Molecular Phylogenetics and Evolution 67 (2013) 95–109 Contents lists available at SciVerse ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Molecular phylogenetics of the species-rich genus Habenaria (Orchidaceae) in the New World based on nuclear and plastid DNA sequences João A.N. Batista a,⇑, Karina S. Borges a, Marina W.F. de Faria a, Karina Proite a, Aline J. Ramalho a, Gerardo A. Salazar b, Cássio van den Berg c a Departamento de Botânica, Instituto de Ciências Biológicas, Universidade Federal de Minas Gerais, Av. Antônio Carlos 6627, Pampulha, C.P. 486, Belo Horizonte, Minas Gerais 31270-910, Brazil Departamento de Botánica, Instituto de Biología, Universidad Nacional Autónoma de Mexico, Apartado Postal 70-367, 04510 Mexico, DF, Mexico c Departamento de Ciências Biológicas, Universidade Estadual de Feira de Santana, Av. Transnordestina s/n, Feira de Santana, Bahia 44036-900, Brazil b a r t i c l e i n f o Article history: Received 17 September 2012 Revised 2 January 2013 Accepted 8 January 2013 Available online 19 January 2013 Keywords: Orchidaceae Habenaria Neotropical Phylogeny ITS matK a b s t r a c t Habenaria is a large genus of terrestrial orchids distributed throughout the tropical and subtropical regions of the world. The integrity and monophyly of this genus have been under discussion for many years, and at one time or another, several genera have been either included in a broadly defined Habenaria or segregated from it. In this study, the phylogenetic relationships of the Neotropical members of the genus and selected groups of African Habenaria were investigated using DNA sequences from the nuclear internal transcribed spacer (ITS) region and the plastid matK gene sampled from 151 taxa of Habenaria from the Neotropics (ca. 51% of the total) as well as 20 species of Habenaria and Bonatea from the Old World. Bayesian and parsimony trees were congruent with each other, and in all analyses, the Neotropical species formed a highly supported group. African species of Habenaria in sections Dolichostachyae, Podandria, Diphyllae, Ceratopetalae and Bilabrellae, and the Neotropical clade formed a highly supported ‘‘core Habenaria clade’’, which includes the type species of the genus from the New World. The topology of the trees indicates an African origin for the Neotropical clade and the low sequence divergence among the Neotropical species suggests a recent radiation of the genus in the New World. Species of Bonatea and Habenaria sections Chlorinae and Multipartitae formed a well-supported clade that was sister to the ‘‘core Habenaria clade’’. The Neotropical clade consists of at least 21 well-supported subgroups, but all Neotropical sections of the current sectional classification are paraphyletic or polyphyletic and will need extensive revision and recircumscription. Most of the Neotropical subgroups formed morphologically uniform assemblage of species, but some cases of morphological divergence within subgroups and convergence between subgroups indicated that morphology alone can be misleading for inferring relationships within the genus. The genera Bertauxia, Kusibabella and Habenella, segregated from New World Habenaria, are not monophyletic and a revision of the sectional classification rather than a generic division seems most appropriate. Our results do not support an extensive generic fragmentation of Habenaria as previously suggested and will provide a framework for revising the infrageneric classification and investigating the patterns of morphological evolution and geographical distribution of the genus in the New World. Ó 2013 Elsevier Inc. All rights reserved. 1. Introduction Habenaria Willd. (Orchidinae, Orchidoideae, Orchidaceae) is a large genus of approximately 876 (Govaerts et al., 2011) terrestrial species distributed throughout the tropical and subtropical regions of the Old and New World (Pridgeon et al., 2001a) with centers of diversity in Brazil, southern and central Africa and East Asia (Kurzweil and Weber, 1992). Most species are perennial, deciduous ⇑ Corresponding author. Fax: +55 31 3409 2671. E-mail address: janb@icb.ufmg.br (J.A.N. Batista). 1055-7903/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ympev.2013.01.008 geophytes with a growth cycle associated with a wet season followed by a dormant period in the form of an underground root tuber during the dry season. In the New World the genus accounts for a large proportion of the Orchidaceae of tropical and subtropical grasslands (Barros, 1987; Toscano de Brito, 1995; Mendonça et al., 1998; Batista and Bianchetti, 2003; Batista et al., 2004; Zappi et al., 2003), including the savannas of the Brazilian Amazon (Batista et al., 2008a,b), and is poorly represented in tropical forests. Brazil and Mexico are the countries with the highest numbers of species in the New World with 167 and 72 species, respectively. In Brazil, the center of diversity of the genus is the cerrado, a species-rich savanna that covers approximately two million square 96 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 kilometers of central Brazil (Ratter et al., 1997), whereas in Mexico, the greatest diversity is found in tropical and subtropical oakconiferous forests (Batista et al., 2011a). Traditionally, Habenaria has been placed in subtribe Habenariinae, which, together with Orchidinae, forms tribe Orchideae (Dressler, 1993). The separation of the two subtribes is based primarily on stigma morphology: Orchidinae have concave, unstalked stigmas, often with confluent lobes, whereas Habenariinae have stalked, convex stigma lobes that are usually distinct, but the separation is questionable (Kurzweil and Weber, 1992; Pridgeon et al., 2001a). The tribe Orchideae, with approximately 62 genera and 1800 species, is particularly well represented in the Afro-Madagascan region (Dressler, 1993; Pridgeon et al., 2001a). Habenaria is by far the largest genus in Orchideae, comprising approximately 45% of the species assigned to the tribe, followed by Platanthera Rich. (200 spp.) and Cynorkis Thouars (125 spp). Only Habenaria is found throughout the American tropics, although a few species of the predominantly north-temperate genus Platanthera extend south to Mexico and Guatemala. The integrity of Habenaria as a genus has been under discussion for many years. Species currently placed in distinct genera such as Platanthera and Coeloglossum Hartm. were formerly placed in Habenaria. Excluding the consistently stalked and strongly convex stigma lobes, which are also found in other Habenariinae, the characters used for defining Habenaria show extensive variation. Habenaria is currently distinguished from other closely related genera by often bifid petals that are not fused to other parts of the flower, a lip that is usually deeply divided and lacking a callus and entire stigma lobes, which are usually free and not adnate to the petals or lip (Pridgeon et al., 2001a). Some African species formerly included in Habenaria were segregated to genera such as Bonatea Willd., Centrostigma Schltr., Platycoryne Rchb.f. and Roeperocharis Rchb.f. However, according to Kurzweil and Weber (1992), they are similar in most characters to Habenaria species and are better treated as specialized forms of the genus at the sectional rank. The only worldwide revisions of Habenaria were those of Kränzlin (1892, 1901) in which 32 sections were recognized. Characterization of the sections was based primarily on the degree of dissection of the petals and lip and on gynostemium structure, particularly the length of the stigmas. After Kränzlin, few authors have treated the infrageneric classification of Habenaria. Schlechter (1915) renamed some of the sections and Summerhayes (1942, 1960, 1962, 1966) and Hunt (1968) proposed new sections, but these works exclusively addressed African species. For the New World, Cogniaux (1893) generally followed Kränzlin’s sectional characterization in his treatment of Habenaria in Flora Brasiliensis, whereas Hoehne (1940) used a different approach in the revision of the genus for Flora Brasilica, dividing the Brazilian species into nine informal groups starting with the vegetative parts and then advancing toward the details of the flowers. In the last major survey of Brazilian orchids, Pabst and Dungs (1975) basically followed the divisions established by Hoehne, using some new characters to distinguish several groups that they called alliances. González-Tamayo (1993) divided the Mexican species into 12 tentative natural groups but did not make any reference to previous sectional treatments of the genus. More recently, Szlachetko recognized three genera within New World Habenaria: Bertauxia Szlach., Kusibabella Szlach. (Szlachetko, 2004a,b) and Habenella Small (Szlachetko and Kras, 2006). However, his work was undertaken on a piecemeal basis based on floral morphological characters, and his genera have not been widely accepted. The floral morphology of Southern African Habenariinae was characterized in detail by Kurzweil and Weber (1992). A similar study is not available for the Neotropical Habenaria, neither is a comparative analysis between the New World and Old World species of the genus. However, an analysis of the literature reveals that the floral morphology of the Old World species is much more diverse than those of the New World. This observation reflects Kränzlin’s sectional classification in which, out of 32 sections, only 12 were from the New World. In fact, compared to the African and Asian groups of the genus, the floral morphology of the New World species is more homogenous, particularly for the gynostemium. Despite being the largest genus in the tribe Orchideae and having a worldwide distribution, Habenaria is underrepresented in molecular systematic studies, especially compared with other genera in the tribe such as Orchis Tourn. ex L., Ophrys L. and Platanthera, which have been extensively investigated (Pridgeon et al., 1997; Bateman et al., 1997, 2003). This is most likely because Habenaria occurs mostly in the Southern Hemisphere and few species are cultivated, rendering access to genetic material difficult. Thus far, the only study addressing the phylogeny of Habenaria using a cladistic approach and DNA sequence data was the phylogenetic analysis of Orchidinae and selected Habenariinae of Bateman et al. (2003). Aside from this work, a few species of Habenaria have been sequenced in the context of general phylogenetic analyses of Orchidaceae or their infrafamilial ranks (Cameron et al., 1999; Douzery et al., 1999; Kores et al., 2001; Bellstedt et al., 2001; Ponsie et al., 2007a) and the DNA barcoding of land plants (Lahaye et al., 2008). The study of Bateman et al. (2003) using ITS sequence data indicated that Habenaria was highly polyphyletic. However, only eight species of Habenaria were sampled (1% of the genus), including only one from the New World. Contrary to the results of Bateman et al. (2003), the relatively homogeneous floral morphology of the New World species, particularly in regard to the gynostemium, suggests that they could be a monophyletic group. The similarities in floral morphology between New World species and some Old World species also suggests a close relationship between some African and American groups of Habenaria, but the currently available taxon sampling for molecular studies is insufficient to address these phylogenetic questions. A phylogenetic hypothesis for the New World species of the genus should also help clarify the infrageneric classification of the genus and be useful for redefining the sections as a step toward a generic revision. In recent years, the use of DNA sequence data has proven useful for inferring phylogenetic relationships in Orchidaceae at several taxonomic levels (Cameron et al., 1999; van den Berg et al., 2000, 2005; Pridgeon et al., 2001b; Williams et al., 2001; Salazar et al., 2003; Freudenstein et al., 2004; review in Cameron, 2007), including the infrageneric relationships in some large genera of the family (Bytebier et al., 2007; Fischer et al., 2007; Russell et al., 2009; Whitten et al., 2007). The internal transcribed spacers (ITSs) of the nuclear ribosomal multigene family and, to a lesser extent, the maturase K (matK) gene of the plastid genome, have provided good resolution in phylogenetic analyses of several groups within Orchidaceae, including the tribe Orchideae (Pridgeon et al., 1997; Bateman et al., 1997, 2003; Ponsie et al., 2007a). Here, we take advantage of the large number of Habenaria species occurring in Brazil and Mexico and the availability of some sequence data from African species in the databanks to investigate the phylogeny of the genus in the New World, using DNA sequences from the nuclear ITS region and the plastid matK gene, focusing on Brazilian and Mexican species, with the following objectives: (1) to establish whether Neotropical species of Habenaria form a monophyletic group; (2) to investigate the relationships between Neotropical and selected groups of African Habenaria and other genera of Habenariinae; and (3) to evaluate the current sectional classification of New World species in light of an explicit phylogenetic hypothesis. J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 2. Materials and methods 2.1. Taxon sampling A total of 217 terminals were used consisting of 180 species, of which 152 were Neotropical Habenaria species, corresponding to 51% of the total number of species known from the Neotropics (Batista et al., 2011a,b). Additionally, the total included ten African Habenaria species, ten species of Bonatea and eight species of Cynorkis, Gennaria Parl., Satyrium L., Stenoglottis Lindl., Platanthera, Orchis and Disa P.J. Bergius from the Old World (Table 1). The latter genus (tribe Diseae) was used as a functional outgroup. The sampling of the New World Habenaria species was concentrated in Brazil and Mexico, included species from all currently recognized sections and informal groups except the monospecific section Pycnostachyae Cogn. (Table 2), and covered most of the morphological variability and geographic distribution of the genus in the Neotropics. Most of the plant material was collected in the field and dried in silica gel, but in a few instances, herbarium specimens were used. Taxa from Brazil were identified by J.A.N. Batista and taxa from Mexico by G.A. Salazar. Several species with broad geographic distributions or significant morphological variability were sampled more than once. In some cases, additional samples were also sequenced to confirm the position of a species in the trees. Voucher information, geographic origin and GenBank accession numbers are listed in Table 1. 2.2. Molecular markers Nucleotide sequences from one nuclear (ITS) and one plastid (matK) genome region were used in the analyses. The ITS region consisted of the 30 and 50 ends of the 18S and 26S ribosomal RNA genes, respectively, the internal transcribed spacers (ITS1 and ITS2) and the intervening gene 5.8S of the nuclear ribosomal multigene family. Amplifications were performed using the primers 17SE and 26SE (Sun et al., 1994). For the matK gene, we amplified an internal fragment of approximately 630 bp using the primers matK-F2 (50 -CTAATACCCCATCCCATCCAT-30 ) and matK-R2 (50 CCCAATACAGTACAAAATTGAGC-30 ). This fragment corresponds approximately to the same region used in the phylogeny of Bonatea (Ponsie et al., 2007a) and for the barcoding of land plants (Chase et al., 2007), and it corresponds to the most variable region of the gene in several orchid groups (e.g., Whitten et al., 2000). Selection of the markers was based on the ease of amplification, availability of sequences from other genera of Old World Orchideae in public data banks and the general use of the markers in phylogenetic studies of Orchidaceae. 2.3. DNA extraction, amplification and sequencing Total DNA was extracted from individual plants using a modified version of the 2  CTAB protocol of Doyle and Doyle (1987). PCR amplifications were performed in a MJ96 Thermocycler. The general PCR system consisted of 20–50 ng of genomic DNA, 1X PCR Buffer, 2 mM MgCl2, 200 lM dNTPs, 0.4 lM of each primer, 2 U of Taq DNA Polymerase (Phoneutria Biotec., Belo Horizonte, Brazil) and water to reach a total volume of 25 ll. Cycling conditions were an initial denaturation at 94 °C for 3 min, 35 cycles of 94 °C for 45 s, 58 °C for 45 s and 72 °C for 1 min, and a final extension for 3 min at 72 °C. For reactions with low yield or unspecific amplification products, the conditions above in the PCR system and cycle parameters were individually adjusted. In contrast with other reports (van den Berg et al., 2005), the use of denaturing reagents such as betaine and DMSO in the ITS amplification did not increase yield or specificity and were therefore not used. PCR prod- 97 ucts with single bands were purified using polyethylene glycol precipitation and sequenced in a MegaBACE 1000 (Amersham Biosciences) automatic sequencer following the manufacturer’s protocol. Some sequences were produced by Macrogen Inc., Korea. All regions were sequenced bi-directionally. Doubtful base calls were, in most cases, verified with a third sequencing reaction. 2.4. Sequence analysis, alignment and pairwise distance calculation DNA sequence electropherograms were edited with the STADEN package software (Bonfield et al., 1995) or Sequencher version 4.8 (GeneCodes Corp., Ann Arbor, Michigan, USA). The edited sequences were aligned with MUSCLE (Edgar, 2004), and the resulting alignment was manually adjusted using MEGA4 software (Tamura et al., 2007). No data were excluded from the analyses due to ambiguous alignment. Individual gap positions were treated as missing data. Pairwise distances between the sequences were calculated using the Maximum Composite Likelihood method in MEGA4 (Tamura et al., 2004, 2007). In the pairwise analyses, the positions containing gaps and missing data were eliminated from the data set (complete deletion option). 2.5. Phylogenetic analyses The data were analyzed using both parsimony and Bayesian inference. Phylogenetic analyses using parsimony were performed in PAUP version 4 (Swofford, 1998) with Fitch parsimony (equal weights, unordered characters; Fitch, 1971) as the optimality criterion. GenBank sequences from representative species of the Old World tribe Diseae (Disa) were used as outgroups. Searches were initially performed separately on each data set, and because no cases of strongly supported incongruence were detected (i.e., no conflicting groups were observed between the two data sets obtaining high internal support), a third search was performed with a combined matrix. Each search consisted of 1000 replicates of random taxon addition with branch swapping using the TBR (tree-bisection and reconnection) algorithm, retaining only up to ten trees per replicate to avoid extensive swapping on suboptimal islands. Internal support was evaluated by character bootstrapping (Felsenstein, 1985) using 1000 replicates, simple addition and TBR branch swapping, retaining up to ten trees per replicate. For bootstrap support levels, we considered bootstrap percentages (BPs) of 50–70% as weak, 71–85% as moderate and >85% as strong (Kress et al., 2002). A model-based phylogenetic analysis using Markov chain Monte Carlo-based Bayesian inference was performed using MrBayes v3.1.2 (Ronquist et al., 2005), treating each DNA region (ITS and matK) as a separate partition. An evolutionary model for each DNA region was selected with MrModeltest 2 (Nylander, 2004). For the both data sets the GTR + I + G model was selected according to the Akaike Information Criterion (AIC) or Hierarchical Likelihood Ratio Tests (hLRTs). Each analysis consisted of two independent runs with four chains for 3,000,000 generations, sampling one tree every 100 generations. In the combined analysis, to improve the swapping of chains, the temperature parameter for heating the chains was lowered to 0.05. After discarding the first 25% of the trees as the burn-in period, the remaining trees were used to assess topology and posterior probabilities (PP) in a majority-rule consensus. Because PP in Bayesian analysis are not equivalent to BP but are generally much higher (Erixon et al., 2003), we used criteria similar to a standard statistical test, considering groups with PP > 95% as strongly supported, PP 90–95% as moderately supported and PP < 90% as weakly supported. 98 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 Table 1 Voucher information and GenBank accession numbers for the sequences analyzed in this study. Taxon Voucher Origina ITS matK Diseae Disa ochrostachya Rchb.f. Disa uniflora P.J.Bergius Satyrium bicorne (L.) Thunb. GenBank GenBank GenBank Africa Africa Africa DQ414966 DQ414864 AY704978 DQ415109 DQ415007 EF612539 Orchideae Bonatea antennifera Rolfe Bonatea boltonii (Harv.) Bolus Bonatea bracteata G.McDonald & McMurtry Bonatea cassidea Sond. Bonatea lamprophylla J. Stewart Bonatea polypodantha (Rchb.f.) L. Bolus Bonatea porrecta (Bolus) Summerh. Bonatea pulchella Summerh. Bonatea saundersioides (Kraenzl. & Schltr.) Cortesi Bonatea speciosa (L.f.) Willd. Cynorkis grandiflora Ridl. Gennaria diphylla (Link) Parl. Habenaria achalensis Kraenzl. Habenaria alata Hook. Habenaria alpestris Cogn. Habenaria anisitsii Kraenzl. Habenaria araneiflora Barb.Rodr. Habenaria aranifera Lindl. Habenaria arenaria Lindl. Habenaria arenaria Lindl. Habenaria armata Rchb.f. Habenaria australis J.A.N. Bat., A.A. Vale and Menini Habenaria ayangannensis Renz Habenaria bahiensis Schltr. Habenaria balansae Cogn. Habenaria balansae Cogn. Habenaria batesii la Croix Habenaria brachyphyton Schltr. Habenaria bractescens Lindl. Habenaria brevidens Lindl. Habenaria brevidens Lindl. Habenaria brevilabiata A. Rich. and Galeotti Habenaria caldensis Kraenzl. Habenaria caldensis Kraenzl. Habenaria calicis R. González Habenaria canastrensis J.A.N. Bat. and B.M.Carvalho Habenaria aff. canastrensis J.A.N. Bat. and B.M. Carvalho Habenaria cardiostigmatica J.A.N. Bat. and Bianch. Habenaria ciliatisepala J.A.N. Bat. and Bianch. Habenaria ciliatisepala J.A.N. Bat. and Bianch. Habenaria clavata (Lindl.) Rchb.f. Habenaria clypeata Lindl. Habenaria coxipoensis Hoehne Habenaria aff. coxipoensis Hoehne Habenaria crassicornis Lindl. Habenaria crucifera Rchb.f. & Warm. Habenaria crucifera var. brevidactyla J.A.N. Bat. and Bianch. Habenaria cryptophila Barb.Rodr. Habenaria culicina Rchb.f. and Warm. Habenaria cultellifolia Barb.Rodr. Habenaria curti-bradei Hoehne Habenaria curvilabria Barb.Rodr. Habenaria curvilabria Barb.Rodr. Habenaria depressifolia Hoehne Habenaria distans Griseb. Habenaria distans Griseb. Habenaria dives Rchb.f. Habenaria edwallii Cogn. Habenaria aff. edwallii Cogn. Habenaria aff. edwallii Cogn. Habenaria egleriana J.A.N. Bat. and Bianch. Habenaria egleriana J.A.N. Bat. and Bianch. Habenaria egleriana J.A.N. Bat. and Bianch. Habenaria eustachya Rchb.f. Habenaria exaltata Barb.Rodr. Habenaria aff. fillifera S. Watson Habenaria aff. flexuosa Lindl. Habenaria fluminensis Hoehne Habenaria glaucophylla Barb.Rodr. var. glaucophylla GenBank GenBank GenBank GenBank GenBank GenBank GenBank GenBank GenBank GenBank GenBank GenBank Batista 2506 (BHCB) Nava 1784 (MEXU) Batista 1576 (BHCB) Pereira-Silva 4794 (CEN) Batista 2521 (BHCB) Batista 2472 (BHCB) Salazar 6407 (K) GenBank Batista 1297 (CEN) Batista 2496 (BHCB) Batista 1919 (BHCB) Batista 2867 (BHCB) Batista 2382 (BHCB) Batista 2336 (BHCB) Pollard 731 (YA) Batista 2515 (BHCB) Batista 2529 (BHCB) Batista 2616 (BHCB) Batista 2617 (BHCB) Nava 1116 (MEXU) Batista 1798 (BHCB) Batista 250 (CEN) Salazar 8184 (MEXU) Batista 1806 (BHCB) Batista 1825 (BHCB) Batista 2939 (BHCB) Batista 1610 (BHCB) Batista 1558 (CEN) GenBank Jacob 461 (MEXU) Batista 372 (CEN) Batista 2523 (BHCB) Salazar 7330 (MEXU) Batista 1826 (BHCB) Batista 3062 (BHCB) Batista 1488 (CEN) Batista 1345 (CEN) Batista 1487 (CEN) Batista 2372 (BHCB) Batista 1573 (BHCB) van den Berg 1267 (HUEFS) Batista 2369 (BHCB) Amaral 18 (CEN) Jiménez 2662 (AMO) GenBank Batista 1717 (BHCB) Batista 2395 (BHCB) Batista 247 (CEN) Batista 535 (CEN) van den Berg 1224 (HUEFS) Batista 2378 (BHCB) Salazar 6239 (PMA) Batista 2771 (BHCB) Salazar 7324 (MEXU) Zárate 401 (MEXU) Mota 3571 (BHCB) Batista 761 (CEN) Africa Africa Africa Africa Africa Africa Africa Africa Africa Africa Africa Africa Brazil, RS Mexico Brazil, DF Brazil, GO Brazil, PR Brazil, RS Africa Africa Brazil, DF Brazil, RS Brazil, MG Brazil, MG Brazil, GO Brazil, MG Cameroon Brazil, RS Brazil, PR Brazil, MG Brazil, MG Mexico Brazil, MG Brazil, GO Mexico Brazil, MG Brazil, MG Brazil, DF Brazil, MG Brazil, GO Africa Mexico Brazil, DF Brazil, PR Mexico Brazil, MG Brazil, DF Brazil, GO Brazil, GO Brazil, GO Brazil, DF Brazil, GO Brazil, GO Brazil, DF Brazil, DF Mexico Africa Brazil, MG Brazil, MG Brazil, GO Brazil, GO Brazil, GO Brazil, GO Panama Brazil, MG Mexico Mexico Brazil, MG Brazil, DF DQ522049 DQ522054 DQ522057 DQ522059 DQ522060 DQ522062 DQ522064 DQ522066 DQ522067 DQ522069 EF079186 AY351380 HM777526 HF560562 HM777655 HM777668 HM777527 HM777626 HF560563 DQ522073 HM777677 HM777724 HM777706 HM777562 HM777682 HM777683 HF560564 HM777557 HM777615 HM777535 KC257473 HF560565 HM777646 HM777645 HF560566 HM777726 HM777708 HM777575 HM777567 HM777565 DQ522074 HF560567 HM777536 HM777543 HF560568 HM777574 KC257471 HM777582 HM777571 HM777675 HM777667 HM777598 HM777599 HM777601 HM777630 HF560569 DQ522075 HM777564 HM777553 HM777554 HM777694 HM777692 HM777695 HF560570 HM777621 HF560571 HF560572 HM777659 HM777631 DQ522082 DQ522083 DQ522084 DQ522085 DQ522086 DQ522087 DQ522088 DQ522089 DQ522090 DQ522091 EF065584 AY368383 HM777794 HF560586 HM777952 HM777920 HM777795 HM777819 HF560587 DQ522092 HM777931 HM777988 HM777934 HM777808 HM777884 HM777883 HF560588 HM777797 HM777839 HM777903 HM777902 HF560589 HM777882 HM777881 HF560590 HM777941 HM777940 HM778018 HM777997 HM777998 DQ522093 HF560591 HM777905 HM777907 HF560592 HM778014 KC257476 HM777870 HM778012 HM777923 HM777918 HM777860 HM777859 HM777868 HM777871 HF560593 DQ522095 HM777803 HM777807 HM777806 HM777963 HM777964 HM777962 HF560594 HM777829 HF560595 HF560596 HM777945 HM777875 99 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 Table 1 (continued) Taxon Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria Habenaria glaucophylla var. brevifolia Cogn. glazioviana Kraenzl. ex Cogn. gonzalez-tamayoi García-Cruz, R. Jiménez and L. Sánchez gourlieana Gill. ex Lindl. guadalajarana S. Watson aff. guadalajarana S. Watson guilleminii Rchb.f. guilleminii Rchb.f. aff. guilleminii Rchb.f. gustavo-edwallii Hoehne hamata Hoehne henscheniana Barb.Rodr. henscheniana Barb.Rodr. heptadactyla Rchb.f. heringeri Pabst hexaptera Lindl. hexaptera Lindl. hieronymi Kraenzl. humilis Cogn. ibarrae R. González imbricata Lindl. imbricata Lindl. aff. imbricata Lindl. itacolumia Garay itatiayae Schltr. jaguariahyvae Kraenzl. johannensis Barb.Rodr. josephensis Barb.Rodr. juruenensis Hoehne kleinii Menini and J.A.N. Bat. laevigata Lindl. lavrensis Hoehne lavrensis Hoehne leprieurii Rchb.f. leprieurii Rchb.f. aff. leprieurii Rchb.f. leptoceras Hook. leucosantha Barb.Rodr. lithophila Schltr. longicauda Hook. ludibundiciliata J.A.N. Bat. and Bianch. ludibundiciliata J.A.N. Bat. and Bianch. macilenta (Lindl.) Rchb.f. macilenta (Lindl.) Rchb.f. macilenta (Lindl.) Rchb.f. macroceratitis Willd. macronectar (Vell.) Hoehne macvaughiana R. González magdalenensis Hoehne magniscutata Catling mannii Hook.f. aff. meeana Toscano melanopoda Hoehne and Schltr. melanopoda Hoehne and Schltr. melanopoda Hoehne and Schltr. melanopoda Hoehne and Schltr. aff. melanopoda Hoehne and Schltr. mello-barretoi Brade and Pabst monorrhiza (Sw.) Rchb.f. montevidensis Spreng. montiswilhelminae Renz montiswilhelminae Renz montiswilhelminae Renz aff. rodriguesii Cogn. mystacina Lindl. nabucoi Ruschi nasuta Rchb.f. and Warm. nemorosa Barb.Rodr. nuda Lindl. aff. nuda Lindl. aff. nuda Lindl. aff. nuda Lindl. aff. nuda Lindl. aff. nuda Lindl. nuda var. pygmaea Hoehne Voucher Origina ITS matK Batista 2428 (BHCB) Pansarin s.n. (BHCB) Salazar 6461 (MEXU) Batista 344 (CEN) Jiménez 2691 (AMO) Salas 6025 (MEXU) Batista 1795 (BHCB) Batista 2414 (BHCB) Batista 2592 (BHCB) Batista 2537 (BHCB) Batista 1519 (CEN) Mota 1584 (BHCB) Batista 2802 (BHCB) Batista 674 (CEN) Batista 1789 (BHCB) Batista 59 (CEN) Batista 2399 (BHCB) Batista 2497 (BHCB) Batista 1901 (BHCB) Nava s.n. (AMO) Batista 1123 (CEN) Batista 2513 (BHCB) Batista 2950 (BHCB) Batista 1380 (CEN) Mota 3566 (BHCB) Batista 1827 (BHCB) Mota 2777 (BHCB) Batista 2452 (BHCB) Batista 1548 (CEN) Klein 13 (UPCB) GenBank Batista 1497 (CEN) Batista 673 (CEN) Batista 1595 (BHCB) Batista 1624 (BHCB) Batista 2954 (BHCB) Batista 2658 (BHCB) Batista 1604 (BHCB) GenBank Batista 1590 (BHCB) Jardim 4529 (HUEFS) Batista 1372 (CEN) Batista 2393 (BHCB) Batista 2378a (BHCB) Batista 2354 (BHCB) Chávez s.n. (MEXU) Batista 2519 (BHCB) Nava s.n. (AMO) Batista 2026 (BHCB) Batista 1227 (CEN) Salazar 6314 (YA) Batista 2028 (BHCB) Batista 1832 (BHCB) Batista 1810 (BHCB) Batista 2471 (BHCB) Batista 2539 (BHCB) Batista 2438 (BHCB) Batista 2666 (BHCB) Salazar 7638A (MEXU) Batista 2479 (BHCB) Batista 1555 (CEN) Batista 2493 (BHCB) Batista 1811 (BHCB) Batista 970 (CEN) Batista 1812 (BHCB) Pivari 549 (BHCB) Batista 1572 (BHCB) Batista 2567 (BHCB) Batista 2869 (BHCB) Batista 1490 (CEN) van den Berg 1238 (HUEFS) Batista 1368 (CEN) Batista 2527 (BHCB) Batista 2091 (BHCB) Batista 939 (CEN) Brazil, MG Brazil, MG Mexico Brazil, DF Mexico Mexico Brazil, MG Brazil, MG Brazil, MG Brazil, MG Brazil, DF Brazil, SC Brazil, MG Brazil, DF Brazil, DF Brazil, DF Brazil, MG Brazil, RS Brazil, MG Mexico Brazil, DF Brazil, RS Brazil, GO Brazil, MG Brazil, MG Brazil, MG Brazil, MG Brazil, MG Brazil, DF Brazil, RS Africa Brazil, DF Brazil, DF Venezuela Brazil, MG Brazil, GO Brazil, RJ Brazil, DF Africa Brazil, PA Brazil, PA Brazil, MA Brazil, MG Brazil, GO Brazil, DF Mexico Brazil, PR Mexico Brazil, MG Brazil, GO Cameroon Brazil, MG Brazil, MG Brazil, MG Brazil, RS Brazil, MG Brazil, MG Brazil, MG Ecuador Brazil, RS Brazil, DF Brazil, RS Brazil, MG Brazil, DF Brazil, MG Brazil, MG Brazil, GO Brazil, MG Brazil, MG Brazil, GO Brazil, GO Brazil, MA Brazil, PR Brazil, MG Brazil, DF HM777632 HM777545 HF560573 HM777612 HF560574 HF560575 HM777539 HM777542 HM777541 HM777529 HM777586 HM777622 HM777623 HM777653 HM777602 HM777537 HM777538 HM777524 HM777581 HF560576 HM777648 HM777650 HM777678 HM777723 HM777663 HM777669 HM777609 HM777596 HM777531 KC257469 DQ522076 HM777710 HM777711 HM777660 HM777661 HM777547 HM777597 HM777568 DQ522077 HM777608 HM777639 HM777638 HM777606 HM777607 KC257472 HF560577 HM777614 HF560578 HM777595 HM777641 HF560579 HM777713 HM777689 HM777686 HM777687 HM777688 HM777681 HM777685 HF560580 HM777619 HM777714 HM777719 HM777693 HM777652 HM777728 HM777611 HM777716 HM777634 HM777718 HM777664 HM777665 HM777715 HM777720 HM777722 HM777651 HM777876 HM777849 HF560597 HM777844 HF560598 HF560599 HM777900 HM777898 HM777899 HM777793 HM777865 HM777827 HM777828 HM777956 HM777915 HM777908 HM777909 HM777924 HM777879 HF560600 HM777926 HM777927 HM777928 HM777975 HM777894 HM777919 HM777841 HM777854 HM777897 KC257477 DQ522096 HM777938 HM777937 HM777951 HM777950 HM777968 HM777855 HM777790 DQ522098 HM777843 HM778007 HM778006 HM777811 HM777814 HM777813 HF560601 HM777833 HF560602 HM777858 HM777880 HF560603 HM777965 HM777890 HM777887 HM777889 HM777888 HM777892 HM777886 HF560604 HM777826 HM777957 HM777960 HM777961 HM777986 HM777970 HM777840 HM777959 HM777872 HM777981 HM777922 HM777921 HM777985 HM777958 HM777982 HM777984 (continued on next page) 100 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 Table 1 (continued) Taxon Voucher Origina ITS matK Habenaria obtusa Lindl. Habenaria cf. odontopetala Rchb.f. Habenaria orchiocalcar Hoehne Habenaria pabstii J.A.N. Bat. and Bianch. Habenaria paranaensis Barb.Rodr. Habenaria parviflora Lindl. Habenaria parviflora Lindl. Habenaria paulensis Porsch Habenaria paulistana J.A.N. Bat. and Bianch. Habenaria petalodes Lindl. Habenaria petalodes Lindl. Habenaria cf. piraquarensis Hoehne Habenaria pleiophylla Hoehne and Schltr. Habenaria praestans Rendle Habenaria pratensis (Salzm. ex Lindl.) Rchb.f. Habenaria psammophila J.A.N. Bat., Bianch. and B.M. Carvalho Habenaria pseudoculicina J.A.N. Bat. and Bianch. Habenaria pseudoglaucophylla J.A.N. Bat., R.C. Mota and N. Abreu Habenaria pseudohamata Toscano Habenaria pubidactyla J.A.N. Bat. and Bianch. Habenaria pubidactyla spp. brasiliensis J.A.N. Bat. and Bianch. Habenaria pubidactyla var. apiculatipetala J.A.N. Bat. and Bianch. Habenaria pungens Cogn. Habenaria quinqueseta (Michx.) A. Eaton Habenaria regnellii Cogn. Habenaria regnellii Cogn. Habenaria repens Nutt. Habenaria repens Nutt. Habenaria aff. repens Nutt. Habenaria rodeiensis Barb.Rodr. Habenaria aff. rodeiensis Barb.Rodr. Habenaria aff. rodeiensis Barb.Rodr. Habenaria rolfeana Schltr. Habenaria rolfeana Schltr. Habenaria roraimensis Rolfe Habenaria rotundiloba Pabst Habenaria rupicola Barb.Rodr. Habenaria cf. rupicola Barb.Rodr. Habenaria rzedoswkiana R. González Habenaria schenckii Cogn. Habenaria schwackei Barb.Rodr. Habenaria schwackei Barb.Rodr. Habenaria secunda Lindl. Habenaria secundiflora Barb.Rodr. Habenaria secundiflora Barb.Rodr. Habenaria setacea Lindl. Habenaria setacea Lindl. Habenaria seticauda Lindl. Habenaria sobraliana J.A.N. Bat., A.A. Vale and Menini Habenaria spanophytica J.A.N. Bat. and Bianch. Habenaria spathulifera Cogn. Habenaria sprucei Cogn. Habenaria strictissima Rchb.f. Habenaria subauriculata Robinson and Greenm. Habenaria subfiliformis Cogn. Habenaria subfiliformis Cogn. Habenaria aff. subfiliformis Cogn. Habenaria aff. subfiliformis Cogn. Habenaria subviridis Hoehne and Schltr. Habenaria subviridis Hoehne and Schltr. Habenaria tamanduensis Schltr. Habenaria tridens Lindl. Habenaria trifida Kunth Habenaria trifida Kunth Habenaria cf. uliginosa Rchb.f. Habenaria umbraticola Barb.Rodr. Habenaria urbaniana Cogn. Habenaria warmingii Rchb.f. and Warm. Habenaria warmingii Rchb.f. and Warm. Habenaria weileriana Schltr. Orchis quadripunctata Cirillo ex Ten. Platanthera chlorantha (Custer) Rchb. Stenoglottis longifolia Hook.f. Batista 291 (CEN) Batista 2037 (BHCB) Batista 1570a (BHCB) Batista 2360 (BHCB) Batista 2436 (BHCB) Batista 1813 (BHCB) Batista 2477 (BHCB) Batista 2481 (BHCB) Pansarin 726 (UEC) van den Berg 1014 (HUEFS) van den Berg 1481 (HUEFS) Batista 1050 (CEN) Batista 2514 (BHCB) GenBank Batista 2686 (BHCB) Batista 1794 (BHCB) Batista 1808 (BHCB) Mota 2818 (BHCB) Batista 2035 (BHCB) van den Berg 1360 (HUEFS) Batista 1785 (BHCB) Batista 1615 (BHCB) Batista 2095 (BHCB) Sánchez s.n. (SERO) Batista 2801 (BHCB) Barfknecht s.n. (BHCB) Batista 2522 (BHCB) van den Berg 929 (HUEFS) Batista 2100 (BHCB) Mota 2824 (BHCB) Batista 2379 (BHCB) Batista 1738 (BHCB) Mota 3563 (BHCB) Batista 2467 (BHCB) Mota 1247 (BHCB) Batista 2684 (BHCB) van den Berg 1279 (HUEFS) Batista 2568 (BHCB) Jacob 234 (MEXU) Batista 2882 (BHCB) Batista 1524 (CEN) Batista 2524 (BHCB) Batista 2640 (BHCB) Batista 2392 (BHCB) Batista 2526 (BHCB) Mota 3019 (BHCB) Batista 1417 (CEN) Batista 1596 (BHCB) Batista 2499 (BHCB) Batista 2408 (BHCB) Without voucher Batista 3086 (BHCB) Leutzi s.n. (MEXU) García-Mendoza 7988 (MEXU) Batista 1597 (BHCB) Batista 2022 (BHCB) Batista 1788 (BHCB) Batista 2808 (BHCB) Batista 1814 (BHCB) Batista 2605 (BHCB) Batista 1784 (BHCB) GenBank Batista 1571a (BHCB) Batista 1783 (BHCB) Batista 1620 (BHCB) Mota 3569 (BHCB) Batista 911 (CEN) Batista 2409 (BHCB) Batista 2584 (BHCB) Salazar 6310 (YA) GenBank GenBank GenBank Brazil, DF Brazil, MG Brazil, GO Brazil, DF Brazil, MG Brazil, MG Brazil, RS Brazil, RS Brazil, SP Brazil, BA Brazil, BA Brazil, DF Brazil, RS Africa Brazil, BA Brazil, MG Brazil, MG Brazil, MG Brazil, MG Brazil, MG Brazil, DF Brazil, MG Brazil, GO Mexico Brazil, MG Brazil, PR Brazil, PR Brazil, BA Brazil, MG Brazil, MG Brazil, GO Brazil, MG Brazil, MG Brazil, MG Brazil, RR Brazil, BA Brazil, MG Brazil, MG Mexico Brazil, BA Brazil, GO Brazil, PR Brazil, RJ Brazil, MG Brazil, PR Brazil, MG Brazil, MG Venezuela Brazil, RS Brazil, MG Brazil, RR Brazil, GO Mexico Mexico Venezuela Brazil, MG Brazil, DF Brazil, MG Brazil, MG Brazil, MG Brazil, DF Africa Brazil, GO Brazil, DF Brazil, MG Brazil, MG Brazil, MG Brazil, MG Brazil, MG Cameroon Europe Europe Africa HM777587 HM777591 HM777662 HM777666 HM777528 HM777560 KC257475 HM777556 HM777610 HM777583 – HM777633 HM777594 DQ522079 HM777546 HM777550 HM777707 HM777590 HM777593 HM777702 HM777690 HM777729 HM777570 HF560581 HM777603 HM777604 HM777627 HM777628 HM777624 HM777577 HM777578 HM777579 HM777727 HM777730 HM777676 HM777717 HM777534 HM777533 HF560582 HM777580 HM777656 HM777657 HM777525 HM777637 HM777636 HM777731 KC257474 HM777584 HM777704 HM777576 HM777544 KC257470 HF560583 HF560584 HM777572 HM777573 HM777705 HM777709 HM777679 HM777680 HM777600 DQ522080 HM777671 HM777672 HM777625 HM777605 HM777658 HM777616 HM777617 HF560585 Z94105/Z94106 AY704975 AF348065 HM777862 HM777867 HM777914 HM777912 HM777796 HM777800 KC257478 HM777798 HM777838 – HM777861 HM777877 HM777857 DQ522100 HM777847 HM778000 HM777943 HM777852 HM777856 HM777972 HM777974 HM777973 HM778011 HF560605 HM777830 HM777831 HM777816 HM777817 HM777818 HM777995 HM777992 HM777991 HM777977 HM777978 HM777925 HM778017 HM777910 HM777911 HF560606 HM777869 HM777954 HM777953 HM777791 HM778004 HM778005 HM777980 HM777979 HM777864 HM777990 HM778016 HM777850 KC257479 HF560607 HF560608 HM778009 HM778008 HM777936 HM778010 HM777929 HM777930 HM777993 DQ522101 HM777916 HM777917 HM777824 HM777874 HM777944 HM777821 HM777820 HF560609 AY368385 DQ522103 AY368387 a Abbreviations for Brazilian states are: BA = Bahia; DF = Distrito Federal; MA = Maranhão; MG = Minas Gerais; GO = Goiás; MT = Mato Grosso; PA = Pará; PR = Paraná; RJ = Rio de Janeiro; RR = Roraima; RS = Rio Grande do Sul; SC = Santa Catarina; SP = São Paulo. 101 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 Table 2 Sections of Neotropical Habenaria following Kränzlin (1892, 1901) and Cogniaux (1893), with the number of taxa in each section sampled for this study, excluding synonyms. Section Abbreviation No. of taxa/no. of taxa sampled % of taxa sampled Clypeatae Macroceratitae Maculosae Micranthae Microdactylae Microstylinae Nudae Odontopetalae Pentadactylae Pratenses Pycnostachyae Quadratae Seticaudae Spathaceae CLY MAC MCU MTH MDA MST NUD ODO PEN PRA PYC QUA SET SPA 10/5 14/10 3/2 14/13 6/4 5/3 12/11 4/2 32/18 6/5 1/0 9/6 5/3 12/9 50 71 67 93 67 60 92 50 56 83 0 67 60 75 Table 4 Taxon sampling, matrix values and parsimony statistics of the ITS, matK and combined data sets. No. taxa No. Neotropical Habenaria No. old world orchidiinae and disinae No. sequences No. Neotropical Habenaria sequences No. old world orchidiinae and disinae sequences Aligned length Variable parsimony-uninformative characters Parsimony-informative characters Best trees length No. of trees Consistency index (CI) Retention index (RI) No. of internal nodes MP/BIa No. nodes with bootstrap >85% ITS matK ITS + matK 180 152 (51%) 28 217 188 29 180 152 (51%) 28 217 188 29 180 152 (51%) 762 97 (12.7%) 365 (48%) 1454 4340 0.53 0.81 109/124 46 (42%) 627 67 (10.7%) 122 (19%) 403 8830 0.60 0.85 62/73 14 (22.6%) 60 (82%) 3. Results No. nodes with posterior probabilities >0.95 3.1. Sequence divergence a Sequence divergence among the Neotropical taxa of Habenaria was low. The mean pairwise distance for the Neotropical ITS sequences (152 taxa) was only 0.018, whereas the mean pairwise distance for the analyzed African sequences of Habenaria (10 species) was 0.142 (Table 3). To analyze whether these results were related to the difference in the number of taxa sampled between the two groups, pairwise distances were calculated for selected taxa within the Neotropical Habenaria; the results were similarly low independent of the number or taxa sampled (data not shown). This divergence among African Habenaria was surprisingly high because it was similar to or greater than that obtained for other groups represented by different genera: 0.083 for Habenariinae (Cynorkis and Stenoglottis) and 0.134 for Orchidinae (Orchis and Plathantera). Among the groups included in the analysis, only the Bonatea species had mean pairwise distances similar to the Neotropical Habenaria species (0.015). The monospecific genus Gennaria showed the greatest divergence from other groups due to its high number of autapomorphic characters. Divergence among matK sequences was similar to that of the ITS region, albeit less. The mean pairwise distance was 0.009 for the Neotropical Habenaria sequences and 0.029 for the African species (data not shown). As a result of low divergence, alignments were straightforward and unambiguous for the Neotropical ITS and all of the matK sequences. In contrast, the Old World ITS sequences were more difficult to align, and some ambiguous positions were present in the alignment. We attempted to include the basal genus Codonorchis Lindl. (Kores et al., 2001) as an outgroup in the phylogenetic analyses, but aligning its ITS sequence proved much more difficult and ambiguous, for which reason it was not used. Accordingly, Table 3 Mean pairwise distances within and between the ITS sequences of the major groups used in this study. 1. 2. 3. 4. 5. 6. Neotropical Habenaria African Habenaria Habenariinae Bonatea Gennaria Orchidinae 1 2 3 4 5 6 0.018 0.132 0.127 0.117 0.217 0.199 0.142 0.133 0.116 0.221 0.196 0.083 0.102 0.185 0.150 0.015 0.199 0.168 n/c 0.246 0.134 86 (69%) 28 217 188 29 1389 164 (11.8%) 487 (35%) 1897 5110 0.53 0.81 131/151 65 (49.6%) 106 (70%) Abbreviations: MP = Maximum parsimony; BI = Bayesian inference. Chemisquy and Morrone (2012) suggested that the published ITS sequence of Codonorchis is most likely a pseudogene, as it lacks conserved motifs and was highly divergent in comparison to other analyzed sequences. 3.2. Parsimony analyses Initially, we performed separate analyses for the ITS and matK data sets. The ITS matrix consisted of 762 characters, of which 365 (48%) were parsimony-informative. The parsimony analysis found 4340 shortest trees with a length of 1454 steps, consistency index (CI) of 0.53 and retention index (RI) of 0.81. The matK matrix did not have a single indel due to its low sequence divergence. The latter matrix had 627 characters, of which 122 (19%) were parsimony-informative. The parsimony analysis found 8830 shortest trees with a length of 403 steps, CI of 0.60, and RI of 0.85 (Table 4). Overall, the resolution of the strict consensus and bootstrap trees of the matK matrix was lower than that obtained with the ITS data set. Because the consensus and bootstrap trees of the matK matrix were largely unresolved and no supported conflict was found between the trees of the individual analyses, only the results of the combined analyses will be presented and described. The general features of the datasets including taxon sampling and parsimony statistics are presented in Table 4. The combined matrix consisted of 1389 characters, of which 487 (35%) were parsimony-informative. A total of 5110 shortest trees were found, with a length of 1897 steps, CI of 0.53 and RI of 0.81 (Table 4). Because the parsimony trees are largely congruent with the Bayesian trees but are less resolved and with weaker overall support, a strict consensus tree of the combined datasets is presented as Supplementary Material (Fig. S1 and S2). The Neotropical taxa of Habenaria formed a highly supported group (99% BS). Habenaria tridens Lindl. from Africa was sister to the Neotropical clade (100% BS), whereas H. batesii, also from Africa, was sister to the Neotropical Habenaria–H. tridens (82% BS). Three other African species of Habenaria (H. dives Rchb.f., H. clavata (Lindl.) Rchb.f. and H. lithophila Schltr.) formed a strongly supported clade (100% BS) that was sister to the Neotropical Habenaria–H. tridens–H. batesii clade (100% BS). All species of Bonatea along with Habenaria 102 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 Fig. 1. Bayesian tree from the combined ITS and matK datasets. Bootstrap percentages from the parsimony analysis and posterior probabilities are shown next to nodes. Neotropical subgroups are boxed and numbered. The generic name for all Neotropical species is abbreviated. The three letter abbreviation to the right of the species name indicates its sectional classification (Table 2). The sectional classifications of African Habenaria are also shown. Alternative recent generic circumscriptions are indicated in brackets. Available somatic chromosome numbers are shown in the terminal branches for a few species. The type species of Habenaria, H. macroceratitis, is highlighted in bold. For species sampled more than once, the two or three letter abbreviation after the species name indicates the geographic origin of the sample. Abbreviations are MEX = Mexico; VEN = Venezuela; and for Brazilian states: BA = Bahia; DF = Distrito Federal; MA = Maranhão; MG = Minas Gerais; GO = Goiás; PA = Pará; PR = Paraná; RS = Rio Grande do Sul; and SC = Santa Catarina. J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 103 Fig. 2. Continuation of the tree of Fig. 1. A single tree with proportional branch lengths is shown in the upper left-hand corner to show the low levels of divergence within the Neotropical clade. For abbreviations, see Fig. 1. laevigata Lindl. formed a strongly supported clade (100% BS). All species of Habenaria along with Bonatea and Gennaria formed a moderately supported clade (78% BS). The Habenaria–Bonatea– Gennaria clade was sister successively to Cynorkis (87% BS), Stenoglottis (66% BS), an Orchis–Platanthera clade (86% BS), and Satyrium (100% BS). 104 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 3.3. Bayesian analyses Similarly to the parsimony analyses, the ITS and matK data sets were initially analyzed separately. Because no significant incongruences were detected between the plastid and nuclear data, a final search was performed with the combined matrices. The resulting Bayesian majority-rule consensus tree was fully congruent with the strict consensus tree of the combined parsimony analyses but was more resolved and with stronger overall support. Because the combined ITS/matK data set included a broad taxonomic sample and is, overall, better resolved and supported than either of the other analyses, it is the tree that best represents our phylogenetic hypothesis and was chosen for presentation and discussion (Figs. 1 and 2). The Neotropical Habenaria formed a monophyletic group with strong support (1.00 PP), with Habenaria tridens (section Dolichostachyae) sister to it (1.00 PP). Habenaria batesii (section Podandria) was sister to the Neotropical Habenaria–H. tridens clade (0.91 PP), whereas Habenaria dives (section Bilabrellae), H. clavata (section Ceratopetalae) and H. lithophila (section Diphyllae) formed a strongly supported clade (1.00 PP) that was sister to the Neotropical–H. tridens–H. batesii clade (1.00 PP). All sampled species of Bonatea along with H. laevigata and H. arenaria Lindl. (section Chlorinae) also formed a strongly supported clade (1.00 PP). The other sampled African Habenaria (H. mannii, H. praestans and H. weileriana) formed with the Bonatea–H. laevigata–H. arenaria clade a strongly supported clade (0.97 PP), which was sister to the Neotropical–H. tridens–H. batesii–H. dives–H. clavata–H. lithophila clade, but with low support (0.66 PP). The Habenaria–Bonatea clade was sister successively to Gennaria (1.00 PP), Cynorkis (1.00 PP), Stenoglottis (0.86 PP), an Orchis–Platanthera clade (0.92 PP), and Satyrium (1.00 PP). The Neotropical species formed several wellsupported subgroups, but many species were either unresolved or weakly supported as sisters to the subgroups. Relationships among the Neotropical subgroups were poorly resolved, as support for the internal nodes of the tree was low overall. 4. Discussion 4.1. Generic limits of Habenaria Irrespective of the method of inference, in all of our analyses of the combined ITS–matK dataset, the Neotropical Habenaria formed a strongly supported monophyletic group (99% BS; 1.00 PP) (Fig. 1, clade A). Furthermore, the African species Habenaria dives, H. clavata and H. lithophila were strongly supported as sister (100% BS; 1.00 PP) to the clade formed successively by the African species H. batesii and H. tridens plus the whole Neotropical clade. These results show that Neotropical Habenaria and African representatives of the genus in the sections Dolichostachyae, Podandria, Bilabrellae, Ceratopetalae, and Diphyllae form a strongly supported ‘‘core Habenaria clade’’ (Fig. 1, clade B) that includes the type species of the genus, H. macroceratitis from the New World. Another group strongly supported by the trees was the ‘‘Bonatea clade’’ formed by the species of Bonatea (formerly a section of Habenaria) plus Habenaria laevigata (section Chlorinae) (100% BS; 1.00 PP). In the Bayesian analysis, this clade also included Habenaria arenaria (section Chlorinae) (1.00 PP), H. mannii and H. praestans (section Multipartitae) plus H. weileriana (section Chlorinae) (0.97 PP) (Fig. 1, clade C), but this result was not supported in the parsimony analysis. Relationships between the ‘‘core Habenaria clade’’ and the ‘‘Bonatea clade’’ were not resolved. In the Bayesian analysis, the ‘‘Bonatea clade’’ was sister to the ‘‘core Habenaria clade’’, rendering all sampled taxa of Habenaria plus Bonatea as monophyletic, but support was low (0.66 PP) (Fig. 1). The monospecific genus Gennaria, formerly included in Habenaria, was highly supported (1.00 PP) as sister to the Habenaria–Bonatea clade. However, the Gennaria sequences were full of autapomorphies, and the position of this species varied in other analyses according to the data sets, species sampled and method of analysis. A division of the African Habenaria into two major clades was first envisioned in the phylogenetic analysis of Bellstedt et al. (2001), albeit with a much lower sampling that used the trnL intron and the trnL-trnF spacer region to investigate phylogenetic relationships in Disa. In that work, H. pseudociliosa Schelpe ex J.C. Manning (section Chlorinae), H. malacophylla Rchb.f. (section Ceratopetalae), and H. laevigata formed one strongly supported clade and H. tysonii Bolus (section Diphyllae) and H. dives formed another. Beyond the division of the Neotropical and African Habenaria into two major clades, in the study of Bateman et al. (2003), Asian species of Habenaria (H. sagittifera Rchb.f. [section Cruciatae], H. tibetica Schltr., H. delavayi Finet) were more closely related to Pecteilis Raf. (Asian) and Herminium L. (Euro–Asian), whereas African (H. arenaria, H. procera (Afzel. ex Sw.) Lindl., H. tridactylites Lindl. [section Tridactylae]) and a single Neotropical species (H. odontopetala Rchb.f.) were grouped with Bonatea (African) and Gennaria (Canary Islands, west and central Mediterranean). A similar result was found in a molecular phylogenetic analysis of Diseae that also used the ITS region (Douzery et al., 1999), where H. sagittifolia (a misspelling of H. sagittifera, from China and Japan) formed a strongly supported group (95% BS) with Herminium, and Bonatea speciosa (L.f.) Willd., Habenaria arenaria and H. procera (section Chlorinae), all from tropical Africa, formed a clade with moderate bootstrap support (77% BS). An exception was the highly supported clade (100% BS) formed by H. repens (Neotropical) and Holothrix, a member of Orchidinae s.s. from tropical and southern Africa. However, this last result is doubtful because the two sequences are identical and according to Bateman et al. (2003), H. repens is either misnamed or misidentified in this work. Although Asian species of Habenaria were not included in our analyses, these results are consistent with ours and indicate a strong relationship between geographical and phylogenetic structure with a separation between African-Neotropical and some Asian groups of Habenaria. Further inferences on the generic limits of Habenaria are limited by the availability of molecular data. Only a few Habenaria from the Old World and eight of the 23 genera of Habenariinae (sensu auct.) listed in Genera Orchidacearum (Pridgeon et al., 2001a) are currently available for molecular analyses. 4.2. Comparison of the New World sectional classification with the phylogenetic analysis The current infrageneric classification of Habenaria is based mostly on Kränzlin’s (1892, 1901) sectional treatments. In his system, the Neotropical species were divided into 12 sections. To facilitate the comparison of this sectional treatment with the results of our cladistic analyses, the sectional assignment of the taxa according to Kränzlin (1892, 1901) and Cogniaux (1893) is indicated after the species’ name in Figs. 1 and 2. Abbreviations for the Neotropical sections, number of species, and species sampled in each section are shown in Table 2. All sections of the current sectional classification are paraphyletic or polyphyletic with regard to the subgroups recovered in the molecular tree. Species from the section Macroceratitae Kraenzl. (MAC) are concentrated in subgroup 4, but other species assigned to this section are dispersed across subgroups 2, 7, 8, 13, 16 and 17. Similarly, many species from section Pratensis Kraenzl. (PRA) are concentrated in subgroup 1, but other species assigned to this section are in subgroup 5 or are unresolved in different positions of the tree. The species in section Clypeatae Kraenzl. (CLY) are dispersed across subgroups 2, 6, 8 and 9 of the tree. Section Pentadactylae Kraenzl. (PEN), the largest among J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 the Neotropical sections, and section Micranthae Kraenzl. (MTH) are also polyphyletic, and the species assigned to these sections are either placed in several different subgroups or are unresolved. The species in sections Quadratae Kraenzl. (QUA) and Seticaudae Kraenzl. (SET) are concentrated in subgroup 10, but this subgroup also includes species assigned to section Micranthae. Even the species belonging to the smaller sections Microstylinae Kraenzl. (MST) and Microdactylae Kraenzl. (MDA) are dispersed across the tree. For the Old World sections of the genus, the current taxon sampling is not sufficient for an evaluation of the infrageneric groups, and they will not be discussed. 4.3. Phylogenetic relationships within New World Habenaria Several terminal nodes of the Neotropical clade formed monophyletic subgroups (Figs. 1 and 2), and approximately 21 such subgroups were recovered that comply with at least one of the following criteria: (1) the group is moderately to strongly supported in the phylogenetic analyses; and (2) the species are morphologically similar. However, several species were unresolved relative to these subgroups, and the relationships between the subgroups are not clear because most of the internal nodes of the trees were poorly resolved or weakly supported. Although the relationships among the subgroups are mostly unresolved, a general characterization and discussion of the subgroups is presented as an attempt to correlate the results of the phylogenetic analyses with the most salient taxonomic, morphological, biogeographical and evolutionary aspects of each group/subgroup of the Neotropical species. 4.3.1. Basal clades: subgroups 1–6 In the Bayesian analysis of the ITS–matK dataset, the Neotropical group formed a polytomy at the base of the clade that included H. monorrhiza, subgroup 1 and the clade formed by all other Neotropical species. Habenaria monorrhiza is a weedy species commonly found by the side of roads along most of Mexico, Central America and northern South America. Subgroup 1 consists of a small group of four species, of which three were sampled here, that differs from all other Neotropical Habenaria in its yellow to orange perianth, laterally expanded segments of the petals and labellum, and diurnal fragrance (Hoehne, 1940). Based on these features, Singer and Cocucci (1997) suggested butterfly pollination for this group, in contrast with most other Habenaria species, which are fragrant at night and have flower syndromes associated with moth pollination. Subgroups 2–6 and some unplaced species form a clade. Although support is low, (0.81 PP) all species in these subgroups have patent, spreading leaves and are concentrated mostly in southern South America, but they vary widely in flower size and morphology. Subgroups 2 and 3 are strongly supported in all analyses and form a well-supported clade including H. leucosantha (0.96 PP). Based on morphological similarities, subgroup 2 is composed of approximately 11 species (five sampled) and subgroup 3 eight species (five sampled). Species in subgroup 2 were assigned to sections Macroceratitae, Micranthae, Pentadactylae and Clypeatae, whereas the species in subgroup 3 were placed in sections Micranthae and Microstylinae in the sectional treatments of Kränzlin (1892, 1901) and Cogniaux (1893). The length of the lateral segments of the petals and labellum, one of the characters used by Kränzlin for the characterization of the sections, is highly variable in subgroup 2, which explains the placement of species in this subgroup in different sections, but this subgroup is otherwise homogeneous in vegetative and floral characters. The center of diversity of both subgroups is southeastern and southern Brazil. The monophyly of subgroups 4–6 was strongly supported in all analyses and in the Bayesian analysis of the combined data sets; 105 they formed a strongly supported clade with H. macilenta sister to the group (Fig. 1) (1.00 PP). However, there are no evident floral morphological similarities between the species, and the inclusion of H. macilenta in this clade was unexpected because this species has a dissimilar morphology and geographic distribution (central and northern Brazil and the Guianas) and on the basis of flower morphology was associated by previous authors (Hoehne, 1940; Pabst and Dungs, 1975) with the species recovered here in subgroup 12. Subgroup 4 consists of a group of nine taxa (seven sampled) with several morphological similarities. A distinctive character of this subgroup is the long (8–14 mm), involute stigma lobes, a character not found in any other Neotropical group of the genus. Also distinct from other Neotropical Habenaria is the short, erect, tooth-like process, originated from the lip, in front of the entrance to the spur found in some species of this subgroup. This feature is one of the characters used to separate Bonatea from Habenaria, but our results show that it is homoplasious. The species in subgroup 4 were previously assigned to section Macroceratitae (Kränzlin, 1901; Cogniaux, 1893; Batista et al., 2006), but H. macroceratitis, the type species of this section, is morphologically distinct and distantly related in the sequence data trees (Fig. 1). Species in subgroup 4 have the longest spurs among Neotropical Habenaria, reaching up to 20 cm in H. longicauda (Batista et al., 2006). Subgroup 5 is composed of approximately seven species (three sampled), found mostly in southern Brazil, Uruguay and Argentina. Subgroup 6 is composed of approximately 12 species (three sampled) including H. repens, the most widespread species of the genus in the Neotropics, extending from the southern USA to northern Argentina. Habenaria repens is typically aquatic, whereas the other species in the subgroup usually grow in water-logged places. Habenaria warmingii has never been formally associated with H. repens and was placed in section Pentadactylae by Kranzlin (1892, 1901) and Cogniaux (1893). However, the two species agree well in their vegetative and floral morphology. 4.3.2. Forest clades: subgroups 7–10 In the Bayesian analysis of the combined ITS and matK datasets, subgroups 7–21 formed a clade (0.96 PP) with subgroups 7–10 placed in a basal polytomy along with a few other unresolved species. Although the relationships among subgroups 7–10 were not resolved, the species in these subgroups are mostly forest-dwelling, in contrast with most other New World Habenaria, which occur in open savanna or grasslands. Forest species are common among African and Asian Habenaria but in the New World, are restricted to these subgroups, suggesting a relationship between them. The species in subgroups 7–10 always have well-developed elliptical, oblong or lanceolate leaves, but flower morphology is variable. Subgroup 7 consists of approximately nine species (six sampled) and in the New World, is the only subgroup formed exclusively by species from tropical forest. The support for this subgroup varied between the analyses (53% BS; 1.00 PP), but the similarities in habitat, habit and flower morphology support a close relationship between the species in this subgroup. Subgroup 8 includes H. macroceratitis, the type species of the genus. This subgroup consists of a small group of approximately five species (two sampled) that are primarily Mexican but also extend over the southern USA, the Caribbean and northern South America. The species in subgroup 9 are all Mexican. This subgroup most likely consists of approximately 56 taxa, mostly from Mexico (10 sampled), with a few species extending to Guatemala and other Mesoamerican countries. The low resolution and limited taxon sampling currently available for this subgroup did not allow a comparison with the 12 informal groups proposed by GonzálezTamayo (1993) for Mexican Habenaria. Considering the high number of species and the variation in flower size and morphology 106 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 displayed by the species of this subgroup, which range from minute flowers a few millimeters in diameter to large green (H. rzedowkiana) or white (H. clypeata) flowers, it is likely that a larger sampling and better resolution might reveal subdivisions. Contrary to most Neotropical species, which inhabit open, grassland habitats and, less frequently, tropical forests, the species in this subgroup primarily inhabit subtropical to warm-temperate conifer-oak forests along the mountains of much of Mexico. In all analyses, subgroup 9 is sister to subgroup 8, and the two form a strongly supported clade (97% BS; 1.00 PP). Subgroup 10 was strongly supported in all combined and individual analyses (98% BS; 1.00 PP). It is composed of a large group of approximately 34 species (16 sampled) distributed throughout the Neotropics. These species primarily inhabit tropical forests, but the group formed by H. seticauda, H. obtusa and H. hamata prefers dry grasslands, indicating a probable reversal. Subgroup 10 is the only one among the Neotropical Habenaria that includes species with entire petals and labellum. Entire flower segments were used to characterize the genera Habenella (Small, 1903; Szlachetko and Kras, 2006) and Platantheroides Szlach. (Szlachetko, 2004b), both segregated from Habenaria. However, our results do not support this division because subgroup 10 also includes two groups of species with short lateral segments on the labellum and petals (H. josephensis–H. leptoceras and H. pleiophylla–H. pseudohamata), indicating that entire segments evolved independently in two or three species groups within subgroup 10 (e.g., H. strictissima and H. brevilabiata; H. curvilabria and H. magdalenensis; and H. odontopetala through H. seticauda). On the basis of chromosome numbers and structure, Felix and Guerra (1998) were the first to suggest a close relationship between these two groups of species, e.g., those with entire segments and with short lateral segments. In accordance with this relationship, the morphology of the gynostemium is homogeneous across all of subgroup 10. Our results indicate that the recently described H. pseudoglaucophylla J.A.N. Bat., R.C. Mota and N. Abreu (Batista et al., 2008b), similar in gynostemium morphology but with long lateral segments of the labellum, is sister to subgroup 10, which suggests a progressive reduction in the size of the lateral segments within the subgroup. This subgroup has a unique triplet insertion mutation in positions 293–295 of ITS1. However, this insertion is absent in H. pseudoglaucophylla. All species sampled in subgroup 9 and some species of Bonatea have different insertions at the same position. Two species, Habenaria schenckii and H. depressifolia, are of particular interest among the species in the forest clades because they are vegetatively identical to each other and differ from all other Neotropical species in having 1–2 basal, orbiculate, fleshy leaves that lay adpressed to the ground. These vegetative characters are common among Old World species of Habenaria and characterize sect. Diphyllae, but in the New World are found only found in these two species. A close relationship between the two species was not confirmed in our analyses, but the unresolved positions in the trees indicate that the available molecular data has not been sufficient to resolve the relationship between them. 4.3.3. Subgroups 11–15 Similarly to subgroups 1–10, most species in subgroups 11–15 have elliptical to lanceolate spreading leaves, whereas flower morphology is variable. Some species in these subgroups are widespread, but most occur in central and southeastern Brazil. Subgroup 11 consists of approximately eight species (five sampled), usually occurring in high-altitude grasslands or in temperate areas at lower altitudes. Subgroup 12 consists of six species, of which all were sampled here. Habenaria trifida, distributed from Mexico to northern Argentina, has one of the broadest geographic ranges among the Neotropical species in the genus, but other species of this group are concentrated in the cerrado vegetation of cen- tral Brazil. They typically occur in grasslands and are characterized by a few-flowered inflorescence, medium to large flowers, a white corolla and a long pedicel. Subgroup 13 is composed of approximately 11 species (eight sampled), most of them occurring in central and southeastern Brazil. The inclusion of H. rupicola and H. coxipoensis in this subgroup was unexpected, as these two species have flower morphologies remarkably similar to H. repens (subgroup 6) and H. subviridis (subgroup 11), most likely indicating pollinator-driven homoplasy. Subgroup 14 consists of two morphologically dissimilar species with similar geographical distributions, concentrated on the rocky fields of the Espinhaço range in Minas Gerais and the highlands of Central Brazil. Subgroup 15 was not strongly supported, but the four species in the subgroup comprise a morphologically uniform assemblage. 4.3.4. Cerrado clades: subgroups 16–21 The species in subgroups 16–21 formed a well-supported clade (Clade D, Fig. 2) in the Bayesian analysis of the combined data sets (1.00 PP). In contrast with subgroups 1–15, most species in subgroups 16–21 have linear, grass-like leaves, which are commonly adpressed to the stem. Based on these characters, these species best correspond to sect. Nudae Cogn. They are primarily Brazilian and are mostly concentrated in the cerrado and campos rupestres (rocky fields) of the central and southeastern regions of the country. Regarding their vegetative parts, subgroup 16 is the only exception in clade D, as most species in this subgroup have lanceolate, spreading leaves. Subgroup 18 consists of a group of approximately eight taxa (five sampled). Some species in this subgroup were recently revised by Batista and Bianchetti (2010), but their circumscription of the H. crucifera group conflicts with the results of the molecular trees, indicating that the morphological characters they used to define the group are homoplasic. Habenaria pungens Cogn. from the cerrado of central Brazil and Bolivia is unequivocally placed in subgroup 18. This is one of the most interesting and distinct Habenaria species in the Neotropics because its flowers form a dense, umbel-like inflorescence and are bright-yellow and non-resupinate, resembling in their general aspect the African genus Platycoryne Rchb.f., formerly segregated from Habenaria. Unfortunately, no Platycoryne material was available for molecular study and it was therefore not possible to test whether such a floral similarity is indicative of a close relationship between H. pungens and Platycoryne or resulted from convergence. Nevertheless, the position of H. pungens deep in the Neotropical clade and the distant position of all African taxa sampled in the molecular trees suggest that a direct relationship between the two taxa is unlikely. Subgroups 19–21 along with some unresolved species formed a well-supported clade in the Bayesian analysis of the combined data sets (0.99 PP), but the relationships within this clade were unclear. Subgroup 20 is composed of a small group of four species (all sampled) concentrated in the cerrado biome of central Brazil. Species in this subgroup belong to four sections in the sectional classification of Kränzlin (1892, 1901) and Cogniaux (1893), but are morphologically cohesive and well-supported in the molecular tree (1.00 PP). Support for subgroup 21 was low, and relationships within the subgroup were poorly resolved. This subgroup currently consists of approximately 23 taxa (13 sampled) distributed from northeastern to southern Brazil but concentrated in the cerrado and rocky fields of central and southeastern Brazil. 4.4. Taxonomic implications The results of Bateman et al. (2003) indicated that Habenaria is highly polyphyletic, and those authors envisioned an extensive dismantling of Habenaria into smaller monophyletic genera. However, only eight species of Habenaria (approximately 1% of the genus) J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 were included in that analysis. At about the same time, Szlachetko (2003a,b, 2004a,b; Szlachetko and Kras, 2006) begin a worldwide division of the genus, recognizing three genera segregated from New World Habenaria: Bertauxia Szlach., Kusibabella Szlach. and Platantheroides Szlach. (a synonym of Habenella Small). In our phylogenetic analysis, Bertauxia is polyphyletic, with the three species assigned to that genus dispersed among clades 4 (H. vaupellii Rchb.f. and Warm. = H. johanennsis Barb.Rodr.), 16 (H. rodeiensis Barb.Rodr.) and 21 (H. nasuta Rchb.f. and Warm.) (Figs. 1 and 2). Even on morphological grounds, this genus has no support (Batista et al., 2006). In turn, Kusibabella and Habenella are paraphyletic. Kusibabella includes most species of subgroup 4 but with H. cryptophila Barb.Rodr. embedded in it and H. johannensis not included, whereas Habenella (formerly described as superfluous Platantheroides Szlach.) includes some of the species in subgroup 10 along with other African and Asian species. Although more narrowly circumscribed genera such as Kusibabella and Habenella could become monophyletic with some adjustments, we do not favor a generic fragmentation of the New World Habenaria on the basis of the following arguments: (1) the Neotropical Habenaria are monophyletic; (2) many lineages are composed of one or few taxa and consequently many genera with one or a few species would have to be created; (3) phylogenetic relationships between the Neotropical clade and many African and Asian clades of the genus are unresolved, and a subdivision of the Neotropical clade will require a corresponding extensive generic fragmentation of the African and Asian groups, for which there is limited molecular data available; and (4) the creation of new genera will not provide any additional information when compared to a sectional subdivision and will require extensive nomenclatural changes, whereas a sectional classification would not require major nomenclatural changes, just realignments of species as required to comply with monophyly and a morphological recircumscription. Rather than a generic fragmentation, we favor a revision of the current sectional classification of the Neotropical species together with a morphological, cytogenetic and biogeographic characterization of the subgroups, work that is already underway. Regarding the genus as a whole, it is clear that, as currently circumscribed, Habenaria is polyphyletic relative to Asian species in sections Cruciatae and Peristyloideae. However, a massive segregation of new genera is unlikely. Our results indicate that the Neotropical Habenaria form a strongly supported clade with African species in sections Diphyllae, Dolichostachyae, Ceratopetalae, Podandria and Bilabrellae, and they are likely to be kept together. However, it should be noted that the sectional assignment of the African species sampled in the molecular analyses was based primarily on Kränzlin’s sectional classification, which is highly artificial, at least for the Neotropical species. Nevertheless, despite the correct sectional classification of the African species sampled, our results indicate that some of them are more closely related to the Neotropical species than to other African species. On the other hand, the relationships of the Neotropical clade with other African sections of the genus such as Chlorinae and Multipartitae and other former sections now treated at the generic rank (Bonatea) are unresolved or weakly supported and will require a more comprehensive sampling of these groups. With the sampling currently available, the recognition of Bonatea at the generic level renders Habenaria paraphyletic and requires either the recognition of other African groups of Habenaria in sects. Chlorinae and Multipartitae as independent genera or the inclusion of Bonatea in a broadly circumscribed Habenaria. In this context, the recent transfer of Bonatea bracteata and B. tentaculifera to Habenaria (Ponsie et al., 2007b) is provisional, and a decision on the taxonomic status of Bonatea and other genera formerly placed in Habenaria such as Centrostigma, Platycoryne and Roeperocharis will 107 have to wait for a more comprehensive sampling, as our understanding of these groups is currently limited by the availability of molecular data. 4.5. Evolutionary trends in Habenaria Compared to African species, sequence polymorphism in the ITS and matK regions among New World Habenaria was much less frequent. The morphological variability of the Neotropical species is correspondingly smaller than that found in African and Asian groups of the genus. Furthermore, the generic diversity of tribe Orchideae in the Neotropics is much lower than tropical Africa and Asia. These results, together with the paraphyletic position of African species of Habenaria in relation to the Neotropical clade, indicate an African origin for the Neotropical clade and suggest a recent dispersal and radiation of the genus in the New World. A time calibration of our reconstructed phylogeny will provide insight into this hypothesis. The only chromosome counts for Neotropical Habenaria indicated a diploid number of 42 with other values such as 44, 50, 80 and 84 (Daviña et al., 2009; Felix and Guerra, 1998). When the available counts are plotted in our trees, the chromosome number increases in the derived lineages, suggesting that polyploidization and other forms of genome evolution may be related to the evolution and speciation of the genus in the New World. However, chromosome counts are still few, and only in one case (subgroup 10) was more than one species in the same group counted, making it difficult to extrapolate the number to each clade. If chromosome numbers are constant within each clade, chromosome number could be a useful character for the characterization of some clades, but additional data are necessary to confirm this possibility. Similarly to other large genera, the extent to which molecular phylogenies will be translated into systematic classifications is unclear. The relationships between the groups recovered in the molecular trees and morphological traits were not always clear. Many of the groups in the trees formed uniform assemblages of species, but in several instances, morphologically dissimilar species where grouped, whereas in other cases, species with similar morphological traits were dispersed among the subgroups, indicating homoplasy of these characters. These results indicate that morphological resemblance alone can be misleading for inferring relationships within the genus. 4.6. Conclusions and perspectives Our results establish unambiguously that Neotropical Habenaria are monophyletic and closely related to some African species of the genus. Furthermore, the topology of the trees indicates an African origin and the low divergence among the Neotropical sequences suggest a recent radiation of the genus in the Neotropics. Further work with a molecular dating approach will provide more insight into the timing of the radiation and other aspects of this issue. The precise relationships between Neotropical and African Habenaria and other groups of African Habenariinae could not be confidently resolved because sampling of African taxa is still limited. Currently, only a few African Habenaria and eight of the 23 genera of Habenariinae listed in Genera Orchidacearum (Pridgeon et al., 2001a) are available in GenBank. The inclusion of more African species and sections of Habenaria as well as the closely related genera Centrostigma, Platycoryne and Roeperocharis will be necessary to define generic limits. Among the Neotropical species, the identification of several well-supported subgroups will provide the basis for a revision of the current sectional classification. Taxon sampling of the Neotropical Habenaria has increased greatly, but additional fieldwork is re- 108 J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 quired in some groups. Relative to the number of species, the Mexican clade (subgroup 9) is still poorly sampled, and the addition of more taxa is necessary to resolve the relationships in this large group. Taxon sampling of some Andean species also needs to be improved, particularly for a few morphologically distinct species restricted to this region. On the other hand, analyses to resolve relationships between and within subgroups 16–21, which are now well-sampled, will require the addition of more DNA regions in the analyses, preferably nuclear genes with high variation. The few chromosome counts available suggest that karyotype evolution is related to the evolution and diversification of the genus in the New World, but more studies with a higher number of species are necessary. The inclusion of other features such as chromosome morphology may also provide insight into this question. Finally, for the first time, we have an explicit phylogenetic hypothesis for the Neotropical species of Habenaria that will provide a basis for investigating the patterns of morphological evolution, diversification and distribution of the genus in the New World. Acknowledgments The authors thank Rubens C. Mota, Nara F.O. Mota, E. Pansarin, Marco O.D. Pivari, Geraldo Barfknecht, João B.A. Bringel, Eric Smidt, Jacques Klein and Rolando J. Machorro for providing samples, Benny Bytebier, Graham Grieve and Rolando J. Machorro for some of the photographs used in graphical abstract and Fig. S1, IBAMA and IEF Minas Gerais for providing scientific collection permits and two anonymous reviewers who provided useful comments and corrections. G.A.S. thanks Laura Vázquez Valedelamar for her assistance with DNA sequencing. This research was financially supported by the Fundação de Amparo a Pesquisa do Estado de Minas Gerais – FAPEMIG, Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPq, and Pró-Reitoria de Pesquisa da Universidade Federal de Minas Gerais – UFMG. CvdB and JANB also acknowledge scholarships received from CNPq (Pq-1D and Pq-2). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ympev.2013.01. 008. References Barros, F., 1987. Orchidaceae. In: Giulietti, A.M., de Menezes, N.L., Pirani, J.R., Meguro, M., Wanderley, M.G.L. (Eds.), Flora da Serra do Cipó, Minas Gerais: caracterização e lista das espécies. Bol. Bot. Univ. São Paulo, vol. 9, pp. 125–130, 1–151. Bateman, R.M., Pridgeon, A.M., Chase, M.W., 1997. Phylogenetics of subtribe Orchidinae (Orchidoideae, Orchidaceae) based on nuclear ITS sequences. 2. Infrageneric relationships and reclassification to achieve monolyphyly of Orchis sensu stricto. Lindleyana 12, 113–141. Bateman, R.M., Hollingsworth, P.M., Preston, J., Yi-Bo, L., Pridgeon, A.M., Chase, M.W., 2003. Molecular phylogenetics and evolution of Orchidinae and selected Habenariinae (Orchidaceae). Bot. J. Linn. Soc. 142, 1–40. Batista, J.A.N., Bianchetti, L.B., 2003. Lista atualizada das orchidaceae do distrito federal. Acta Bot. Bras. 17, 183–201. Batista, J.A.N., Bianchetti, L.B., Nogueira, R.E., Pellizzaro, K.F., Ferreira, F.E., 2004. The genus Habenaria (Orchidaceae) in the Itacolomi state park, Minas Gerais, Brazil. Sitientibus ser. Cienc. Biol. 4, 25–36. Batista, J.A.N., Bianchetti, L.B., Miranda, Z.J.G., 2006. A revision of Habenaria section Macroceratitae (Orchidaceae) in Brazil. Brittonia 58, 10–41. Batista, J.A.N., Silva, J.B.F., Bianchetti, L.B., 2008a. The genus Habenaria (Orchidaceae) in the Brazilian Amazon. Rev. Bras. Bot. 31, 105–134. Batista, J.A.N., Mota, R.C., Abreu, N.L., Menini Neto, L., 2008b. Habenaria pseudoglaucophylla (Orchidaceae), a new species from Minas Gerais, Brazil. Novon 18, 409–414. Batista, J.A.N., Bianchetti, L.B., 2010. Taxonomy, distribution and new taxa from the Habenaria crucifera (section Nudae, Orchidaceae) aggregate from Brazil and the Guianas. Brittonia 62, 57–79. Batista, J.A.N., Bianchetti, L.B., González-Tamayo, R., Figueroa, X.M.C., Cribb, P.J., 2011a. A synopsis of new world Habenaria (Orchidaceae) I. Harvard Pap. Bot. 16 (1), 1–47. Batista, J.A.N., Bianchetti, L.B., González-Tamayo, R., Figueroa, X.M.C., Cribb, P.J., 2011b. A synopsis of new world Habenaria (Orchidaceae) II. Harvard Pap. Bot. 16 (2), 233–273. Bellstedt, D.U., Linder, H.P., Harley, E.H., 2001. Phylogenetic relationships in Disa based on non-coding trnL–trnF chloroplast sequences: evidence of numerous repeat regions. Am. J. Bot. 88, 2088–2100. Bonfield, J.K., Smith, K.F., Staden, R., 1995. A new DNA sequence assembly program. Nucleic Acids Res. 24, 4992–4999. Bytebier, B., Bellstedt, D.U., Linder, H.P., 2007. A molecular phylogeny for the large African orchid genus Disa. Mol. Phylogenet. Evol. 43, 75–90. Cameron, K.M., 2007. Molecular phylogenetics of Orchidaceae: the first decade of DNA sequencing. In: Cameron, K.M., Arditti, J., Kull, T. (Eds.), Orchid Biology: Reviews and Perspectives, IX. New York Botanical Garden Press, New York, pp. 163–200. Cameron, K.M., Chase, M.W., Whitten, W.M., Kores, P.J., Jarrel, D.C., Albert, V.A., Yukawa, T., Hills, H.G., Goldman, D.H., 1999. A phylogenetic analysis of the Orchidaceae: evidence from rbcL nucleotide sequences. Am. J. Bot. 86, 208–224. Chase, M.W., Cowan, R.S., Hollingsworth, P.M., van den Berg, C., Madriñán, S., Petersen, G., Seberg, O., JØrgsensen, T., Cameron, K.M., Carine, M., Pedersen, N., Hedderson, T.A.J., Conrad, F., Salazar, G.A., Richardson, J.E., Hollingsworth, M.L., Barraclough, T.G., Kelly, L., Wilkinson, M., 2007. A proposal for a standardised protocol to barcode all land plants. Taxon 56, 295–299. Chemisquy, M.A., Morrone, O., 2012. Molecular phylogeny of Gavilea (Chloraeinae: Orchidaceae) using plastid and nuclear markers. Mol. Phylogenet. Evol. 62, 889– 897. Cogniaux, A., 1893. Orchidaceae. Habenaria. In: Martius, C.F.P., Eichler, A.G., Urban, I. (Eds.), Flora Brasiliensis 3(4), F. Fleischer, Munich, pp. 18–102. Daviña, J.R., Grabiele, M., Cerutti, J.C., Hojsgaard, D.H., Almada, Rubén.D., Insaurralde, I.S., Honfi, A.I., 2009. Chromosome studies in Orchidaceae from Argentina. Genet. Mol. Biol. 32, 811–821. Douzery, E.J.P., Pridgeon, A.M., Kores, P., Linder, H.P., Kurzweil, H., Chase, M.W., 1999. Molecular phylogenetics of diseae (Orchidaceae): a contribution from nuclear ribosomal ITS sequences. Am. J. Bot. 86, 887–899. Doyle, J.J., Doyle, J.S., 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19, 11–15. Dressler, R.L., 1993. Phylogeny and Classification of the Orchid Family. Dioscorides Press, Portland. Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797. Erixon, P., Svennblad, B., Britton, T., Oxelman, B., 2003. Reliability of Bayesian posterior probabilities and bootstrap frequencies in phylogenetics. Syst. Biol. 52, 665–673. Felix, L.P., Guerra, M., 1998. Cytogenetic studies on species of Habenaria (Orchidoideae: Orchidaceae) occurring in the Northeast of Brazil. Lindleyana 13, 224–230. Fitch, W.M., 1971. Toward defining the course of evolution: minimum change for a specific tree topology. Syst. Zool. 20, 406–416. Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39, 783–791. Fischer, G.A., Gravendeel, B., Sieder, A., Adriantiana, J., Heiselmayer, P., Cribb, P.J., Camargo Smidt, E., Samuel, R., Kiehn, M., 2007. Evolution of resupination in Malagasy species of Bulbophyllum (Orchidaceae). Mol. Phyl. Evol. 45, 358– 376. Freudenstein, J.V., van den Berg, C., Goldman, D.H., Kores, P.J., Molvray, M., Chase, M.W., 2004. An expanded plastid DNA phylogeny of Orchidaceae and analysis of jackknife branch support strategy. Am. J. Bot. 91, 149–157. González-Tamayo, R., 1993. Algunas consideraciones sobre el género Habenaria (Orchidaceae) en México. Bol. Inst. Bot. (Guadalajara) 1 (7), 485–511. Govaerts, R., Pfahl, J., Campacci, M.A., Holland Baptista, D., Tigges, H., Shaw, J., Cribb, P., George, A., Kreuz, K., Wood, J., 2011. World Checklist of Orchidaceae. The Board of Trustees of the Royal Botanic Gardens, Kew. <http://apps.kew.org/ wcsp/> (accessed 09.09.11). Hoehne, F.C., 1940. Orchidaceas, Habenaria. In: Hoehne, F.C. (Ed.), Flora Brasilica 12(1). Secretaria da Agricultura, Indústria e Comércio de São Paulo, São Paulo, pp. 52–254. Hunt, P.E., 1968. African orchids: XXXII. Kew Bull. 22, 489–492. Kores, P.J., Molvray, M., Weston, P.H., Hopper, S.D., Brown, A.P., Cameron, K.M., Chase, M.W., 2001. A phylogenetic analysis of Diurideae (Orchidaceae) based on plastid DNA sequence data. Am. J. Bot. 88, 1903–1914. Kränzlin, F., 1892. Beitrage zu einer monographie der gattung Habenaria Willd. Engl. Bot. Jahrb. 16, 52–223. Kränzlin, F., 1901. Orchidacearum genera et species. vol. I. Habenaria. Mayer and Müller, Berlin. Kress, W.J., Prince, L.M., Williams, K.J., 2002. The phylogeny and a new classification of the gingers (Zingiberaceae): evidence from molecular data. Am. J. Bot. 89, 1682–1696. Kurzweil, H., Weber, A., 1992. Floral morphology of southern African Orchideae. II. Habenariinae. Nord. J. Bot. 12, 39–61. Lahaye, R., van der Bank, M., Bogarin, D., Warner, J., Pupulin, F., Gigot, G., Maurin, O., Duthoit, S., Barraclough, T.G., Savolainen, V., 2008. DNA barcoding the floras of biodiversity hotspots. Proc. Natl. Acad. Sci. USA 105, 2923–2928. Mendonça, R.C., Felfili, J.M., Walter, B.M.T., Silva Junior, M.C., Rezende, A.V., Filgueiras, T.S., Nogueira, P.E., 1998. Flora vascular do cerrado. In: Sano, S.M., J.A.N. Batista et al. / Molecular Phylogenetics and Evolution 67 (2013) 95–109 Almeida, S.P. (Eds.), Cerrado: ambiente e flora. EMBRAPA–CPAC, Planaltina, pp. 289–556. Nylander, J.A.A., 2004. MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University. Pabst, G.F.J., Dungs, F., 1975. Orchidaceae Brasilienses, vol. 1. Brucke-Verlag Kurt Schmersow, Hildesheim. Ponsie, M.E., Mitchell, A., Edwards, T.J., Johnson, S.D., 2007a. Phylogeny of Bonatea (Orchidaceae: Habenariinae) based on molecular and morphological data. Plant Syst. Evol. 263, 253–268. Ponsie, M.E., Edwards, T.J., Johnson, S.D., 2007b. A taxonomic revision of Bonatea Willd. (Orchidaceae: Orchidoideae: Habenariinae). S. Afr. J. Bot. 73, 1–21. Pridgeon, A.M., Bateman, R.M., Cox, A.V., Hapeman, J.R., Chase, M.W., 1997. Phylogenetics of subtribe Orchidinae (Orchidoideae, Orchidaceae) based on nuclear ITS sequences. 1. Intergeneric relationships and polyphyly of Orchis sensu lato. Lindleyana 12, 89–109. Pridgeon, A.M., Cribb, P.J., Chase, M.W., Rasmussen, F.N., 2001a. Genera Orchidacearum vol. 2, Orchidoideae, Part 1. Oxford University Press Inc., New York. Pridgeon, A.M., Solano, R., Chase, M.W., 2001b. Phylogenetic relationships in Pleurothallidinae (Orchidaceae): combined evidence from nuclear and plastid DNA sequences. Am. J. Bot. 88, 2286–2308. Ratter, J.A., Ribeiro, J.F., Bridgewater, S., 1997. The Brazilian cerrado vegetation and threats to its biodiversity. Ann. Bot. 80, 223–230. Ronquist, F., Huelsenbeck, J.P., van den Mark, P., 2005. MrBayes 3.1 Manual. <http:// mrbayes.net>. Russell, A., Samuel, R., Rupp, B., Barfuss, M.H.J., Safran, M., Besendorfer, V., Chase, M.W., 2009. Phylogenetics and cytology of a pantropical orchid genus Polystachya (Polystachyinae, Vandeae, Orchidaceae): evidence from plastid DNA sequence data. Taxon 59, 389–404. Salazar, G.A., Chase, M.W., Soto, M.A., Ingrouille, M., 2003. Phylogenetics of cranichideae with emphasis on spiranthinae (Orchidaceae, Orchidoideae): evidence from plastid and nuclear DNA sequences. Am. J. Bot. 90, 777–795. Schlechter, R., 1915. Orchidaceae Stolzianae; ein Beitrag zur Orchideenkunde des Nyassa Landes. Engl. Bot. Jahrb. 53, 477–605. Singer, R.B., Cocucci, A.A., 1997. Eye attached hemipollinaria in the hawkmoth and settling moth pollination of Habenaria (Orchidaceae): a study on functional morphology in five species from subtropical South America. Bot. Acta 110, 328– 337. Small, J.K., 1903. Flora of the Southeastern United States. New York. Summerhayes, V.S., 1942. African orchids, XII. Bot. Mus. Leafl. Harvard Univ. 10, 257–299. Summerhayes, V.S., 1960. African orchids, XXVII. Kew Bull. 14, 126–157. Summerhayes, V.S., 1962. African orchids, XXVIII. Kew Bull. 16, 253–314. Summerhayes, V.S., 1966. African orchids, XXX. Kew Bull. 20, 165–199. 109 Sun, Y., Skinner, D.Z., Liang, G.H., Hulbert, S.H., 1994. Phylogenetic analysis of Sorghum and related taxa using internal transcribed spacers of nuclear ribosomal DNA. Theor. Appl. Genet. 89, 26–32. Swofford, D.L., 1998. PAUP 4.0. Phylogenetic Analysis using Parsimony. Sinauer Publications, Sunderland, Massachusetts. Szlachetko, D.L., 2003a. Matériaux pour la révision de Habenaria (Orchidaceae, Orchidoideae) – 2. Richardiana 3, 153–157. Szlachetko, D.L., 2003b. Matériaux pour la révision de Habenaria (Orchidaceae, Orchidoideae) – 3. Richardiana 3, 158–162. Szlachetko, D.L., 2004a. Matériaux pour la révision des Habenariinae (Orchidaceae, Orchidoideae) – 4. Richardiana 4, 52–65. Szlachetko, D.L., 2004b. Matériaux pour la révision des Habenariinae (Orchidaceae, Orchidoideae) – 5. Richardiana 4, 103–108. Szlachetko, D., Kras, M., 2006. Notes sur le genre Habenella. Richardiana 6, 33–39. Tamura, K., Nei, M., Kumar, S., 2004. Prospects for inferring very large phylogenies by using the neighbor-joining method. Proc. Natl. Acad. Sci. USA 101, 11030–11035. Tamura, K., Dudley, J., Nei, M., Kumar, S., 2007. MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol. Biol. Evol. 24, 1596–1599. Toscano de Brito, A.L.V., 1995. Orchidaceae. In: Stannard, B.L. (Ed.), Flora of the Pico das Almas: Chapada Diamantina – Bahia, Brazil. Royal Botanic Gardens, Kew, pp. 725–767. van den Berg, C., Higgins, W.E., Dressler, R.L., Whiten, W.M., Soto Arenas, M.A., Culham, A., Chase, M.W., 2000. A phylogenetic analysis of Laeliinae (Orchidaceae) based on sequence data from internal transcribed spacers (ITS) of nuclear ribosomal DNA. Lindleyana 15, 96–114. van den Berg, C., Goldman, D.H., Freudenstein, J.V., Pridgeon, A.M., Cameron, K.M., Chase, M.W., 2005. An overview of the phylogenetic relationships within Epidendroideae inferred from multiple DNA regions and recircumscription of Epidendreae and Arethuseae (Orchidaceae). Am. J. Bot. 92, 613–624. Whitten, W.M., Williams, N.H., Chase, M.W., 2000. Subtribal and generic relationships of Maxillarieae (Orchidaceae) with emphasis on Stanhopeinae: combined molecular evidence. Am. J. Bot. 87, 1842–1856. Whitten, W.M., Williams, N.H., Blanco, M.A., Endara, L., Neubig, K., Koehler, S., Carnevali, G., Singer, R., 2007. Molecular phylogenetics of Maxillaria and related genera (Orchidaceae: Cymbidieae) based upon combined molecular data sets. Am. J. Bot. 94, 1860–1889. Williams, N.H., Chase, M.W., Fulcher, T., Whitten, W.M., 2001. Molecular systematic of the Oncidiinae based on evidence from four DNA regions: expanded circumscriptions of Cyrtochilum, Erycina, Otoglossum and Trichocentrum and a new genus (Orchidaceae). Lindleyana 16, 113–139. Zappi, D.C., Lucas, E., Stannard, B.L., Lughadha, E.N., Pirani, J.R., Queiroz, L.P., Atkins, S., Hind, D.J.N., Giulietti, A.M., Harley, R.M., Carvalho, A.M., 2003. Lista das plantas vasculares de Catolés, Chapada Diamantina, Bahia, Brasil. Bol. Bot. Univ. São Paulo 21, 345–398.