Phylogeny and evolution of Dyckia (Bromeliaceae) inferred from chloroplast and nuclear sequences

Elton Leme; kurt weising

In this contribution the rediscovery of the bromeliad Dyckia excelsa in Mato Grosso do Sul state, central-western Brazil, is reported. This species was first described based on a single individual from a particular collection in 1993 with no precise locality. Illustrations and a brief discussion on the leaf anatomy of the species are provided. One additional species, Dyckia gracilis, which has been overlooked by previous authors, is included in the list of Bromeliaceae of Mato Grosso do Sul State. An updated key for Dyckia species from the region of D. excelsa is presented. The status of D. excelsa in Mato Grosso do Sul State, Brazil, is discussed, and recommendations for its conservation are offered.

Plant Syst Evol (2014) 300:1591–1614 DOI 10.1007/s00606-014-0985-0 ORIGINAL ARTICLE Phylogeny and evolution of Dyckia (Bromeliaceae) inferred from chloroplast and nuclear sequences Florian Krapp • Diego Sotero de Barros Pinangé Ana Maria Benko-Iseppon • Elton M. C. Leme • Kurt Weising • Received: 11 November 2013 / Accepted: 8 January 2014 / Published online: 29 January 2014 Ó Springer-Verlag Wien 2014 Abstract The genus Dyckia (Bromeliaceae) comprises more than 150 terrestrial or epilithic species with a strongly xeromorphic habit. Most of its members belong to the azonal rock vegetation of Neotropical savannas and forests of Brazil and adjacent countries. Dyckia is relatively species-rich compared with its closest relatives Encholirium (27 species) and Deuterocohnia (17 species). Here, we present the first molecular phylogenetic analysis of Dyckia using DNA sequence data from six plastid loci (matK gene, rps16 intron, petD intron, rpl32-trnL, rps16-trnK and trnDtrnT) and a portion of the nuclear gene phyC. A total of 124 accessions were included, corresponding to 79 taxa from six genera. Phylogenetic trees were generated using parsimony, likelihood and Bayesian methods. DNA sequence variation among Dyckia species turned out to be extremely low, and phylogenies were poorly resolved. The monophyly of Dyckia is supported, whereas evidence is provided that Encholirium is paraphyletic. Based on a dated plastid DNA tree, Dyckia experienced a recent radiation starting around 2.9 million years ago. Four major clades could be identified that roughly correspond to the geographic origin F. Krapp K. Weising (&) Plant Molecular Systematics, Department of Sciences, Institute of Biology, University of Kassel, Heinrich-Plett-Str. 40, 34132 Kassel, Germany e-mail: weising@uni-kassel.de D. S. de Barros Pinangé A. M. Benko-Iseppon Genetics Department, CCB, Universidade Federal de Pernambuco (UFPE), Av. Prof. Moraes Rego, 1235, Recife, PE 50670-420, Brazil E. M. C. Leme Herbarium Bradeanum, C.P. 15005, Rio de Janeiro, RJ 20031-970, Brazil of the samples. A parsimony network based on plastid DNA haplotypes shows a star-like pattern, indicating recent range expansions. Our data are compatible with a scenario where Dyckia and Encholirium diverged in northeastern Brazil, whereas one lineage of Dyckia dispersed to southern Brazil from where a rapid colonization of suitable habitats was initiated. We discuss our results in relation to species delimitation in Dyckia. Keywords Dyckia Encholirium Bromeliaceae Biogeography Molecular phylogeny Dated plastid DNA tree Introduction According to the current classification, the family Bromeliaceae comprises 58 genera with more than 3,350 species that are almost exclusively confined to tropical and subtropical zones of the New World (Luther 2012). Bromeliads have radiated into extreme environments that range from mesic to xeric, from terrestrial to epiphytic and from sea level to high altitudes. The family is therefore often viewed as a prime example for adaptive radiation (Benzing 2000). Based on plastid sequence data, Givnish et al. (2007, 2011) divided Bromeliaceae into eight monophyletic subfamilies: Tillandsioideae, Bromelioideae, Pitcairnioideae s.str., Puyoideae, Brocchinioideae, Navioideae, Hechtioideae and Lindmanioideae. In their plastid phylogeny based on eight loci, subfamily Pitcairnioideae s.str. is sister to a clade consisting of Bromelioideae plus Puyoideae. Pitcairnioideae s.str. comprises the three xerophytic genera Dyckia (159 species), Encholirium (27 species) and Deuterocohnia (17 species) as well as the more mesophytic genera Fosterella 123 1592 F. Krapp et al. A B Dyckia Encholirium D Deuterocohnia F C E G H Fig. 1 Distribution, habit and features of Dyckia. a–d Flowers of different species of Dyckia. a D. goehringii, b D. remotiflora, c D. leptostachya and d D. brevifolia. e Habit of D. remotiflora. f Distributional ranges of the genera Deuterocohnia (orange, Schütz 2012), Dyckia (green) and Encholirium (violet, Forzza 2005). g Scales and armed leaf margin of D. marnier-lapostollei, h Laterally inserted inflorescence of D. goehringii (31 species) and Pitcairnia (*400 species). Recent taxonomic revisions are available for Encholirium (Forzza 2005), Fosterella (Peters 2009) and Deuterocohnia (Schütz 2012), but are still missing for the large and diverse genera Dyckia and Pitcairnia. Dyckia species are terrestrial or saxicolous, rosetteleaved, tank-less perennial plants that are characterized by a strongly xeromorphic habit (Smith and Downs 1974; Fig. 1). Leaves are usually coriaceous with fierce spiny margins, show different levels of succulence and are often densely lepidote. The simple or branched inflorescences are inserted laterally, a character that discriminates Dyckia against its close relative Encholirium (Forzza 2005). Flowers are nearly sessile to pedicellate. Petals, sepals and often also the inflorescence axis typically show brilliant red, orange or yellow colours. The filaments are basally connate and form a tube, which comprises another important feature for the delimitation of Dyckia against Encholirium (Smith and Downs 1974). Hummingbirds, butterflies, and other insects including the neobiotic Apis mellifera are commonly observed pollinators of Dyckia flowers (e.g. Bernadello et al. 1991). Some species also bear extrafloral nectaries that attract unspecific ant visitors (Vesprini et al. 2003). Fruits are capsules that release winged, anemochorous seeds upon dehiscence. Wind dispersal of Dyckia seeds is probably less efficient due to their larger size, when compared to those of Deuterocohnia, Fosterella and Pitcairnia. All Dyckia, Encholirium and Deuterocohnia species investigated so far follow the Crassulacean acid metabolism (CAM) photosynthetic pathway and share a number of leaf anatomical synapomorphies (Givnish et al. 2007; Santos-Silva et al. 2013), whereas the mesophytic Fosterella and Pitcairnia species perform standard C3 photosynthesis (Martin 1994; Crayn et al. 2000, 2004). Dyckia has its diversity centre in mountainous regions of the central Brazilian Cerrado biome, from where it 123 Phylogeny and evolution of Dyckia (Bromeliaceae) ranges into the adjacent Mata Atlântica and Caatinga biomes in the east, to the Chacos of Bolivia and Paraguay in the west, and also to Uruguay and the northern Argentinean Pampas in the south (Smith and Downs 1974). The highest levels of species richness and a large proportion of narrow endemics are encountered in the so-called Campos Rupestres (‘‘rocky fields’’) of the Espinhaço Range in the Brazilian state of Minas Gerais, and of the Chapada Diamantina in the state of Bahia (Versieux and Wendt 2006, 2007; Versieux et al. 2008; Fig. 1). Within their distributional range, Dyckia species are an important component of the azonal vegetation. Typical Dyckia habitats are dry, rocky outcrops, cliffs, slopes and inselbergs that are generally characterized by poor soils, lack of nutrients, little water supply and high sunlight exposure (Smith and Downs 1974; Barthlott et al. 1993). Some Dyckia species can be found on rocky or sandy ground along the coast (e.g. Dyckia maritima; Winkler 1980), whereas others grow as rheophytes in riverbeds with alternating seasons of submergence and desiccation (e.g. Dyckia ibiramensis; Hmeljevski et al. 2011). In his most recent list of bromeliad binomials, Luther (2012) accepted 147 Dyckia species, but this number is steadily growing (e.g. Leme et al. 2012) and has now risen to 159. Many species are rare and narrow endemics, and it is therefore no surprise that very little is known yet about infrageneric relationships. Baker (1889) divided Dyckia into five subgenera, i.e. Dyckia, Prionophyllum, Navia, Cephalonavia and Encholirion. However, this division has later been abandoned, and the latter three taxa turned out to belong to (or represent) other genera after renaming. Since these early days, there have been no more serious attempts to subdivide the genus. More recently, a few Dyckia species have been included in molecular systematic investigations of Bromeliaceae based on plastid DNA sequences (Horres et al. 2000; Crayn et al. 2004; Rex et al. 2009; Givnish et al. 2011). In these studies, Dyckia, Encholirium and Deuterocohnia together commonly formed a wellsupported xerophytic clade, being sister to monophyletic Fosterella (Givnish et al. 2007, 2011; Rex et al. 2009). The relationships among the three xeromorphic genera as well as their monophyly remained, however, ambiguous, especially since Deuterocohnia came out as paraphyletic in chloroplast trees (Rex et al. 2009; Schütz 2012). A long-standing problem of Dyckia taxonomy is the poor delimitation of key species, many of which are known from incomplete type specimens or few supplementary collections only (Smith and Downs 1974; Versieux and Wendt 2006). The paucity of consistent taxonomic studies based on living, well-documented field-collected specimens makes species identification and discrimination notoriously difficult. As early as 1934, Smith already stated that ‘‘Dyckia is a true nightmare to the systematist because 1593 of the extremely fluid way in which species pass into one another’’. More than three decades later, Smith (1967) considered only a handful of Dyckia species as being clearly distinguishable, whereas he suspected the remainder of the genus to be the result of a recent and rapid evolution. Still in 2007, Versieux and Wendt stressed that many Dyckia species are poorly known, with overlapping diagnostic features and sometimes represented in herbaria by only a single leaf and part of the inflorescence. Clearly, the genus needs more attention. In the present study, we undertook a first phylogenetic analysis of Dyckia and related genera, based on six plastid DNA regions and a portion of the nuclear phytochrome C gene. The aims of our study were (1) to determine the position of Dyckia within the subfamily Pitcairnioideae s.str. and to assess its monophyly; (2) to assess infrageneric relationships; (3) to compare the topology of phylogenetic trees based on the nuclear phyC gene on the one hand and plastid loci on the other hand; and (4) to reconstruct the historical biogeography and evolutionary history of Dyckia. Materials and methods Taxon sampling A total of 124 accessions of Dyckia and closely related genera were included in this study, corresponding to 79 taxa from six genera of Bromeliaceae (Table 1). Dyckia was represented by 97 accessions, covering 58 of the 159 accepted species. Five species each were included of Deuterocohnia, Fosterella and Pitcairnia. Encholirium, which is considered the closest relative of Dyckia, was represented by four species in ten accessions. Two species of Puya (Puyoideae) were chosen as outgroups. Plant materials were derived from the well-documented living collections of the Botanical Gardens of Heidelberg, Vienna, Berlin and Bonn, the private collection of one of the authors (Elton Leme, Teresópolis, Brazil) and field collections. Voucher specimens and duplicates have been deposited in various herbaria and living collections (for details see Table 1). DNA extraction and amplification Total DNA was extracted from 300 to 500 mg of fresh or 50 mg of silica-dried leaf material using a cetyltrimethylammonium bromide (CTAB) procedure modified for succulent plants (Tel-Zur et al. 1999; Krapp 2013). Final DNA concentrations were estimated by electrophoresis on ethidium bromide-stained agarose gels along with known amounts of phage k-DNA. 123 1594 123 Table 1 Studied plant material, voucher specimens, geographical provenances, DNA isolate and GenBank accession numbers Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC N. Schütz 06/060 (FR)/KAS NiSch 06/060 Bolivia, Tarija, Aniceto Arce FK0074 KF784162 KF784286 KF784658 KF784534 KF783915 KF784410 KF784038 N. Schütz 06/028 (FR)/KAS NiSch 06/028 Bolivia, Santa Cruz, Samaipata FK0071 KF784159 KF784283 KF784655 KF784531 KF783912 KF784407 KF784035 Deuterocohnia glandulosa E. Gross N. Schütz 06/019 (FR)/KAS NiSch 06/019 Bolivia, Santa Cruz, Ipati FK0072 KF784160 KF784284 KF784656 KF784532 KF783913 KF784408 KF784036 Deuterocohnia cf. haumanii Castellanos N. Schütz 06/122 (FR)/KAS NiSch 06/122 Argentina, Salta, Cafayate FK0075 KF784163 KF784287 KF784659 KF784535 KF783916 KF784411 KF784039 Species Abbreviation used for network Deuterocohnia brevifolia (Grisebach) M. A. Spencer & L. B. Smith Deuterocohnia brevispicata Rauh & L. Hromadnik Dbr Dme N. Schütz 06/009 (FR)/KAS NiSch 06/009 Bolivia, Chuquisaca, Monteagudo FK0073 KF784161 KF784285 KF784657 KF784533 KF783914 KF784409 KF784037 Dyckia aurea L. B. Smith aur E. Leme 6445 (Leme)/ Leme 6445 Brazil, Goiás, Cristalina FK0194 KF784224 KF784348 KF784720 KF784596 KF783976 KF784472 KF784100 Dyckia beateae E. Gross & Rauh bea P. Braun 560 (BONN)/ BONN 4338 Brazil, Mato Grosso, Araquainha FK0044 KF784145 KF784269 KF784641 KF784517 KF783898 KF784393 KF784021 Dyckia beateae E. Gross & Rauh bea E. Estenes s. n. (Leme)/Leme 1961 Brazil, Mato Grosso do Sul, Coxim FK0091 KF784179 KF784303 KF784675 KF784551 KF783932 KF784427 KF784055 Dyckia brachyphylla L. B. Smith brc P. Braun 836 (BONN)/ BONN 3644 Brazil, Minas Gerais, Itacambira FK0045 KF784146 KF784270 KF784642 KF784518 KF783899 KF784394 KF784022 Dyckia braunii Rauh bra P. Braun 690 (BONN)/ BONN 4339 Brazil, Goiás FK0042 KF784143 KF784267 KF784639 KF784515 KF783896 KF784391 KF784019 Dyckia aff. brevifolia Baker bre P. Braun 840 (HEID)/ BGHD 130223 Brazil, Minas Gerais, Itacambira FK0010 KF784117 KF784241 KF784613 KF784489 KF783870 KF784365 KF783993 Dyckia brevifolia Baker bre Unknown (KAS)/KAS FK0067 Brazil FK0067 KF784156 KF784280 KF784652 KF784528 KF783909 KF784404 KF784032 Dyckia choristaminea Mez cho Unknown (HEID)/ BGHD 130018 Brazil, Rio Grande do Sul FK0021 KF784128 KF784252 KF784624 KF784500 KF783881 KF784376 KF784004 Dyckia choristaminea Mez cho Unknown (BONN)/ BONN 2410 Brazil, Rio Grande do Sul FK0040 KF784142 KF784266 KF784638 KF784514 KF783895 KF784390 KF784018 Dyckia cinerea Mez cin Unknown (BONN)/ BONN 4348 Brazil, Minas Gerais FK0047 KF784147 KF784271 KF784643 KF784519 KF783900 KF784395 KF784023 F. Krapp et al. Deuterocohnia meziana Kuntze ex Mez Species Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Dyckia dawsonii L. B. Smith Dyckia delicata Larocca & Sobral daw Unknown (BONN)/ BONN 16540 E. Leme 6492 (Leme)/ Leme 6492 Brazil, Goiás FK0043 KF784144 KF784268 KF784640 KF784516 KF783897 KF784392 KF784020 Brazil, Rio Grande do Sul FK0184 KF784217 KF784341 KF784713 KF784589 KF783969 KF784465 KF784093 del den E. Leme 4249 (Leme)/ Leme 4249 Brazil, Minas Gerais, Serra da Piedade FK0213 KF784231 KF784355 KF784727 KF784603 KF783983 KF784479 KF784107 Dyckia dissitiflora Schult. f. dis A.M. Iseppon, Pinangé, D. & Cruz, G. 1605 (UFP) Brazil, Bahia, Morro do Chapéu FK0141 KF784215 KF784339 KF784711 KF784587 KF783967 KF784463 KF784091 Dyckia distachya Hassl. dit Ahlgrimm s. n. (B)/B 236-17-85-23 Paraguay, Paraguari, Chololo FK0126 KF784214 KF784338 KF784710 KF784586 KF783966 KF784462 KF784090 Dyckia elisabethae S. Winkl. eli E. Leme 4461 (Leme)/ Leme 4461 Brazil, Rio Grande do Sul, Barra do Ribeiro FK0219 KF784233 KF784357 KF784729 KF784605 KF783985 KF784481 KF784109 Dyckia encholirioides (Gaudichaud) Mez enc E. Leme s. n. (Leme)/ Leme s. n. Brazil FK0095 KF784183 KF784307 KF784679 KF784555 KF783936 KF784431 KF784059 Dyckia espiritosantensis Leme et al. esp E. Leme 6930 (Leme)/ Leme 6930 Brazil, Espı́rito Santo, São Roque do Canaã FK0207 KF784229 KF784353 KF784725 KF784601 KF783981 KF784477 KF784105 Dyckia estevesii Rauh est Brazil FK0001 KF784111 KF784235 KF784607 KF784483 KF783864 KF784359 KF783987 Dyckia estevesii var. braunii nom. nud. Rauh bra P. Braun s. n. (HEID)/ BGHD 105188 P. Braun s. n. (HEID)/ BGHD 130025 Brazil, Goiás FK0004 KF784112 KF784236 KF784608 KF784484 KF783865 KF784360 KF783988 Dyckia estevesii Rauh est E. Esteves Pereira s. n. (HEID)/BGHD 105012 Brazil, Goiás, Caiaponia FK0005 KF784113 KF784237 KF784609 KF784485 KF783866 KF784361 KF783989 Dyckia estevesii Rauh est P. Braun s. n. (BONN)/BONN 1477 Brazil, Goiás FK0033 KF784137 KF784261 KF784633 KF784509 KF783890 KF784385 KF784013 Dyckia ferox Mez fer W. Rauh 64237 (HEID)/BGHD 130031 Argentina, Cordoba, Cerro Colorado FK0020 KF784127 KF784251 KF784623 KF784499 KF783880 KF784375 KF784003 Dyckia ferox Mez fer L. Horst 375 (HEID)/ BGHD 130028 Brazil, Bahia, Morro do Chapéu FK0027 KF784132 KF784256 KF784628 KF784504 KF783885 KF784380 KF784008 1595 123 Dyckia densiflora Schult. f. Phylogeny and evolution of Dyckia (Bromeliaceae) Table 1 continued 1596 123 Table 1 continued Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Dyckia ferox Mez fer W. Till 6016 (WU)/ WU AB 60 Paraguay, Paraguari, Paraguari FK0056 KF784155 KF784279 KF784651 KF784527 KF783908 KF784403 KF784031 Dyckia ferruginea Mez feu E. Estenes s. n. (Leme)/Leme 1958 Brazil, Mato Grosso do Sul FK0090 KF784178 KF784302 KF784674 KF784550 KF783931 KF784426 KF784054 Dyckia floribunda Griseb. flo W. Till 5069 (WU)/ WU 115/90 Argentina, La Rioja, Patquia FK0052 KF784151 KF784275 KF784647 KF784523 KF783904 KF784399 KF784027 Dyckia floribunda Griseb. flo W. Till 5012 (WU)/ WU 52/90 Argentina, Cordoba, Dept. Jesus Maria, Ascochinga FK0107 KF784195 KF784319 KF784691 KF784567 KF783948 KF784443 KF784071 Dyckia floribunda Griseb. flo W. Till 5144 (WU)/ WU AB 17/90 Argentina, San Juan, Chucuma FK0108 KF784196 KF784320 KF784692 KF784568 KF783949 KF784444 KF784072 Dyckia fosteriana L. B. Smith fos E. Leme 6461 (Leme)/ Leme 6461 Brazil, Pará FK0094 KF784182 KF784306 KF784678 KF784554 KF783935 KF784430 KF784058 Dyckia goehringii E. Gross & Rauh goe W. Rauh 67622 (HEID)/BGHD 105013 Brazil, Minas Gerais, Diamantina FK0031 KF784135 KF784259 KF784631 KF784507 KF783888 KF784383 KF784011 Dyckia grandidentata Braun, P.J. & E. Esteves Pereira grn E. Leme 6840 (Leme)/ Leme 6840 Brazil, Mato Grosso do Sul, Sete Quedas FK0183 KF784216 KF784340 KF784712 KF784588 KF783968 KF784464 KF784092 Dyckia granmogulensis Rauh gra W. Rauh 56484 (HEID)/BGHD 130019 Brazil, Minas Gerais, Grão Mogol FK0007 KF784115 KF784239 KF784611 KF784487 KF783868 KF784363 KF783991 Dyckia hebdingii L. B. Smith heb Unknown (HEID)/ BGHD 103913 Brazil FK0013 KF784120 KF784244 KF784616 KF784492 KF783873 KF784368 KF783996 Dyckia hebdingii L. B. Smith heb Unknown (BONN)/ BONN 3259 Brazil, Rio Grande do Sul FK0038 KF784140 KF784264 KF784636 KF784512 KF783893 KF784388 KF784016 Dyckia hebdingii L. B. Smith heb Unknown (WU)/WU 130030 Brazil FK0112 KF784200 KF784324 KF784696 KF784572 KF783953 KF784448 KF784076 Dyckia aff. ibiramensis Reitz ibi L. Horst 1287 (HEID)/ BGHD 130023 Brazil, Minas Gerais, Diamantina FK0025 KF784131 KF784255 KF784627 KF784503 KF783884 KF784379 KF784007 Dyckia jonesiana Strehl jon E. Leme 2959 (Leme)/ Leme 2959 Brazil, Rio Grande do Sul, Caçapava do Sul FK0217 KF784232 KF784356 KF784728 KF784604 KF783984 KF784480 KF784108 F. Krapp et al. Species Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Dyckia aff. leptostachya Baker lep W. Rauh s. n. (HEID)/ BGHD 130017 Argentina, Cordoba, Cerro Colorado FK0016 KF784123 KF784247 KF784619 KF784495 KF783876 KF784371 KF783999 Dyckia aff. leptostachya Baker lep J. Piltz s. n. (BONN)/ BONN 4355 Argentina, Cordoba, Dean Funes FK0035 KF784138 KF784262 KF784634 KF784510 KF783891 KF784386 KF784014 Dyckia leptostachya Baker lep H. Amerhauser s. n. (WU)/WU B 395/96 Paraguay, Cordillera, Caacupé FK0053 KF784152 KF784276 KF784648 KF784524 KF783905 KF784400 KF784028 Dyckia cf. leptostachya Baker lep H. Amerhauser 6 (WU)/WU B 266/95 Bolivia, Santa Cruz, Mobote FK0115 KF784203 KF784327 KF784699 KF784575 KF783956 KF784451 KF784079 Dyckia lindevaldae Rauh lin P. Braun BR 691 (HEID)/BGHD 108614 Brazil, Goiás, Alto Paraiso FK0019 KF784126 KF784250 KF784622 KF784498 KF783879 KF784374 KF784002 Dyckia lunaris Leme lun E. Leme 4951 (Leme)/ Leme 4951 Brazil, Goiás, Alto Paraı́so FK0193 KF784223 KF784347 KF784719 KF784595 KF783975 KF784471 KF784099 Dyckia macedoi L. B. Smith mac R.B. Louzada, Sotero, D. & Medeiros, M. 151 (RB) Brazil, Minas Gerais, Santana do Riacho FK0099 KF784187 KF784311 KF784683 KF784559 KF783940 KF784435 KF784063 Dyckia macedoi L. B. Smith mac R.B. Louzada, Sotero, D. & Medeiros, M. 151 (RB) Brazil, Minas Gerais, Santana do Riacho FK0100 KF784188 KF784312 KF784684 KF784560 KF783941 KF784436 KF784064 Dyckia macedoi L. B. Smith mac R.B. Louzada, Sotero, D. & Medeiros, M. 153 (SP) Brazil, Minas Gerais, Santana do Riacho FK0101 KF784189 KF784313 KF784685 KF784561 KF783942 KF784437 KF784065 Dyckia macedoi L. B. Smith mac R.B. Louzada, Sotero, D. & Medeiros, M. 153 (SP) Brazil, Minas Gerais, Santana do Riacho FK0102 KF784190 KF784314 KF784686 KF784562 KF783943 KF784438 KF784066 Dyckia machrisiana L. B. Smith mah E. Esteves s. n. (Leme)/Leme 3291 Brazil, Goiás FK0189 KF784219 KF784343 KF784715 KF784591 KF783971 KF784467 KF784095 Dyckia maracasensis Ule maa G. Martinelli s. n. (Leme)/Leme 0274 Brazil, Bahia, Maracás FK0087 KF784175 KF784299 KF784671 KF784547 KF783928 KF784423 KF784051 Dyckia maritima Baker mar E. Leme 3319 (Leme)/ Leme 3319 Brazil, Rio Grande do Sul, Tôrres FK0092 KF784180 KF784304 KF784676 KF784552 KF783933 KF784428 KF784056 Dyckia maritima Baker mar Unknown (WU)/WU Genf Brazil FK0113 KF784201 KF784325 KF784697 KF784573 KF783954 KF784449 KF784077 1597 123 Species Phylogeny and evolution of Dyckia (Bromeliaceae) Table 1 continued 1598 123 Table 1 continued Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Dyckia marnierlapostollei var. estevesii L. B. Smith/ Rauh man L. Horst 5 (HEID)/ BGHD 130151 Brazil, Goiás, Goiânia FK0030 KF784134 KF784258 KF784630 KF784506 KF783887 KF784382 KF784010 Dyckia marnierlapostollei var. marnier-lapostollei L. B. Smith man L. Horst 4 (HEID)/ BGHD 130234 Brazil, Goiás, Cristalina FK0029 KF784133 KF784257 KF784629 KF784505 KF783886 KF784381 KF784009 Dyckia microcalyx Baker mic Paraguay, Paraguari, Acahay FK0009 KF784116 KF784240 KF784612 KF784488 KF783869 KF784364 KF783992 Dyckia cf. microcalyx Baker mic W. & S. Till 6021 (HEID)/BGHD 102970 W. Till 6066 a (WU)/ WU AB 55 Paraguay, Paraguari, Yaguarón FK0051 KF784150 KF784274 KF784646 KF784522 KF783903 KF784398 KF784026 Dyckia microcalyx Baker mic W. Till 6020 (WU)/ WU AB 57 Paraguay, Paraguari, Acahay FK0054 KF784153 KF784277 KF784649 KF784525 KF783906 KF784401 KF784029 Dyckia microcalyx Baker mic W. Till 6066 (WU)/ WU AB 54 Paraguay, Paraguari, Yaguarón FK0111 KF784199 KF784323 KF784695 KF784571 KF783952 KF784447 KF784075 Dyckia milagrensis Leme mil E. Leme s. n. (Leme)/ Leme s. n. Brazil, Bahia, Milagres FK0096 KF784184 KF784308 KF784680 KF784556 KF783937 KF784432 KF784060 Dyckia mirandiana Leme & Z. J. G.Miranda mir E. Leme 6379 (Leme)/ Leme 6379 Brazil, Goiás, Alto Paraı́so FK0202 KF784227 KF784351 KF784723 KF784599 KF783979 KF784475 KF784103 Dyckia monticola L. B. Smith & Reitz mon E. Leme 1664 (Leme)/ Leme 1664 Brazil, Santa Catarina, Campo Alegre FK0088 KF784176 KF784300 KF784672 KF784548 KF783929 KF784424 KF784052 Dyckia nana Leme et al. nan E. Leme 7485 (Leme)/ Leme 7485 Brazil, Minas Gerais, Diamantina FK0191 KF784221 KF784345 KF784717 KF784593 KF783973 KF784469 KF784097 Dyckia niederleinii Mez nie Unknown (MB)/MB 1982-166 Argentina, Misiones FK0103 KF784191 KF784315 KF784687 KF784563 KF783944 KF784439 KF784067 Dyckia niederleinii Mez nie Unknown (WU)/WU 184/95 Argentina, Misiones FK0110 KF784198 KF784322 KF784694 KF784570 KF783951 KF784446 KF784074 Dyckia paraensis L. B. Smith par E. Leme 7647 (Leme)/ Leme 7647 Brazil, Pará, Guarantan do Norte FK0190 KF784220 KF784344 KF784716 KF784592 KF783972 KF784468 KF784096 Dyckia pectinata L. B. Smith & Reitz pec E. Leme 6490 (Leme)/ Leme 6490 Brazil, Minas Gerais FK0188 KF784218 KF784342 KF784714 KF784590 KF783970 KF784466 KF784094 F. Krapp et al. Species Species Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Dyckia pernambucana L. B. Smith per Diego Pinangé Dyckia Pe-1 (UFP) Brazil, Pernambuco, Brejo da Madre de Deus FK0097 KF784185 KF784309 KF784681 KF784557 KF783938 KF784433 KF784061 Dyckia pernambucana L. B. Smith per Diego Pinangé Dyckia Pe-2 (UFP) Brazil, Pernambuco, Pesqueira FK0098 KF784186 KF784310 KF784682 KF784558 KF783939 KF784434 KF784062 Dyckia platyphylla L. B. Smith pla E. Leme s. n. (Leme)/ Leme s. n. Brazil, Bahia FK0209 KF784230 KF784354 KF784726 KF784602 KF783982 KF784478 KF784106 Dyckia pulquinenses Wittm. pul William Backer s. n. (Leme)/Leme 2415 Bolivia FK0199 KF784226 KF784350 KF784722 KF784598 KF783978 KF784474 KF784102 Dyckia aff. pumila L. B. Smith pum P. Braun BR 696 (HEID)/BGHD 104592 Brazil, Mato Grosso, Ponte Branca FK0017 KF784124 KF784248 KF784620 KF784496 KF783877 KF784372 KF784000 Dyckia pumila L. B. Smith pum E. Leme 4706 (Leme)/ Leme 4706 Brazil, Goiás, Caiaponia FK0093 KF784181 KF784305 KF784677 KF784553 KF783934 KF784429 KF784057 Dyckia rariflora Schult. f. rar Unknown (BONN)/ BONN 2411 Brazil, Minas Gerais FK0039 KF784141 KF784265 KF784637 KF784513 KF783894 KF784389 KF784017 Dyckia aff. reitzii L. B. Smith rei A. Hofacker 386 (WU)/WU B 02/62-1 Brazil, Rio Grande do Sul, Cambará do Sul FK0050 KF784149 KF784273 KF784645 KF784521 KF783902 KF784397 KF784025 Dyckia remotiflora Otto & A. Dietr. Dyckia remotiflora Otto & A. Dietr. rem L. Horst s. n. (HEID)/ BGHD 130009 L. Horst 345 (HEID)/ BGHD 130010 Brazil FK0011 KF784118 KF784242 KF784614 KF784490 KF783871 KF784366 KF783994 Brazil FK0015 KF784122 KF784246 KF784618 KF784494 KF783875 KF784370 KF783998 rem rem Unknown (WU)/WU AB 3/86 Unknown FK0055 KF784154 KF784278 KF784650 KF784526 KF783907 KF784402 KF784030 Dyckia remotiflora Otto & A. Dietr. Dyckia rojasii Mez rem Uruguay FK0068 KF784157 KF784281 KF784653 KF784529 KF783910 KF784405 KF784033 roj Unknown (KAS)/KAS FK0068 E. Leme 6465 (Leme)/ Leme 6465 Brazil, Paraná, Rio Branco do Ivaı́ FK0195 KF784225 KF784349 KF784721 KF784597 KF783977 KF784473 KF784101 Dyckia saxatilis Mez oli W. Barthlott 10327 (BONN)/BONN 4346 Brazil, Minas Gerais, Serro FK0036 KF784139 KF784263 KF784635 KF784511 KF783892 KF784387 KF784015 Dyckia secunda L. B. Smith sec E. Leme 3682 (Leme)/ Leme 3682 Brazil, Bahia, Contendas do Sincorá FK0192 KF784222 KF784346 KF784718 KF784594 KF783974 KF784470 KF784098 1599 123 Dyckia remotiflora Otto & A. Dietr. Phylogeny and evolution of Dyckia (Bromeliaceae) Table 1 continued 1600 123 Table 1 continued Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Dyckia spec. sp. W. Rauh 64142 (HEID)/BGHD 130026 Argentina, Salta, Salta FK0023 KF784129 KF784253 KF784625 KF784501 KF783882 KF784377 KF784005 Dyckia spec. sp. Leuenberger, Arroyo & Eggli 4254 a (HEID)/BGHD 104100 Argentina, Catamarca, Belen FK0024 KF784130 KF784254 KF784626 KF784502 KF783883 KF784378 KF784006 Dyckia spec. sp. H. Amerhauser 96-15 (WU)/WU 401/96 Paraguay, Boquerón, Fortı́n Capitan Demattei FK0049 KF784148 KF784272 KF784644 KF784520 KF783901 KF784396 KF784024 Dyckia spec. sp. Unknown (WU)/WU 1088 Unknown FK0116 KF784204 KF784328 KF784700 KF784576 KF783957 KF784452 KF784080 Dyckia spec. sp. L. Horst 386 (HEID)/ BGHD 130035 Brazil, Minas Gerais, Diamantina FK0086 KF784174 KF784298 KF784670 KF784546 KF783927 KF784422 KF784050 Dyckia tobatiensis Hassl. tob W. & S. Till 6050 (WU)/BGHD 102969 Paraguay, Cordillera, Tobati FK0018 KF784125 KF784249 KF784621 KF784497 KF783878 KF784373 KF784001 Dyckia tomentella Mez tom Unknown (WU)/WU 210/91 Paraguay FK0114 KF784202 KF784326 KF784698 KF784574 KF783955 KF784450 KF784078 Dyckia aff. tuberosa (Vellozo) Beer tub E. Leme 6837 (Leme)/ Leme 6837 Brazil, Paraná, São Gerônimo da Serra FK0206 KF784228 KF784352 KF784724 KF784600 KF783980 KF784476 KF784104 Dyckia ursina L. B. Smith urs Unknown (HEID)/ BGHD 103809 Brazil FK0012 KF784119 KF784243 KF784615 KF784491 KF783872 KF784367 KF783995 Dyckia ursina L. B. Smith urs E. Leme 1837 (HB)/ Leme 1837 Brazil, Minas Gerais, Jaboticatubas FK0089 KF784177 KF784301 KF784673 KF784549 KF783930 KF784425 KF784053 Dyckia velascana Mez vel W. & S. Till 5012 (WU)/BGHD 103740 Argentina, Cordoba, Ascochinga FK0006 KF784114 KF784238 KF784610 KF784486 KF783867 KF784362 KF783990 Dyckia velascana Mez vel Unknown (WU)/WU 68/88 Argentina FK0104 KF784192 KF784316 KF784688 KF784564 KF783945 KF784440 KF784068 Dyckia velascana Mez vel Unknown (WU)/WU B 183/95 Argentina FK0105 KF784193 KF784317 KF784689 KF784565 KF783946 KF784441 KF784069 Dyckia cf. velascana Mez vel W. Till 10245 (WU)/ WU B 13/93 Argentina, Tucumán, Famaillá FK0106 KF784194 KF784318 KF784690 KF784566 KF783947 KF784442 KF784070 F. Krapp et al. Species Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Dyckia vestita Hassl. ves W. & S. Till 6018 (WU)/BGHD 103741 Paraguay, Paraguari, Paraguari FK0032 KF784136 KF784260 KF784632 KF784508 KF783889 KF784384 KF784012 Dyckia vestita Hassl. ves W. Till 6019 (WU)/ WU AB 59 Paraguay, Paraguari, Paraguari FK0109 KF784197 KF784321 KF784693 KF784569 KF783950 KF784445 KF784073 Encholirium erectiflorum L. B. Smith Eer P. Braun 4072 (HEID)/BGHD s. n. Brazil FK0123 KF784211 KF784335 KF784707 KF784583 KF783963 KF784459 KF784087 Encholirium horridum L. B. Smith Eho W. Schindhelm s. n. (HEID)/BGHD 108213 Brazil, Minas Gerais, Pedra Azul FK0070 KF784158 KF784282 KF784654 KF784530 KF783911 KF784406 KF784034 Encholirium magalhaesii L. B. Smith Ema Unknown (BONN)/ BONN 4344 Brazil FK0122 KF784210 KF784334 KF784706 KF784582 KF783962 KF784458 KF784086 Encholirium maximum Forzza & Leme Emx P. Braun 4063 (HEID)/BGHD s. n. Brazil FK0124 KF784212 KF784336 KF784708 KF784584 KF783964 KF784460 KF784088 Encholirium spec. Esp R. Schulz s. n. (HEID)/BGHD 125585 Brazil FK0078 KF784166 KF784290 KF784662 KF784538 KF783919 KF784414 KF784042 Encholirium spec. Esp R. Schulz s. n. (HEID)/BGHD 143704 Brazil FK0079 KF784167 KF784291 KF784663 KF784539 KF783920 KF784415 KF784043 Encholirium spec. Esp R. Schulz s. n. (HEID)/BGHD 112920 Brazil FK0080 KF784168 KF784292 KF784664 KF784540 KF783921 KF784416 KF784044 Encholirium spec. Esp Unknown (B)/B 232-39-94-60 Brazil FK0125 KF784213 KF784337 KF784709 KF784585 KF783965 KF784461 KF784089 Encholirium spec. Esp W. Rauh 56468 (HEID)/BGHD 130033 Brazil, Bahia, Brumado FK0014 KF784121 KF784245 KF784617 KF784493 KF783874 KF784369 KF783997 Encholirium spec. Esp M. G. L. Wanderley & Sousa, G. M. 2630 (SP) Brazil, Piauı́ FK0257 KF784234 KF784358 KF784730 KF784606 KF783986 KF784482 KF784110 Fosterella albicans (Grisebach) L. B. Smith J. Peters 06.0005 (HEID)/KAS JP06.0005 Bolivia, Santa Cruz, Pampagrande FK0117 KF784205 KF784329 KF784701 KF784577 KF783958 KF784453 KF784081 Fosterella penduliflora (C. H. Wright) L. B. Smith J. Peters 06.0054 (HEID)/KAS JP06.0054 Bolivia, Tarija, Aniceto Arce FK0081 KF784169 KF784293 KF784665 KF784541 KF783922 KF784417 KF784045 1601 123 Species Phylogeny and evolution of Dyckia (Bromeliaceae) Table 1 continued 1602 123 Table 1 continued Species Abbreviation used for network Collector (herbarium)/ living plant Geographic provenance (State, Dept., Prov.) DNA no. rpl32trnL rps16trnK matK rps16 intron petD intron trnD-trnT phyC Fosterella spectabilis H. Luther J. Peters 06.0046 (HEID)/KAS JP06.0046 Bolivia, Chuquisaca, Monteagudo FK0118 KF784206 KF784330 KF784702 KF784578 KF783959 KF784454 KF784082 Fosterella villosula (Harms) L. B. Smith J. Peters 06.0105 (HEID)/KAS JP06.0105 Bolivia, Cochabamba, Cochabamba FK0076 KF784164 KF784288 KF784660 KF784536 KF783917 KF784412 KF784040 Fosterella weddelliana (Brongniart ex Baker) L. B. Smith M. Miyagawa s. n. (HEID)/BGHD 104866 Bolivia, Solacana FK0077 KF784165 KF784289 KF784661 KF784537 KF783918 KF784413 KF784041 Pitcairnia feliciana (A. Chevalier) Harms and Mildbraed I. Ebert & D. Bangoura s. n. ex coll. P. Bak (WU)/ WU s. n. Guinea, Prefecture de Kindia, Kindia FK0119 KF784207 KF784331 KF784703 KF784579 KF783960 KF784455 KF784083 Pitcairnia fuertesii Mez W. Till 18087 (WU)/ WU s. n. Dominican Republic, Prov. Puerto Plata, Puerto Plata FK0120 KF784208 KF784332 KF784704 KF784580 n/a KF784456 KF784084 Pitcairnia heterophylla (Lindley) Beer K. Senghas O-11230 (HEID)/BGHD 104945 Mexico, Guerrero, Cruz de Ocotte FK0083 KF784171 KF784295 KF784667 KF784543 KF783924 KF784419 KF784047 Pitcairnia pipenbringii Rauh & E. Gross W. Rauh 67430 (HEID)/BGHD 103786 Brazil, Bahia, Barreiras FK0085 KF784173 KF784297 KF784669 KF784545 KF783926 KF784421 KF784049 Pitcairnia pungens Link, Klotzsch & Otto W. Rauh 69140 (HEID)/BGHD 130648 W. Rauh s. n. (HEID)/ BGHD 130165 Peru, Lambayeque, valley of Olmos river Peru, Valley of Rio Marañon FK0084 KF784172 KF784296 KF784668 KF784544 KF783925 KF784420 KF784048 FK0082 KF784170 KF784294 KF784666 KF784542 KF783923 KF784418 KF784046 T. Krömer 6581 (HEID)/BGHD 105240 Bolivia, Cochabamba, Carrasco FK0121 KF784209 KF784333 KF784705 KF784581 KF783961 KF784457 KF784085 Puya ferruginea (Ruiz & Pavón) L. B. Smith Puya herzogii Wittm. B, Botanical Garden Berlin-Dahlem; BGHD, Botanical Garden of the University of Heidelberg; BONN, Botanical Garden of the University of Bonn; KAS, living collections of the University of Kassel; Leme, Private collection of Elton Leme (Teresópolis, Brazil); MB, Botanical Garden of the University of Marburg; WU, Botanical Garden of the University of Vienna F. Krapp et al. Phylogeny and evolution of Dyckia (Bromeliaceae) Primer pairs used for the PCR were derived from Demesure et al. (1995) for trnD-trnT, Watts et al. (2008) for the petD intron and Oxelman et al. (1997) for the rps16 intron. For the amplification of the matK gene the primers matK5 F (Crayn et al. 2000) and BROM1 R (Schulte et al. 2005) were used. The spacer rpl32-trnL was amplified in two overlapping portions, using primers designed by Shaw et al. (2007) and two additional internal primers (Schütz 2012). The rps16-trnK spacer was amplified in three portions, using primers from Shaw et al. (2007) and additional internal primers specifically designed for this study (rps16-trnK a rev: 50 -AAG AAA AAG GAA AAT GGT GTG-30 , rps16-trnK b fwd: 50 GGG AAG GGT AAC TAG TAT C-30 , rps16-trnK b rev: 50 -AAA GAC TTG TGT TGG ATT GGC-30 , rps16-trnK c fwd: 50 -CTT TTC CTT AAT TTT TTC CAT TCC-30 ). All primers carried an M13-tail at their 50 end to facilitate subsequent sequencing (see below). The nuclear locus phyC was amplified using the primers phyc515f-br and phyc1699r-br (Barfuss 2012) without M13-tail. PCR products generated by these primers were separated on agarose gels, bands were picked and subjected to a bandstab PCR for purification (Bjourson and Cooper 1992), using the internal primer pairs phyc524f and phyc1145r for fragment A and phyc974f2 and phyc1690r for fragment B (Barfuss 2012). These nested primers carried an M13-tail at their 50 ends. All primers were purchased from Metabion GmbH (Martinsried, Germany). PCRs were performed in total volumes of 20 lL using a Biometra T1-cycler (Biometra GmbH, Göttingen, Germany). Each assay contained 1–5 ng of template DNA for plastid loci and 5–25 ng for phyC, 19 Mango-Taq reaction buffer (Bioline, Taunton, USA), 1.5 mM MgCl2, 0.2 mM of each dNTP, 0.25 lM of each primer, 0.5 lg/lL BSA (for plastid loci only), 2.5 % DMSO (for phyC only) and 0.1 U Mango-Taq DNA polymerase (Bioline, Taunton, USA). All plastid loci were amplified with the same PCR program, consisting of an initial denaturation at 80 °C for 300 s, followed by 30 cycles of denaturation at 94 °C for 60 s, primer annealing at 52 °C for 60 s and elongation at 65 °C for 120 s, and a final extension at 65 °C for 10 min. The phyC locus was amplified with the two external primers, applying an initial denaturation at 94 °C for 120 s, followed by 15 cycles at 94 °C for 15 s, 59 °C for 30 s and 70 °C (ramp 1 °C/s) for 120 s, then 25 cycles of 94 °C for 15 s, 59 °C for 30 s and 70 °C (ramp 1 °C/s) for 120 s, increasing by 10 s per cycle. Nested primers were then used for the second PCR, with an initial denaturation at 94 °C for 120 s, followed by 25 cycles of 94 °C for 30 s, 50 °C for 30 s and 72 °C for 120 s. Final extension was at 72 °C for 8 min. To check for the presence and quality of amplicons, aliquots of PCR products were electrophoresed on agarose gels and stained with ethidium bromide. 1603 DNA sequencing All PCR products were sequenced without further purification using the dideoxynucleotide chain termination method. Both strands were sequenced simultaneously with a Thermo Sequenase Primer Cycle Sequencing Kit (GE Healthcare, Little Chalfont, UK), IRDye700-labelled forward sequencing primer (M13f: 50 -TGT AAA ACG ACG GCC AGT-30 ) and IRDye800-labelled reverse sequencing primer (M13r: 50 -CAG GAA ACA GCT ATG ACC-30 ), both purchased from Metabion GmbH (Martinsried, Germany). The sequencing reaction was performed in a Biometra T1-cycler following the protocol of the kit manufacturer. After initial denaturation at 95 °C for 300 s, 25 cycles were run consisting of 95 °C for 30 s, 57 °C for 30 s and 72 °C for 60 s. After final elongation at 72 °C for 490 s the products were mixed with the same volume of formamide buffer [98 % (v/v) formamide, 0.025 % (v/v) basic fuchsine, 10 mM EDTA] and denatured at 85 °C for 300 s. The products were separated on denaturing polyacrylamide gels (SequaGel XR, National Diagnostics, Atlanta, USA) in an automated LI-COR 4200 DNA sequencer (LI-COR Biosciences, Bad Homburg, Germany). Data analysis The forward and reverse sequences were edited and assembled using e-Seq 2.0 and AlignIR 1.2 software (both LI-COR Biosciences, Bad Homburg, Germany). Consensus sequences were aligned with the help of PhyDE (Müller et al. 2011) followed by manual adjustments. Indels, simple sequence repeat (SSR) regions and inversions were excluded from all analyses. All phylogenetic reconstructions were performed separately for the combined plastid data set and for the phyC sequence alignment. No combined analysis of nuclear and plastid data sets was attempted due to the considerable degree of incongruence. Bayesian relaxed phylogenetics and network analyses were applied only to the plastid data set. Maximum parsimony (MP) analyses were performed using PAUP* 4.0b (Swofford 2003). Consensus trees were generated from a heuristic search with 100 stepwise random addition replicates using tree bisection and reconnection (TBR) branch swapping with steepest-descent modification and MulTrees option in effect. To evaluate the extent of homoplasy in the data set the consistency (CI) and retention (RI) indices were calculated. Statistical support values were estimated running 1,000 bootstrap (BS) replicates with 10 random addition replicates, each using TBR branch swapping with the same modifications as above. No more than 100 trees were saved per bootstrap replicate. Maximum likelihood (ML) analyses and ML bootstrapping for 1,000 replicates each were performed 123 1604 F. Krapp et al. with RAxML 7.2.6, assuming a GTRCAT substitution model (Stamatakis 2006). For Bayesian inference (BI) analyses the data sets were tested for the best-fit model of evolution with MrModeltest v2.3 (Nylander 2004) using the Akaike information criterion (AIC). To this end, the combined chloroplast data set was divided into six partitions (matK, rps16 intron, petD intron, rpl32-trnL, rps16-trnK and trnD-trnT) and each locus was evaluated individually. BI analyses were done using MrBayes 3.2.1 (Huelsenbeck and Ronquist 2001), and all parameters except branch length and topology were allowed to vary among the individual partitions. Two independent runs were initiated, with 1,000,000 generations sampled every 100th generation including branch length. Each run consisted of three heated chains using a heating parameter of 0.2 and one cold chain. After plotting the likelihood-by-generation values, the first 10 % of the runs were discarded as burn-in. The remaining trees were used for the construction of a 50 % majority-rule consensus tree. The software package BEAST 1.6.1 (Drummond and Rambaut 2007) was used to generate a dated phylogeny for the plastid data set via Bayesian relaxed phylogenetics. The time scale was calibrated using the age of the Pitcairnioideae crown group as estimated by Givnish et al. (2011) as single calibration point. Therefore, we used a normal prior distribution with 11.8 million years ago (Mya) as mean value and a standard deviation of 0.5 million years (My). GTR ? G with empirically determined base frequencies was used as a substitution model, and a relaxed clock with uncorrelated rates following a log-normal distribution was chosen as clock model. The tree prior was set to the Yule process. Four independent analyses were run for 10,000,000 generations, each with every 1,000th tree sampled. To control convergence and to ensure a sufficient effective sample size of 200 and above for all parameters, the log files were analysed with Tracer 1.5 (Rambaut and Drummond 2009). All resulting trees were combined using LogCombiner 1.6.1, discarding the first 10 % of each run as burn-in. From the remaining trees a 50 % majority-rule consensus tree was generated using PAUP* 4.0b. Node ages determined by the BEAST runs were added onto this tree using TreeAnnotator 1.6.1. Haplotype networks were calculated using the statistical parsimony approach implemented in TCS 1.21 (Clement et al. 2000). Deuterocohnia was used to root the network of Dyckia and Encholirium plastid haplotypes. Gaps were treated as missing data, and a connection limit of 95 % was used. The results were compared to various single most parsimonious trees from the MP analysis to evaluate incidences of homoplasy assumed by TCS. Results Alignment and sequence statistics An overview of alignment and sequence statistics is given in Table 2. The concatenated alignments of the six investigated plastid loci comprised of a total of 6,009 characters. Of these, 613 (10.2 %) were variable, and 293 (4.9 %) were parsimony-informative in the full alignment that included all outgroups. Only a small portion of the overall variation was, however, contributed by the ingroup (Dyckia only), in which 119 plastid characters (2.0 %) were variable, and 61 (1.0 %) were parsimony-informative. Amplicon sizes as well as numbers and percentages of polymorphic sites varied considerably between individual plastid DNA regions, with petD intron and rps16-trnK representing the least and the most variable regions, respectively (for details see Table 2). The overall alignment contained 76 indels and 22 polymorphic mononucleotide repeats (SSRs). Within Dyckia, 16 indels and seven variable SSRs were observed. Gaps were treated as missing data in all analyses. A 6-bp region within the rpl32-trnL spacer showed signs of frequent and recurrent inversion, and was also excluded from the analysis. Table 2 Sequence statistics of the investigated plastid loci and the nuclear locus phyC Region Alignment length Size range (bp) All taxa (including outgroups) Variable sites PI Dyckia only Indels/SSR Variable sites PI Indels/SSR 3/3 rpl32-trnL 1,008 854–946 124 (12.4 %) 62 (6.2 %) 10/4 26 (2.6 %) 15 (1.5 %) rps16-trnK 896 830–861 119 (13.3 %) 66 (7.4 %) 14/5 33 (3.7 %) 18 (2.0 %) 6/2 matK 803 775–796 74 (9.2 %) 36 (4.5 %) 4/1 16 (2.0 %) 6 (0.7 %) 0/0 rps16 intron 976 810–884 92 (9.4 %) 42 (4.3 %) 18/4 20 (2.0 %) 10 (1.0 %) 4/1 petD intron 719 703–767 76 (9.6 %) 25 (3.2 %) 9/1 7 (0.9 %) 6 (0.8 %) 1/0 17 (1.1 %) 119 (2.0 %) 6 (0.4 %) 61 (1.0 %) 2/1 16/7 66 (5.7 %) 34 (2.9 %) 0/0 trnD-trnT All plastid loci 1,538 6,009 845–1376 4939–5449 128 (8.3 %) 613 (10.2 %) 62 (4.0 %) 293 (4.9 %) 21/7 76/22 phyC 1,159 1159 189 (16.3 %) 114 (9.8 %) 0/0 PI parsimony informative characters, SSR polymorphic simple sequence (mononucleotide) repeats 123 Phylogeny and evolution of Dyckia (Bromeliaceae) For phyC the final alignment had a length of 1,159 characters (Table 2). Of these, 189 (16.3 %) were variable, and 114 (9.8 %) were parsimony-informative in the full alignment including outgroups. Within Dyckia, 66 phyC characters (5.7 %) were variable, and 34 were parsimonyinformative (2.9 %). A high incidence of intra-individual allelic variation was observed at the phyC locus, as indicated by the occurrence of double peaks at the respective sequence position. Only 25 of the 96 investigated Dyckia samples were fully homozygous, whereas 24 samples were heterozygous at a single sequence position, and 47 Dyckia samples were heterozygous at 2–11 sequence positions. One particular sample showed allelic variation at 21 positions. Phylogenetic analyses of plastid loci For the combined plastid sequence alignment all applied algorithms (MP, ML, BI and BEAST) produced trees with congruent topologies, but slightly variable support values. Only the tree resulting from Bayesian relaxed phylogenetic analysis using BEAST 1.6.1 is shown here (Fig. 2), with posterior probabilities given above branches, and bootstrap values resulting from an MP and an ML analysis (RAxML 7.2.6) shown below branches. The BEAST tree was also used for evaluating the spatio-temporal evolution of Dyckia, which is treated in detail in the Discussion section. When rooted with Puya, Pitcairnioideae s.str. is clearly monophyletic (BEAST posterior probability PP = 1), and is split into an unresolved tritomy consisting of two separate lineages of Pitcairnia and the monophyletic (PP = 0.99) remainder of the subfamily. Within the latter clade, Fosterella is monophyletic (PP = 1) and sister to a monophylum (PP = 1) formed by the three xerophytic genera Dyckia, Deuterocohnia and Encholirium. Deuterocohnia is clearly paraphyletic (PP = 1), with one of its two lineages being sister to a clade formed by Encholirium and Dyckia (PP = 0.95). Dyckia comes out as a monophyletic group that arises from within Encholirium, rendering the latter genus paraphyletic. An early branching lineage (Dyckia beateae 2 from Mato Grosso do Sul, BR) is excluded from the monophyletic remainder by PP = 1, followed by a lineage comprising Dyckia espiritosantensis and Dyckia mirandiana. These two taxa are sister to all other accessions of the genus, which form the large, monophyletic core Dyckia (PP = 0.98). Core Dyckia comprises four well-supported larger clades together with one smaller clade and numerous single accessions in a poorly resolved polytomy. The four large clades follow a geographic rather than taxonomic pattern, and we refer to these clades as Central Brazilian, Paraguayan, Southern Brazilian, and Argentinean clade, respectively (Fig. 2). 1605 The unexpectedly low levels of infrageneric plastid DNA variation and hence limited resolution of our trees prompted us to attempt a statistical parsimony network analysis of plastid haplotypes using the TCS computer program (Clement et al. 2000). Given that a large number of mutational steps separated Dyckia and Encholirium from the remaining taxa, we limited our analysis to these two genera. Results are shown in Fig. 3. A total of 63 different haplotypes were found among the 97 Dyckia accessions, whereas each of the 10 Encholirium accessions had its own haplotype. Dyckia haplotypes proved to be closely spaced, with only 39 ‘‘empty’’ haplotypes not represented in the dataset (Fig. 3). In contrast, all accessions of Encholirium were separated from each other by numerous base substitutions, resulting in a scattered distribution. When rooted with Deuterocohnia, the network suggests paraphyly for Encholirium and monophyly for Dyckia. The two lineages of Dyckia that were early branching in the trees (i.e. D. beateae, D. espiritosantensis and D. mirandiana, see Fig. 2) take an intermediate position between Encholirium and core Dyckia. The haplotypes of all members of core Dyckia are arranged in a star-like pattern and appear to have originated from one central haplotype which was found in two plants from southern Brazil (one accession each of Dyckia reitzii and Dyckia jonesiana). The four major, geographically defined Dyckia clades retrieved in the plastid phylogenetic trees correspond to the four major lineages originating from this central haplotype, i.e. the Central Brazilian clade, the Argentinean clade, the Southern Brazilian clade that also harbours one plant from Uruguay, and the Paraguayan clade that also comprises a few plants from Southern Brazil, Bolivia and Argentina. The relatively large Central Brazilian also shows a star-like pattern, with a central haplotype found in several plants from different species. Phylogenetic analysis of the nuclear phyC locus In preliminary experiments, we had tested the applicability of phyC for phylogenetic analysis in a larger set of Bromeliaceae that included samples from five subfamilies. Unexpectedly, the two species of Puya included in our sample set were found to be nested inside Pitcairnioideae s.str. in all these analyses (not shown). This situation did not change when we included GenBank-derived phyC sequences from additional Puya species published by Jabaily and Sytsma (2010): the phyC sequences from Puya were consistently nested within the Pitcairnioideae. Given that Puya was not the focus of our study, we decided to use midpoint rooting for our phyC analyses instead. The results from the Bayesian analysis are shown in Fig. 4. No major incongruities were observed when MP and ML were used instead. 123 1 89/87 0,98 64/61 1 62/- 1 68/96 1 56/94 0,99 62/83 1 98/100 0,60 -/- 0,99 63/63 1 0,99 88/95 51/56 1 63/80 1 63/82 0,99 0,90 -/52 67/69 0,99 -/65 1 62/92 1 -/55 0,99 -/- 0,50 1 -/78 85/87 1 85/87 1 0,52 61/77 -/1 -/- 0,99 1 85/93 59/58 0,98 53/- 0,98 -/- 1 65/80 1 93/94 1 97/96 1 93/95 1 99/100 0,68 -/54 0,55 -/- 0,95 84/91 1 100/100 1 100/100 1 74/81 1 89/94 1 100/100 0,99 88/87 0,91 61/67 0,91 -/75 1 100/100 1 100/100 1 100/100 0,99 -/- 0,95 64/66 1 90/93 1 66/80 0,99 64/67 1 100/100 1 98/100 1 99/100 1 99/100 1 100/100 million years ago 13 12 123 11 10 9 8 7 6 5 4 3 2 1 0 Paraguayan clade 1 74/78 Southern Brazilian clade 0,95 -/- 1 77/86 Dyckia dissitiflora Dyckia maracasensis Dyckia secunda Dyckia granmogulensis Dyckia pernambucana 1 Dyckia pernambucana 2 Dyckia milagrensis Dyckia nana Dyckia ursina 2 Dyckia macedoi 1 Dyckia macedoi 2 Dyckia macedoi 3 Dyckia macedoi 4 Dyckia estevesii var. braunii 2 Dyckia lindevaldae Dyckia braunii Dyckia aurea Dyckia platyphylla Dyckia dawsonii Dyckia cinerea Dyckia lunaris Dyckia estevesii 1 Dyckia estevesii 2 Dyckia marnier-lapostollei var. mar. Dyckia marnier-lapostollei var. est. Dyckia goehringii Dyckia saxatilis Dyckia brevifolia 2 Dyckia delicata Dyckia beateae 1 Dyckia brachyphylla Dyckia monticola Dyckia ferruginea Dyckia fosteriana Dyckia densiflora Dyckia pectinata Dyckia aff. tuberosa Dyckia vestita 1 Dyckia vestita 2 Dyckia tomentella Dyckia ferox 3 Dyckia cf. microcalyx 2 Dyckia microcalyx 4 Dyckia microcalyx 1 Dyckia microcalyx 3 Dyckia distachya Dyckia ferox 2 Dyckia spec. 3 Dyckia tobatiensis Dyckia niederleinii 2 Dyckia cf. leptostachya 4 Dyckia spec. 5 Dyckia ferox 1 Dyckia rojasii Dyckia pulquinenses Dyckia choristaminea 1 Dyckia hebdingii 2 Dyckia remotiflora 1 Dyckia hebdingii 1 Dyckia remotiflora 2 Dyckia choristaminea 2 Dyckia hebdingii 3 Dyckia maritima 2 Dyckia machrisiana Dyckia elisabethae Dyckia aff. brevifolia 1 Dyckia rariflora Dyckia remotiflora 4 Dyckia aff. ibiramensis Dyckia spec. 4 Dyckia floribunda 1 Dyckia floribunda 3 Dyckia cf. velascana 4 Dyckia ursina 1 Dyckia spec. 1 Dyckia spec. 2 Dyckia velascana 1 Dyckia remotiflora 3 Dyckia niederleinii 1 Dyckia velascana 2 Dyckia velascana 3 Dyckia floribunda 2 Dyckia aff. leptostachya 1 Dyckia aff. leptostachya 2 Dyckia aff. pumila 1 Dyckia grandidentata Dyckia leptostachya 3 Dyckia maritima 1 Dyckia encholirioides Dyckia paraensis Dyckia aff. reitzii Dyckia jonesiana Dyckia estevesii 4 Dyckia pumila 2 Dyckia mirandiana Dyckia espiritosantensis Dyckia beateae 2 Encholirium spec. 6 Encholirium erectiflorum Encholirium spec. 1 Encholirium maximum Encholirium horridum Encholirium spec. 2 Encholirium spec. 4 Encholirium spec. 3 Encholirium magalhaesii Encholirium spec. 5 Deuterocohnia brevispicata Deuterocohnia meziana Deuterocohnia brevifolia Deuterocohnia cf. haumanii Deuterocohnia glandulosa Fosterella albicans Fosterella spectabilis Fosterella villosula Fosterella weddelliana Fosterella penduliflora Pitcairnia heterophylla Pitcairnia pungens Pitcairnia pipenbringii Pitcairnia fuertesii Pitcairnia feliciana Puya ferruginea Puya herzogii 0,52 0,68-/62 -/1 0,99 -/51 62/92 Central Brazilian clade F. Krapp et al. Argentinean clade 1606 Phylogeny and evolution of Dyckia (Bromeliaceae) 1607 b Fig. 2 Dated plastid phylogeny of Pitcairnioideae, based on a Encholirium and Dyckia, with Puya being sister to Encholirium plus Dyckia. Encholirium comes out as paraphyletic, and all samples of Dyckia, including some of Encholirium form a large unresolved polytomy. We made several attempts to improve the disappointingly poor infrageneric resolution within Dyckia, by (1) excluding samples with more than two heterozygous positions; (2) excluding all positions that are heterozygous in the majority of samples; or (3) excluding all genera but Dyckia and Encholirium from the dataset. However, the trees resulting from these attempts showed more or less the same topology and the same low resolution as those resulting from the complete analyses. Removal of all heterozygous positions even led to a complete breakdown of resolution, also regarding the deeper nodes that receive good support with the full dataset. Bayesian relaxed phylogenetic analysis of the combined plastid data set with BEAST 1.6.1. The tree was rooted with Puya. The age of Pitcairnioideae (11.8 million years) estimated by Givnish et al. (2011) was used as a single calibration point. Posterior probabilities are given above branches, bootstrap values from 1,000 replicates of a maximum parsimony (PAUP* 4.0b) and a maximum likelihood analysis (RAxML 7.2.6) are given below branches (MP/ML) The use of midpoint rooting suggests that monophyletic Pitcairnia is sister to a clade comprising the remaining genera of Pitcairniodeae s.str. plus Puya. The latter clade is divided into two subclades that are sister to each other. One subclade consists of Deuterocohnia and Fosterella (both monophyletic with PP = 1) in a weakly supported sister position to each other (PP = 0.76). The second subclade is a weakly supported (PP = 0.65) monophylum of Puya, mac mac daw mac mac dis est cin lun lin bra bra aur pla ves ves tom feu pec man, man bea, mon fos, den sax fer mic mic dit urs mil per per nan goe esp tob, nie lep, sp. roj mic mic fer Esp Emx mir Esp est par pul ibi lep lep pum Esp pum Ema Dbr rei jon lep Paraguayan clade sp. mar grn bea bre rar enc Dme sp. urs Eer Esp brc Central Brazilian clade bre fer Esp tub del sp. gra maa sec est rem, rem cho, heb mah, eli vel sp. flo Argentinean clade flo vel, rem nie, vel, vel, flo rem cho heb Esp mar heb Fig. 3 Plastid haplotype network constructed using statistical parsimony implemented in TCS 1.21 (Clement et al. 2000) and crossverified with Bayesian and maximum parsimony analyses. Only accessions of Dyckia and Encholirium (violet-coloured circles) are included. Two samples of Deuterocohnia (grey-coloured circles) Southern Brazilian clade Eho were used for rooting. Colours of Dyckia haplotypes reflect geographical origins as indicated on the map (white unknown). Small empty circles represent hypothetical sequence variants that were presumed by the program but not observed in the data set. For species abbreviations see Table 1 123 1608 F. Krapp et al. 0,99 51/59 1 93/99 0,65 58/0,99 74/81 1 100/100 1 97/98 0,76 62/65 1 99/100 1 99/100 1 100/100 1,0% 123 Dyckia cinerea Dyckia aurea Dyckia lindevaldae Dyckia saxatilis Dyckia brachyphylla Dyckia ursina 2 Dyckia macedoi 1 0,66 Dyckia macedoi 2 Dyckia macedoi 3 -/Dyckia macedoi 4 Dyckia nana Dyckia secunda Dyckia aff. tuberosa Dyckia espiritosantensis Dyckia densiflora Dyckia floribunda 1 Dyckia floribunda 3 Dyckia cf. velascana 4 0,68 Dyckia ursina 1 Dyckia spec. 1 57/75 Dyckia spec. 2 Dyckia velascana 1 Dyckia floribunda 2 Dyckia estevesii 1 Dyckia goehringii 0,67 Dyckia estevesii 4 Dyckia beateae 1 -/Dyckia ferruginea 0,79 Dyckia mirandiana -/Dyckia dawsonii Dyckia paraensis 0,82 Dyckia tobatiensis Dyckia vestita 1 -/74 Dyckia microcalyx var. indet. 1 0,52 Dyckia microcalyx 3 -/Dyckia vestita 2 Dyckia distachya Dyckia granmogulensis Dyckia milagrensis 0,68 Dyckia pernambucana 1 -/Dyckia pernambucana 2 Dyckia dissitiflora Dyckia remotiflora var. indet. 1 0,95 Dyckia remotiflora var. indet. 2 Dyckia choristaminea 2 0,54 63/64 Dyckia cf. microcalyx 2 -/Dyckia microcalyx 4 0,99 Dyckia marnier-lapostollei var. mar. Dyckia marnier-lapostollei var. est. -/75 0,66 Dyckia aff. reitzii -/Dyckia platyphylla Dyckia beateae 2 Dyckia aff. ibiramensis 1 Dyckia remotiflora 3 0,98 94/95 Dyckia spec. 4 -/62 Dyckia jonesiana Dyckia hebdingii 3 0,64 Dyckia maritima 2 -/Dyckia delicata Dyckia elisabethae 0,56 Dyckia brevifolia 2 1 Dyckia niederleinii 1 74/98 64/60 Dyckia velascana 3 0,67 Dyckia estevesii var. braunii 2 Dyckia estevesii 3 -/0,61 Dyckia aff. leptostachya 2 Dyckia cf. leptostachya 4 63/65 0,57 Dyckia pumila 2 Dyckia grandidentata -/Dyckia aff. brevifolia 1 Dyckia hebdingii 1 Dyckia aff. leptostachya 1 Dyckia aff. pumila 1 0,90 Dyckia ferox 1 -/Dyckia choristaminea 1 Dyckia ferox 2 Dyckia hebdingii 2 Dyckia rariflora Dyckia braunii Dyckia spec. 3 Dyckia leptostachya 3 Dyckia ferox 3 Dyckia remotiflora 4 Dyckia maracasensis Dyckia monticola Dyckia maritima 1 Dyckia fosteriana 0,83 Dyckia encholirioides -/90 Dyckia velascana 2 Dyckia niederleinii 2 Dyckia tomentella Dyckia spec. (FK0116) Dyckia pectinata Dyckia machrisiana 0,59 Dyckia lunaris Dyckia rojasii -/63 Dyckia pulquinenses Encholirium spec. 5 Encholirium spec. 2 Encholirium spec. 3 Encholirium spec. 4 Encholirium magalhaesii Encholirium spec. 6 Encholirium horridum 0,70 Encholirium spec. 1 Encholirium erectiflorum 70/Encholirium maximum Puya ferruginea Puya herzogii 0,81 Deuterocohnia brevispicata Deuterocohnia meziana -/Deuterocohnia brevifolia 0,88 0,87 Deuterocohnia cf. haumanii -/62 Deuterocohnia glandulosa -/69 0,73 Fosterella weddelliana Fosterella albicans 0,91 58/66 Fosterella villosula 66/66 Fosterella penduliflora Fosterella spectabilis 0,98 Pitcairnia heterophylla 0,94 Pitcairnia pungens 62/64 83/94 Pitcairnia fuertesii Pitcairnia pipenbringii Pitcairnia feliciana Phylogeny and evolution of Dyckia (Bromeliaceae) b Fig. 4 Bayesian analysis of phyC sequences from of all pitcairnioid samples and two accessions of Puya using MrBayes 3.2.1. The tree was rooted by midpoint rooting. Posterior probabilities are given above, bootstrap values from 1,000 replicates [maximum parsimony (PAUP* 4.0b)/maximum likelihood (RAxML 7.2.6)] below branches. Branch lengths reflect the estimated number of base substitutions as indicated on the bar below Discussion Phylogenetic relationships suggested by plastid DNA trees Phylogenetic relationships among the pitcairnioid genera derived from our six-locus plastid data set are largely congruent with results from earlier studies and support the monophyly of Pitcairnioideae s.str. (Terry et al. 1997; Horres et al. 2000; Crayn et al. 2004; Rex et al. 2009; Givnish et al. 2007, 2011). Pitcairnia is the earliest branching genus within the subfamily, but is deeply split into two separate lineages that apparently diverged more than 10 Mya (Fig. 2). In most of our analyses, Pitcairnia is paraphyletic with generally low support levels (BI PP = 0.74, MPBS = 78, MLBS = 92), but the three branches leading to the Pitcairnia lineages and the remainder of the subfamily collapse into a tritomy in consensus trees (Fig. 2). The status of Pitcairnia was also ambiguous in previous investigations. The genus was paraphyletic in the trnL tree of Horres et al. (2000) as well as in the four-locus plastid tree of Rex et al. (2009), whereas Givnish et al. (2011) identified the genus as being monophyletic in their eight-locus plastid tree. However, support values were weak, and the number of putative synapomorphies defining a monophyletic Pitcairnia was small also in the study of Givnish et al. (2011). The consensus from all plastid DNA-based studies published so far, including the present one, suggests that Pitcairnioideae s.str. experienced two early divergence events within a short time period, one separating Pitcairnia from the rest of the subfamily, and one separating the two principal Pitcairnia lineages from each other. A much better sampling of Pitcairnia species will be required for disentangling the relationships between these three major clades, which may perhaps not be achievable by plastid data alone. The positions of Fosterella, Deuterocohnia and Dyckia/ Encholirium in the plastid tree also match previous findings. The mesophytic genus Fosterella is clearly monophyletic and became separated from the xeric clade, composed of the latter three genera, around 10.5 Mya. Deuterocohnia is deeply paraphyletic, as was also found in all previous plastid phylogenies with sufficient sampling (Horres et al. 2000; Crayn et al. 2004; Rex et al. 2009; Schütz 2012). One of the two separate Deuterocohnia 1609 lineages branched off 8.5 Mya (as in the study of Givnish et al. 2011), the second one 6.2 Mya (Fig. 2). The paraphyly of Deuterocohnia could be explained by an early chloroplast capture event, resulting in the introgression of a Dyckia/Encholirium-type chloroplast into one lineage of Deuterocohnia (Schütz 2012). A well-supported monophylum is formed by Dyckia plus Encholirium, which started to diversify around 5.6 Mya and also share a number of leaf anatomical synapomorphies (Santos-Silva et al. 2013). The relationship of Dyckia and Encholirium remained ambiguous in all previous molecular systematic studies due to limited sampling (e.g. only two Encholirium accessions were included in the work of Givnish et al. 2011). In our plastid trees, Dyckia comes out as a monophyletic clade that arose from within Encholirium around 4.6–4 Mya, thereby rendering the latter genus paraphyletic. The paraphyletic state of Encholirium is also in accordance with essential morphological characters as has been stated before by Forzza (2005). Thus, Encholirium and Deuterocohnia are both characterized by non-connate stamens and terminal inflorescences, which presumably represent the ancestral states of these two characters. On the contrary, in Dyckia stamens are connate, and inflorescences are inserted laterally. Although we sequenced more than 6,000 nucleotides of plastid DNA, infrageneric relationships within Dyckia remain largely unresolved. On the one hand, this lack of phylogenetic resolution is a consequence of limited cpDNA sequence variation in this very young genus. On the other hand, resolution may have been further reduced by hybridization and introgression events, a view that is also supported by our phyC results (see below). Weakly resolved trees or clades were also found in other speciesrich and fast-evolving bromeliad genera (e.g. Tillandsia; Barfuss et al. 2005; Puya; Jabaily and Sytsma 2010). Our BEAST analysis suggests that the Dyckia stem group arose from within Encholirium approximately 4.6 Mya, i.e. in the Pliocene, and diversified mainly in the Pleistocene. Except for two early branching lineages from central Brazil, the vast majority of species, which we here coin ‘‘core Dyckia’’, form a large polytomy, consisting of five reasonably well-supported clades together with numerous single accessions (Fig. 2). These groups appear to be arranged according to geographic criteria rather than following species assignment, a phenomenon which will be discussed below. Network analyses and the spatio-temporal evolution of Dyckia plastids We also determined plastid haplotypes and arranged these into a parsimony network (Clement et al. 2000; Fig. 3). 123 1610 Similar approaches were taken also in other genera where phylogenetic trees remained unresolved, and the resulting networks often allowed additional insights especially into biogeographical patterns (e.g. Bänfer et al. 2006). In our study, the plastid haplotype network showed a much better resolution than the corresponding phylogenetic tree. Its topology supports the paraphyly of Encholirium as well as the monophyly of Dyckia. Whereas all plastid haplotypes found in Encholirium were separated from each other by numerous mutational steps, all members of core Dyckia can be traced back to a single origin, and proved to be arranged in a star-like pattern that largely reflects their geographical distribution (Fig. 3). The central haplotype is found in plants from southern Brazil. Somewhat disturbing is the observation that the same haplotype is sometimes found in more than one species. This is especially true for the centres of the stars. On the other hand, some species carry more than one plastid haplotype, and these are sometimes only distantly related. Especially plastids from widely distributed species like Dyckia remotiflora and Dyckia leptostachya are found in distant positions within the phylogeny. Apparently, the plastid trees and networks do not reflect the species assignment of the investigated Dyckia samples. In part, e.g. for D. remotiflora, cryptic species or misidentifications could be responsible for some unexpected results. However, a geographic rather than taxonomic pattern of plastid haplotype distribution has been encountered in numerous other genera, including Eucalyptus (McKinnon et al. 2001), Hordeum (Jakob and Blattner 2006) and Macaranga (Bänfer et al. 2006), to name just a few. It has also been found in the close relative Deuterocohnia (Schütz 2012). Such observations are commonly explained by hybridization and introgression events, and/or by incomplete sorting of ancient plastid lineages. Our chloroplast network provides some new insights into the historical biogeography of Dyckia (see Figs. 2, 3, 5a–f). In accordance with Givnish et al. (2011), the ancestor of Dyckia and Encholirium presumably became separated from Deuterocohnia between 8.5 and 5.6 Mya, migrated from the Bolivian Andes to the east and started to proliferate in Brazil between 5.6 and 4.6 Mya (Fig. 5a, b). But what exactly happened after this? It was so far an open question whether Dyckia and/or Encholirium are paraphyletic, or each monophyletic and sister to each other. Given that Dyckia is the much larger genus, with a distributional range that overlaps that of Deuterocohnia in the west and almost fully engulfs that of Encholirium in the east, one would intuitively assume that Dyckia were the paraphyletic ancestor of Encholirium. Based on morphological characters, Forzza (2005), however, suspected that the opposite is the case: i.e. Dyckia is a monophyletic group that arose 123 F. Krapp et al. from within Encholirium. Our molecular study clearly supports this view. According to our current model illustrated in Fig. 5c–f, Encholirium and Dyckia separated from each other between 4.6 and 4.1 Mya, most likely within the area of current co-distribution of the two genera close to the eastern cape of Brazil. The occurrence of a 4.0-Mya-old early branching plastid lineage found in D. beateae from Mato Grosso do Sul witnesses Dyckia’s migration towards the south after its separation from Encholirium and remnants of early Dyckia lineages are still found in central Brazil. In the network (Fig. 3) the central haplotype is found in D. reitzii and D. jonesiana from the southernmost Brazilian state Rio Grande do Sul. It is therefore conceivable that the ancestor of the core Dyckia group arrived in southern Brazil around 3.4–2.9 Mya (Fig. 5c), from where a massive radiation and successive dispersal took place into all directions (Fig. 5d). Around 2.5 Mya, secondary radiations began and gave rise to four major clades that now inhabit distinct areas (Fig. 5e). Particularly interesting is the Central Brazilian clade. The pattern within this group is also star-like, with the central haplotype found in five different species. This indicates that another massive radiation could have taken place during the conquest of this area. Central Brazil and especially the states of Bahia and Minas Gerais are nowadays the centres of diversity for Dyckia, with more than 40 species coexisting in a relatively small area, many of them as narrow endemics (e.g. Versieux and Wendt 2007). The lineage made up of plants mostly from Paraguay also shows a starlike pattern but with less diversity. Diversification is probably still ongoing, but the geographic arrangement of plastid haplotypes indicates that only few chloroplast lineages were able to migrate to other areas by long-distance dispersal (arrows in Fig. 5f). Assuming a maternal inheritance of plastids in Dyckia as in most other angiosperms, a strong geographic pattern of plastid haplotypes would implicate week seed dispersal abilities. Climatic changes and the evolution of Dyckia The separation of Deuterocohnia from the common ancestor of Dyckia and Encholirium took place in the Pliocene, when the Andes had almost reached their present altitude and the climate became cooler and dryer (Ghosh et al. 2006; Ehlers and Poulsen 2009). This may have promoted the expansion of dry areas and the appearance of arid corridors or at least stepping-stones for xerophytic plants, including the ancestor(s) of Dyckia and Encholirium that may have used such corridors for their migration to the eastern Brazilian mountain ranges. Phylogeny and evolution of Dyckia (Bromeliaceae) 1611 A 8.5-5.6 Mya B 5.6-4.6 Mya C 4.6-2.9 Mya D 2.9-2.5 Mya E 2.5 Mya-today F 2.5 Mya-today A-D: Deuterocohnia Encholirium Dyckia E-F: major clades Fig. 5 Current hypothesis for the historical biogeography of Dyckia. a–d Assumed distributional patterns of the three xeric genera Dyckia, Encholirium and Deuterocohnia at different time intervals. e– f Geographical pattern of four major clades within core Dyckia as it occurs today. Asterisks indicate cores of secondary radiations. Arrows indicate presumed dispersal/migration events The vast majority of extant Dyckia lineages obviously arose during the Pleistocene. This period was characterized by strong oscillations of global temperature, which in turn caused continuous changes of environmental conditions (Lisiecki and Raymo 2005). Antonelli et al. (2010) suggested that climatic and environmental changes could have had a strong impact on speciation processes of xerophytic plants, exemplified by the orchid genus Hoffmannseggella which grows in rocky areas of the Brazilian Campos Rupestres. The areas occupied by Campos Rupestres are nowadays mostly restricted to altitudes above 1,000 m and are therefore strongly fragmented. Their island-like occurrence presumably limits gene flow between populations of xeric plants, and may therefore favour allopatric speciation. Antonelli et al. (2010) hypothesized that during cooler and drier periods the lower altitudinal limits of Campos Rupestres may have shifted down to 800 m, resulting in the connection of previously separated habitat islands by numerous arid corridors. Alternating periods of isolation and fusion of suitable habitats could have triggered allopatric speciation not only in Hoffmannseggella, but could also be one of the major forces driving the speciation in Dyckia and Encholirium that inhabit the same environments. A similar mechanism of alternating expansion and contractions of forests separating suitable habitats for xerophytes was proposed for lower elevations, where Dyckia species also occur (Winkler 1986; Versieux and Wendt 2007). The above hypothesis does, however, not explain why the number of species described for Dyckia is much higher than that observed for Encholirium, especially since molecular data suggest that Dyckia is the younger genus. For example, in Minas Gerais, the diversity centre of both genera, Dyckia shows a threefold higher species number than Encholirium (Versieux and Wendt 2006, 2007). There is still no convincing explanation for this phenomenon. On the one hand, Dyckia tends to grow on slightly higher elevations than Encholirium and also reaches further south, which may have caused a stronger insulation of suitable habitats in Dyckia. It is also possible that extinction rates 123 1612 are higher in Encholirium than in Dyckia, and that Encholirium species, maybe due to their higher specialization (e.g. bat pollination vs. insect/bird pollination; Sazima et al. 1989), have a poorer capacity to colonize new habitats/areas. Finally, the large number of described Dyckia species may also in part result from difficulties of species delimitation (Smith 1934) and the taxonomist’s desire to split and describe new species. Phylogenetic utility of the phyC gene Most phylogenies for Bromeliaceae have been founded on plastid DNA variation, and only few nuclear markers have so far been tried with variable success. For example, Schulte et al. (2009) successfully used the phosphoribulokinase (PRK) marker for phylogenetic reconstruction in Bromelioideae, whereas Jabaily and Sytsma (2010) applied the phytochrome C (phyC) marker for generating a phylogeny of Puya. On the contrary, the nuclear ribosomal internal transcribed spacer (ITS) proved to be experimentally reluctant and of little informativeness (Barfuss 2012). Our previous attempts to use the PRK marker for phylogenetic reconstruction in Pitcairnioideae were of limited success (Schütz 2012), and we therefore decided to use the phyC marker which indeed provided a much higher level of variation than the plastid markers (5.7 % within Dyckia). On the negative side, we also observed an unusually high incidence of intra-individual allelic variation that interfered with phylogenetic reconstructions. The phylogenies derived from nuclear phyC data contradict the plastid data in many ways. First of all, the alleles of this gene found in the Pitcairnioideae do not form a monophylum, because alleles found in Puya species clearly group within the subfamily Pitcairnioideae. The reasons for this are unclear, but it may be assumed that Puyoideae and Pitcairnioideae exchanged at least some genetic material long after their split. Other obvious differences between phyC and plastid phylogenies concern the status of Deuterocohnia and Pitcairnia, which are monophyletic in the phyC tree but paraphyletic in the plastid trees. Also interesting and against the odds is the, although weakly supported, monophyletic group formed by the phyC alleles found in Deuterocohnia and Fosterella. These findings are more thoroughly discussed in a separate manuscript focusing on subfamilial relationships with a broader sampling (Schütz, Krapp, Wagner and Weising, in preparation). Regarding Dyckia and Encholirium, the paraphyly of the latter is supported by the phyC tree. However, the strikingly low resolution in spite of the relatively high variability of phyC is a strong evidence for a large amount of contradictions in the data set. In part, this may be caused by hybrids (although rarely reported from the field) and could perhaps have been avoided by cloning before sequencing. 123 F. Krapp et al. Despite the origin of our samples mostly from living collections, the inclusion of artificial hybrids is quite unlikely, because Dyckia plants are usually propagated vegetatively in Botanical Gardens (Timm Stolten, pers. comm.) and seeds are regularly discarded in Leme’s collection. Moreover, removing all accessions that show high levels of heterozygosity did not improve tree resolution. This could be explained by different factors. First, the extent of homoplasy may be higher for phyC than for the plastid data set because of the higher variability in combination with the focus of substitutions on third codon positions. This may have led to a partial saturation. Second, the extant alleles of phyC found in Dyckia may be derived from a set of alleles that was already present in the most recent common ancestors. Third, ancient hybridization and recombination events may have contributed to the more or less random distribution of alleles among the genus. The observed phylogenies of plastids on the one hand and of the nuclear gene phyC on the other hand tell rather different stories about the evolution of Dyckia. The plastid lineages apparently reflect the geographical distribution of their host plants, and are most likely inherited via seeds. Species of Dyckia and Encholirium distribute their seeds by wind, which is comparatively ineffective for long-distance dispersal (e.g. Versieux and Wendt 2006). Once the plants established a stand in a certain region, it is very unlikely that plastids from elsewhere become dominant there. This means that the plastids were distributed during a certain time period in the past and stayed there till today, with few exceptions. In contrast, phyC alleles are also inherited and dispersed via pollen, and therefore their mobility may be disproportionately higher. Even large distances between populations and species may be bridged by hummingbirds and flying insects regularly. Major incongruencies between a plastid and a phyC phylogeny were also found for the genus Puya (Jabaily and Sytsma 2010), but the genetic admixture between different clades was not that high in their study, and was explainable by assuming only a limited number of hybridization and backcrossing events. Species boundaries and species concepts in Dyckia Excoffier et al. (2009) simulated the effect of range expansions on the genetic composition of species. They observed unexpectedly high degrees of introgression from local plants into the gene pool of species invading this area. In the case of Dyckia the recurrent isolation and fusion of suitable habitats expectedly lead to frequent secondary contacts of formerly isolated species or populations. Many Dyckia species are capable of spawning fertile interspecific hybrids, at least in cultivation (Timm Stolten, pers. comm.). Secondary contacts may therefore have promoted Phylogeny and evolution of Dyckia (Bromeliaceae) some genetic admixture, which according to our molecular data can be considerable. Of the 19 Dyckia species where two or more accessions were included in our analysis, only two carried plastids of monophyletic origin. While in eight cases there was a lack of resolution, nine species carried plastids of clearly paraphyletic origin. The same situation was observed for phyC, where only the two varieties of Dyckia marnier-lapostollei carried alleles of monophyletic origin, whereas 14 species carried phyC alleles of paraphyletic origin. In four cases the resolution did not allow a statement on this issue. Even if we cannot rule out occasional misidentification of plant material, these findings strongly support the hypothesis of poor reproductive isolation between the Dyckia species. The introgression of genes that have an impact on the phenotype could at least in part be responsible for the difficult delimitation of many Dyckia species. What is referred to as plasticity of morphological characters may reflect the actual genetic situation. Nevertheless, and in spite of the observed genetic admixture, most Dyckia morphospecies seem to be more or less cohesive. This may be explained by the idea of porous genetic isolation barriers (Wu 2001; Morjan and Rieseberg 2004; Wu and Ting 2004). According to this concept, only a set of co-adapted alleles is responsible for the capability of a plant to compete successfully in a certain environment. The remainder of the genome, possibly also including the plastids is exchangeable via hybridization and backcrossing, unless there is selection. Comparable observations were recently made for sympatric species of Pitcairnia growing on inselbergs around Rio de Janeiro (Palma-Silva et al. 2011). The Cerrado biome shows an enormous diversity of pollinators, soils and habitats in general. So the assumption of manifold adaptations to narrow ecological niches for Dyckia species is quite likely. In conclusion, many Dyckia species may still exchange genetic material but still remain cohesive. One potential strategy to infer species boundaries in such a difficult situation is to use a set of more sensitive markers like nuclear microsatellites (e.g. Caddah et al. 2013). Studies in this direction are underway in our laboratory (Wöhrmann et al. 2013) and will hopefully shed more light on the question whether Dyckia morphospecies are not only morphologically, but also genetically distinct from each other. Acknowledgments The authors thank J. Peters, N. Schütz and the Botanical Gardens of Heidelberg, Bonn, Berlin-Dahlem, Marburg and Vienna for providing plant material as well as R.B. Louzada, G. Cruz and A. M. Wanderley for help during fieldwork. F. Krapp and D. Pinangé are supported by PhD fellowship grants of the Otto-BraunFonds (Melsungen) and the Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco (FACEPE), respectively. This work was supported by DAAD (German Academic Exchange Service), by CAPES (Brazilian Coordination for the Improvement of Higher Education Personnel) in the frame of the PROBRAL and PNADB 1613 programs, an also by CNPq (Brazilian Counsel of Technological and Scientific Development). References Antonelli A, Verola CF, Parisod C, Gustafsson ALS (2010) Climate cooling promoted the expansion and radiation of a threatened group of South American orchids (Epidendroideae: Laeliinae). Bot J Linn Soc 100:597–607 Baker JG (1889) Handbook of the Bromeliaceae. George Bell & Sons, London Bänfer G, Moog U, Fiala B, Mohamed M, Weising K, Blattner FR (2006) A chloroplast genealogy of myrmecophytic Macaranga species (Euphorbiaceae) in Southeast Asia reveals hybridization, vicariance and long-distance dispersals. Mol Ecol 15:4409–4424 Barfuss MHJ (2012) Molecular studies in Bromeliaceae: implications of plastid and nuclear DNA markers for phylogeny, biogeography, and character evolution with emphasis on a new classification of Tillandsioideae. PhD Thesis, University of Vienna Barfuss MHJ, Samuel R, Till W, Stuessy T (2005) Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on DNA sequence data from seven plastid regions. Amer J Bot 92:337–351 Barthlott W, Porembski S, Szarzynzki J, Mund P (1993) Phytogeography and vegetation of tropical inselbergs. In: Guillaumet J-L, Belin M, Puig H (eds) Phytogéographie tropicale. Réalités et perspectives. Colloques & Seminaires ORSTOM, Paris, pp 15–24 Benzing DH (2000) Bromeliaceae: profile of an adaptive radiation. Cambridge University Press, New York Bernadello LM, Galetto L, Juliani HR (1991) Floral nectar, nectary structure and pollinators in some Argentinean Bromeliaceae. Ann Bot 67:401–411 Bjourson AJ, Cooper JE (1992) Band-stab PCR: a simple technique for the purification of individual PCR products. Nucleic Acids Res 20(17):4675 Caddah MK, Campos T, Zucchi MI, Pereira de Souza A, Bittrich V, do Amaral MCE (2013) Species boundaries inferred from microsatellite markers in the Kielmeyera coriacea complex (Calophyllaceae) and evidence of asymmetric hybridization. Plant Syst Evol 299:731–741 Clement M, Posada D, Crandall KA (2000) TCS: a computer program to estimate gene genealogies. Mol Ecol 9(10):1657–1660 Crayn DM, Terry RG, Smith JAC, Winter K (2000) Molecular systematic investigations in Pitcairnioideae (Bromeliaceae) as a basis for understanding the evolution of crassulacean acid metabolism (CAM). In: Wilson KL, Morrison DA (eds) Monocots: systematics and evolution. CSIRO, Melbourne, pp 569–579 Crayn DM, Winter K, Smith JAC (2004) Multiple origins of crassulacean acid metabolism and the epiphytic habit in the neotropical family Bromeliaceae. Proc Natl Acad Sci USA 101:3703–3708 Demesure B, Sodzi N, Petit RJ (1995) A set of universal primers for amplification of polymorphic noncoding regions of mitochondrial and chloroplast DNA in plants. Mol Ecol 4(1):129–131 Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 7:214 Ehlers TA, Poulsen CJ (2009) Influence of Andean uplift on climate and paleoaltimetry estimates. Earth Planet Sc Lett 281:238–248 Excoffier L, Foll M, Petit RJ (2009) Genetic consequences of range expansions. Annu Rev Ecol Evol Syst 40:481–501 Forzza RC (2005) Revisão taxonômica de Encholirium Mart. ex. Schult. & Schult. f. (Pitcairnioideae––Bromeliaceae). Bol Bot Univ São Paulo 23(1):1–49 123 1614 Ghosh P, Garzione CN, Eiler JM (2006) Rapid uplift of the altiplano revealed through 13C–18O bonds in paleosol carbonates. Science 311:511–515 Givnish TJ, Millam KC, Berry PE, Sytsma KJ (2007) Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data. Aliso 23:3–26 Givnish TJ, Barfuss MHJ, Ee BV, Riina R, Schulte K, Horres R, Gonsiska PA, Jabaily RS, Crayn DM, Smith JAC, Winter K, Brown GK, Evans TM, Holst BK, Luther H, Till W, Zizka G, Berry PE, Sytsma KJ (2011) Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: insights from an eightlocus plastid phylogeny. Am J Bot 98(5):872–895 Hmeljevski KV, Reis A, Montagna T, dos Reis MS (2011) Genetic diversity, genetic drift and mixed mating system in small subpopulations of Dyckia ibiramensis, a rare endemic bromeliad from Southern Brazil. Conserv Genet 12:761–769 Horres R, Zizka G, Kahl G, Weising K (2000) Molecular phylogenetics of Bromeliaceae: evidence from trnL (UAA) intron sequences of the chloroplast genome. Plant Biol 2:306–315 Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17(8):754–755 Jabaily RS, Sytsma KJ (2010) Phylogenetics of Puya (Bromeliaceae): placement, major lineages, and evolution of Chilean species. Am J Bot 97(2):337–356 Jakob SS, Blattner FR (2006) A chloroplast genealogy of Hordeum (Poaceae): long-term persisting haplotypes, incomplete lineage sorting, regional extinction, and the consequences for phylogenetic inference. Mol Biol Evol 23(8):1602–1612 Krapp F (2013) Phylogenie und Evolution der Gattung Dyckia (Bromeliaceae). PhD thesis, University of Kassel Leme EMC, Ribeiro OBC, Miranda ZJG (2012) New species of Dyckia (Bromeliaceae) from Brazil. Phytotaxa 67:9–37 Lisiecki LE, Raymo ME (2005) A Pliocene–Pleistocene stack of 57 globally distributed benthic d18O records. Paleoceanography 20:PA1003 (1–17) Luther HE (2012) An alphabetical list of bromeliad binominals, 13th edn. Marie Selby Botanical Gardens, and The Bromeliad Society International Martin CE (1994) Physiological ecology of the Bromeliaceae. Bot Rev 60:1–80 McKinnon GE, Vaillancourt RE, Jackson HD, Potts BM (2001) Chloroplast sharing in the Tasmanian eucalyptus. Evolution 55:703–711 Morjan CL, Rieseberg LH (2004) How species evolve collectively: implications of gene flow and selection for the spread of advantageous alleles. Mol Ecol 13:1341–1356 Müller J, Müller K, Quandt D (2011) PhyDE––Phylogenetic data editor. Version 0.9971. Program distributed by the author Nylander JAA (2004) MrModeltest v2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University Oxelman B, Lidén M, Berglund D (1997) Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Plant Syst Evol 206:393–410 Palma-Silva C, Wendt T, Pinheiro F, Barbará T, Fay MF, Cozzolino S, Lexer C (2011) Sympatric bromeliad species (Pitcairnia spp.) facilitate tests of mechanisms involved in species cohesion and reproductive isolation in Neotropical inselbergs. Mol Ecol 20:3185–3201 Peters J (2009) Revision of the genus Fosterella L. B. Sm. Bromeliaceae. PhD thesis, University of Kassel Rambaut A, Drummond AJ (2009) Tracer 1.5. Program distributed by the author Rex M, Schulte K, Zizka G, Peters J, Vásquez R, Ibisch PL, Weising K (2009) Phylogenetic analysis of Fosterella L. B. Sm. (Pitcairnioideae, Bromeliaceae) based on four chloroplast DNA regions. Mol Phylogenet Evol 51:472–485 123 F. Krapp et al. Santos-Silva F, Saraiva DP, Monteiro RF, Pita P, Mantovani A, Forzza RC (2013) Invasion of the South American dry diagonal: what can the leaf anatomy of Pitcairnioideae (Bromeliaceae) tell us about it? Flora 208:508–521 Sazima I, Vogen S, Sazima M (1989) Bat pollination of Encholirium glaziovii, a terrestrial bromeliad. Plant Syst Evol 168:167–179 Schulte K, Horres R, Zizka G (2005) Molecular phylogeny of Bromelioideae and its implications on biogeography and the evolution of CAM in the family (Poales, Bromeliaceae). Senckenb Biol 85:113–125 Schulte K, Barfuss MHJ, Zizka G (2009) Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily. Mol Phylogenet Evol 51:327–339 Schütz N (2012) Systematics, morphology and taxonomy of the genus Deuterocohnia L. B. Sm. Bromeliaceae. PhD thesis, University of Kassel Shaw J, Lickey EB, Schilling EE, Small RL (2007) Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: the tortoise and the hare III. Am J Bot 94:275–288 Smith LB (1934) Geographical evidence on the lines of evolution in the Bromeliaceae. Bot Jahrbuch 66:446–468 Smith LB (1967) Notes on Bromeliaceae. Phytologia 14(8):457–491 Smith LB, Downs RJ (1974) Pitcairnioideae (Bromeliaceae). Fl Neotrop Monogr 14 Part 1:1–662 Stamatakis A (2006) RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22(21):2688–2690 Swofford DL (2003) PAUP*. Phylogenetic analysis using parsimony (*and other methods). Version 4.0 Beta 10. Sinauer Associations, Sunderland, Massachusetts Tel-Zur N, Abbo S, Myslabodski D, Mizrahi Y (1999) Modified CTAB procedure for DNA isolation from epiphytic cacti of the genera Hylocereus and Selenicereus (Cactaceae). Plant Mol Biol Rep 17:249–254 Terry RG, Brown GK, Olmstead RG (1997) Examination of subfamilial phylogeny in Bromeliaceae using comparative sequencing of the plastid locus ndhF. Am J Bot 84:664–670 Versieux LM, Wendt T (2006) Checklist of Bromeliaceae of Minas Gerais, Brazil, with notes on taxonomy and endemism. Selbyana 27(2):107–146 Versieux LM, Wendt T (2007) Bromeliaceae diversity and conservation in Minas Gerais state. Brazil Biodivers Conserv 16:2989–3009 Versieux LM, Wendt T, Louzada RB, Wanderley MGL (2008) Bromeliaceae da Cadeia do Espinhaço. Megadiversidade 4:98–110 Vesprini JL, Galetto L, Bernadello G (2003) The beneficial effect of ants on the reproductive success of Dyckia floribunda (Bromeliaceae), an extrafloral nectary plant. Can J Bot 81:24–27 Watts CD, Fisher AE, Shrum CD, Newbold WL, Hansen S, Liu C, Kelchner SA (2008) The D4 set: primers that target highly variable intron loops in plant chloroplast genomes. Mol Ecol Resour 8:1344–1347 Winkler S (1980) Ursachen der Verbreitungsmuster einiger Bromeliaceae in Rio Grande do Sul (Südbrasilien). Flora 170:371–393 Winkler S (1986) Differenzierungen und deren Ursachen innerhalb der Bromeliaceen. Beitr Biol Pflanzen 61:283–314 Wöhrmann T, Pinangé D, Krapp F, Benko-Iseppon AM, Huettel B, Weising K (2013) Development of 15 nuclear microsatellite markers in the genus Dyckia (Pitcairnioideae; Bromeliaceae). Conservation Genet Resour 5:81–84 Wu C-I (2001) The genic view of the process of speciation. J Evol Biol 14:851–865 Wu C-I, Ting C-T (2004) Genes and speciation. Nat Rev Genet 5:114–122

RELATED PAPERS

RELATED TOPICS

Log In

Phylogeny and evolution of Dyckia (Bromeliaceae) inferred from chloroplast and nuclear sequences

Phylogeny and evolution of Dyckia (Bromeliaceae) inferred from chloroplast and nuclear sequences

Related Papers

RELATED PAPERS

RELATED TOPICS