Perezia. Thus, the segregation of this group into the separate genus Calorezia (Panero, 2007) is justified, but our data (not shown) for several samples of both species suggest that only one species rather than two may be involved. Calopappus (often treated as a synonym of Nassauvia; Hind, 2007), one of the genera closest to Calorezia according to Panero (2007), occurs above the treeline in the central Chilean Andes. Nassauvia itself is primarily southern South American. Although we sampled only five of the 38+ species, our data suggest that Nassauvia is paraphyletic with respect to Triptilion and that the latter should be included within Nassauvia. Our data confirm that Burkartia (Perezia) lanigera, as suggested by Vuilleumier (1970), is not a Perezia. Moreover our data indicate that it should either be considered a monophyletic genus (Burkartia, cf. Crisci, 1976) sister to the North American genus Acourtia or perhaps a member of Acourtia because, according to some of our phylogenetic reconstructions, Acourtia is paraphyletic with respect to Burkartia (Fig. 3: B). Morphological characters that B. lanigera shares with Acourtia but not with Perezia include a shrubby habit, subsessile capitula, capitula with six to 14 florets, pubescent corolla, wooly trichomes, and a Trixis Lag. type of pollen exine (Crisci, 1974). Until Acourtia is more fully studied, we retain Burkartia as a monophyletic genus, but note that this is the only clade in the Perezia group that has a New World amphitropical disjunct distribution. These new data enable the clarification of the evolution of Perezia. Contrary to Vuilleumier’s (1970) suggestions that the Perezia pungens group is “basal” (= sister) to the rest of the genus, the Perezia multiflora group is clearly the sister to the rest of Perezia s.s. 2009 Table 4 Chromosome numbers of members of the Perezia clade Chromosome Source∗ number Taxon Acourtia belizeana B. L. Turner n = 18 1 A. carpholepis (A. Gray) Reveal & R. M. ∼27 pairs 1 King A. nana (A. Gray) Reveal & R. M. King n = 27 1 A. microcephala DC. 2n = 54 3 A. rigida DC. n = 26 1 A. scapiformis (Bacig.) B.L. Turner ∼27 pairs 2 A. thurberi (A. Gray) Reveal & R. M. King 2n = 54 1 A. wrightii (A. Gray) Reveal & R. M. King n = 27 1 Nassauvia aculeata var. robusta (Cabrera) n = ∼44 1 Cabrera N. axillaris D. Don n = 11 1 N. chubutensis Speg. n = 11 1 N. darwinii (Hook. & Arn.) O. Hoffm. & n = 11 1 Dusén N. gaudichaudii Cass. 2n = 22 2 N. glomerulosa D. Don n = 11 1 N. lagascae F. Meigen n = 11 1 N. magellanica J. F. Gmel. n = 11 2 N. pygmaea (Cass.) Hook. f. n = 11 1 N. revoluta D. Don n = 11 1 N. serpens d’Urv. n = 11 2 N. uniflora (D. Don) Hauman n = 11 1 Panphalea bupleurifolia Less. n=8 1 Perezia calophylla (Phil.) Reiche 2n = 24 3 P. carduncelloides Griseb. 2n = 24 3 P. ciliaris Hook & Arn. 2n = 24 3 P. ciliosa (Phil.) Reiche 2n = 24 2 P. coerulescens Wedd. 2n = 24 3 Perezia magellanica (L.f.) Lag. n = 24 2 P. multiflora (Humb. & Bonpl.) Less. n=8 1, 3 Perezia “nivalis” Wedd. 2n = 24 4 P. pedicularidifolia Less. 2n = 24 1 P. pilifera (D. Don. & Arn.) Hook. n = 16 1 P. pungens (Humb. & Bonpl.) Less. 2n = 28 1, 3 P. recurvata (Vahl) Less. 2n = 24 1, 3 P. squarrosa ssp. cubaetensis (Less.) Vuill. n=4 2 Triptilion gibbosum Remy 2n = 20 1 ∗ Sources for the information as are follows: 1, Index to Plant Chromosome Numbers (IPCN; available at http://mobot.mobot.org/ W3T/Search/ipcn.html, accessed 15 September 2008) and/or batch files at that URL; 2, Crisci (1974); 3, Vuilleumier (1970); 4, Bolkhovskikh et al. (1969). (Fig. 4). This placement is consistent with the chromosome numbers (Table 4) known for species of this group: 2n = 8 for P. squarrosa subsp. cubaetensis and 2n = 16 for P. multiflora compared with 2n = 24 for the remaining species studied. Like most members of the multiflora group, species of Panphalea (the sister genus to Perezia) are also large (20–100 cm tall), have leafy stems, and inflorescences of numerous small heads (Cabrera, 1953). In addition, the only chromosome count made for that genus (Panphalea bupleurifolia) is n = 8 (Table 4). Most of the Perezia species (those sister to the multiflora group) form a clade with either a grade of species of southern humid and alpine habitats (P. fonkii, followed by a clade of P. magellanica plus P. megalantha) sister to the remaining species (Fig. 3: A) or with  C 2009 Institute of Botany, Chinese Academy of Sciences SIMPSON et al.: Phylogeny of the Andean genus Perezia a basal polytomy of southern species (Fig. 3: B). Thus, most of the species of the magellanica group of Vuilleumier (1970) constitute a grade rather than a group or clade. The combined analysis supports the recurvata group (Fig. 4: d) of Vuilleumier (1970) plus P. viscosa. The members of the recurvata group are small loose cushion-forming, narrow-leaved (often spiny) plants (Vuilleumier, 1970), with P. viscosa anomalous morphologically in this group with its basal rosette of broad oblanceolate leaves reminiscent of other members found in Nothofagus forest. This clade and several other taxa placed in the magellanica and pungens groups (Fig. 1) form a polytomy with a large clade of species that occur from northwestern Argentina to Ecuador. This northern clade contains a mixture of species placed in the pungens and coerulescens groups by Vuilleumier (1970). 3.2 Biogeographic implications Considering the results as a whole, the confirmation that the prenanthoides group is sister to Calopappus and Nassauvia plus Triptilion strongly suggests that the Perezia group (the clade of Calopappus, Nassauvia, Panphalea, Perezia, and Triptilion) arose in southern, probably southwestern, South America. However, Panphalea, the sister genus to Perezia, occurs in eastern Argentina and adjacent Uruguay and Paraguay. The species of the multiflora clade, sister to the majority of Perezia also occur predominantly in southeastern South America, with P. kingii occurring in northeastern Argentina and Uruguay, P. squarrosa Less. in extreme southeastern Brazil and Uruguay, and both P. eryngioides and P. catharinensis in Santa Catarina, Brazil. Within this group of five species, only P. multiflora itself extends its distribution westward into dry high regions of the central Andes (and accounts for the seemingly broad distribution of this group; Fig. 2: A). Therefore, it would appear that following the origination of the Perezia group, there was a spread or dispersal into eastern subtropical South America with either a recolonization of the southwestern Andes or a later radiation of an ancestral stock in southwestern South America. Although the cladogram in Fig. 4 suggests the former, it is possible that inclusion of more species of both Panphalea and the multiflora group will show these to be sister and consistent with the second pattern. The three species that branch sequentially to the multiflora clade form a grade of taxa that occur in relatively low-elevation moist forest or humid steppe habitats south of 30◦ S. Although there is little resolution among the high Andean species (Fig. 4: c), it is evident that the high-elevation species are the most derived and that the genus is of temperate South American  C 2009 Institute of Botany, Chinese Academy of Sciences 441 origin, radiating in western South America from south to north and from low to high elevation (Fig. 4). As far as we can tell, the pattern we found in Perezia does not completely match that of any temperate/Andean plant genus studied phylogenetically to date. The most similar pattern to that of Perezia was shown in a recent study by Hershkovitz et al. (2006a) for the 20 species of Tropaeolum L. sect. Chilensia Sparre (Tropaeolaceae). Although plagued by problems of uncertain rooting, the data of Hershkovitz et al. (2006a) also indicate an east–west split in a basal clade (eastern Argentina, southern Chile). The remaining species are primarily Chilean and show a pattern of diversification from southern mesophytic areas to central Mediterranean scrub habitats to northern deserts. However, it should be noted that species of this section rarely reach the elevations of many Perezia taxa (over 4000 m a.s.l.) and,, unlike Perezia, Tropaeolum consists of 90 species, most of which (70 species) occur in tropical America. A molecular study of Chaetanthera Ruı́z & Pavón, an Andean/Patagonian genus of the Mutisieae (Hershkovitz et al., 2006b), found some high-elevation species to be derived from lowland stock, whereas others were interpreted to be relictual, migrating upward on account of increasing aridity in the central Andes. In addition, a number of the lower-elevation species in southcentral Chile were found to be more derived. Regardless of differences, Chaetanthera, like Perezia, underwent a major species radiation in high-elevation habitats in the more arid areas of the central Chilean Andes and puna. In the small Andean genus Schizanthus (Solanaceae), lowland species were shown to have diverged more recently than the few alpine species in that genus (Perez et al., 2006). Yet another pattern is seen in Hamadryas, a small dioecious, predominantly alpine genus of four species in the Patagonian component of the Ranunculus grade of Ranunculaceae. Its closest relatives are found in the alpine of southern South America, Asia, North America, and South Africa (Hoot et al., 2008). Finally, a study of Ourisia (Plantaginaceae) by Meudt (2006) and Meudt and Simpson (2007) indicated an origin in south-central America (∼34◦ S) at mid elevation (800– 2400 m a.s.l.) and a subsequent spread both south and north in the Andes as well as to New Zealand, where a subsequent significant radiation occurred. Clearly, rigorous phylogenies of many more groups need to be performed before we can say whether the Perezia or one of the other patterns is most commonly found in genera with a predominance of species in temperate and high Andean regions of South America. It is possible that the evolutionary patterns of radiation are related to the location of the ancestors of the groups 442 Journal of Systematics and Evolution Vol. 47 No. 5 studied. Genera (clades) with tropical Andean ancestors may commonly show the basal members of the clade to be mesophytic with lowland xerophytic species derived, whereas groups originating in temperate areas may commonly show patterns of late radiation in the very high supraforest habitats of the tropical Andes. Acknowledgements The authors are grateful for the sample of Perezia lanigera supplied by Serge AUBERT (Directeur du Jardin Alpin du Lautaret, Lautaret, France) and thank José PANERO (Integrative Biology, The University of Texas, Austin, TX, USA) for generously providing a portion of his sample of Calopappus acerosus. Jun WEN (Smithsonian Institution, Washington DC, USA) kindly supplied the specimen for material of Perezia nutans. The figures in the present study were drawn by JM (coauthor) and Gwen GAGE (Computer Illustrator, The University of Texas). The authors also thank the curators of GH, LPB, NY, and TEX for allowing sampling of material used in the present study. MTKA acknowledges funding from ICM P05-002. References Bacigalupi R. 1931. A monograph of the genus Perezia sect. Acourtia with a provisional key to the section Euperezia. Contributions to the Gray Herbarium 97: 1–81. Bolkhovskikh Z, Grif V, Matvejeva T, Zakharyeva O. 1969. In: Federov An A ed. Chromosome numbers of flowering plants. Nauka: Leningrad. Cabrera AL. 1953. Las especies del género Panphalea. Notas del Museo, Universidad Nacional de Eva Peron 16(Botánica No. 82): 225–237. Cabrera AL. 1971. Compositae. In: Correa MN ed. Flora Patagónica (República Argentina). Parte VII. Buenos Aires: I.N.T.A. 378–396. Cabrera AL. 1978. Perezia. In: Cabrera A ed. Flora de la Provincia de Jujuy 8(10). Buenos Aires: I.N.T.A. 652–669. Cassini H. 1825. Nassauviées. In: Cuvier GF ed. Dictionnaire des Sciences Naturelles 34. Paris: Le Normant. 208– 238. Crisci JV. 1974. A numerical-taxonomic study of the subtribe Nassauviinae (Compositae, Mutisieae). Journal of the Arnold Arboretum 55: 568–610. Crisci JV. 1976. Burkartia: Nuevo género de Mutisieae (Compositae). Boletin de la Sociedad Argentina de Botánica 17: 241–246. Crisci JV. 1980. Evolution in the subtribe Nassauviinae (Compositae, Mutisieae): a phylogenetic reconstruction. Taxon 29: 213–224. Edgar RC. 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32: 1792–1797. Farris JS, Kallersjo M, Kluge AG, Bult C. 1995. Testing significance of incongruence. Cladistics 10: 315–319. GeneCodes Corp. 2005. Sequencher 4.5: software for DNA sequence assembly. Ann Arbor, MI: Gene Codes Inc. 2009 Hershkovitz MA, Hernández-Pellicer CC, Arroyo MTK. 2006a. Ribosomal DNA evidence for the diversification of Tropaeolum sect. Chilensia (Tropaeolaceae). Plant Systematics and Evolution 260: 1–24. Hershkovitz MA, Arroyo MTK, Bell C, Hinojosa LF. 2006b. Phylogeny of Chaetanthera (Asteraceae: Mutisieae) reveals both ancient and recent origins of the high elevation lineages. Molecular Phylogenetics and Evolution 41: 594– 605. Hind DJN. 2007. II Tribe. Mutisieae. In: Kadereit JW, Jeffrey C eds. VII. Flowering Plants. Eudicots. Asterales. The families and genera of flowering plants. New York: Springer. 90– 122. Hoot SB, Kramer J, Arroyo MTK. 2008. Phylogenetic position of the South American dioecious genus, Hamadryas, and related Ranunculeae (Ranunculaceae). International Journal of Plant Sciences 169: 433–443. Huelsenbeck JP, Ronquist F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755. Katinas L, Crisci JV, Jabaily RS, Williams C, Walker J, Drew B, Bonifacino JM, Sytsma KJ. 2008. Evolution of secondary heads in Nassauviinae (Asteraceae, Mutisieae). American Journal of Botany 95: 229–240. Kim H-G, Loockerman D, Jansen RK. 2002. Systematic implications of ndhF sequence variation in the Mutisieae (Asteraceae). Systematic Botany 27: 598–609. Kim K-J, Jansen RK. 1994. Comparison of phylogenetic hypotheses among different data sets in dwarf dandelion (Krigia: Asteraceae): additional information from internal transcribed spacer sequences of nuclear ribosomal DNA. Plant Systematics and Evolution 190: 157–185. Loockerman DJ, Jansen RK. 1996. The use of herbarium material for DNA studies. In: Sohmer SH ed. Sampling the green world. New York: Columbia University Press. 205– 220. Maddison DR, Maddison WP. 2005. MacClade 4.08. Sunderland: Sinauer Associates. Meudt HM. 2006. Monograph of Ourisia (Plantaginaceae). Systematic Botany Monographs 77: 1–188. Meudt HM, Simpson BB. 2007. The biogeography of the austral, subalpine genus Ourisia (Plantaginaceae) based on molecular phylogenetic evidence: South American origin and dispersal to New Zealand and Tasmania. Biological Journal of the Linnean Society 87: 479–513. Nylander JAA. 2004. MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University, Sweden. Panero J. 2007. Calorezia, a new genus of tribe Nassauvieae (Asteraceae, Mutisioideae). Phytologia 89: 198–201. Perez F, Arroyo MTK, Medel R, Hershkovitz MA. 2006. Ancestral reconstruction of flower morphology and pollination systems in Schizanthus (Solanaceae). American Journal of Botany 93: 1029–1038. Reveal JL, King RM. 1973. Re-establishment of Acourtia D. Don (Asteraceae). Phytologia 15: 228–232. Shaw J, Lickey EB, Schilling EE, Small RL. 2007. Comparison of whole chloroplast genome sequences to choose non-coding regions for phylogenetic studies in angiosperms: the tortoise and the hare III. American Journal of Botany 94: 275– 288.  C 2009 Institute of Botany, Chinese Academy of Sciences SIMPSON et al.: Phylogeny of the Andean genus Perezia Swofford DL. 2002. PAUP∗ : Phylogenetic Analysis Using Parsimony (∗ and other methods), v. 4.0b10. Sunderland, MA: Sinauer Associates. Timme RE, Kuehl JV, Boore JL, Jansen RK. 2007. A comparative analysis of the Lactuca and Helianthus (Asteraceae) plastid genomes: Identification of divergent regions and categorization of shared repeats. American Journal of Botany 94: 302–312. Turner BL. 1978. Taxonomic study of the scapiform species of Acourtia (Asteraceae-Mutisieae). Phytologia 38: 456– 468.  C 2009 Institute of Botany, Chinese Academy of Sciences 443 Vuilleumier BS. 1970. The systematics and evolution of Perezia sect. Perezia (Compositae). Contributions from the Gray Herbarium 199: 1–163. Werle E, Schneider C, Renner M, Völker M, Fiehn W. 1994. Convenient single-step, one tube purification of PCR products for direct sequencing. Nucleic Acids Research 20: 4354– 4355. Zwickl DJ. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. Dissertation. Austin: The University of Texas at Austin.