P. 1
2007 Lucas Myrtaceae Filogenia

2007 Lucas Myrtaceae Filogenia

|Views: 6|Likes:

More info:

Published by: Raquel Moura Machado on Sep 17, 2013
Copyright:Attribution Non-commercial


Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less





TAXON 56 (4) • November 2007: 1105–1128

Lucas & al. • Suprageneric phylogenetics of Myrteae

Suprageneric phylogenetics of Myrteae, the generically richest tribe in Myrtaceae (Myrtales)
Eve J. Lucas1, Stephen A. Harris2, Fiorella F. Mazine3, Stephen R. Belsham4, Eimear M. Nic Lughadha1, Annika Telford1, Peter E. Gasson1 & Mark W. Chase1

Royal Botanic Gardens, Kew, Richmond, Surrey, TW9 3AE, U.K. e.lucas@kew.org (author for correspondence) 2 University of Oxford, South Parks Road, Oxford OX1 3RB, England, U.K. 3 Herbário, Departamento de Botânica, Escola Superior de Agricultura Luiz de Queiroz, Universidade de São Paulo, Caixa Postal 9, 13418-900 Piracicaba, São Paulo, Brazil 4 Botany Department, University of Otago, P.O. Box 56, Dunedin 9001, New Zealand To clarify relationships within the predominantly Neotropical and exclusively fleshy-fruited Myrteae (49 genera and c. 2,500 species), we provide a phylogenetic hypothesis for evolutionary relationships between 31 of these genera by analyzing nuclear ITS and ETS ribosomal DNA, and plastid psbA-trnH and matK DNA sequences from 75 Myrteae species and 13 outgroup taxa using parsimony and Bayesian inference. Four morphological characters are epitomized on the resulting trees, and biogeographical analyses are also performed. Myrteae are monophyletic, comprising seven clades plus two isolated taxa of unclear relationships. Morphological characters exhibit homoplasy, although in combination are useful for clade diagnosis. Biogeographical analyses are inconclusive regarding the ancestral area of the tribe, but South American colonization before northern radiation via the Andes appears likely. The largest genera, Eugenia and Myrcia s.l., have western and southeastern South American origins, respectively.

KEYWORDS: ETS, Eugeniinae, ITS, matK taxonomy, morphology, Myrciinae, Myrtaceae, Myrteae,
Myrtinae, psbA-trnH

Tribe Myrteae DC. (sensu Wilson & al., 2005; Myrtaceae) are trees or occasionally shrubs with a Pantropical distribution. They comprise 49 genera and c. 2,500 species (World Checklist of Myrtaceae, 2006), including the large genera Eugenia and Myrcia (estimated at > 550 and 250–500 species, respectively; Mabberley, 1997). The tribe is distinguished from the 14 other tribes in subfamily Myrtoideae (Wilson & al., 2005) by their indehiscent fleshy fruits, transeptal vascular system, and, commonly, presence of uni- or multi-cellular hairs (Schmid, 1972; Wilson & al., 2001). Previous classifications of Myrtaceae (De Candolle, 1828; Berg, 1855–56, 1857–59) used subfamily Myrtoideae and tribe Myrteae interchangeably; however, following reclassification by Wilson & al. (2005), the two ranks have different circumscriptions and are no longer considered synonymous. The most enduring subfamilial and subtribal classifications of Myrtaceae and Myrteae are reviewed in Lucas & al. (2005), with additional interpretations discussed here (Tables 1, 2). The highest species diversity in Myrteae occurs in South America, particularly along the eastern coast of Brazil, the Guayana Highlands, and the Caribbean (McVaugh, 1968).

De Candolle’s (1826) concept of Myrtaceae encompassed five tribes (Schlechtendahl, 1827): Leptospermeae, Myrteae, Chamaelaucieae, Barringtonieae, and Lecythideae. This concept of Myrtaceae was narrowed by Niedenzu (1893) into two subfamilies; a mostly Neotropical fleshy-fruited Myrtoideae, consisting of tribe Myrteae, and an almost exclusively Paleotropical, dry, capsularfruited Leptospermoideae, consisting of tribes Leptospermeae and Chamaelaucieae. Niedenzu also transferred tribes Barringtonieae and Lecythideae to Lecythidaceae, a family now known to be only distantly related to Myrtaceae (in Ericales; Morton & al., 1996; APG, 2003). Until recently, most authors recognized Niedenzu’s subfamilies, although not always at this rank; Schmid (1980) reinstated Chamelaucioideae (Table 1), whereas Kausel (1956, 1966) reclassified Myrteae and Leptospermeae as families Myrtaceae and Leptospermaceae. Myrteae. — The most commonly recognized previous classifications of Myrteae are the three, embryologybased subtribes of Berg (1855–56, 1857–59) and the six “informal groupings” of McVaugh (1968; Table 2; Fig. 1). Kausel (1956, 1966) divided Myrtaceae (Myrteae sensu Wilson & al., 2005) into two groups of five subfamilies based on patterns of seed germination: those with hypogeal germination, Eugenioideae (Eugenia, Stenocalyx,

Lucas & al. • Suprageneric phylogenetics of Myrteae

TAXON 56 (4) • November 2007: 1105–1128

Table 1. Subfamilial groupings of Myrtaceae according to different authors. Taxa associated with subfamilies Myrtoideae, Leptospermoideae and Psiloxyloideae are indicated by light, intermediate and dark shading respectively.

De Candolle ex Schlechtendahl (1826–7) Tribe Leptospermeae Tribe Chamaelaucieae Tribe Myrteae Tribe Barringtonieae Tribe Lecythideae

Niedenzu (1893) Subfamily Leptospermoideae

Kausel (1956) Family Leptospermaceae

Johnson & Briggs (1984) Subfamily Leptospermoideae

Schmid (1980) Subfamily Leptospermoideae Subfamily Chamaelaucoideae

Wilson & al. (2005)

Subfamily Myrtoideae Family Lecythidaceae (excluded from Myrtaceae)

Family Myrtaceae

Subfamily Myrtoideae

Subfamily Myrtoideaea

Subfamily Myrtoideaeb

Subfamily Psiloxyloideae

Subfamily Psiloxyloideae

Including 3 groups: Osbornia


Including 15 tribes: Eucalypteae, Kanieae, Leptospermeae, Lindsayomyrteae, Lophostemoneae, Melaleuceae, Metrosidereae, Osbornieae, Syncarpieae, Tristanieae, Xanthostemoneae, Chamelaucieae, Backhousieae Myrteae, Syzygieae

“Myrtoideae sensu stricto” The “Acmena” alliance

Fig. 1. Embryo variation in tribe Myrteae. A, myrcioid embryo sensu Berg (Myrcia); B, eugenioid embryo sensu Berg (Myrcianthes); C, embryo of the Campomanesia complex sensu Landrum (Campomanesia); D, myrtoid embryo sensu Berg (Psidium); E, Luma; F, Lagrizea; G, Blepharocalyx cruckshanksii; H, B. salicifolius. Labels: c, cotyledon; h, hypocotyl. Reprinted by permission; A–D from Brittonia 49(4) 508–536, M.L. Kawasaki, © 1997, The New York Botanical Garden, Bronx, New York; E, G, H from Fl. Neotrop. 45, B. Angell, © 1986, The New York Botanical Garden, Bronx, New York; F from Syst. Bot. 31(2), P. da Costa e Silva Assumpção, © 2006, The American Society of Plant Taxonomists.


Myrciaria. • Suprageneric phylogenetics of Myrteae Calycorectes. Marlierea. Ugni. The “Myrcioid” genera Group 2. having reduced those four containing Neotropical taxa into a single alliance.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. Myrteola) and Myrcioideae (Aulomyrcia. often containing a visible vascular core that in the past had been mistaken for the radicle and the surrounding cortex interpreted as fused cotyledons in several genera. Myrtus. 1921. (2005). and those Austromyrtus species that Snow & al. Landrum. pers. Pseudomyrcianthes). Myrcianthes and related genera Myrtus Principal group of Subtribe genera associated Pimentoideae with Myrtus (later Myrtinae) Psidium Subfamily Myrtoideae Group 5. intermediate and dark shading respectively. comm. “Myrtoideae sensu stricto” (in their sense: Austromyrtus. Psidium and related genera Group 6. Lophomyrtus. no longer recognizing Myrtoideae. Kausel. Myrtus. Landrum (1988) conceptualized a further group. Myrtastrum. They proposed a “Campomanesia complex” in which the seed coat is usually soft and the hypocotyl swollen (usually much wider than the cotyledons). Sterocaryum. Psidium. whereas Eugenioideae have undivided cotyledons indistinguishable from the hypocotyl. Feijoa). Eugeniinae and Myrtinae are indicated by light. Myrrhinium. Britoa. 1966) Subfamily Myrcioideae Subfamily Plinioideae Subfamily Eugenioideae McVaugh (1968) Group 1. Campomanesia Pseudocaryophylcomplex lus and Pimenta Subfamily Cryptorhizoideae Group 4.R. The “Eugenioid” genera Landrum (1986) Principal group of Subtribe genera associated Myrcioideae with Myrcia (later Myrciinae) Principal group of Subtribe genera associated Eugenioideae with Eugenia (later Eugeniinae) Plinia Eugenia Myrtus Group 3.). swollen cotyledons. Kausel was later shown to be wrong in the case of Myrcianthes which has hypogeal germination (L. the Neotropical genera falling into four of these. Blepharocalyx) versus those with epigeal germination. Campomanesia and related genera Imperfectly known Cryptorhiza Subtribe Myrrhinioideae (later Myrrhinieae) Subtribe Orthostemonoideae (later Feijoinae) Unplaced genera Psidium Campomanesia 1107 . Paramyrciaria) and Cryptorhizoideae (Campomanesia. Eugenia. Eugenia and Cryptorhiza alliances. (2003) later distinguished as Austromyrtus sensu stricto. the Myrcia. such as Cryptorhiza (Urban. Landrum & Stevenson (1986) studied embryo variability in Myrtinae. Neomyrtus. Myrtus. Later. Uromyrtus and Xanthomyrtus). Fenzlia. linking Myrteola. according to different authors. Plinioideae (Myrcianthes. Myrcia. Myrceugenia. Amomyrtella. Myrtella. Linnaeus (1753) Sprengel (1825) Calyptranthes De Candolle (1826) Berg (1855–56. Siphoneugena. 1857–59) Kausel (1956. Briggs & Johnson (1979) divided Myrtoideae into six informal groups. Cryptorhizoideae were erroneously thought by Kausel to have large cotyledons surrounding the hypocotyl. Myrcia. Taxa associated with Berg’s subtribes Myrciinae. Pilothecium. 1956. 1966). Calyptranthes. Myrceugenella. Myrtoideae (Amomyrtus. Johnson & Briggs (1984) erected three fleshy-fruited groups (Table 1). Plinioideae have separate. Acreugenia. Psidium. This association is based on the shared characteristics of C-shaped embryos with cotyledons roughly the Table 2. Groupings of Neotropical Myrteae sensu Wilson & al. Ugni. Meteoromyrtus.

smooth. the generically richest tribe.. (2) to assess. 1974. 50 ma (Sytsma & al. however. Raven & Axelrod (1974) suggested instead that original differentiation between the recognized subfamilies Leptospermoideae and Myrtoideae. such an understanding will also provide insight into historical factors that have influenced the modern distribution of the family and the tribe in South America.. ITS. with species selected to encompass as much morphological diversity and geographical variation as possible. primers used are listed in Table 3. 1857–59) also noted the division of the Myrtinae genera between those with bony and soft testae. The mostly tropical. A new forward primer was designed for ETS and designated ETS MyrtF (5′-CTCCGTGCTGGTGCATCGAACTGC-3′). as are the number of colonization events. — Eighty-eight Myrtaceae accessions were sequenced for the ETS. extending from bp 1. 2002).. when Australia was connected to South America via warm-temperate Antarctic land bridges.1993). 1984. with Myrteae originating and diversifying in Australasia between 77–56 ma.. This section of matK is approximately 750 base pairs (bp) long. MATERIALS AND METHODS Sampling. It was suggested that this was followed by at least two separate dispersal events bringing Eugenia from South America to Africa (van der Merwe & al. The area of origin of Myrteae in South America is unconfirmed. Pigg & al. 2001. Sequencing reactions were carried out with the Taq DyeDeoxy Terminator Cycle Sequencing Kit™ (Applied Biosystems. Sytsma & al. is necessary for prioritization of conservation initiatives. according to the manufacturer’s protocol. 1914.or air-dried leaf material using the CTAB extraction protocol of Doyle & Doyle (1987) but purified by equilibrium centrifugation in CsCl-ethidium bromide gradients (1. Inc). Electropherograms were assembled and edited using Sequencher™ version 4. 2005). C-shaped embryos and parietal placentation.. The fossil genus Paleomyrtinaea (known only from fruit. initial hypotheses (Berry.. ten outgroup genera (13 species) were selected from seven other tribes of Myrtoideae. and Myrtus from an unknown locality to the Mediterranean c. the phylogenetic hypothesis produced by the three-region analysis was used to select a subset of 46 taxa.Lucas & al. 2001. followed by dialysis to remove the caesium chloride. Given the environmental threats to these and other Myrtaceae-rich areas. 2004). hard seed coats opening by means of opercula and uniflorous peduncles in the leaf axils. disjunct distribution of Myrtaceae has created debate over their historical biogeography (Berry. Biogeography.55 g × ml–1) after precipitation in –18°C 100% ethanol.. Sytsma & al. 2004).3 g of fresh.822 relative to the tobacco trnK map (Sugita & al. 1108 has been thought to be most closely related to Psidium or Mosiera) from the late Palaeocene of North Dakota and early Eocene of British Columbia (Pigg & al. 1993. — Myrtaceae are one of the most species-rich families in the Neotropics.. raising the possibility that at least some species arrived in North America via Africa/Asia. Wilson & al. Sequences were read on an ABI 3100 Genetic Analyzer. 1915) were of a fleshy-fruited plesiomorphic state and North American origin for Myrteae. comprising representatives from each clade. silica gel. 2005.. Target regions were amplified in an Applied Biosystems Inc. Johnson & Briggs. and (4) to investigate the biogeographical relationships of the major clades. This has been interpreted to indicate that past Myrteae had a wider distribution. To improve bootstrap support.075 to 1. 0. 2000). • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 same length as the hypocotyl and not folded back over it. Tepualia from Australasia to South America. Raven & Axelrod. took place from ancestral stock that occupied west Gondwana (Myrtoideae) and Australasia (Leptospermoideae) and that modern distributions are mainly the result of early Cretaceous Gondwanan vicariance with long-distance dispersal resulting in disjunct distributions for some taxa. sequencing and alignment. Ethidium bromide was removed with butanol extraction. making up 10 to 15 percent of total tree species in the wet forests and cerrado of eastern Brazil—more than any other family (OliveiraFilho & Fontes. 1915. with probably five locules. This study has four goals: (1) to clarify sister relationships between the (included) genera of Myrteae as currently defined and to determine their monophyly. Berg (1855–56. — Total genomic DNA was extracted from c. McVaugh (1968) concluded that Myrteae initially established in southern South America and subsequently spread throughout the continent. 1985). PCR products were purified using QIAGEN® QIAquick™ Spin Columns.5 (Gene Codes . morphological characters as potential synapomorphies for delimiting groups of genera in the light of these results.. and psbA-trnH studies (Appendix). DNA extraction. an understanding of the evolutionary history of Myrteae. 60 ma (Davis & al. This sampling represents 31 Myrteae genera. Syzygium and Metrosideros from Australasia to Africa. 2004) suggested a Gondwanan origin of Myrtaceae. Based on generic morphology and modern distributions. Phylogenetic analyses of morphological data (Briggs & Johnson 1979.. perhaps via North Atlantic land connections during periods of Arctic warming that began c. for which a portion of matK was also sequenced (Appendix). (ABI) Gene Amp 9700 PCD system (Table 3). (3) to evaluate previously proposed suprageneric classifications of Myrteae. Following putative Myrteae fossils finds in North American Tertiary deposits. although he did not create any formal taxonomy for these groups. Johnson & Briggs 1984) and later DNA sequences (Wilson & al.

1 min at 72° C Corporation). Four such ing to the rescaled consistency indices of the best trees. 1991). Overlapping indels were not included recognized by the presence of inconsistent groupings of as they occurred only in sequence regions where it was impossible to obtain an unambiguous alignment of all taxa with moderate to high bootstrap support ( > 75 BP) in taxa. according to the simple method of Simmonds were analyzed together.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. and parsimony analysis was carried out of sequence divergence and incongruence was not expected from the uniparentally inherited plastid genome. The heuMyrtaceae RNA secondary structure 8 psbA-trnH 892–902 ristic search settings were: 1. Biffin & al. 2 min 30 seconds at 72°C matK 390 (F). Successive approximations weighting was used in combination with these model parameters fitted in(SW. ing the replicates.000 predictions to assess the use of struc9 psbA-trnH 936–958 bootstrap replicates with taxa tural partitioning in ITS.0b2 (Swofford. netic analyses. 30 seconds at 50°C. 1 min at 48°C. All models require six substitution trees for a further search using TBR and branch-swapping types (NST = 6) and a gamma model of among-site rate with MulTrees on.0b4 (Huelsenbeck & Ronquist (Appendix). TVM + G. Heuristic searches consisted of 1. Trees generated were starting trees to collect all shortest plotted against likelihoods in MicrosTable 4. Models of nucleotide substitution were selected for random taxon entry order. with incongruence and presence as 1.000 heterogeneity (Rates = gamma). ETS. for ITS. 1998). Datasets were analyzed trees. Bootstrap support percentages (BP) of 50–74 are considered weak support. 2000). characters were reweighted accorddependently to the separate data partitions. (1994) 2 min at 94°C followed by 30 cycles of 1 min at 94°C.000 replicates using 2001). (2007) use 7 psbA-trnH 858–883 the reweighting scheme. Informative indels scored. Phylogenetic analyses. 1998). with absence of a gap coded as 0 combined was evaluated by inspection. where a majority 3 psbA-trnH 215–220 rule consensus tree was produced disbootstrap (Felsenstein. Their study 10 psbA-trnH 977–988 11 ETS 58–59 added by simple stepwise addition. GTR + I + G. Plastid formations are treated as equally likely and unordered regions were analyzed together as both exhibit low levels (Fitch. (2001) 4 min at 94°C followed by 30 cycles of 1 min at 94°C. 1 min at 52°C. PCR primers. this time saving no more than 10. • Suprageneric phylogenetics of Myrteae Table 3.500 trees 1 ITS 327 these were the successive approxiwas deleted. this set of 7. Thirteen potentially informative. 1971). 1 min at 72°C ETS Myrt (F) see text. and ETS datasets were as separate.. 1 min at 72°C psbA-trnH psbA (F). trees. TVM + G. trnH (R) Hamilton (1999) 4 min at 94°C followed by 30 cycles of 1 min at 94°C. Congruence among pair-wise combinations of the three datasets and for all datasets & Ochoterena (2000).06 (Posada & (TBR) and branch-swapping with MulTrees on but only Crandall. in” set of trees. — All character transand 85–100 strong support (Chase & al. reagents and conditions. psbAoptimal islands of trees (Maddison. The following models were selected unsaving 10 trees per replicate to reduce time searching subder the Akaike information criterion. non-overlapping indels were scored The combined plastid regions. Optimal trees of 10 to ensure the analysis did not stall on suboptimal were again collected and used as tree islands. concluded that bias in the mutational 12 ETS 495–505 TBR swapping with MulTrees on dynamics of stems and loops and non13 matK 267–275 independence amongst nucleotides in and saving no more than 10 trees 1109 . 18S (R) Wright & al. AB 102 (R) Sun & al. analyses were conducted with 500. All sequences are deposited in GenBank Bayesian analysis was performed on the combined dataset in MrBayes v3. The remaining trees were 2 ITS 340–342 mation weighted (SW) trees. 1326 (R) Johnson & Soltis (1994) 2 min at 94° C followed by 30 cycles of 1 min at 94° C. Sequences were aligned by eye following per replicate. binary characters at the end of the data matrix initially analyzed separately before the three datasets (Table 4). all such regions were excluded from the phylogethe component matrices when analyzed alone. once tree length remained oft Excel to determine the post “burn Indel Positions affected the same in two successive rounds. respectively. After complettrnH and matK. tree bisection reconnection the sequence partitions using Modeltest 3. 1 min at 50°C. 1985) was 4 psbA-trnH 359–369 5 psbA-trnH 783–829 used to assess support for individplaying the posterior probabilities (PP) 6 psbA-trnH 842–857 ual clades without weights from of all nodes. the resulting trees were used as starting TVN + G. 75–84 moderate support the guidelines of Kelchner (2000).000 generations of and after each round of reweighting a heuristic search Monte Carlo Markov chains and a sampling frequency of ten random replicates was performed. The imported into PAUP. 1969). ITS. ITS Primers PCR conditions AB101 (F). using PAUP* 4. Farris.

McVaugh. 1968). plano-convex cotyledons (Fig. Outgroup genera were pruned from the MPT.g. possibly due to rapid and recent speciation in the family (Landrum.. We accept the assertion of Snow & al. 1857–59) found the ovaries of Myrciinae and Eugeniinae to be constant. placing little emphasis on them while providing a detailed discussion on ovary variability . regardless if partitioned or not. embryo. H). modified according to subsequent authors. PP ≥ 95%). results of a survey of the wood anatomy of Myrteae by Telford & al. homogenous or fused (Fig. 2005. the condition of the outgroup is scored.e. the intention here is to observe the fit of the original tripartite system. (1) > 120 (Fig.0 (Maddison & Maddison. 1 and 3) that correlate with the original three embryo types. 2000). or inapplicable to all taxa. 1A). Results presented here and in Lucas & al. crumpled cotyledons (Fig.. the partitioning of the ITS data matrix recommended by Biffin & al. (2003). used independently in distinguishing between genera of Myrteae. was not implemented here. Wilson. Myrcia is here used in a broad sense. The four potentially useful characters were optimized onto a single most parsimonious tree (MPT) resulting from the successively weighted parsimony analyses. androecium. see Proença & al. 1968. (1) hypocotyl much reduced. Luma and Algrizea. or spiralled embryos (character state 2). and from herbarium material. seeds and inflorescence. Grifo (1992). If embryo type is inconsistent within a genus. usually wider than the cotyledons. 1E). 2001). Optimization of morphological characters was performed with MacClade 4. sometimes spiralled (Fig. Character coding was developed here based on ranges of states and characters taken from the generic descriptions of Bentham (1869). Characters. Our analyses concur that phylogenetic signal in Myrtaceae ITS is insufficient to resolve relationships among short internal branches. In all. including these other genera. with the majority proving impossible to score sensibly throughout the tribe. (0) < 70. The approach applied was to select key characters for optimization onto the molecular tree without including them in the tree-building process. hypocotyl linear or C-shaped (Fig. Austromyrtus. perianth. Gossia can have typical “myrtoid” embryos (character state 3). 1969). (5) short hypocotyl enclosed by lobed and thickened cotyledons (Fig. and conclusions in this study are therefore only considered reliable for those well-supported relationships (i.. and monophyletic clades of individual genera reduced to single terminals. (2) hypocotyl swollen. Embryo. 1986. (3) cotyledons inconspicuous. (unpubl.. favoring partitioning the three conditions into several characters more likely to be homologous. Landrum & Kawasaki (1997). and all other members of their clade have the “myrtoid” condition leading us to assume that character state 3 is plesiomorphic in Gossia and to assign it as such. — Morphological data in Myrtaceae are notoriously homoplasious. invariant. (0) hypocotyl wrapped around well-developed. G. with Snow & al. Snow & al.). 206). Mean number of ovules per ovary. Cladistic analysis of morphological characters was inappropriate at this scale due to low numbers of putative homologous characters available and the limitations of using place-holders for genera. character selection. Particular heed was therefore paid to Stevens’ (2000) recommendation that before embarking on a morphological phylogenetic exercise all aspects of methodology. and Fig. respectively. distribution of characters and variation patterns within the group should be carefully considered. (4) short hypocotyl enclosed by entire. (2005) suggest that the four core genera of subtribe Myrciinae form a poorly resolved monophyletic group with low internal 1110 support in which Myrcia and Marlierea are paraphyletic with Gomidesia and Calyptranthes nested within them. states 2. 2). For these reasons and also due to the relatively small effect that similar weighting schemes of loops versus stems for 18S rDNA data had on analysis of that region (Soltis & Soltis. in addition to those states (0. (2006) for discussion of the Luma and Algrizea embryos. — 1. cotyledons swollen. 1981a. The embryo is the most frequently used feature for grouping genera in previous Myrteae taxonomy (see Lucas & al. 1C. 1D). and still are. However. Proença (1990). Holst (2003). wood anatomy. 1B). 4 and 5 correspond to those of the Campomanesia complex sensu Landrum & Stevenson (1986). To that aim. (2003) rejecting the three traditional embryo types on the basis that they constitute several independent characters. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 stem-pairing sites is such that partitioning significantly improves the likelihood of more weakly supported (i. 1988). Numbers of locules and ovules are characters that have been. 1F). in light of this. 2. Berg (1855–56. 1 for more details of embryo types). PP < 95%) relationships in the resulting topology whereas well-supported clades [nodes] were generally found to be consistent with those inferred from other DNA regions regardless of the evolutionary model or partitioning used. the sister group to Gossia. There has recently been a move away from such heavy dependence on embryo type (e. until ultimately four characters were selected. from the gynoecium.Lucas & al. with trends in both characters associated with specific groups. 1998).e. (2003) that intrageneric variation in Myrteae embryos is such that ultimately the complete abandonment of traditional embryonic archetypes may be necessary. to our DNA tree to provide a starting point for future subdivision of embryo characters based on phylogenetic relationships. Snow & Guymer (1999) successfully developed such a system for Myrtella and Lithomyrtus but suggested it might “be of limited use in the Myrtaceae as a whole” (p. McVaugh (1963. twelve characters were considered. Morphological optimization. Landrum (1981b. for example.

1857–59) discussed placental diversity when defining Myrtus and Psidium whereas Kausel (1947. 2. we assume that the axile condition is not significantly different from the parietal one and that it is a homoplasious condition arising more than once.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. A similar view is taken here. Ranked means of ovules per ovary. and structure of the latter. 1869). and the character was coded according to the single break observed in the data (indicated by a dashed line). 3A. 1956) emphasized position and shape of the placenta in his definition of Argentinian genera. Placentation. incomplete septal division is reported in Myrteola (Landrum. 3F) are found in Rhodamnia and Neomyrtus . it is here hypothesized that the developmental trend in Myrteae has been a decrease in number of ovules per ovary and that numbers of either locules or of ovules are of secondary taxonomic importance. He concluded that “generic distinction based on [ovary characters] among species that are otherwise so strikingly alike. McVaugh (1963). 1988) and Myrtus (Bentham. (2) ovules in one or more series around a protruding placenta (Fig. • Suprageneric phylogenetics of Myrteae in Myrtinae. Seeking a new approach to compare these characters to a molecular tree. Fig. (1) ovules in one or more series along a non-peltate placenta (Fig. McVaugh (1968) concluded that number of ovules per locule is probably taxonomically significant although to an unknown extent. would seem to be speciose”. Kausel (1956) suggested that taxonomic decisions might be made based on a combination of ovule number. Genera were ranked according to this value (Fig. E). Berg (1855–56. (0) ovules mostly arising at a single point on septum (Fig. including mean maximum and minimum values. McVaugh (1968: 360) doubted that “valid conclusions as to generic limits can yet be drawn on the basis of [placentation]”. 3D. Dashed line represents delimitation of character states. Minimum and maximum locule and ovule per locule numbers were recorded for each genus. This study found that infrageneric variation in locule and ovule number is such that neither character can be divided into discrete states. 3. 2. with minimum and maximum values indicating range). Unilocular ovaries with parietal placentation (Fig. position of the placenta. 3C). 1991). and mean number of ovules per ovary calculated. 1111 . Bentham (1869: 119) believed “modifications of the ovules and their placentation afford the best generic characters”. considered relationships within a group of Psidium now accepted as Chamguava (Landrum. one of which possesses a unilocular ovary. B).

and Patel (1995) have reported helical thickenings of vessel elements. 3. Predominant placentation in A. scalariform. L. scalariform. Campomanesia.. with one bifurcating bar (RBGE 1992-3445 ). J–L. 1978). D–F. Amomyrtus.. 1997) was used to calculate likely distributions of the now extinct ancestors represented by the internal nodes of the Myrteae tree and examine biogeographical relationships of major Myr- Fig. of these. Rhodamnia. from radial sections. Ugni . I = 10 µm.1. Fig. simple in D. 4C. — Dispersal vicariance analysis (DIVA 1. L). Vessel element perforation plates in tribe Myrteae. simple (RBGK 2000-3513). A. (1) scalariform perforation plates (Fig. C. D–F. Vessel elements. Octamyrtus. C. reticulate in B (RBGE 1996-1223). placentation is in a single series on a non-peltate placenta. Despite the sister group to Psidium having non-peltate placentation. I). (0) simple perforation plates (Fig. Within the unilocular ovary of Rhodamnia and Neomyrtus. B.) suggest that. Placental variation in tribe Myrteae. Myrteola nummularia. with one bifurcating bar (RBGE 1996-1096). Kew ’51). 4G. J. Siphoneugena. Ugni molinae. B. those genera with incomplete septal development demonstrate a common but unrelated process. reticulate in E and F (Hort. character state 2 has been assigned. D. scalariform with only 3 bars in I (de Lima 2560). (2) mixed simple. I. Myrteola. the most and possibly only taxonomically useful character of the wood for distinguishing Myrteae genera is occurrence of scalariform perforation plates. Luma apiculata. G. D–G. and F. H. Additional wood anatomy data (Telford & al. therefore. Legrandia. Lophomyrtus. the majority of further outgroup genera have peltate placentas leading us to assume that character state 2 is plesiomorphic in Psidium and to assign it as such. Myrciaria . A. Acca. salicifolius. E. Pimenta. H. Myrcianthes. simple (Lucas 78). simple in A. Scale bars. H. Psidium . 4A. K. Placentation in Psidium varies with a non-peltate condition (character state 1) somewhat less common than the peltate (character state 2). scalariform. Meylan & Butterfield (1975. Myrcia. Blepharocalyx cruckshanksii. C. Lophomyrtus obcordata. 4. scalariform plates and tracheids in the wood of Myrteae. K). J. 1996. simple in H. Eugenia . 1112 . Ronquist.Lucas & al. reticulate or scalariform perforation plates (Fig. Plinia. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 with the incomplete septum and unilocular conditions arising from incomplete or “non-development of the placenta-bearing dissepiments” (Bentham. B. B. Decaspermum. Neomitranthes . Neomyrtus. 4. Myrceugenia alpigena. = 20 µm. unpubl. becoming reticulate towards top (RBGK 1958-50308). B. 1869: 119). Biogeography. Schmid & Baas (1984). Ragonese (1976). Myrtus communis. Wood anatomical surveys of Myrtaceae by Ingle & Dadswell (1953).

Maranhão. separating areas with most species diversity and endemism. Myrteae distribution with biogeographical divisions used for dispersal-vicariance (DIVA) analysis. 5). Central America to Panama. Adapted with permission from Heywood (1993). Pará. Rio Grande do Sul. 1981b. 1980). Peru. Mexico. North/Central America (NCA). western South America (WSA). Mato Grosso. Rio Grande do Norte. Bolivia Venezuela. Paraíba. (2001). Ecuador. without remarkable Myrtaceae species diversity but with distinct phytogeographical compositions (Cabrera & Willink. The two remaining areas (North/Central America. North Africa/Mediterranean (NAM). Australia (AUS). 2000). 1980). and sub-Saharan Africa (AFR). Roraima. South American regions (Table 5) are amalgamations of the phytogeographically distinct divisions of Cabrera & Willink (1980) made in consideration of Myrteae distribution. Goiás. Mato Grosso do Sul. • Suprageneric phylogenetics of Myrteae teae clades. The method maximizes the probability of vicariance events (no cost) and minimizes the probability of dispersal and extinction (cost of 1). yet tree-independent regions (Fig. 1986). Amapá.S. southern South America (SSA). the Guianas (McVaugh. eastern South America (ESA). Ceará.A. north/ central Brazil (NCB). Malesia (MAL). Argentina and Paraguay east of the Paraná and Paraguai rivers. Sergipe. Brazil north of 1°N. not only hierarchical ones. with Syzygium serving as outgroup due to its position as sister to tribe Myrteae in Fig. including the states of Acre. Indochina (INC). Bahia. Amazonas. Brazilian states of Alagoas. 1976). Rio de Janeiro. Definitions of Neotropical areas based on established phytogeographic centers (Cabrera & Willink. Colombia. Espírito Santo. the Guianas. the Amazon basin (Prance & al. Greater Antilles. Rondônia. Sergipe Chile. Myrteae distribution outside the Americas was divided according to the phytogeographically distinct areas proposed by Brummitt & al.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. (Florida). northern South America (NSA). western South America) are areas of similar size to the others. Papuasia (PAP). Minas Gerais. Pacific (PAC). 5. São Paulo. allowing the existence of reticulate relations among the areas. Brasília. Paraná. 1969) and the central to southern Andes (Landrum. Myrteae distribution was divided into 14 phytogeographically meaningful. Argentina and Paraguay west of the Paraná and Paraguai rivers 1113 Southern South America (SSA) . Santa Catarina. such as the Atlantic coast and cerrado of Brazil (Oliveira-Filho & Fontes. New Zealand (NZ). Lesser Antilles Brazil south of 1°N.. Table 5. Area North/Central America (NCA) Western South America (WSA) Northern South America (NSA) North/central Brazil (NCB) Eastern South America (ESA) Definition U. The MPT on which the biogeographical analyses were based was the same as for the morphological optimizations. Tocantins Uruguay.

7) and the sister relationship of Myrtus communis with the rest of the tribe (BP 61). The combined analysis provides better resolution and higher bootstrap support for clades than any of the three separate analyses with every group receiving > 50% BP in any of the individual analysis receiving increased support in the combined analysis. Myrceugenia and Myrcianthes. the plastid analysis comprises 503 and 806 aligned bp from psbA-trnH and matK. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 the molecular analysis. variable characters (%) No.77 Combined 2.59 0. 1995–1997) and available literature.52 ETS 448 292 (65. Myrceugenia and Myrcianthes receive > 50 BP in the ETS topology alone as do Pimenta and Psidium in the plastid topology. clade A. Austromyrtus) Decaspermum).003 (41) 600 (24. At the suprageneric level. Fig.71 0. rather than evidence of hybridization or horizontal transfer. ITS Characters No. of the “Myrcia group” (BP 64.000 > 10. Campomanesia. — The “soft incongruence” of the partitioned analyses is attributed to lack of phylogenetic signal.. potentially parsimony informative characters (%) Trees Trees (SW) Length (L) L (SW) (Fitch length of the SW result) Consistency Index (CI) CI (SW) Retention Index (RI) RI (SW) 1114 689 308 (44.77 0. — Statistics from the parsimony analyses are summarized in Table 6. and Myrcia. and these data are considered suitable for combined analysis (Sheahan & Chase.59 Plastid 1. ITS ((Rhodomyrtus.000 > 10. the relationship (BP 55) between Blepharocalyx cruckshanksii and Luma and the relationship of Gossia and Austromyrtus (BP 89).42 0. Gomidesia. The ITS and ETS topologies each receive > 50 BP for the additional following genera. with an association (BP 66) of Gossia + Austromyrtus with the remaining genera of the Australasian clade (clade G. Fig.000 1. hard incongruities) were present between Decaspermum. Five genera (Amomyrtus. three further clades consistently receive > 50 BP. Fig. Dataset and parsimony-based tree characteristics for ITS. The strict consensus tree from the combined analysis is shown in Fig. Marlierea. the “Myrcia group”). Siphoneugena.60 . Siphoneugena.286) 0. plastid and combined analyses plus indels.121 265 (1. Octamyrtus) Decaspermum).8) 190 (14. Table 6. Myrciaria. with the exception of the well-supported (BP 100) tribe Myrteae (clade H.52 0.125) 0. A tree indicating branch lengths of a single MPT (Fig.446 1.5) > 10. we provide informal subtribal groupings in quotation marks (e.47 0.000 1. Rhodamnia) receive > 50% bootstrap support (BP) in all three topologies.128 310 (1.000 674 396 (675) 0. A clade comprising the core Myrciinae genera plus Neomitranthes. Calyptranthes.Lucas & al. Siphoneugena.8 (1.50 0.40 0. and Algrizea also receives > 50 BP in the plastid analysis as do clades containing Eugenia + Myrcianthes and Austromyrtus + Gossia in the ETS analysis. Genera were scored as present or absent in each area with the distribution of a genus taken as the sum of areas that it currently inhabits.987 467.40 0. unrelated to Austromyrtus.70 0. these are a clade containing all Myrteae genera.3) > 10. 6) from the combined analysis demonstrates that the length of the branch separating Myrtus communis from the rest of Myrteae is one of the longest in the analysis. 7) comprising the four core genera of the Myrciinae.g. Gossia. 7) comprising Neomitranthes.e.58 0.000 2. unrelated to Austromyrtus. Calyptranthes.840 > 10. Octamyrtus. ETS. The combined topology still lacks bootstrap support along the spine of the tree however. Fig. Distributional data were collected from Kew herbarium collections (Brummitt & Brummitt. 7). RESULTS Sequence data. clade B.7) 195 (28. plastid ((Octamyrtus.610 1.134) 0.5) 3. Phylogenetic analysis. whereas Eugenia recieves > 50 BP in the plastid and ETS topologies. before and after successive weighting (SW). Other suprageneric relationships receiving bootstrap support in the combined topology are: monophyly of the “Plinia group” (BP 95. Gomidesia. whereas other branch lengths within Myrteae are relatively short.43 0. 7 and comprises seven clades including all but three isolated taxa. Plinia and Myrciaria. 2000).2) 215 (48) > 10. The only incongruity between groupings with > 50 BP in the three topologies (i. Rhodomyrtus) Octamyrtus). Siphoneugena + Neomitranthes and Neomyrtus + Lophomyrtus. Pinia. Rhodamnia. respectively..309 403 (30. Luma. Rhodamnia and Austromyrtus which emerged with > 50 BP at each of the following nodes: ETS ((Decaspermum. The consensus trees from the SW analyses of individual data partitions (not shown) demonstrate low bootstrap support along the spine of the trees and low resolution in the in-group.

• Suprageneric phylogenetics of Myrteae Fig.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. Psidium cinereum Psidium cattleianum Acca sellowiana Rhodamnia rubescens Rhodamnia argentea Decaspermum humile Rhodomyrtus psidioides Octamyrtus pleiopetala Gossia inophloia Gossia hillii Austromyrtus dulcis Blepharocalyx salicifolia Myrtus communis Syzygium jambos Syzygium maire Anetholea anisata Acmena smithii Eucalyptus tetragona Eucalyptus perriniana Leptospermum scoparium Callistemon comboynensis Xanthomyrtus montivaga Xanthomyrtus compacta Tepualia stipularis Metrosideros perforata Lophostemon confertus 10 changes 1115 . A single tree of 1. and matK data.427 minimum length trees generated by the heuristic search of combined ITS. floribunda Calyptranthes concinna Calyptranthes lanceolata Calyptranthes kiaerskovii Calyptranthes thomasiana Gomidesia tijucensis Gomidesia affinis Gomidesia schaueriana Gomidesia flagellaris Myrcia pubipetala Myrcia venulosa Myrcia laricina Marlierea eugeniopsoides Myrcia rostrata Myrcia fallax Myrcia coumeta Marlierea suaveolens Marlierea obscura Myrcia multiflora Myrcia laruotteana Myrcia tomentosa Myrcia saxatilis Myrcia acuminatissima Myrcia bicarinata Algrizea macrochlamys Blepharocalyx cruckshanksii Luma apiculata Luma chequen Myrceugenia lanceolata Myrceugenia leptospermoides Myrceugenia planipes Myrceugenia myrcioides Myrceugenia alpigena Myrceugenia ovata Myrteola nummularia Ugni molinae Neomyrtus pedunculata Lophomyrtus obcordata Lophomyrtus bullata Amomyrtus meli Amomyrtus luma Legrandia concinna Pimenta dioica Pimenta racemosa Pimenta pseudocaryophyllus Eugenia stictosepala Eugenia latifolia Eugenia punicifolia Eugenia florida Eugenia uniflora Eugenia sulcata Eugenia langsdorfii Myrcianthes pungens Myrcianthes pseudomato Campomanesia guazumifolia Campomanesia pubescens Campomanesia sp. psbAtrnH. ETS. aff. Neomitranthes cordifolia Siphoneugena densiflora Siphoneugena guifolyleana Plinia pauciflora Myrciaria cauliflora Myrciaria sp. 6.

7. clades that receive Bayesian probabilities greater than 0. Strict consensus tree resulting from parsimony analysis with successive weighting and Bayesian analysis of the ITS. Bootstrap percentages greater than 50 are shown above branches.Lucas & al. floribunda Algrizea macrochlamys Calyptranthes concinna Calyptranthes lanceolata Calyptranthes kiaerskovii Calyptranthes thomasiana Marlierea eugeniopsoides Myrcia rostrata Myrcia fallax Myrcia coumeta Gomidesia tijucensis Gomidesia affinis Gomidesia schaueriana Gomidesia flagellaris Myrcia laricina Myrcia pubipetala Myrcia venulosa Marlierea suaveolens Marlierea obscura Myrcia multiflora Myrcia acuminatissima Myrcia laruotteana Myrcia tomentosa Myrcia saxatilis Myrcia bicarinata Blepharocalyx cruckshanksii Luma apiculata Luma chequen Myrceugenia lanceolata Myrceugenia leptospermoides Myrceugenia planipes Myrceugenia myrcioides Myrceugenia alpigena Myrceugenia ovata Myrteola nummularia Ugni mollinae Neomyrtus pedunculata Lophomyrtus obcordata Lophomyrtus bullata Amomyrtus meli Amomyrtus luma Legrandia concinna Campomanesia guazumifolia Campomanesia pubescens Campomanesia sp. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 Fig. clades not recovered in the Bayesian analysis are marked with solid arrowheads. 100 A 77 95 71 97 100 100 98 L 100 100 100 89 B K 64 66 53 100 100 55 100 77 C I 96 61 84 85 D 100 97 100 J E 94 100 100 63 70 80 53 100 61 F 81 90 100 100 H 99 71 99 65 89 100 G 100 100 59 55 100 95 85 Neomitranthes cordifolia Siphoneugena densiflora Siphoneugena guifolyleana Plinia pauciflora Myrciaria cauliflora Myrciaria aff. psbAtrnH and matK data.95 are marked in bold. ETS. Psidium cinereum Psidium cattleianum Acca sellowiana Pimenta dioica Pimenta racemosa Pimenta pseudocaryophyllus Eugenia stictosepala Eugenia latifolia Eugenia punicifolia Eugenia florida Eugenia uniflora Eugenia sulcata Eugenia langsdorffii Myrcianthes pungens Myrcianthes pseudomato Blepharocalyx salicifolius Rhodamnia rubescens Rhodamnia argentea Decaspermum humile Octamyrtus pleiopetala Rhodomyrtus psidioides Gossia inophloia Gossia hillii Austromyrtus dulcis Myrtus communis Syzygium jambos Syzygium maire Anetholea anisata Acmena smithii Eucalyptus tetragona Eucalyptus perriniana Leptospermum scoparium Callistemon comboynensis Xanthomyrtus montivaga Xanthomyrtus compacta Tepualia stipularis Metrosideros perforata Lophostemon confertus Plinia group Myrcia group South American Myrteae Myrceugenia group Myrteola group Pimenta group Eugenia group Australasian group Outgroups 1116 .

WSANCB-ESA. NSA. NZ. Clades identified by letters refer to Fig. This suggests that the distributions of some large genera with centers of diversity in eastern or northeastern South America. NCA-NCB-ESA. North Africa/Mediterranean. 8). ESA-SSA WSA. NCB-ESA. WSA-NCB-ESA-SSA. each displaying some level of polymorphism when optimized over the entire Myrteae tree. NSA-WSA-NCB-ESA-SSA. Characters based on embryo. INC. PAP. 7. NCA-NSAWSA-NCB-ESA-SSA NCA-NSA-WSA-NCB-ESA-SSA “Plinia group” + Algrizea “Myrceugenia group” (clade C) “Myrteola group” (clade D) “Pimenta group” (clade E) “Eugenia group” (clade F) “Australasian group” (clade G) Clade I Clade J Clade K Clade L South American Myrteae AUS. NCA-NCB-ESA-SSA. latifolia clade (PP 91). and (4) E. E. NCA-NSA-NCB-ESA. New Zealand. Malesia. Papuasia. 1981a). SSA. Australia. NSA-WSA-NCB-ESA. allowing no precise conclusions to be drawn regarding the area of origin of the ancestor of the “Plinia group” + Algrizea + “Myrcia group” association and the “Myrceugenia group”. suggesting an Andean origin for the “Myrteola”. whereas the four nodes incongruent with the SW analysis are indicated by arrows. Areas of ancestral distributions of key Myrteae clades as inferred from DIVA. NCA. perhaps reaching the east via temperate to subtropical forests covering what is now Chile and southern Argentina in the Oligocene (Landrum. eastern South America. supporting Landrum (1981a). Identical DIVA results indicate an origin of both the South American Myrteae clade and clade I (South American Myrteae excluding Blepharocalyx sal- cifolius) in any or all of the South American areas used here. ESA. North/Central America. Malesia. and L. however. combinations of characters may often be used to diagnose clades. B.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. NCA-NSA-NCB-ESA-SSA. Table 7. The four incongruent arrangements in the Bayesian topology are: (1) Myrtus communis sister to the Australasian clade (PP 71). — Of the morphological characters under study (Fig. MAL. Morphological optimization. (2) Rhodomyrtus and Decaspermum with a sister relationship (PP 96). stictosepalaE. western South America. Australia. NCA-WSA-NCB-ESA-SSA. punicifolia as sister to the E. a scenario supported by the suggestion of the same areas of origin for all three of these last clades (D. PAC. may have migrated from southwestern to north or northeastern South America and speciated there. and the Pacific. NAM. Nodes supported by ≥ 95% posterior probability (PP) are indicated by bold branches in Fig. Indochina. DIVA suggests that of the two main clades into which the ancestor of clade I diverged (J and K). ovary and placentation correlate less well. “Pimenta” and “Eugenia” groups. Node “Plinia group” (clade A) Area of ancestral distribution of key Myrteae clades inferred from DIVA NCB. NCA-NSA-WSA-NCB-ES. (3) the “Eugenia group” emerging as sister to the “Myrteola group” (PP 39). NCB-ESA SSA WSA. NCB-ESA-SSA. Despite this. albeit with exceptional taxa in several cases (see discussion). Biogeography. this supports a movement of one or more ancestral taxa from present-day Australasia to South America. NSA-NCB-ESA. SSA INC-MAL-AUS-PAP-PAC NCA-NSA-WSA-NCB-ESA-SSA WSA. ESA. 7. Pacific. K. respectively) with the additional inclusion of eastern South America in the case of the “Pimenta group”). NCB. — DIVA results (Table 7) suggest that the early branching “Australasian group” (clade G) evolved in situ in Indochina. 1117 . southern South America. Any or all South American divisions are suggested as areas of origin for clades A. NSANCA-WSA-NCB-ESA-SSA NCB. such as Eugenia and Psidium. only the nature of the perforation plates of the vessel elements provided unambiguous support for two clades. the area of origin of the “Myrceugenia group” is exclusively suggested as southern South America. SSA NCA-NSA-NCB-ESA-SSA. NCA-WSA-NCB-ESA. and F. • Suprageneric phylogenetics of Myrteae A 50% majority rule consensus of the trees produced by the Bayesian analysis was almost entirely congruent with the SW tree with an apparent increase in support. SSA WSA-SSA. the origin of clade J was in either western or southern South America. NCA-NSA-NCB-ESA-SSA. NCA-NSA-WSA-NCB-ESA-SSA NCA-NSA-NCB-ESA. NCA-NSA-WSA-NCB-ESA. NSA-NCB-ESA-SSA. Papuasia. The same lack of certainty applies to the areas of origin of the “Plinia” and “Myrcia” groups and Algrizea. north/central Brazil. WSA. northern South America.

cotyledons inconspicuous short hypocotyl enclosed by thick plano-convex cotyledons . vessel elements. mean number of ovules per ovary. hypocotyl enclosed by well . usually wider than the cotyledones. lobed cotyledons equivocal B.salicifolius Rhodamnia Decaspermum Octamyrtus Rhodomyrtus Gossia Austromyrtus Myrtus A. sometimes spiralled hypocotyl linear or c-shaped. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 Neomitranthes Siphoneugena Plinia Myrciaria Algrizea Myrcia group B. B. 1118 . A. crumpled cotyledons hypocotyl much reduced. Vessel elements with simple perforation plates with scalariform perforation plates with mixed simple. 8. homogenous or fused hypocotyl swollen. developed.Lucas & al. embryo. Mean ovules per ovary < 70 > 120 C. reticulate or scalariform perforation plates Fig. cotyledons swollen. Embryo hypocotyl wrapped around well developed. cruckshanksii Luma Myrceugenia Myrteola Ugni Neomyrtus Lophomyrtus Amomyrtus Legrandia Campomanesia Psidium Acca Pimenta Eugenia Myrcianthes B. placentation. Placentation ovules mostly arising at a single point on septum ovules in one or more series along a non-peltate placenta ovules in one or more series around a protruding placenta equivocal D. Character optimization over a single tree from the combined analysis. D. C.

The SW and Bayesian analyses presented here strongly support (BP 100.. Temu will need to be reinstated. 2002) or PP is excessively liberal and BP more reliable (Suzuki & al. 2004). plus Eugenia and Myrcianthes. However. Clade J (BP < 50. emerging as in Sytsma & al.. the “Myrceugenia group” (clade C) comprises B. Blepharocalyx salicifolius is the type species with dichasial inflorescences. 2002) is a subject of debate. Ugni. each considered a separate genus by previous authors. Octamyrtus. Despite wide sampling (including 31 of 49 Myrteae genera) and including sequences from four DNA regions. In this combined analysis. an Archirhodomyrtus-Uromyrtus clade comparable to the Australasian clade observed here was supported by a 36bp deletion in matK. The “Australasian group” (clade G) is split into two clades with the two species of Gossia sampled appearing monophyletic (BP 100. The position and status of Blepharocalyx is particularly perplexing. a trend observed in other studies (Karol & al. In Landrum’s (1986) account of Blepharocalyx. homogenous embryo with little distinction between the cotyledons. In a weakly supported sister relationship to the “Myrteola group”.. and with a different position in the Bayesian topology mean that its position is still unclear. Low bootstrap support. the species was resequenced for each region and always gave the same result. Lophomyrtus and Neomyrtus. salicifolius as used here emerged associated with Luma and Myrceugenia. but also by the degree of congruence of the SW topology with the Bayesian analysis. Performing the SW search on the taxonomic sample used here. we prefer the latter position because the low BP better reflects the short branch lengths observed (Fig. with this clade sister to Myrceugenia (BP < 50%. PP 100). cruckshanksii sister to Luma (BP 61. PP 100) the monophyly of Myrteae (clade H. it was described as Marlieriopsis Kiaersk. with this clade sister to the other Australasian genera (BP 65. such as transitions and third positions of protein-coding genes (Albert & al. PP 100) for Rhodamnia as sister to a clade comprising single samples each of Decaspermum..TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. PP 99. PP 94). 2001. an open calyx with four persistent lobes and a C-shaped embryo. Gossia and Austromyrtus were recently segregated by Snow & al. Blepharocalyx eggersii has a paniculate inflorescence. Blepharocalyx eggersii was not sampled here and only its inclusion in a future analysis will indicate if it is a true Blepharocalyx or if Marlieriopsis will require resurrection. again. salicifolius in this analysis is a result of increased sampling. B. but with low BP in both cases. in which B. salicifolius retains the position it has in the present analysis (BP < 50). (2005) the same sequence of B. plastid. Whether BP is too conservative and PP a better indicator of confidence (Wilcox & al. In the combined analysis of Lucas & al. whilst characters such as placentation unite these three species. The MrBayes analysis retrieves eight clades with PP > 95 that received BP < 50 in the parsimony analysis. • Suprageneric phylogenetics of Myrteae DISCUSSION Phylogenetic relationships in Myrteae. Should the results presented here stand up to greater taxonomic sampling and addition of characters. whereas in an unsupported arrangement B. This arrangement is not found in the single region analyses. salicifolius emerges in three different positions. an open calyx with lobes that fall at about anthesis and a spiralled embryo. PP 100) and sister to the single sample of Austromyrtus (BP 89. — The successive weighting (SW) of all characters based on their performance (RC). is a clade (BP < 50. their monophyly remaining to be evaluated by broader sampling. (2005). (2004) placed Mediterranean Myrtus sister to the Neotropical genera. PP 100) but no support for their sister relationship to Ugni or the sister relationship of Myrteola to the rest. Sytsma & al. excluding Blepharocalyx salicifolius. the genus contains three species. there are low levels of DNA divergence and low bootstrap support along the spine of the tree. Within the tribe. none of which are with the “Myrceugenia group”. but a decrease in the rate of sequence divergence is another possibility. a closed calyx splitting into four lobes that fall at anthesis and a large. however. This deletion is not present in the portion of matK sequenced in this analysis. salicifolius is sister to all other South American Myrteae. South American Myrteae. The “Myrteola group” (BP < 50. Clarkson & al. (2004) in a matK analysis or sister to all other Myrteae when data from the spacer between matK and the 5′ exon of trnK was included. fall into two groups (clades J and K). (2003). 1119 . other characters vary. The strongest evidence for the position of Myrtus in this analysis is from the parsimony analyses. with strong support for the monophyly of the New Zealand genera (BP 100. suggesting that the position of B. 1993). Berg. and combined analyses and in an unresolved position within the Myrteae clade in the ETS analysis. (2005) varied. There is support (BP 71.. (2005). and Rhodomyrtus . this high genetic similarity supports the suggestion of recent and rapid speciation in the tribe. Fig. in which it emerges as sister to the rest of Myrteae in the ITS. PP 100) comprises the South American Myrtinae sensu DC. PP 100). is justified not only by the observation that CI and RI are consistent between sites considered under frequent change. 6). In Wilson & al. rather than only those estimated to have a low frequency of change. 7). PP 98) composed of two subclades. using only the ITS and psbAtrnH sequence data as in Lucas & al. a more detailed study of these groups will be necessary to clarify relationships within this group. Blepharocalyx cruckshanksii has dichasial inflorescences. whereas its position in Wilson & al. clade D) comprises Myrteola. it was originally described as Temu O. PP 100). and the two New Zealand genera.

Future ontogenetic studies seeking developmental differences between seemingly identical states with independent origins (e. the “Myrceugenia group”). and that the swollen or folded embryos of Eugeniinae and Myrciinae evolved independently to increase the capacity of the embryo to store food. whereas they are almost always entirely so in Eugenia although in Myrcianthes the cotyledons are nearly always free and plano-convex (McVaugh. Ultimately. 1986. In addition. apparently monophyletic Eugenia (BP 100. comprises Amomyrtus (BP 100. as would similar studies in clades in which more than one embryo type is present (e. 2004. This contrasts with previous results (Proença & al. Whereas the association of Algrizea with the other genera of clade L is clear. Developmental studies will be necessary before such reversals or parallelisms can be confirmed.g. Landrum. and the position of this taxon nested between otherwise myrtoid clades suggests that this species has long been isolated from the rest of the South American Myrteae. therefore. with genera of the “Australasian group”. PP 100). Algrizea. undeveloped cotyledons and although enlarged cotyledons are found in genera of tribe Syzygieae. “Myrteola”.. Cotyledons are less likely to be 1120 fully fused in Plinia and associated genera (Kausel. clade L) between the “Plinia group” (clade A) of genera previously ascribed to Eugeneniinae.. in which the ovary is occasionally tetralocular. Wilson & al. each having arisen more than once. PP 100) sister relationship to monophyletic (BP 100. 1986). Psidium. and “Eugenia” groups). 2003) unambiguous partitioning of the embryo into several.. 2004. homologous characters aiding the natural classification of the Myrteae genera.Lucas & al. Luma. Results indicate that the apparently homogenous eugenioid embryo has arisen independently in the “Plinia” and “Eugenia” groups. Optimization of mean numbers of ovules per ovary (Fig. The sister relationship (BP < 50. Poorly supported clade K suggests a sister relationship of the “Plinia group” plus Algrizea and the “Myrcia group” (clade L) to the “Myrceugenia group” (clade C). PP 100. Snow & al. Salywon & al. Optimization of a modified tripartite embryo arrangement confirms that states relied on by previous authors are not suitable for the grouping of Myrteae genera. with this last clade sister (BP < 50. PP 100) Myrcianthes. The Campomanesia-Pimenta-Acca association is supported by overall morphology. in conjunction with other morphological characters. spiralled embryos or crumpled cotyledons are unique to Myrteae. This is supported here. its fluctuating position between analyses and lack of support renders this another taxon for which exact phylogenetic position requires further investigation. 1963). The modified tripartite embryo characterization can.g. The foliaceous myrcioid embryo occurs throughout the “Myrcia group” and also in Myrceugenia. PP 100) sister to Legrandia (BP < 50. The spiralled embryo of Blepharocalyx salicifolius. Pimenta. Morphological optimizations. 1968). and in these cases. The “Myrcia group” appears monophyletic (BP 64. and Blepharocalyx cruckshanksii or its independent evolution in the “Myrcia group” and Myrceugenia. Bentham (1869: 150) described finding “nothing in habit or character to separate Acca [at that time in Orthostemonoideae] from Psidium” other than the presence of endosperm in the seed of Acca (later found to be erroneous) and by Psidium and Campomanesia forming a group distinct from those other Myrtinae genera in which “the ovary [is] never more than tricelled”.g. optimization of the form of the embryo provides potential synapomorphies for these groups. The “Eugenia group” (clade F) comprises two traditional Eugeniinae genera. Algrizea and the “Myrcia group” (clade B) supports previous analyses (e. 8B) indicates that ancestral Myrteae had between 12 and 70 ovules. Within some subtribal clades (the “Myrcia”. and Acca. a generalization that holds true in Neotropical Myrtinae sensu DC. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 one of which. swollen cotyledons in the “Plinia” and “Eugenia” groups) would be helpful. and Acca. except for Ugni. embryo type is largely consistent. similar to that of Ugni (Myrtinae).. the gradual accumulation of such information should allow the much-desired (McVaugh. and unfolded. Optimization of ovule number per ovary may support the paraphyly of Eugeniinae since Eugenia and Myrcianthes of the “Eugenia group” contain . 2006) and the Bayesian analysis (PP 85) presented here that support a sister relationship between Algrizea and the “Myrcia group”. and the “Myrteola group” sharing the myrtoid embryo and these genera successively emerging from the deepest nodes of Myrteae (Fig. suggesting either its repeated loss in the “Plinia group”. thin. this “Myrcia group” will be the focus of a future paper. however.. Myrtus communis. Locular complement in Myrteae varies from unilocular Rhodamnia to up to 20-locular Campomanesia. including the four core genera of Berg’s subtribe Myrciinae. PP 98) to a clade of Campomanesia. PP 100) in a well-supported (BP 90. whereas number of ovules per locule varies from two in the “Myrcia group” and some genera of the “Plinia group” to just under fifty in Acca. the “Pimenta group” (clade E).. Landrum & Stevenson (1986) suggested the ancestral Myrteae embryo was small. “Plinia”. a group now confirmed as artificial with its genera spread widely through the tribe. many of the apparently oldest Myrtaceae genera (Sytsma & al. 8A). 2005) indicating paraphyly of Eugeniinae. developing the spiralled embryo type independently from the other genera of the Campomanesia complex (Landrum & Stevenson. 1968. Sytsma & al. Psidium. with the average number subsequently increasing in Campomanesia. PP 100). all of which appear monophyletic. 1956. 2004) also have linear or slightly incumbent embryos with small. Number of ovules per ovary.. — Embryo. be used to assess the phylogenetic hypothesis presented here. McVaugh.

preventing the delimitation of an additional state within the character. Nature of placentation has limited use as a character for defining groups within Myrteae. after more striking differences (such as unilocular vs. nested between the “Plinia” and “Myrcia” groups.or multilocular ovaries) have been disregarded. Mixed scalariform and simple plates are also found in Lenwebbia (Australia) and Myrtastrum (New Caledonia).. 8D) suggests that they developed twice in the history of Myrteae. Scalariform plates are usually considered to be associated with more ancient taxa (Bailey & Tupper. with an at least partial reversal to the plesiomorphic state apparent in Neomitranthes and Siphoneugena. 1988. values closer to the range of the majority of genera in the “Pimenta” and “Myrteola” groups to which the “Eugenia group” appears more closely allied. There is some additional overlap with Rhodomyrtus and Pimenta of the “Australasian” and “Pimenta” groups. The function of scalariform plates may be to trap air bubbles during the thawing of ice. or they may be a relictual state in taxa in warmer environments that have lost scalariform plates to provide a more efficient hydraulic system during the wet season (Schmid & Baas. averages eleven ovules per locule. 8C). non-simple perforation plate states. Results indicate that placentation in ancestral Myrteae was bi. the character apparently of limited use as currently employed. when the genera of the “Plinia group” with ovules arranged in rows along at least part of the septum are considered. The link is unclear. Myrtus and the “Australasian group”. have simple perforation plates.. an arrangement more similar to that of the “Myrcia group” (with a mean of 4 to 5 ovules per ovary) than to the “Eugenia group” with which it was previously associated. 2004). in the “Myrteola group” and the “Myrceugenia group”. however. they are probably retained as an adaptation to cold and are therefore more prevalent at high altitude (Jansen & al. with arrangements of ovules on peltate or otherwise extruded placentas having arisen independently in the “Australasian group” and the “Myrteola” and “Pimenta” groups and Myrcianthes. particularly as in each case the ovules emerge from below the midpoint of the septum. respectively. This fine splitting of characters. a trend not supported here as the earliest branching Myrteae lineages. is tenuous. Carlquist. 1857– 1121 .or multi-seriate on flat or parietal placentas covering at least part of the septum (Fig.. bi. • Suprageneric phylogenetics of Myrteae an average of 22 and 50 ovules per ovary respectively. Jansen & al. whereas Algrizea. 1918. Assessment of previous supra-generic arrangements.. It appears that the subsequent transition to the arrangement in which ovules arise at a single point on the septum occurred independently in Eugenia and the “Plinia group” + Algrizea + “Myrcia group” association. Placentation. 2005). 1996). only one more than Siphoneugena. Algrizea and those genera of the “Plinia group” in which the ovules arise at a single point. South American Myrteae with scalariform perforation plates are consistently from (sub)montane Andean localities. entirely scalariform proved unnecessary with the “Myrceugenia and “Myrteola groups” equally supported if the character is divided into simple vs. The hypothesis that an increased number of ovules per ovary has more taxonomic significance than number of locules or ovules is not supported except in the case of the Campomanesia-PsidiumAcca association. Jansen & al. The presence of scalariform perforation plates in Neomyrtus and Lophomyrtus supports the suggestion of the phylogenetic topology that these genera arrived in New Zealand after dispersal from western South America. The perforation plate character is the most robust of those evaluated here. The division of the character into mixed scalariform vs. 1984. 2004). 2004). — De Candolle (1828) and Berg’s (1855–56. challenging the weakly supported. 2004) supports a scenario whereby taxa in Myrteae and Myrtaceae speciated in the Andes and lost their scalariform perforation plates as they spread into the tropical lowlands. whereas in the latter group placentation of ovules usually occurs at the middle of the septum or above. the interval between the highest value of the “Plinia group” and the lowest values of the next ranked taxon is not discrete. however. The presence of scalariform perforation plates in Andean Tepualia of tribe Metrosidereae (Jansen & al. but an equally valid although less parsimonious explanation could be that they are relictual in these two groups and there have been multiple losses in the other genera (Baas & Wheeler. Vessel elements. The optimization of scalariform plates (Fig. and higher than the range in the “Plinia group”.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. both averaging ten ovules per ovary. suggesting that these genera result from similar dispersal events by (an) ancestor(s) of the “Myrteola group” and supports the placement of Lenwebbia outside of the “Australasian group” (Wilson & al. it provides potential synapomorphies only if a high and unlikely number of reversals and parallelisms are assumed to have taken place and further use of this character will require developmental investigation of the gynoecium similar to that undertaken by Bohte & Drinnan (2005). The average number of ovules per ovary in the “Plinia group” is 4 to 9. Should placentation at a single point be shown to be synapomorphic for clade L. molecular-based grouping of the former with the “Plinia group” + Algrizea + “Myrcia group” association. In the tropical taxa where scalariform perforation plates do occur. It is tempting to hypothesize a link between placentation in the “Myrcia group”. However. clearly defining two clades and associating the apparently less closely related “Myrceugenia group” with the “Myrteola group”. it might be distinguished from the ovules of the “Eugenia group” by its position below the midpoint of the septum..

bony seeds and ovaries mostly trilocular or less. instead Myrcianthes is associated with Eugenia. McVaugh (1968) recognized six informal groups of American Myrteae with a further eight unaligned genera. However. united by possession of fused cotyledons. Wilson & al. embryo-based viewpoint. McVaugh’s group 4 unites Campomanesia. but it is assumed to be based on shared possession of a myrtoid embryo. Plinia. respectively. Psidium. and bilocular ovary. A relationship between Myrcianthes and Luma is refuted by our analysis. Plinia and Siphoneugena with Legrandia allied. or Myrtus. Reichea. tentatively ascribing their affinities to the “myrtoid” genera. Grifo (1992) recognized that Pseudanamomis is closely related to Myrcianthes (including Reichea). as Legrand considered it a Eugenia (McVaugh. Blepharocalyx. Britoa. Our analysis indicates McVaugh’s group 1 to be monophyletic. McVaugh left Acca and Myrrhinium with no placement in his scheme. McVaugh 1968.g. it is clear that further research on these groups is required. 2005). pyriformis (Cambess. not homogenous or fused as they are in Eugenia. Within Myrcioideae sensu Kausel. their emergence together here supports the monophyly of an inclusive Myrcianthes. Myrciaria. McVaugh appearing to have correctly associated these genera. McVaugh’s group 1 comprised the four core Myrciinae genera with Myrceugenia and Nothomyrcia (a synonym of Myrceugenia) allied. Myrtus. and Legrandia in the “Pimenta group”. these are characters that support their association in this study with Legrandia and the “Pimenta group”. Eugenia in its own group. spiralled embryos. Unpublished data from Wilson & al. The subfamilial groupings of Kausel (1956.) Kausel is included. Cryptorhizoideae sensu Kausel is artificial for reasons already discussed regarding the Campomanesia complex of Landrum & Stevenson (1986). McVaugh’s group 5 unites Calycolpus. swollen cotyledons. with Amomyrtus and Amomyrtella allied. Biochemical similarities suggested that Amomyrtella is closely related to Amomyrtus (Weyerstahl & al. anomalous in McVaugh’s group due to a multi-locular ovary. (including Myrceugenia) are polyphyletic. Ugni. reflexed cotyledons of Acca. Kausel transferred those Myrciaria species (including Myrciaria cauliflora) with free cotyledons to Plinia (Plinioideae). Myrceugenia. the former has yet to be sequenced. of a possible common origin for Myrteola nummularia and Lophomyrtus obcordata. McVaugh (1968) did not explain his assumption of an affinity of Amomyrtus and Amomyrtella with the core group. P. and Temu based on possession of soft testa and enlarged. 1966) are to date the only arrangement that did not accept a direct relationship between Plinia and Eugenia . and Siphoneugena emerging in the “Plinia group”. based on the dichasial inflorescence.Lucas & al. Barrie. based mainly on embryo morphology. a grouping noted by Landrum’s (1986) subsequent synonymization of the former with the latter and supported by this analysis. McVaugh’s group 2 comprised Eugenia. 1992). solitary or three flowered peduncles.l. (pers. Kausel included what remained of Myrciaria in Eugenioideae.). whereas floral morphology and dynamics in Ugni differ from the New Zealand species. excluding Myrceugenia. 2004) and here challenged by the grouping of two species of Myrciaria separated from the only included species of Plinia . including some but not all members of the “Myrteola” and “Pimenta” groups. In the first realignment attempting to depart from the three subtribe. with Luma allied. Our analysis indicates that this group is paraphyletic. Luma (as Myrceugenella) and Acca (as Feijoa) do not appear directly related to the “Myrcia group”.. fragrans. Snow & Guymer (2001) have shown similarities between floral morphology in Ugni and the Australian genus Uromyrtus . Myrtoideae sensu Kausel is also artificial. In addition to these six groups. Sobral. an association based on Kausel’s belief that the folded embryo of Myrceugenia and the “Myrcia group” and the large. A phylogenetic study of the “Eugenia group” by Mazine & al. and Pseudanamomis. comm. a suite of characters demonstrated to be homoplasious. Results within the “Myrteola group” support developmental floral morphology-based observations by Belsham (unpubl. is not directly related to Ugni. Myrtinae are paraphyletic with genera previously ascribed to Eugeniinae nested within it. Calycorectes. had a common origin. as Kausel noted that Plinia cotyledons are discrete and plano-convex. (unpubl. Acca emerges with Psidium in the “Pimenta group”. McVaugh’s group 3 unites Myrcianthes. and its position remains unconfirmed.. based on possession of a “myrtoid” embryo. involute. Myrteola.) 1122 indicates a paraphyletic Eugenia including Calycorectes. and Myrteola. 1968). Pseudomyrcianthes and Acreugenia (now treated within Myrcianthes). with Myrciaria.) suggest that in matK-based analyses Calycolpus is sister to Myrtus. McVaugh’s group 6 unites Pseudocaryophyllus and Pimenta. Our study indicates that Psidium. were placed separately in Eugenioideae and Plinioideae. the shared possession of scalariform perforation plates supports the inclusion of Ugni in the “Myrteola group”. although this could change when the type species of Pseudomyrcianthes. Paivaea. an inclusion questioned by subsequent authors (e. 2001. These last genera possess solitary (Amomyrtus. Amomyrtella) or racemose (Amomyrtus) inflorescences and subapical protruding placentas. six of which were nevertheless allied to a group. whereas the latter and subtribe Myrciinae s. and Landrum (1986) suggested that Myrrhinium shares significant . and this congruence with our analysis suggests that their inclusion in McVaugh’s fifth group is erroneous. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 59) division of Neotropical Myrtaceae into three subtribes based on embryo morphology is artificial (Landrum 1986. 1993. These last genera are represented here by Myrcianthes pseudomato and M.

a mean of 20 to 70 ovules per ovary (except Rhodomyrtus). Foliaceous cotyledons. Myrcianthes. testa and inflorescence characters suggesting their inclusion in Landrum’s group. usually below the mid-point. Relationships between these groups do not have high support. scalariform plates absent. but tentatively allied with them. which on this basis. only three (groups 1. testa bony (except some species of Gossia). seven subtribal groupings within Myrteae s. rather than Luma. a mean of 20 to 70 ovules per ovary. and Luma. support for their exclusion comes from their placentation and lack of scalariform perforation plates. If the position of Blepharocalyx salicifolius is correct. ovules arising at a single point on the septum. to a certain extent. ovules inserted along the length of the septum. scalariform plates present. agree with the results presented here. scalariform plates present. Landrum (1981b) suggested that anomalous characters preventing the straightforward placement of Myrceugenia in a subtribal classification may be plesiomorphic and shared by Blepharocalyx. they remain two of the most frequently cited and accurate classifications of Myrteae genera to date.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. Myrrhinium was not sampled in this analysis. ovules at a single point on the septum. are without exception erroneously associated with the majority of genera in whichever group they were associated. 6). placenta subapical and protruding. inflorescence racemose (except some species of Gossia and Decaspermum). if. and Luma and that the three former genera are similar to ancestors of the three previously accepted subtribes of tribe Myrteae whereas Luma is an early offshoot of the tribe. protruding placentation as do two extra-South American genera. Informal subtribal groupings of tribe Myrteae. and its affinities remain unconfirmed. Myrceugenia. In summary. then this genus will share this position. • Suprageneric phylogenetics of Myrteae morphological similarities with Acca . then. Blepharocalyx salicifolius. C-shaped embryo or sometimes spiralled. are identified. that is the South American taxon most distantly related to the remainder of tribe Myrteae. Foliaceous or much reduced cotyledons. Although Myrtus and Austromyrtus possess embryo. C-shaped embryo. and uniflorous peduncles. Plano-convex or fully fused cotyledons. based on clades generated by the combined molecular analyses (Fig. testa soft. based on this analysis. a mean of between 20 and 70 ovules per ovary. 3. a mean of 20 to 120 ovules per ovary (except Pimenta). 1123 “Myrcia group” “Myrceugenia group” “Myrteola group” “Pimenta group” “Eugenia group” “Australasian group” .l. inflorescence racemose (sometimes reduced and appearing fasciculate). Myrceugenia. Landrum (1988) suggested a group associated with Myrteola including genera with myrtoid embryos. Informal group “Plinia group” Character combinations Plano-convex fleshy cotyledons not completely fused. this correlates with our “Myrteola group”. testa soft. placentation protruding or along the length of the septum. with several morphological characters suggesting that this last group in particular might have closer associations with the subtribal groups of clade J. 1966) and McVaugh (1968). and the “Myrceugenia group”. scalariform plates absent. which are not precursors of independent groups. and of all. ovules at a single point on the septum. a mean of between 20 and 70 ovules per ovary. testa soft. usually below the mid-point or inserted along its lower part. however. be distinguished by combinations of morphological characters (Table 8). Of McVaugh’s six groups. McVaugh’s hypothesis that the unplaced genera plus Acca and Myrhinnium were evolutionary lines from less successful proto-myrtaceous ancestors that today remain morphologically similar to ancestral groups is refuted by their emergence here nested in terminal clades. bony testae. those considered most likely to change in future analyses are the positions of Myrtus communis. a mean of nine ovules per ovary. Genera in the “Myrteola group” have scalariform perforation plates and subapical. testa bony or soft. scalariform plates absent. The six genera not satisfactorily placed in any of McVaugh’s groups. The weakly supported placement of the “Myrceugenia group” in the overall topology prevents conclusions from being drawn as to whether it is an early branching lineage of Myrteae. scalariform plates absent. The seven subtribal groups distinguished by the molecular analysis can. testa bony. Despite the shortcomings of the systems of Kausel (1956. identifying subtribal groups that follow outlines of groups identified here. it is this species. taken in their strict sense. scalariform plates absent. Calycolpus is sister to Myrtus. Increased taxonomic sampling representing all Myrteae genera is desirable and further sequence data are required before the tribe can be formally reclassified along these Table 8. Myrtastrum (New Caledonia) and Lenwebbia (Australia). 7) and combinations of morphological characters useful for their circumscription. a mean of five ovules per locule. testa soft. C-shaped embryo. This study suggests that the plesiomorphic characters suggested by Landrum are in fact synapomorphic for Blepharocalyx cruckshanksii. are predicted to fall within it.

with the explanation for the presence of Paleomyrtinaea in North America at such an early period likely due to an extended range of the tribe during warmer periods in the Paleocene and Eocene.R. with long-distance dispersals more likely to explain at least some of the intercontinental movements of Myrteae species than vicariance but the higher costing of these dispersal events in DIVA. the DIVA analysis was performed with Calycolpus in this position. 2004). it is impossible to accurately determine the optimal size of an area in which a taxon might have arisen. Curitíba). and a further eight represented by three or more. Bot.Lucas & al. coarse and 1124 debatable areas into which the distribution of Myrteae has been divided. Gard. pers. The important role of dispersal in the history of Southern Hemisphere biogeography is currently experiencing a re-acceptance (Sanmartín & Ronquist. eight by two. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Curitíba). Character-state weighting for cladistic analysis of protein-coding DNA sequences. Amer. P. Vinicius Souza (University of São Paulo. Additional samples were provided by Martin Gardner (Royal Botanic Gardens.A. Soc. I. — The biogeographical scenario resulting from DIVA supports the hypothesis that Myrteae originated in Gondwana and migrated to South America. with the exception of Blepharocalyx. Floral Development of Selected Myrtaceae . vicariance may be assessed using molecular dating techniques. Linn. Biogeography. Size variations in tracheary cells: I. Gert Hatschbach (Botanical Museum. As a result of the indication (Wilson. & Tupper. 141: 399–436.W. ACKNOWLEDGEMENTS Collecting expeditions critical to this project were supported by the British Airways Community and Conservation programme and relied on Cássia Sakuragui (University of Maringá). Peter Wilson (Royal Botanic Gardens. such as merosity. The molecular results do not conflict with modern concepts of genera. M. 1996. Carolyn Proença (Univeristy of Brasília). potentially removing an unknown number of optimal ancestral area combinations from the analysis. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 lines. and Dion Devey. Steven Jansen and Daniela Zappi. 1993. Joel M.A. further data are required to establish monophyly. Proc. Angiosperm Phylogeny Group (APG II). Jomar Jardim (University of Feira da Santana). and Bruce Holst (Marie Selby Botanical Garden). The consequence of the limitations of our analysis and those resulting from the assumptions of the DIVA program are that biogeographical scenarios presented here can be considered only a starting point upon which future. gymnosperms and angiosperms. the majority are represented by one sample only. The results of DIVA may be further influenced by the geographic scale of this analysis and the large. Acad.A. Efforts to ensure that these samples were from widely separated ends of the taxonomic spectrum add weight to suggestions of their monophyly. 2003.) of a likely relationship between Myrtus and Calycolpus. Barrie. E. Jim Clarkson. Piracicaba).W. Baas. Missouri Bot. We thank IBAMA and CNPQ (Brazil) for granting permits and support. Belsham S. and it should be emphasized that results are based on a topology in which some nodes may be subject to future change. Jean Jacques de Granville (Cayenne Herbarium). W.A. & Mishler. Helpful discussion and guidance at Kew was provided by Félix Forest. prolongation of the calyx or hypanthium and nature of the calyx lobes are of use at the specific level. more refined investigations can be based and from which the probabilities of taxa moving by dispersal vs. with ancestors of several currently widespread genera originating in the south of the continent and moving northwards. de Silva (Botanical Museum. Marie-France Prévost (Cayenne Herbarium). J. 80: 752–766. Bailey. 2003. Parallelism and reversibility in xylem evolution. Journal 17: 351–364. but when sampling is low and. and Marcos Sobral (University of Belo Horizonte). Edinburgh). Novon 14: 380–400. Ancestors of Eugenia and the New Zealand genera. particularly in the case of the Australasian genera. I. but not at the level of genus or subtribe. V. LITERATURE CITED Albert. Chase. also Leslie Landrum (Arizona State University).W. in the first case to South East Asia and Africa and in the latter two genera. to New Zealand.D. comm. Sydney). DIVA also assumes that ancestral species distributions occupied only one unit area. Sydney). Comparison of the molecular results with morphology of the tribe supports McVaugh’s opinion (1968) that the most obvious characters with which to classify Myrteae genera.. 2004. Synopsis of Plinia (Myrtaceae) in Mesoamerica. F. This addition did not significantly change the resulting biogeographical scenario. have apparently originated in southern South America before dispersing. 1918. Rafaela Forzza (Rio de Janeiro Botanical Garden). Of the Neotropical genera sampled. six are represented by one accession only. Arts 54: 149–204. increasing or sometimes decreasing the number of combinations of potential ancestral areas at a given node but not altering the regions involved. Logistical support was provided by Laura Jennings (Royal Botanic Gardens. Lophomyrtus and Neomyrtus.W. B. and although efforts were made to use phytogeographically distinct areas. There is no evidence of a Laurasian colonization of the Americas. A comparison between the secondary xylems of vascular cryptogams. Kew). This pattern is supported here. Ann. Marcos Sobral and Peter Wilson (Royal Botanic Gardens. Labwork at Kew was facilitated by Martyn Powell. & Wheeler.

10: 101–166. A Revision of Myrcianthes Berg (Myrtaceae). Bot. Natl. Evolution 39: 783–791. & Stevenson. 1953.V. L. J.. N. Felsenstein. Bohte. Bot. Laurasian migration explains Gondwanan disjunctions: evidence from Malpighiaceae.str. Landrum. Garcia V. B. New York Botanical Garden. A rapid isolation procedure for small quantities of fresh leaf tissue. Molec. 1: 353–401. Landrum.L.. A. Baas..D. 18: 274–385. J. vol. E. L. & Dadswell. 13: 120–132. Phylog. 8: 521–523. Bot. Flora of the Venezuelan Guayana. Evolution in the Myrtaceae—evidence from inflorescence structure..R. & Johnson. Hollis. Ingle.P. S. 2002.D. Rey et Gravier. vol. Comparative Wood Anatomy: Systematic.P. 19: 143–156. Syst. Phylogenetics of Asphodelaceae Asparagales: an analysis of plastid rbcL and trnL-F DNA sequences. M. A. B. 33: 75–90. 2005. Landrum. & Eguiarte. Treuttel and Würtz.R..A. & Delwiche. Crisp. Proc. 2003. O. Cornell University. thesis. General Secretariat of the Organization of American States. Lista de las Mirtáceaes y Leptospermáceaes Argentinas. A. 1998. Cimino. matK sequences and phylogenetic reconstruction in Saxifragaceae s. Ecological.K. A. F. New South Wales 102: 157–256. Strasbourg. Oxoford University Press.D. F. Missouri Botanical Garden Press. Olmstead. Dunedin [unpublished]. Ann. 1993. Johnson. Ontogeny. Gard.. 1984.. Weighted ancestral area analysis and a solution of the redundant distribution problem. Ph. 11: 155–162. Prof.T. In: von Martius. Hunt Institute for Botanical Documentation. Lilloa 13: 125–149. Myrrhinium. Berry. M.J.K. 1869. Database produced by the Royal Botanic Gardens. Lens.W.S. Davis. O. Four primer pairs for the amplification of chloroplast intergenic regions with intraspecific variation. L. J. De Candolle. Craven. Cabrera. M. Brittonia 49: 508–536. Ark.T.A. & Holst. E. 1828. 2001. Berg.G. C. 1987.K.S. 1985. Landrum..G. Science 294: 2351–2353.N.D. M. 1986. and Luma (Myrtaceae). The anatomy of the timbers of the south-west Pacific area III. Cox. Natl. 71: 700–756. Myrtaceae. Toward defining the course of evolution: minimum change for a specific tree topology. Brummitt. thesis. University of Otago.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. H. F. Berlin. Bull. 11. 2.A. E. & Soltis. 1999. 3: 491–516. Fitch. & Drinnan. The affinities and distribution of the Lower Eocene flora of southeastern North America.R.A. and Evolutionary Aspects of Dicotyledon Wood. Myrtaceae).W. Chase. 53: 129–250. Johnson. Soc.. S.G. MRBAYES: Bayesian inference of phylogeny. Biol. J. M. Jansen. Evol. U. Bioinformatics 17: 754–755.F. Kelchner. Biffin. 1857–1859.W. Huelsenbeck. McCourt. Farris. Biogeografía de América Latina. J. Soc. A successive approximations approach to character weighting. P. R.A. 1947 [1948]. Gard. Kausel. Geol. Pl. 91: 1–481. & Chase. Proc.A. 1988. Linn. Myrtaceae...E. 101: 8833–8837.R. & Briggs. Flowering Plants of the World. De Bruijn. 2000. A. Sci. Legrandia. 1971.. 87: 482–498. Holst.. Lilloa 32: 323–368.C.S. Springer-Verlag.. 99: 6833–6837. Bot. Berry. L. 1981b. vol. Carlquist. 2000. Myrtaceae. Johnson..E. H. Zool. Bot. Landrum. A. & Brummitt.).. Confidence-limits on phylogenies—an approach using the bootstrap. 1988. Linn. Molec. Acad. L. & Ronquist. Louis. Zool. anatomy and systematic significance of ovular structures in the ‘eucalypt group’ (Eucalypteae. 2001. & Willink. U. 16: 21–29. 14 part 1. Sci.J. E. Syst. 1956. [Serie de Biología 13].. A. Syst. Revisio Myrtacearum Americae. Ann.T. Molec. B. The Lower Eocene floras of southeastern North America. Pp. B.. Acad. 207–296 in: Prodromus systematis naturalis regni vegetabilis.. 1125 . Pap. S. R. Bot. Briggs. Bentham. J. Phylogenetic relationships in Nicotiana (Solanaceae) inferred from multiple plastid DNA regions. Berry. Myrtacaeae – Plumbaginaceae. Leitch. 2007. Campomanesia. 1–99 in: Steyermark. U. 2001. Ithaca. 1994.A. C. L. 3. Systematics of Myrteola (Myrtaceae).R. Blepharocalyx. 1986. London.. Syst. & Brummitt. Syst.Y. E.. Phytochem.M. P. A. 2004. (eds. Missouri Bot. New York Botanical Garden..F. D. Yatskievych. Variation in xylem structure from tropics to tundra: evidence from vestured pits. F. G. 1991.G. L. L. no. B. Washington. 1980. Grifo.A. & Kawasaki. E. Surv. Pittsburgh.J.E. Rudall. Evol. Lipsiae. R.R. Paris. Hausdorf. P. Linnaea 27: 1–472. Bell.. 255: 17–39. Ph. Landrum. Harrington. Dictionaire classique d’histoire naturelle. Kausel. K. Heywood. 2001. 2nd ed. E.S. Bronx. • Suprageneric phylogenetics of Myrteae with Emphasis on the Fleshy-fruited Taxa. The phylogeny and geography of Myrceugenia (Myrtaceae). Karol. Myrtales and Myrtaceae—a phylogenetic analysis.G. C. Knapp.E. Proc. Syst. Kew. De Candolle. G. N. Distributions of Vascular Plant Families & Genera. Kausel. Acca. Soc. St.L. Frid.S. M. Bronx. M. Austral. Paris. 1826. & Gadek. Pimenta. S. 1992. 19: 11–15. W. Fleischer. World Geographical Scheme for Recording Plant Distributions. Notes on Myrtaceae. 2004. C. Berg. 1979. Bot.). Ann. Syst. P. Bot. R. Doyle. paired-sites models and evolution of the ITS transcript in Syzygium and Myrtaceae.R. A. L. M. V. D. Ecol. The evolution of non-coding chloroplast DNA and its application in plant systematics. Hamilton. New York [Flora Neotropica Monographs 45]. L. Landrum. 47: 445–456. Pp. Clarkson.J.. Beitrag zur Systematik der Myrtaceen. 86: 935–951. D.P.M. 2nd ed. S. J. 1997.H. Proc. Structural partitioning. Reeves.S. Pando. DC. Gasson.W. S. Myrtaceae. New York [Flora Neotropica Monographs 29]. Evol. A monograph of the genus Myrceugenia (Myrtaceae). 7. & Donoghue. Flora Brasiliensis. A. 43: 124–139. Notas Mirtológicas.K. 20: 406–416. Missouri Bot. vol. L. Variability of embryos in the subtribe Myrtinae (Myrtaceae).B. The genera of Myrtaceae in Brazil: an illustrated synoptic treatment and identification keys. P. Amer. & Smets.F.A. 1966. Philos. K. Chamguava: a new genus of Myrtaceae (Myrtinae) from Mesoamerica. Seeing phylogenetics through the trees. Brummitt. 1914. M. (ed. Brittonia 33: 105–129. & Doyle.A. Mathews. 1855–1856.R. Plant Taxonomic Database Standards. Phylog. 1969. 1981a. J.A.P. New York. 1915. L. Syst.L.

E. Bot. S. Syst.. Gaps as characters in sequence-based phylogenetic analyses. The occurrence of scalarifrom perforation plates and helical vessel wall thickenings in wood of Myrtaceae.E.S. Snow. Sinauer. Syst. Lucas. 13: 1–18. & Prantl. W. a new genus of permineralized myrtaceous fruits and seeds from the Eocene of British Columbia and Paleocene of North Dakota. . M. Molecular evolution of 18S rDNA in angiosperms: implications of character weighting in phylogenetics analysis. & Axelrod D. Phylogenetic relationships of Myrtaceae as inferred from nrDNA ITS sequence data [Abstract 795]. multiple. Mueller and Lithomytrus F. Prance.B. Tropical American Myrtaceae.systbot. (eds. P.L. & Woodgyer.W. Nic Lughadha.H.. Available at: http:// mrbayes.. Utah. 1991. 1999. K. Syst.S. Myrtoideae (part 1). Phylogenetic relationships within Zygophyllaceae based on DNA sequences of three plastid regions. P. 1997.L.M. A study of the origin of central Brazilian forests by the analysis of plant species distribution patterns. uu. 1827. A resolution of the Eugenia-Syzygium controversy (Myrtaceae). DIVA. & Baas. Conspectus of Australasian Myrtinae (Myrtaceae). C. Belsham. C. Bot. 1993. J.E. G. K. Gard. Canad. 1974.F. Wojciechowski. Die natürlichen Pflanzenfamilien. P. 16th ed. Algrizea (Myrteae. N.M. Linnaeus. Simmonds. 2006. Bioinformatics 14: 817–818. Salywon. Computer program and manual available by anonymous FTP from Uppsala University.edu/manual. Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography. vol.M.F. I. 1976. Myrtaceae – subfam. Orlovich. 1998. 25: 371–384. Missouri Bot. 21: 567–586.A.). The Plantbook. 2000. Myrtaceae. & Swensen.R.R. 2005. New York Bot. C. & Butterfield.A.. S.. Chase. 1968. Biol.A. D. Sanmartín. Pp. notes on generic concepts and descriptions of previously unrecognized species. Consideraciones sobre el problema de la clasificación de los elementos traqueales no perforados de las diccótiledoneas y en especial de algunas Mirtáceas. M.T. Rodrigues. 61: 539–673. P. 52: 141–194.fsu. 33: 541–555.T. & Crandall.W. MrBayes 3. K. Darwiniana 20: 476–490. D. Linnaea 2: 504. and combination perforation plates in the vessels of New Zealand woods.E. J. Amer. & Wilson. Southern Hemisphere biogeography inferred by event-based models: plant versus animal patterns. Lucas. Kluwer Academic Publishers. 31. F. Molecular Systematics of Plants. MacClade Version 4. R.G. 188–210 in Soltis. Systematics of Austromyrtus.uu. London]. Lenwebbia and the Australian species of Gossia (Myrtaceae).A.W. D. & Soltis. J. K. Bot. N. R. W.A. Kron. vol. Csizmadi. Bull. (eds.P. Ann. Syst. Oliveira-Filho. B..T. Gottingae. F. Bot. Comparative anatomy and morphology of Psiloxylon and Heteropyxis. Stockholm [Facsimile edition. Cambridge University Press. Snow. 1957–1959.. & Sawvel. Occurrence of simple. F. Stockey. McVaugh. A. Ronquist. E. N. 1978. 25. 1126 Raven.G.. Draft 5/26/2005.php. Sumtibus Librariae Dieterichianae.J. Evol.A.J. & Doyle. 1893.A.1 Manual. Wood anatomy of the dicotyledons indigenous to New Zealand. C. E. P.) 1825. 1998. Ray Society.1. with special emphasis on Zygophylloideae.B. Systematic and cladistic studies of Myrtella F.1. Schmid. Ragonese. D. 2004. Schmid.A. Edinburgh J. & Fontes. Botany of the Guayana Highland—part III. D. & Ochoterena. 1993. Oliveira-Filho. Sheahan. & Ronquist. Sakuragui. A. Morton. 2000. & Chase.P. Snow. & da Silva. D. Engelmann. B. Syst.G. 1995. R. Snow. The discovery and importance of multiple islands of most-parsimonious trees. Occurrence of helical thickenings in the vessels of New Zealand woods. (ed. Sobral. Available at: http://www.P. J. I.J. 2000. Syst. 47: 239–271. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 Linnaeus. 2004.E. McVaugh. Massachusetts. A molecular evaluation of the monophyly of the order Ebenales based upon rbcL sequence data. A. F. vol. E.W. 57–105 in: Engler.. G. M. Zool. M. 1975. Maddison.R. J. De plantis in expeditione speculatoria Romanzoffiana observatis disserere pergunt Adelbertus de Chamisso et Diedericus de Schlechtendal. 32: 320–326. Guymer. R. Acta Amazônica 6: 9–35. 46: 195–203.Lucas & al. C. Leipzig. A. Pl. Biol. Angiosperm biogeography and past continental movements.. C. Paleomyrtinaea. MA. Syst. 1969. A.. Patel. & Maxwell. 2003.org/. 3 (7). Syst.A. & Butterfield.L. Inventário florestal de um hectare de mata de terra firme km 30 das estrada Manaus-Itacoatiara. Meylan.M. The genera of American Myrtaceae—an interim report. Norwall. & Maddison. & Ratter. Sprengel. Taxon 17: 354–418. Myrtaceae. Bot. Meylan. 26: 733–742. Napaea 9: 13–41. B. ftp.J. Posada.P. 1990. Sinopse de Myrciaria (Myrtaceae). Botany 2004. Ronquist. M. H. Nic Lughadha. M.. E.se.. Syst. Schmid. Gard. 1980. S. 53: 216–243. Phylogenetic patterns in the fleshy-fruited Myrtaceae—preliminary molecular evidence. 59: 423–436. 18: 79. Bot.botanyconference.P. 1753 Species plantarum. and the subfamilial and tribal classification of Myrtaceae. Mem. M. Schlechtendahl. R. Huelsenbeck.R. D. Proença.0 for Macintosh. 1963.csit. Snowbird. Cambridge. Snow. 1976. & van der Mark. 2001.se or ftp.. & Guymer. Edinburgh J. Sunderland. M. Bot. New Zealand J.A. Ronquist. R. 1984. Soltis. L. M.). 2000. McVaugh. 2005. Bot. Proença... 2. v. Soltis. 2nd ed. G. Mabberley.M. Monogr.N. D. 1996. Pigg.2004. 81: 139–46. G. B. N. 71: 1–9. MODELTEST: testing the model of DNA substitution. 251: 35–51. Systema vegetabilium. Chase.J.B. 49: 369–381. 2: 471–488. Niedenzu. Biotropica 32: 793–810. Bot. J. 5: 197–215. 2000. Kew Bull. Revision of Australian species of Uromyrtus (Myrtaceae) and two new combinations from New Caledonia. Myrtaceae): a new genus from the highlands of Brazil. Patterns of floristic differentiation among Atlantic forests in SE Brazil and the influence of climate. New Zealand J. 1972. F. Biol.M.P. Fieldiana 29: 145–228. W. Bot.P. 65: 1–95. 40: 315–328. 1996. Maddison. New Phytol. P. G. A revision of Siphoneugena Berg. K. N. 22: 647–654.I.F. Bot. Mueller (Myrtaceae). Pp.M. Taxon 29: 559–595.. July–5 August 2004. R. Salvius. 1997. & Guymer. & Landrum. R. Syst. 1995. Austrobaileya 5: 173–207.A.

K (AM234086. Bailey & C. AM489975). Brazil. Eugenia punicifolia (Kunth) DC. AM489889.V.S. G. OTA (AM234138. K (AM234139. Dawson. Lucas 201. AM489834). AM489900. & Quinn. Syst.) D. RBG Kew. 28: 769–774.J. AM489901.J. AM489817. AM489904. Fench Guiana. 1998. Cult. 2002. Berg. S. 89: 26–32. J.) Nied. AM489807. AM489826). AM489911. • Suprageneric phylogenetics of Myrteae Stevens. Blepharocalyx cruckshanksii (Hook. Accessed April 2006. Calyptranthes concinna DC. Pires. U. Eugenia uniflora L. M.C.. Cult. Brazil. Gadek. Flavour and Fragrance Journal 7: 247–251. 2001.L. D. Regni. Eugenia stictosepala Kiaersk. AM489811). Berg. Eugenia sulcata Spring ex Mart. Calyptranthes lanceolata O. AM489979). Acmena smithii (Poir. 251: 21–34 Weyerstahl. Sci. Swofford.. K (AM234104. psbA-trnH. I. Lucas 283. RBG Melbourne.. Campomanesia pubescens O. AM489810.A. Acad. K (AM234078. XIII. Fench Guiana.. R. 99: 16138–16143. AM489919. Australia. AM489916.. AM489837). Cult. Pl.J. Kew. K (AM234077. & L. RBG Kew. AM489830). Udovicic 177. AM489982). Belsham M78.C.P. Nov.) D. Sertum antillanum. K (AM234090. AM489821. Wilson. Evol. Gomidesia tijucensis (Kiaersk. & Sugiura. Cult.. E (AM234070. AM489836.A. Glazko. K (AM234092. Eugenia florida DC. RBG Melbourne. & Botha. 81–105 in: Scotland. Farinaccio & Costa 1669. AM489895. Cult. Legrand & Kausel. clocks. 165: 85–105. AM489822). Amomyrtus luma (Molina) D. Berg. RBG Kew. 82: 3557–3561.. Eucalyptus tetragona F.A. Lucas 68.H.0b2. Campomanesia guazumifolia (Cambess. OTA (AM234132. Brazil. AM489920. Belsham M37.. Cult. Sm. AM489888. Gomidesia schaueriana O. AM489907. AM489825. J. O’Brien. RBG Melbourne.. Tobacco chloroplast tRNA Lys (UUU) gene contains a 2. Belsham M85. C. 1921. K (AM234067. (Myrtaceae). Callistemon comboynensis Cheel.W. Pl. M. Y. Theor.). 2006. Gomidesia affinis (Cambess. K (AM234113. Brazil. voucher. & Landrum. AM489988). AM489980). Coding Characters of Phylogenetic Analysis.. Lucas 106.S. AM489902. On characters and character states: do overlapping and non-overlapping variation. Sun. AM489989). Eugenia latifolia Aubl. AM489814.. Appl. World Checklist of Myrtaceae. L. Sci. Relationships within Myrtaceae sensu lato based on a matK phylogeny..) Snow & Guymer.. U. AM489917.M. Belsham M86. RBG Edinburgh. Zwickl. P. Marschall. Molecular phylogenetic analysis of Eugenia L.. Campomanesia sp.. AM489812). Berg. White) Snow & Guymer. Zappi 305. AM489819.G.. Amomyrtus Luma (Molina) Legrand & Kausel and of Amomyrtella guili (Speg. E. Dunedin Botanic Gardens. K (AM234088. GenBank accession numbers (ITS. Syst. Brazil. AM489912. Natl. Victoria. Lucas 78. AM489986). AM489906. & Gardner.. and continents: historical and biogeographical analysis of Myrtaceae. T. Acad. Lucas 83. A..J. & Wilson. Eucalyptus perriniana F.) Proença & Nic Lughadha. van Wyk. AM489892. (eds. Taylor and Francis. Heath. Brazil. AM489910. AM489809.W.T.. Blepharocalyx salicifolia O. AM489985). 1127 . 2000. K. Cult. S. Belsham M79. Wichman. Sinauer. with emphasis on southern African taxa. Lucas 62. Y. K (AM234140. K (AM234110. Austromyrtus dulcis (C. Clades. AM489987).G. C. Algrizea macrochlamys (DC. Brazil. Myrtaceae revisited: a reassessment of infrafamilial groups. AM489824). Wilson. S. Suzuki. OTA (AM234137.TAXON 56 (4) • November 2007: 1105–1128 Lucas & al. PAUP*. Amomyrtus meli (Phil. AM489813. Sunderland. Evol. Repert. & Masatoshi. Shinosaki. Berg. AM489908. AM489897. M. K (AM234084. Wright. 1985.A. Stepping stones to Hawaii: a transequatorial dispersal pathway for Metrosideros (Myrtaceae) inferred from nrDNA (ITS + ETS). 2001. Fench Guiana. P. K (AM234111.L. Perry. The Board of Trustees of the Royal Botanic Gardens. Decaspermum humile (G. RBG Kew.. Zjhra. AM489981). Lucas 207.. Muell. RBG Melbourne. Eugenia langsdorffii O. & Quinn. Sugita. P. v. British Virgin Islands. AM489893. E (AM234073. AM489820). & Arn. Berg) Burret. Appendix.) D.D.M. 2005. 25: 361–371.R. Vouchers for accessions used in the phylogenetic analysis. Bot. AM489913. British Virgin Islands.M. Legrand. AM489976). AM489909.R. RBGE 1996-1065. RBGE 1996-1083. Cult. AM489905. Genet. Berg.. 2005.M. Lucas 64. Brazil. J. 2004. matK ).. Massachusetts. AM489918. T. A. Calyptranthes kiaerskovii Krug.M. Natl. & Hulbert. G. RBG Edinburgh. AM489974). AM489890.) Merr.M. & Urb.. AM489828. AM489833). Prevost 4724. Belsham M77. AM489983).J. AM489891. Conti. K (AM234076. Giulietti & Harley 1648.org/ wcsp/myrtaceae. OTA (AM234133. AM489831). M. AM489899. Constituents of the leaf extract of Amomyrtus meli (R. E (AM234069.G. AM489808. AM489823. AM489977). RBGE 1998-073D. Spec. 1992.. Phylogenetic Analysis Using Parsimony (*and Other Methods). Don) A. AM489903. Phylogenetic relationships of the dwarf boas and a comparison of Bayesian and bootstrap measures of phylogenetic support. Cult. AM489914. P. M. Da Silva & Farias 4528. RBG Sydney. Urban. Wilson. Legrand. K (AM234103. R. F. Calyptranthes thomasiana O. Prevost 4707. provenance. D. AM489973). A. AM489835.F.5-kilobase-pair intron: An open reading frame and a conserved boundary sequence in the intron. London. J. Heslewood. T. Brazil. Acca sellowiana (O. OTA (AM234128. Amer. AM489832). AM489818). ETS. 251: 3–19. and relatives in the Southern Hemisphere. 88: 2013–2025. AM489898.Z. K. AM489815. D. http://www. M. M.. Gomidesia flagellaris D. Liang. AM489816). AF190381. K (AM234126.M. Lucas 206. White) L. Phylogenetic analysis of Sorghum and related taxa using internal transcribed spacers of nuclear ribosomal DNA. Brazil.kew.T. Van der Merwe. Litt. with GenBank numbers. Proc. Yong. Legrand & Kausel. Brazil. H. Pollard 1195. Phylog. Homology and Systematics. K (AM234112. MELU (AF190364. AM489827).) O. & Hillis. Brazil. Biogeogr.4. Overcredibility of molecular phylogenies obtained by Bayesian phylogenetics. OTA (AM234131. P. Evol. ex Rodway. O’Brien.H. Cult. & Pennington. Legrand. K (AM234091. Skinner. 1994. Lucas 74. K (AM234106. Pollard 1194.F. C.. K (AM234089. Int. 17: 402–404.. Zappi 406. AM489894.. K (AM234087. Cult.E. Sci. AM489896.. Cult. Berg. A. M. Brazil. Species. Gossia hillii (Benth. Nepokroeff. Proc. Muell. Pl. Lucas 84. RBG Sydney. 2002. Gossia inophloia (J.S. Molec... J.) Kausel. AM489978). N. Cult. Lucas 205.. RBG Edinburgh. AM489984). Philippi) Legrand & Kausel. Cult. K (AM234105. Scott. Anetholea anisata (Vickery) Peter G. herbarium. D. Veg. Belsham M82. Sytsma. Vochysiaceae. AM489829. P. morphology and molecules all yield data of the same value? Pp. Wilcox. AM489915.

Dunedin Botanic Gardens. E (AM234068. Myrcia laruotteana Cambess.W. Myrcia rostrata DC. Myrcia saxatilis (Amsh. K (AM234108. Irian Jaya. AM489869. Legrand. Udovicic 337. AM489970. Berg) D. Continued. Cult. Mazine 796. Berg. AM489858. RBGE 1989-1714C. Rhodamnia argentea Benth. Pimenta pseudocaryophyllus (Gomes) Landrum. AM489852). Siphoneugena densiflora O. Tepualia stipularis (Hook. AF190368. Brazil. AM489937. AM489859). AM489932. AM489959. Neomyrtus pedunculata (Hook. K (AM234124.) Burret. Psidium cinereum Mart. AM489991). K (AM234094. AM489413. AM489880). Holst. ex J. K (AM234148.R. Myrcianthes pungens (O. Cult.. AM489994). Cunn. AM489997). AM489967. French Guiana. Metrosideros perforata (J. OTA (AM234143. Cult.) W. Lucas 73. Lucas 210. K (AM234114. AM489958. RBG Kew. Waterh. K (AM234081. Belsham M83. AM489842. RBG Kew. AM489927.) O. RBG Edinburgh.. Brazil. AM489872. AM489870). K (AM234093. AM489874. AM489962. Tressens & al 5481. E (AM234071. Myrciaria cauliflora O. Cult.) Landrum. Legrand. Brazil. Fench Guiana. AM490010). E (AM234074.R. AM489877). AM489839.. West ex Willd. Belsham M81. AM490637.. Cult. Myrcianthes pseudomato (D. AM489949... AM489955.) McVaugh. AM489855.Lucas & al. Berg) D. Legrand. K (AM234130. AM489957. K (AM234109. RBG Melbourne. Myrcia bicarinata (O. AM489940. K (AM234120. AM490003). AM489863).. AM489934. AM489569). AM489950. Garnock-Jones.) Alston. Proença. E (AM234072.. RBG Kew. Holst 8866. Brazil. Berg. AM489862). ESA (AM489411. Berg. AM489926.. AM489925.) J. Lucas 98. Neomitranthes cordifolia (D. Legrand) McVaugh. K (AM234100. AM489885. RBG Edinburgh. Berg. Dunedin Botanic Gardens. AM489948. Syzygium jambos (L. Cunn. AM489849).) DC. AM489878.. Cult. Lucas 211. AM489936. Myrcia pubipetala Miq. Belsham M31. RBG Edinburgh. K (AM234080. AM490012). AM489921. Forst.) Merr. AM489570)..) Miq. Ugni molinae Turcz. AM490001). K (AM234117. Cult. Lucas 107. Lucas 208. AM489963.. AM489996). AM489860. RBG Kew.. AM489995). RBGE 1995-2370A. Brazil. OTA (AM234146. Cult.. Johns s. AM489572. Brazil. Psidium cattleianum Sabine. AM489882. Cult. Myrcia fallax (Rich. AM490018). Myrciaria aff... floribunda (H.) Kausel. Myrceugenia planipes O. Marlierea obscura O. Prevost 4716. Dunedin Botanic Gardens. St.L. Victoria. Brazil. AM489972. Cruz & Cordeiro 195. Xanthomyrtus montivaga A. Lucas 65. ESA (AM489412. AM489866). K (AM234101. Moore. AM489886). Lucas 85. RBG Edinburgh. AM490016). AM489838). Mello Silva 1713. AM490015). Berg.T.) Kausel. Plinia pauciflora M. Belsham M72. & Arn.) DC. Lucas 82. ex A. AM489998). AM489876. Lucas 284. AM489840. Lophomyrtus obcordata (Raoul) Burret. Myrcia laricina (O. Forst. Lucas 71. AM490005). Rhodomyrtus psidioides (G. AM490009).) Kausel. K (AM234107. Argentina. Irian Jaya.) Burret. Lophostemon confertus (R. AM490004). AM489960. Cult. Beck 9667.. RBGE 1996-1096. Cult. AM489956. K (AM234098. AM489935. AM489844). Kawas. Da Silva & Farias.) O. Cult. AM489864). AM490014). AM489966. Forster 1011. Belsham M69. Sykes & P. OTA (AM234134. AM489954. Xanthomyrtus compacta Diels. E (AM234075. Lucas 212. • Suprageneric phylogenetics of Myrteae TAXON 56 (4) • November 2007: 1105–1128 Appendix. Brazil. Australia. Cult. Bolivia. K (AM234116.) A. K (AM234123. AM489851). Lucas 63. RBG Kew. RBG Edinburgh. RBG Edinburgh. OTA (AM234144. Belsham M41. Cult.) Peter G. Lucas 209. Lucas 70. Berg. AM489952.) DC. Lophomyrtus bullata (Soland. AM489922. RBG Kew. 4535. Brazil. Cult. Lucas 61. AM489865. K (AM234149.) Burret. Lucas 88. E (AM234096. AM489854). MELU (AF048897. Lucas 214. Dunedin Botanic Gardens. Brazil. Myrcia venulosa DC. Brazil. Brazil. AM489968. Berg. Brazil.. & G. AM489951...) DC. AM489845. RBG Edinburgh. Lucas 213. AM489928.R. Brazil. AM489992).) Burret. K (AM234115.) Griseb. AM489931. Marlierea eugeniopsoides (Kausel) D. RBGE 1998-0725G. AM489933. AM489945. Lucas 161. AM489884). Octamyrtus pleiopetala Diels. AM490007). Cult. AM489964. AM489881). AM489938. AM489887). Brazil. Legrandia concinna (Phil. Myrteola nummularia (Lam. E (AM234102. RBG Melbourne. Berg. AM489942. ESA (AM489410. AM489879. AM489939.J. Brazil. AM490000). K (AM234099. K (AM234125. AM489875. OTA (AM234129. K (AM234121. Berg. Syzygium maire (A. Belsham M42. Dunedin Botanic Gardens. RBGE 1998-2353C. Myrceugenia alpigena (DC. Lucas 16. Lucas 86. AM489971. RBG Kew. Marlierea suaveolens Cambess. AM489846). RBG Edinburgh. Myrcia tomentosa (Aubl. f. K (AM234141. AM489861. Cult. AM490638. AM489847. AM489993). K (AM234079. Pimenta dioica (L. Cult. Mazine 957. K (AM234085. AM489853. Cult. AM489961. AM490013). AM489946. Legrand) D. RBG Kew. OTA (AM234127. AM489843.J. Siphoneugena guilfoyleana C. Pimenta racemosa (Mill.. AM489944. Brazil. 1128 . AM490017). AM489941. Rhodamnia rubescens (Benth. Myrcia acuminatissima O. Luma chequen (DC. Brazil. Brazil. AM489868). AM489969. Cult. AM489943. Rich. Br. RBGE 1998-0662. Belsham M84. AM489856. & G. K (AM234135. AM489867). Cult. K (AM234118. Cult.n. AM489965. Don) Benth. Scott. K (AM234142. Myrcia multiflora (Lam. AM489999). Lucas 167.. K (AM234122. Berg) Burret ex Luetzelb. Myrcia coumeta (Aubl. AM489414. Cult. AM489990). OTA (AM234145. AM490011). AM489883). AM489571). Wilson & J. French Guiana. Myrceugenia myrcioides O. Mello Silva 1705. Forst. AM489850. AM490002). AM489871. Irian Jaya. K (AM234119. AM489857). K (AM234082.. Soares 752. Brazil. AM490006). OTA (AM234136. E (AM234095. AM489929. F. AM490008) Myrtus communis L. Brazil. Myrceugenia ovata O. RBGE 1999-0656. & B. AM489947. Edwards 4213A. K (AM234097. Dunedin Botanic Gardens. Cult. AM489848. ex DC. K (AM234147. AM489923. Myrceugenia leptospermoides (DC. Cult. Mazine 1050. AM489873). AM489841. AM489930. RBGE 1998-1561B. Luma apiculata (DC.-Hil. Legrand. Myrceugenia lanceolata (Juss. Leptospermum scoparium J. AM489953. K (AM234083. AM489924.

You're Reading a Free Preview

/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->