| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics and Phytogeography |
2 Academy of Natural Sciences, 1900 Benjamin Franklin Parkway, Philadelphia, Pennsylvania 19103 USA 3 Washington University, Department of Biology, One Brookings Drive, Campus Box 1229, St. Louis, Missouri 63130 USA
Received for publication 27 April 2007. Accepted for publication 14 November 2007.
ABSTRACT
Swartzia (ca. 180 spp.) is a characteristic and diverse element of neotropical rainforest tree communities. As such, it has been identified as a focal group in studies of evolutionary diversification and community assembly in rainforests. However, progress is impeded by the lack of a phylogenetic hypothesis for the genus and its close relatives, which together constitute the descendents of one of the first branches of the papilionoid phylogeny. Here we present a molecular phylogenetic study with extensive sampling of species of Swartzia and with complete sampling of genera of the "swartzioid" clade. The results, based on analysis of chloroplast (atpB–rbcL, trnL intron, and trnL–F) and nuclear (AAT1 and ITS) DNA sequences, add substantially to our understanding of relationships within this diverse group and call for taxonomic changes, particularly within Swartzia. The monophyly of a redefined Swartzia is weakly to moderately supported. Within Swartzia, the analyses identify 11 nonoverlapping subclades, few of which correspond exactly to previously published taxa. The analyses support the recent segregations of Bobgunnia from Swartzia and of Trischidium from Bocoa, as well as the resurrection of the monospecific genus Fairchildia. The analyses identify a "core swartzioid" clade comprising the remainder of Swartzia, Bocoa, and Candolleodendron.
Key Words: Fabaceae • Leguminosae • neotropics • Papilionoideae • phylogeny • Swartzia • swartzioid • taxonomy
The neotropical tree genus Swartzia Schreb. (Papilionoideae) contains approximately 180 species distributed throughout the lowland neotropics from southern Mexico and the Caribbean islands to southern Brazil and Bolivia. Although it occurs in a variety of habitats, Swartzia is especially prevalent in lowland rainforests. In parts of lowland Amazonia, for example, 10 or more species of Swartzia can be found growing in close sympatry (Vásquez-Martínez, 1997
; De Oliveira and Mori, 1999; Mesquita and Hopkins, 1999). The genus is characterized morphologically by its unusual flowers, which in the majority of species possess an entire calyx, a single white or yellow petal, a stalked ovary, and dimorphic stamens.
Elucidation of the phylogeny of Swartzia and closely related genera of the so-called "swartzioid" clade of legumes is particularly important for at least two reasons. First, because of its species richness and ubiquity in lowland neotropical rainforests, Swartzia has been selected as a focal group in ongoing studies of the evolution of the neotropical flora (e.g., Torke, 2006; R. T. Pennington, Royal Botanic Garden, Edinburgh, personal communication; P. V. A. Fine, University of California–Berkley, personal communication), which in the wet lowlands is dominated by massively radiated woody plant clades, including Swartzia (De Oliveira and Mori, 1999; Richardson et al., 2001; Erkens et al., 2007). Phylogenies of these species-rich clades are needed to assess the degree to which phenomena such as phylogenetic niche conservatism, ecological diversification, vicariance, long-distance dispersal, and key innovations have impacted speciation rates and community assembly in neotropical rainforests (Richardson et al., 2001; Erkens et al., 2007). Second, the swartzioid genera constitute the living descendents of one of the first branches of papilionoid phylogeny (Ireland et al., 2000; Pennington et al., 2001; Wojciechowski et al., 2004; Lewis et al., 2005). As such, the study of phylogenetic relationships among swartzioid genera is likely to have important implications for understanding the early evolution of subfamily Papilionoideae, by far the largest and most economically important of the three traditionally recognized legume subfamilies (Lewis et al., 2005).
Phylogenetic relationships within Swartzia are poorly understood. In the most recent monograph of the genus, Cowan (1968) divided Swartzia into two sections, two subsections, and eight series (Table 1). Although these infrageneric groupings were created mostly to facilitate species identifications and were not necessarily intended to correspond to monophyletic groups, Cowan did attempt to order species in his monograph along "putative phylogenetic lines" (i.e., species complexes). He also made a number of speculations on morphological character polarities within Swartzia but was unable to propose a hypothesis of relationships among the major morphological complexes of the genus. In addition, he left a certain amount of ambiguity as to the limits of these complexes in his keys to the genus, which sometimes placed the same species in more than one equally ranked infrageneric group.
|
; Hutchinson, 1964; Cowan, 1981b; Polhill, 1994; Ireland, 2005), the genus forms the chief element of tribe Swartzieae, which has been repeatedly shifted between subfamilies Caesalpinioideae and Papilionoideae (see Ireland et al., 2000; Ireland, 2005). Molecular phylogenetic studies support the placement of Swartzieae in Papilionoideae but show that the tribe as traditionally defined is polyphyletic, with some of the constituent genera being most closely related to members of tribes Dalbergieae, Dipterygeae, and Sophoreae (Doyle et al., 1997; Ireland et al., 2000; Pennington et al., 2001; Wojciechowski et al., 2004).
However, these studies, albeit with uneven sampling of genera, have consistently resolved a "swartzioid" clade, composed of Swartzia and five much smaller genera: Ateleia (20 species), Bobgunnia (two species), Bocoa (three species), Cyathostegia (one species), and Trischidium (five species recently segregated from Bocoa, see Ireland, 2007). With the addition of the monospecific genus Candolleodendron, which although not sampled in molecular studies has strong morphological ties to Swartzia and Bocoa (Cowan, 1966; Herendeen, 1994; Polhill, 1994; Ireland, 2005), the "swartzioid" clade comprises six neotropical genera and a single African genus, Bobgunnia. Molecular phylogenetic analyses have resolved Trischidium as the sister taxon to a clade grouping Ateleia and Cyathostegia and suggest a close relationship between Swartzia and Bocoa, but other relationships among swartzioid genera have not been fully resolved (Ireland et al, 2000; Pennington et al., 2001; Wojciechowski et al., 2004).
The current study employs phylogenetic analysis of DNA sequence data to test the monophyly of Swartzia, to clarify relationships among its species, and to determine the phylogenetic positions of Swartzia and closely related genera within the swartzioid clade. Taxonomic implications and morphology are discussed in the context of a new phylogenetic hypothesis.
MATERIALS AND METHODS
Taxon sampling and molecular markers
Taxon sampling spanned the morphological and taxonomic diversity of the swartzioid legumes, with particular emphasis on Swartzia. DNA sequences were gathered from 110 individuals, representing 15 genera and 96 species (Appendix). We sampled 76 of ca. 180 species of Swartzia, including multiple exemplar species of each of the infrageneric groupings of Cowan (1968) (Table 1). Within these groupings, species were selected to maximize coverage of morphological diversity. One to six species of all other genera of the swartzioid clade were sampled, including the previously unsampled genus Candolleodendron. In total, 88 of ca. 212 (42%) swartzioid species were sampled in at least one of the molecular data sets. Also included were eight outgroup genera (each represented in the sample by a single species) traditionally placed near the various swartzioid genera in tribes Swartzieae and Sophoreae but falling outside the swartzioid clade in recent molecular studies (Doyle et al., 1997; Ireland et al., 2000; Pennington et al., 2001; Wojciechowski et al., 2004).
DNA sequence data were collected from the atpB–rbcL spacer, the trnL intron, and the trnL–F spacer, three noncoding chloroplast regions, from the low-copy nuclear gene encoding aspartate aminotransferase (AAT1), and from the internal transcribed spacer region of nuclear ribosomal DNA (ITS: ITS1, 5.8s, and ITS2). With the exception of AAT1, these regions have been shown to be useful at resolving relationships among and/or within papilionoid genera (e.g., Hurr et al., 1999; Wojciechowski et al., 1999; Ireland et al., 2000; Pennington et al., 2001; Doi et al., 2002; Ellison et al., 2006).
For assessing sequence variation among reiterated ITS copies and possible multiple copies of AAT1, molecular cloning of the two loci was performed. ITS was cloned in one or two individuals of each of 11 species of Swartzia, (S. aptera, Torke 213; S. arborescens, Torke 308; S. bannia, Torke 347; S. grandifolia, Redden 3204; S. latifolia, Torke 329; S. leiocalycina, Torke 304; S. panamensis, Torke 381; S. pendula, Redden 3558, Torke 266; S. polyphylla, Torke 233, Torke 256; S. sp. nov. G, Torke 301; S. sprucei var. tesselata, Redden 3205, plus one individual of Candolleodendron brachystachyum, Torke 170). The sample for AAT1 was identical, except that it included only one individual of S. polyphylla, Torke 256, and did not include S. panamensis (see Appendix for ITS and AAT1 GenBank accession numbers).
DNA sequence data and sequence alignment
Genomic DNA was extracted from leaf material field-collected in silica or from herbarium specimens using a modification of the CTAB method of Doyle and Doyle (1987) or a VIOGENE Plant Genomic DNA Extraction Miniprep System (Viogene USA, Sunnyvale, California), following the manufacturer’s protocol. PCR amplifications were performed in 20 µL reactions on a PTC-100 Thermocycler (MJ Research, Waltham, Massachusetts, USA) using Taq or KlenTaq-LA DNA polymerases. Cycling conditions for the three chloroplast regions were 2-min denaturation at 94°C; followed by 40 cycles of 1 min at 94°C, 1 min at 50°C, and 2 min at 72°; and ending with a 7-min extension at 72°C. Those for the two nuclear regions were 2-min denaturation at 92.5°C; followed by 35 cycles of 45 s at 92.5°C, 45 s at 50–55°C, 55 s at 70.5°C; and ending with a 7-min extension at 70.5°C. Amplification of the atpB–rbcL spacer used the BO1 and BO2 primers of Hurr et al. (1999), or the BO2 primer and an internal primer (5'–GCAAGTTGATCGGTTAATTC–3') were used to amplify only the portion included in the analyses. The trnL intron and trnL–F spacer were amplified with the c and d primer pair and the e and f primer pair of Taberlet et al. (1991), respectively. Amplification of ITS employed the ITS1, ITS2, ITS3, and ITS4 primers of White et al. (1990). AAT1 was amplified with the AATX5F and AATX7R primers of Strand et al. (1997). PCR products were purified using a Geneaid Gel/PCR DNA Fragments Extraction Kit (Geneaid Biotech, Bade City, Taiwan) or ExoSap and were labeled with a Big Dye Terminator Kit (Applied Biosystems, Foster City, California, USA). Sequencing was carried out on an ABI 3130xl Genetic Analyzer (Applied Biosystems) or on a MJ Research BaseStation with the primers mentioned previously. Sequences are deposited in GenBank (Appendix).
Sequences were aligned manually in Se-Al (Rambaut, 1996). Indel characters that could be scored unambiguously were treated as presence/absence characters using the "simple indel coding" method of Simmons and Ochoterena (2000) as implemented in the program GapCoder (Young and Healy, 2003). Taxon sampling overlap among data partitions was high within a core swartzioid group including Swartzia, Bocoa, and Candolleodendron. Because of alignment problems and difficulties encountered in sequencing the two nuclear regions, sampling of most other genera in the data matrix was restricted to cpDNA sequences, and the phylogenetic analysis of all swartzioid genera was carried out using only cpDNA data. The aligned data matrix, including cpDNA, AAT1, ITS, and indel data partitions, is available in TreeBASE (website http://www.treebase.org).
Molecular cloning
Ligations were performed with the pGEM-T Easy Vector System (Promega, Madison, Wisconsin, USA) following the manufacturer’s protocol. The resulting products were transformed to JM-109 Competent Cells (Promega). Colonies were cultured overnight at 37°C on LB (Luria-Bertani medium) ampicillin/IPTG (isopropyl β-D-1-thiogalactopyranoside)/X-gal (5-bromo-4-chloro-3-idolyl beta-D-galactoside) selective plates. For ITS, eight insert-containing colonies (identified by white color) were sequenced from each individual. Six colonies per individual were sequenced for AAT1 (only four cloned sequences were obtained from Torke 256).
Phylogenetic analyses
Phylogenetic analyses employed maximum parsimony and Bayesian approaches. Portions of the data matrix with alignment ambiguities and/or poor-quality sequence data were excluded from the analyses. Based on preliminary results (data not shown), which tended to show small to moderate increases in resolution and clade support when indel characters were added to the data matrix, the analyses presented here used both substitution and indel characters. Parsimony analyses were performed in the program PAUP* (version 4.0b10; Swofford, 2002). All characters were treated as unordered and were equally weighted. Each analysis consisted of a heuristic search with 1000 random sequence addition replicates (saving 100 trees per replicate), stepwise addition, MULTREES, and tree-bisection-reconnection (TBR) branch swapping. Maximum parsimony bootstrap (MPB) percentages were calculated from 1000 bootstrap replicates, each comprising 100 random sequence addition replicates, saving 10 trees per replicate.
Bayesian analyses were performed in the program MrBayes 3.1 (Ronquist and Huelsenbeck, 2003) using models of sequence evolution independently selected for cpDNA, AAT1, and ITS data sets by hierarchical likelihood ratio tests (hLRT) and the Akaike information criterion (AIC) in MrModeltest 2.0 (Nylander, 2004). An F81-like model for simple binary data were used to model the evolution of indel characters (Ronquist and Huelsenbeck, 2005). A combined analysis (described later) employed a partitioned Bayesian approach, incorporating independently chosen models for cpDNA, AAT1, ITS, and indel data partitions (see Ronquist et al., 2005). For all Bayesian analyses, at least two simultaneous searches were conducted comprising four Markov chains started from a randomly chosen tree and run for 6 000 000 generations, with sampling every 50 000 generations. The "burn-in" was estimated by examination of log likelihood scores of sampled trees, and posterior probabilities (PP) were calculated from postburn-in trees. The following descriptors were assumed to indicate convergence on a common phylogenetic topology by separate Bayesian searches: similarity in log likelihood scores at stationarity, similarity in consensus tree topologies and PP values for supported nodes, and a final average standard deviation of split frequencies (ASDSF) for simultaneous searches approaching zero.
The cpDNA data set, which contained sequences from all taxa in the data matrix, formed the basis of an analysis focusing on relationships among swartzioid genera. That analysis supported the monophyly of a "core swartzioid" group composed of Bocoa, Candolleodendron, and Swartzia (excluding S. panamensis). To examine relationships within the core swartzioid clade, we conducted individual analyses of the AAT1 and ITS data sets of sequences obtained by direct sequencing. In the ITS analysis, Bobgunnia madagascariensis and Swartzia panamensis were used as outgroups to the "core swartzioids." We were unable to obtain AAT1 sequences for any taxa outside of the core swartzioids, so the AAT1 analysis was rooted with Bocoa viridiflora and Candolleodendron brachystachyum, which together formed a clade resolved as the sister taxon to Swartzia in the ITS analysis. For assessing the degree to which AAT1 and ITS allelic variation coalesced within individuals and species, matrices of aligned AAT1 and ITS clone sequences were analyzed independently using a maximum parsimony criterion (data matrices and trees available in TreeBASE; website http://www.treebase.org).
Finally, parsimony and Bayesian analyses of the combined cpDNA, AAT1, and ITS data (excluding clone sequences) were conducted, also focusing on the core swartzioid group, with particular emphasis on identifying major subgroups within Swartzia. Phylogenetic congruence among cpDNA, AAT1, and ITS data partitions (substitution and indel characters) was tested with incongruence length difference tests (ILD; Farris et al., 1994) performed in PAUP*, with analysis parameters identical to those given above for maximum parsimony analyses but with MULTREES turned off. Congruence among data partitions was further assessed by inspecting tree topologies from the independent analyses. For the most part, only those individuals and taxa that were scored for at least 85% of the characters in the data matrix were included in the combined analyses, but several species for which we were unable to obtain AAT1 sequences were included because of their importance as outgroups (B. madagascariensis and S. panamensis) or as type species (S. recurva and S. leptopetala) of Cowan, 1968 infrageneric groups (Table 1).
RESULTS
cpDNA sequences and phylogenetic analyses
The three cpDNA regions were sequenced for all 110 individuals in the data matrix, representing 101 taxa, though several taxa were represented by partial sequences (to a maximum of 112 missing cpDNA aligned positions). Unaligned sequences (not including partial sequences) of the portion of the atpB–rbcL spacer included in the analyses varied in length from 630 to 688 bp. Those of the sequenced portion of the trnL intron and trnL–F spacer ranged from 366 to 528 bp and from 81 to 364 bp, respectively. The combined cpDNA data matrix contained 1919 aligned base pair positions and included numerous inferred indels, of which 22 indels (1–35 bp long) were identified that could be scored unambiguously as separate phylogenetic characters. Removal of sequence regions with ambiguous alignment and/or poor-quality sequence left 1403 aligned base pair positions, 421 variable sites, and 158 parsimony informative sites (not counting indels). Under both hLRT and AIC, MrModeltest selected the GTR + G model for the combined cpDNA data.
Substantial convergence between two simultaneous Bayesian searches of the aligned cpDNA data was indicated by a high degree of similarity in postburn-in log likelihood scores and posterior probabilities for resolved nodes and by an ASDSF = 0.030. Topologies of 50% majority-rule consensus trees generated from each of the searches were nearly identical. Parsimony analysis of the same data matrix yielded 35 550 shortest trees of 699 steps (CI = 0.671; RI = 0.778). Although a greater degree of resolution was apparent in the Bayesian trees, a high level of congruence between parsimony and Bayesian analyses was indicated by a lack of conflict among topologies for nodes with at least 50% MPB or PP support. All cpDNA analyses strongly resolved a monophyletic ingroup (99% MPB, 100% PP), the swartzioid clade, as well as a number of smaller clades within the swartzioid clade (Fig. 1A). One of these clades, heretofore called the core swartzioid clade (98% MPB, 100% PP), is composed of Bocoa viridiflora, Candolleodendron, and all sampled species of Swartzia, except for S. panamensis. Many of the deepest nodes within the core swartzioid clade were not resolved by the analyses.
|
). However, a gene-duplication scenario, possibly through polyploidization, is particularly likely for the sampled individual of S. latifolia, which had five alleles, differing from one another by as much as 1.2% sequence divergence. Cloning of ITS yielded one to six alleles per individual. ITS alleles usually differed by less than 1% sequence divergence within individuals. In one notable exception, Torke 304 (S. leiocalycina), a single ITS allele, perhaps a pseudogene (Álvarez and Wendel, 2003; Bailey et al., 2003), differed from other alleles from the same individual by ca. 4% sequence divergence. Separate phylogenetic analyses of sequences of AAT1 and ITS clones revealed coalescence of alleles within individuals and species (data not shown), consistent with the characterization of AAT1 as a single-copy diploid locus in most of the individuals and species examined and suggesting rapid homogenization of reiterated ITS copies by concerted evolution (Baldwin et al., 1995). However, it is important to note that very little is known about genome size and organization in Swartzia. Apparently, only a single published chromosome count exists for the entire genus (2n = 28 for S. laevicarpa: Goldblatt, 1981). The AAT1 data set (obtained by direct sequencing) contained sequences from 86 individuals that varied in length from 442 to 527 bp (for the portion of AAT1 included in the analyses). The sample included 76 taxa of Swartzia and one individual each of Bocoa viridiflora and Candolleodendron brachystachyum. The aligned data set contained 531 bp positions. Several individuals were missing portions of the alignment (to a maximum of 92 missing positions). Removal of areas of ambiguous alignment left 472 bp positions and 138 variable sites, of which 66 were parsimony informative. Five indel characters, 1–72 aligned bp long, were also included in the analyses. The HKY + G model of sequence evolution was selected by both hLRT and AIC for the AAT1 data.
The ITS data set (obtained by direct sequencing) contained 91 individuals and 82 taxa. The sample also included the noncore swartzioid taxa Bobgunnia madagascariensis and S. panamensis. Unaligned sequences varied in length from 535 to 589 bp. The aligned data set contained 657 bp positions. One individual was missing four aligned positions. Because of substantial alignment ambiguities, only 352 aligned sites (102 variable) were included in the analyses. There were 78 parsimony informative substitutions and four unambiguous indel characters (1–3 bp long). The GTR + I + G model of sequence evolution was selected for the ITS data.
Independent Bayesian and parsimony analyses of the AAT1 and ITS data sets produced topologies largely lacking in resolution of deeper nodes within the core swartzioid clade (Fig. 2). The monophyly of Swartzia was moderately supported (93% PP) in the Bayesian analysis of ITS data. The nesting of Bocoa within Swartzia rendered Swartzia paraphyletic in the Bayesian analysis of AAT1 data, but the latter result was poorly supported (65% PP). A number of subclades of Swartzia were supported in one or both analyses. For each DNA region, two simultaneous Bayesian searches displayed high levels of convergence (e.g., ASDSF = 0.028 for AAT1; 0.029 for ITS) and produced highly similar consensus topologies with similar PP values. Parsimony analysis of the AAT1 data found 56 891 shortest trees of 214 steps (CI = 0.707, RI = 0.741), while 6523 shortest trees of 483 steps (CI = 0.189, RI = 0.340) were found in a parsimony analysis of the ITS data. There was very little difference between parsimony and Bayesian topologies for the two DNA regions, although the Bayesian were generally better resolved.
|
Combined analyses
The combined analyses took advantage of strong taxonomic overlap of core swartzioid taxa among cpDNA, AAT1, and ITS data partitions and included 89 individuals, representing 81 taxa. As mentioned previously, Bobgunnia madagascariensis, Swartzia leptopetala, S. panamensis, and S. recurva, though not sequenced for AAT1, were included in the combined analyses. Previous analyses (data not shown) showed that omitting these four taxa did not affect overall tree topology. In total, the combined alignment contained 3107 bp positions, of which 2227 positions were included in the analyses. There were 439 variable sites, 228 parsimony informative substitutions, and 22 indel characters.
Two separate Bayesian searches produced identical topologies, with very similar PP values for supported nodes. Convergence between the two searches was also apparent in similar log likelihoods for trees sampled at likelihood stationarity and an ASDSF = 0.025. Parsimony analysis of the combined data produced 3861 trees of 1063 steps (CI = 0.522, RI = 0.683). There was no conflict between Bayesian and parsimony analyses for nodes with more than 50% PP or MPB support, although Bayesian topologies were generally better resolved. The consensus of postburn-in trees from one of the Bayesian searches (Fig. 3) is largely concordant with trees obtained in the individual analyses (Figs. 1A, 2). However, the combined analyses produced considerably greater resolution of and within subclades of Swartzia. Better supported in the individual analysis of ITS data, the monophyly of Swartzia (excluding S. panamensis) received limited support (56% MPB, 75% PP) in the combined analyses. As in other analyses, most of the deepest nodes within Swartzia were not resolved.
|
The analyses presented meet the principal objectives of this study, which were to test the monophyly of Swartzia and to achieve a greatly improved understanding of relationships among species of Swartzia and among genera of the swartzioid clade. The analyses also demonstrate the applicability of four widely used DNA markers (the chloroplast atpB–rbcL spacer, the trnL intron, the trnL–F spacer, and nuclear ITS) to the study of swartzioid phylogeny. To our knowledge, this is the first phylogenetic investigation within Leguminosae to employ sequence data from the low-copy nuclear gene AAT1, a commonly used marker in allozyme studies. Our confidence in the phylogenetic results is enhanced by substantial phylogenetic congruence among DNA regions and by molecular cloning results, which showed coalescence of AAT1 and ITS alleles within individuals and species. Next, we discuss the major clades that were recovered by the phylogenetic analyses, potential morphological synapomorphies and diagnostic characters, and taxonomic implications.
Swartzioid clade
The swartzioid clade, as outlined by Ireland (2005), is strongly supported as monophyletic (Fig. 1A). One possible morphological synapomorphy for the group is the lack of unidirectional order in the initiation of the stamens. This condition has been documented in several swartzioid taxa (a single species each of Ateleia, Bobgunnia, and Cyathostegia and three species of Swartzia), and it appears very rarely in other legumes (Tucker, 1990, 2003; Tucker and Douglas, 1994). The stamens in these taxa are initiated basipetally or acropetally, sometimes in irregular whorls, from a ring meristem, a rare and taxonomically restricted feature in legumes (Tucker, 1987, Tucker, 1990, Tucker, 2003). Other morphological features that characterize the swartzioid clade, but are probably not synapomorphic for it, include complete or near complete fusion of sepals resulting from intercalary growth early in development, relatively numerous stamens, and a single or no petal, with other petals not at all apparent in development (Tucker, 2003).
Ateleioid clade
In agreement with previous studies (Ireland et al., 2000; Pennington et al., 2001), the analyses recovered a well-supported "ateleioid" clade (Fig. 1A) positioned as the sister taxon to the remainder of the swartzioid clade and composed of the genera Ateleia, Cyathostegia, and Trischidium. The ateleioid clade differs from most other swartzioid taxa in having an actinomorphic androecium, basifixed anthers, exarillate seeds, and a tendency toward alternate leaflets. Ireland, 2007 resurrection of Trischidium Tulasne for five species placed in Bocoa by Cowan (1974) is justified by both molecular data (Ireland et al., 2000; this study, Fig. 1A) and by morphological evidence. Trischidium differs from the three species that Ireland retained in Bocoa (B. prouacensis, B. ratteri H. E. Ireland, and B. viridiflora) in having a single (vs. no) petal, a tendency toward alternate (vs. opposite) leaflets, reticulate-veined (vs. smooth-surfaced) fruits, and exarillate (vs. arillate) seeds. Trischidium is the sister taxon to a strongly supported clade grouping a monophyletic Ateleia and monospecific Cyathostegia (Fig. 1A). The latter grouping is also supported by apparent morphological synapomorphies, including valvate calyx lobes and one or two ovules per carpel, and it is reinforced by the presence of unusual pollination syndromes, including probable beetle pollination in Cyathostegia (Lewis et al., 2003) and wind-pollination in Ateleia, most species of the latter genus being dioecious (Janzen, 1989; Ireland, 2001).
Fairchildia
In a novel result, Swartzia panamensis of Central America and the Pacific lowlands of Colombia was supported as the sister taxon to a clade composed of Bobgunnia, Bocoa, Candolleodendron, and all other species of Swartzia (Fig. 1A). Cowan (1968, p. 99) commented that his placement of S. panamensis in the ordering of species in his monograph of Swartzia was "totally lacking phylogenetic implications." He acknowledged that the pendulous racemes and large, flattened pods of S. panamensis set it apart from other species of Swartzia but questioned the creation of the monospecific genus Fairchildia by Britton and Rose (1930) based on fruit characters. The present result, based on the first molecular data for the species, is corroborated by obvious morphological differences between S. panamensis and other species of Swartzia: pendant (vs. erect) inflorescences, elastic (vs. inert) fruit dehiscence, and flattened, exarillate (vs. ellipsoid, arillate) seeds. Partial sequences from two other individuals of S. panamensis (data not shown) confirm this result.
Bobgunnia
The analyses strongly support a sister-group relationship between Bobgunnia and the core swartzioid clade (i.e., Bocoa, Candolleodendron, and Swartzia; Fig. 1A), consistent with Kirkbride and Wiersema’s (1997) erection of Bobgunnia for the two African species formerly included in Swartzia. The possession of arillate seeds is an apparent synapomorphy for the grouping of Bobgunnia and the core swartzioids. The flowers of Bobgunnia are similar to those of most species of Swartzia in having an entire calyx, a single petal, and a zygomorphic androecium, with the stamens mostly of two obvious size classes, though with some intergradation between the two sizes. Bobgunnia differs from Swartzia in having alternate (vs. opposite) leaflets, pendulous inflorescences with the flowers arranged in fascicles along the axes (vs. erect and spirally arranged), prolate-shaped pollen with a thin exine (vs. mostly spheroidal, with the exine relatively thicker, Ferguson and Skvarla, 1991), indehiscent pods with two large, transversely reniform-shaped canals in the mesocarp containing a mucilage-like substance (vs. dehiscent and lacking such canals), and seeds with typically faboid characteristics, including a hard, thick testa with hourglass-shaped cells, a tracheid bar beneath the hilar groove, and a curved embryo (vs. seeds laking faboid characteristics; see Corner, 1976; Kirkbride and Wiersema, 1997).
Core swartzioid clade
A core swartzioid group comprising Bocoa, Candolleodendron, and Swartzia (excluding Fairchildia), was strongly resolved as monophyletic (Figs. 1A, 2, 3). The same clade was recovered in previous phylogenetic studies based on morphological (Herendeen, 1994) and molecular data (Ireland et al., 2000), although the latter study did not sample Candolleodendron and included only four species of Swartzia. Despite numerous molecular synapomorphies in the cpDNA (Fig. 1B) and ITS data sets, there appear to be few if any morphological synapomorphies for the core swartzioid clade; Herendeen, 1994 result was based on two characters, a coarsely granular colpus membrane and the possession of a single petal, both of which originated earlier in the evolution of the swartzioid legumes or even before the origin of the swartzioid clade.
Candolleodendroid clade
In agreement with the morphological analysis of Herendeen (1994), the analyses recovered a "candolleodendroid" clade (Figs. 1A, 2, 3), composed of Bocoa and monospecific Candolleodendron. This small clade, as redefined by Ireland (2005, Ireland, 2007), contains four species restricted to lowland forests of the Guiana shield and Amazon basin. It differs from its sister taxon, Swartzia, in having an actinomorphic androecium, with monomorphic stamens and basifixed (or nearly so) anthers. (Ireland [2007] reports that the anthers are basally dorsifixed in Bocoa). These features, which also characterize the ateleioid clade, apparently trace to an independent origin in the ancestor of the candolleodendroids.
Swartzia
Our analyses provide the first phylogenetic evidence for the monophyly of Swartzia (exluding Fairchildia) based on a broad sampling of species (Figs. 2, 3). The weakly supported nesting of Bocoa viridiflora within Swartzia in the individual Bayesian analysis of AAT1 data (Fig. 2) is in conflict with all other analyses. The limited support for the monophyly of Swartzia is consistent with an apparent scarcity of morphological synapomorphies for the genus. One possible synapomorphy, the presence of a watery sap in the stems and branches that turns red or pinkish upon exposure, occurs in a variety of legume genera and in diverse species of Swartzia but has not been reported in other taxa of the swartzioid clade. Most other morphological features that characterize Swartzia (entire calyx, corolla usually comprising a single petal, dimorphic stamens, and dorsifixed anthers) are apparently plesiomorphic, having originated earlier in the evolution of the swartzioid clade, or even before its origin. In a phylogenetic analysis of morphological data, Herendeen, 1994 finding of a nonmonophyletic Swartzia was based on a single character, the presence of arils, which he scored as variable for two groups of Swartzia. In our examination of the genus, we have found only one species, S. panamensis (i.e., Fairchildia), in which arils are entirely absent. [V. Mansano, Jardim Botânico do Rio de Janeiro (personal observation) reports that S. glazioviana (Taub.) Glaziou also lacks arils].
The analyses failed to resolve many of the early branching events within Swartzia (Figs. 1A, B, 2, 3). The most interior node within Swartzia that was recovered defines a possible clade including most species and nine of the 11 supported subclades of Swartzia discussed later (Figs. 1A, 3). Deep polytomies in phylogenies of other species-rich, neotropical, rainforest tree genera have been linked to rapid taxonomic diversifications (e.g., Inga: Richardson et al., 2001; Guatteria: Erkens et al., 2007). This possibility should be further investigated in Swartzia.
The lack of resolution of early branching events prevents most inferences regarding the polarities of morphological characters within Swartzia. In addition, many of the morphological characters used in previous subgeneric classifications (e.g., Cowan, 1968), when examined across the entire genus, vary continuously, have multiple origins, or else are synapomorphic for shallow phylogenetic lines. As such, our discussion of morphological variation among subclades of Swartzia will focus on characteristics, not necessarily synapomorphies. The character combinations specified in Table 2 are considered diagnostic of nonoverlapping subclades of the genus and can be used to predict the phylogenetic placement (within subclades) of Swartzia species that were not sampled in the molecular analyses.
|
). Although for some of the subclades the taxonomic discrepancies are fairly large, they generally contain the type species of Cowan’s similarly named infrageneric group.
Unifoliolatoid, recurvoid, and possiroid clades
As suggested previously (Torke, 2004), the white-flowered portion of Cowan's (1968) circumscription of section Possira forms a well-supported "unifoliolatoid" clade (Figs. 1–3). The unifoliolatoid species have stipulate bracts, a condition also found in the "possiroid" clade and in S. aptera of the "recurvoid" clade (the bracts of nearly all other species of Swartzia are estipulate, see Table 2). A close relationship between these three clades is also suggested by similarity in the overall form of the gynoecia and fruits and by the presence of unifoliolate leaves in some species of all three clades. Leaflet number was the basis of Cowan’s series Unifoliolatae, an artificial assemblage of unifoliolate species of the unifoliolatoid and possiroid clades. The unifoliolatoid clade is mostly restricted to lowlands of the Amazon basin and Guiana shield.
Also strongly supported, the recurvoid clade (Figs. 1–3) largely corresponds to Cowan’s series Recurvae plus series Racemosae. Both series were described as having glabrous gynoecia, a relatively unstable character in the evolutionary history of Swartzia. With the addition of several species with pubescent gynoecia, the recurvoid clade is characterized by having bracteolate pedicels and a relatively elongate ovary stipe (Table 2). One notable result of the analyses is the placement in the recurvoid clade of S. aptera, one of the few species of Swartzia combining stipulate bracts with bracteolate pedicels. Cowan (1968) placed S. aptera in section Possira, near a group of species that constitute the possiroid clade of the current study. However, its bracteolate pedicels and obtuse (vs. capitellate) stigma convey its true affinities with the recurvoid clade. The recurvoid clade is also found mostly in the Amazon basin and Guiana shield.
The strongly supported possiroid clade (Figs. 1–3) groups together the yellow-flowered species of Cowan’s circumscription of section Possira, with the exception of S. aptera. This closely knit group is the most widespread and common of the presently recognized subclades of Swartzia: its distribution is nearly as large as that of the entire genus (Table 2). Morphological variation is nearly continuous among species of the possiroid clade, making this one of the more taxonomically challenging groups of Swartzia. Historical reticulation or incomplete lineage sorting may account for discrepancies between nuclear and chloroplast topologies within the possiroid clade and several other clades of Swartzia (see Rieseberg and Soltis, 1991; Doyle et al., 1999; Jakob and Blattner, 2006; Figs. 1A, 2).
Tounateoid, terminaloid, and pittierianoid clades
A "tounateoid" clade was recovered containing the majority of species in Cowan’s circumscription of series Tounateae (Figs. 1, 3). This clade is a morphologically cohesive group distributed primarily in South America and characterized by small apetalous flowers (for other diagnostic characters, see Table 2). The tounateoids differ from other, primarily Central American, apetalous Swartzia species most noticeably in having much less elongated gynoecia with a more obviously laterally positioned style. They also differ in their mostly single-seeded, broadly ellipsoid or globose fruits (vs. multiseeded and moniliform or if single-seeded, then obviously reduced from a moniliform type). The tounateoid clade is highly diversified in lowland parts of the Guiana shield, with secondary centers of diversity in Amazonia and eastern Brazil.
Corresponding to Cowan's (1968) series Pittierianae, a small clade composed of S. jorori and S. pittieri was consistently recovered (Figs. 1–3). The recently discovered western Amazonian species, S. juruana Torke, a close relative of S. jorori, brings the number of "pittierianoid" species to three (Torke, 2004). All three species are characterized by relatively small, ebracteolate flowers with a yellow petal (Table 2). In S. pittieri, the calyx splits regularly into two hemispheric lobes (vs. splitting irregularly into three to six segments in most Swartzia species). Among species of Swartzia, this condition is otherwise known only in S. trianae Benth., which although it was not sampled in the phylogenetic analyses, appears to be a fourth pittierianoid species.
A well-supported "terminaloid" clade (Figs. 1–3) groups together a large portion of species of Cowan’s subsection Terminales and most of the remaining apetalous species of Swartzia. Moniliform fruits, ebracteolate pedicels, and a typically elongate ovary characterize this clade (Table 2), within which there appears to be a basal split between the petalous, mostly South American species and the apetalous, mostly Central American species. Both here and in the tounateoid clade the switch to apetaly appears to have been accompanied by an overall reduction in the size of the flowers, with some of the smallest flowers in the genus occurring in tounateioid species, such as S. tessmannii and S. manausensis, and in apetalous terminaloid species, such as S. cubensis and S. nicaraguensis.
Panacocoid and grandifolioid clades
The "panacocoid" clade (Figs. 1–3) is a closely knit assemblage distributed across a large geographic range in northern South America. It is characterized by large, orbicular stipules, semiannular stipular scars, costate twigs, ebracteolate pedicels, and relatively large flowers, with a yellow petal and an elongate gynoecium (Table 2). As mentioned previously, the placement of the typical variety of S. sprucei in the terminaloid clade by the cpDNA analyses (Fig. 1A) conflicts with its placement in the panacocoid clade by all other analyses (Figs. 2, 3). The morphology of S. sprucei var. spucei is like that of other members of the panacocoid clade, thus supporting the latter result.
Similar to the panacocoid clade in morphology and geographic distribution, the strongly supported "grandifolioid" clade (Figs. 1–3) differs primarily in having bracteolate pedicels and ovate to triangular stipules and in lacking semiannular stipular scars (Table 2). Together, these two clades probably contain most of the remaining species of Cowan’s subsection Terminales. Although their phylogenetic placement is poorly supported, the species S. macrosema and S. acreana appear to be related to the panacocoid and/or grandifolioid clades based on overall morphological similarity.
Acutifolioid clade
Nearly matching in content Cowan’s (1968, 1981) series Acutifoliae, the well-supported "acutifolioid" clade (Figs. 1–3) is highly diverse in the Atlantic coastal forest of eastern Brazil and in the cerrados and dry forests of the Brazilian shield. A few species occur in the Amazon basin, mostly along its southern fringe. Along with the apetalous terminaloid group of Central America and perhaps the widely distributed possiroid clade, the acutifolioid clade represents one of the few radiations of Swartzia to have occurred primarily outside the Amazon basin and/or Guiana shield. Among the subclades of Swartzia that have estipulate bracts, the acutifolioid clade is characterized by the morphology of the gynoecium and fruits and by often having relatively numerous leaflets (Table 2).
Orthostyloid and benthamioid clades
The analyses recovered an "orthostyloid" clade (Figs. 2, 3) comprising the core elements of Cowan's (1968) circumscription of series Orthostylae. This clade is characterized by elongate gynoecia, typically with the stipe and style nearly as long or somewhat longer than the ovary (Table 2). The multicellular cupular hairs of several species are also distinctive. Species of the orthostyloid clade are characteristic canopy elements across large portions of the Amazon basin and Guiana shield lowlands. The "benthamioid" clade (Figs. 2, 3), likely containing most of the species of Cowan’s series Benthamianae, was weakly resolved as the sister taxon to the orthostyloid clade (Fig. 3), from which it differs substantially in the morphology of its gynoecium (Table 2).
Several other species had poorly resolved phylogenentic placements within Swartzia (Fig. 3). The overall form of the sericeous gynoecium of S. bombycina sets it apart from other species of Swartzia (Table 2). "Swartzia sp. nov. 3," a narrowly distributed endemic from eastern Brazil, appears to be very closely related to the Amazonian species S. reticulata Ducke and S. calva R. S. Cowan; all three species are characterized by relatively massive pods and seeds, a complete lack of pubescence, and by coriaceous, reticulate-veined leaflets. The distinctive species S. dipetala is one of several distantly related species of Swartzia that typically have multiple pistils in most if not all of their flowers.
Taxonomic implications
Our analyses suggest that Swartzia panamensis should be treated as a separate genus under the preexisting name Fairchildia (Britton and Rose, 1930). Further evidence for the distinctness of Fairchildia is expected from more detailed morphological study. The analyses provide additional support for the segregation of Bobgunnia from Swartzia (Kirkbride and Wiersema, 1997) and for the resurrection of Trischidium (Ireland, 2007). Because it was otherwise resolved as monophyletic, albeit with limited support, Swartzia should be retained without further division, despite its large number of species and considerable morphological diversity. The present resolution of subclades of Swartzia should be used as a basis for revising the highly artificial infrageneric classification of Cowan (1968). We agree with Ireland et al. (2000) and Ireland (2005) that tribe Swartzieae should be redefined to include only genera belonging to the swartzioid clade. It might also be acceptable to adopt a broader tribal circumscription to include members of a clade of Sophoreae and Dipterygeae genera that was weakly resolved as the sister group to the swartzioid clade by Wojciechowski et al. (2004), but this broader circumscription would need to be supported by additional phylogenetic data.
Appendix 1. Taxa, vouchers, localities, and GenBank accession numbers for sampled representatives of the swartzioid clade and outgroups. Asterisks denote accession numbers of unique sequences obtained from AAT1 and ITS clones; sequences used in the broader phylogenetic analyses were obtained by direct sequencing and are not marked with an asterisk. Within Swartzia, species and varietal names generally follow the system of Cowan (1968), with most exceptions being recently published names (e.g., Torke, 2007b) and nomenclatural changes (Torke, 2007a).
|
FOOTNOTES
1 The authors thank G. Fritz, A. Larson, K. Olsen, P. Raven, and M. Richardson for their guidance and M. Lavin, V. Mansano, R. T. Pennington, and two anonymous reviewers for helpful comments on the manuscript. They are grateful to J. Beck, Y.-C. Chiang, K.-F. Chung, L. Sánchez, P. Sweeney, and members of the Schaal Laboratory for help with laboratory work and analyses. The authors also thank M. Alves, D. Augulo, P. Berry, R. Duno, B. Gunn, H. Ireland, S. Mori, K. Redden, R. Rivera García, and Kew Gardens for plant materials, N. Cuello for locating voucher specimens at PORT, and the curators of MO, NY, and US for the use of their specimens. Permissions for field studies were granted by the following agencies: Forest Department (Belize), SERNAP (Bolivia), CNPq (Brazil), MINAE (Costa Rica), CONAP (Guatemala), EPA (Guyana), MARENA (Nicaragua), and INRENA (Peru). The authors thank the following persons, herbaria, and institutions for logistical support of field studies: H. Aldana (HUT), D. Angulo (CICY), G. Carnevali (CICY), A. Carvalho (deceased, CEPEC), K.-F. Chung (HAST), I. Coronado (HULE), R. Duno (CICY), K. Erskine (BRG), S. Fuentes (MO), S. Gonzales (CAY), J.-J. de Granville (CAY), J. Lidell (BRG), A. MacVean (UVAL), D. Mondragon (OAX), V. Medina (HUT), M.-F. Prevost (CAY), R. Rivera Garcia (OAX), E. Rodriguez (HUT), R. Rueda (HULE), J. Ruiz (AMAZ), M. Saldias (USZ), L. Sánchez, D. Santa Maria (INB), B. Simón Pérez, R. Thomas (FDG), W. Thomas (NY), N. Zamora (INB), PN Santa Rosa staff–Costa Rica, OTS–Costa Rica, SI–Biodiversity of the Guianas Program, and TNC Climate Change Project. Funding was provided by the National Science Foundation (DEB-0309162), the Deep Time Project, the Botanical Society of America, the American Society of Plant Taxonomists, the Rupert Barneby Award, a grant from the Mellon Foundation to the Missouri Botanical Garden, and DBBS at Washington University. 
4 Author for correspondence (e-mail: torke{at}ansp.org)
LITERATURE CITED
Álvarez, I., AND J. F. Wendel. 2003. Ribosomal ITS sequences and plant phylogenetic inference. Molecular Phylogenetics and Evolution 29: 417–434.[CrossRef][Web of Science][Medline]
Bailey, C. D., T. G. Carr, S. A. Harris, AND C. E. Hughes. 2003. Characterization of angiosperm nrDNA polymorphism, paralogy, and pseudogenes. Molecular Phylogenetics and Evolution 29: 435–455.[CrossRef][Web of Science][Medline]
Baldwin, B. G., M. J. Sanderson, J. M. Porter, M. F. Wojciechowski, C. S. Campbell, AND M. J. Donoghue. 1995. The ITS region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82: 247–277.[CrossRef][Web of Science]
Barnes, W. M. 1992. The fidelity of Taq polymerase catalyzing PCR is improved by an N-terminal deletion. Gene 112: 29–35.[CrossRef][Web of Science][Medline]
Bentham, G. 1865. Leguminosae. In G. Bentham, and J. D. Hooker [eds.], Genera plantarum, vol. 1, part 2, 434–600. L. Reeve, London, UK.
Britton, N. L., AND J. N. Rose. 1930. Caesalpiniaceae. North American Flora 23: 201–349.
Corner, E. J. K. 1976. The seeds of dicotyledons. Cambridge University Press, Cambridge, UK.
Cowan, R. S. 1966. Candolleodendron, a new genus of Leguminosae (Caesalpinioideae). Rhodora 68: 429–432.
Cowan, R. S. 1968. Flora neotropica monograph 1. Swartzia (Leguminosae, Caesalpinioideae, Swartzieae). Hafner, New York, New York, USA.
Cowan, R. S. 1974. A revision of the genus Bocoa (Caesalpinioideae-Swartzieae). Proceedings of the Biological Society of Washington 87: 95–128.
Cowan, R. S. 1981a. New taxa of Leguminosae-Caesalpinioideae from Bahia, Brazil. Brittonia 33: 9–14.[CrossRef][Web of Science]
Cowan, R. S. 1981b. Swartzieae. In R. M. Pohill, and P. H. Raven [eds.], Advances in legume systematics, vol. 1, 209–212. Royal Botanic Gardens, Kew, UK.
De Oliveira, A. A., AND S. A. Mori. 1999. A central Amazonian terra firme forest. I. High tree species richness on poor soils. Biodiversity and Conservation 8: 1245–1259.[CrossRef][Web of Science]
Doi, K., A. Kaga, N. Tomooka, AND D. A. Vaughan. 2002. Molecular phylogeny of genus Vigna subgenus Ceratotropis based on rDNA, ITS and atpB-rbcL intergenic spacer of cpDNA sequences. Genetica 114: 129–145.[CrossRef][Web of Science][Medline]
Doyle, J. J., AND J. L. Doyle. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15.
Doyle, J. J., J. L. Doyle, J. A. Ballenger, E. E. Dickson, T. Kajita, AND H. Ohashi. 1997. A phylogeny of the chloroplast gene rbcL in the Leguminosae: Taxonomic correlations and insights into the evolution of nodulation. American Journal of Botany 84: 541–554.[Abstract]
Doyle, J. J., J. L. Doyle, AND A. H. D. Brown. 1999. Incongruence in the diploid B-genome species complex of Glycine (Leguminosae) revisited: Histone H3-D alleles versus chloroplast alleles. Molecular Biology and Evolution 16: 354–362.[Abstract]
Ellison, N. W., A. J. Liston, J. J. Steiner, W. M. Williams, AND N. L. Taylor 2006. Molecular phylogenetics of the clover genus (Trifolium–Leguminosae). Molecular Phylogenetics and Evolution 39: 688–705.[CrossRef][Web of Science][Medline]
Erkens, R. H. J., L. W. Chatrou, J. W. Maas, T. van der Niet, AND V. Savolainen. 2007. A rapid diversification of rainforest trees (Guatteria; Annonaceae) following dispersal from Central into South America. Molecular Phylogenetics and Evolution 44: 399–411.[CrossRef][Web of Science][Medline]
Farris, J. S., M. Källersjö, A. G. Kluge, AND C. Bult. 1994. Testing significance of incongruence. Cladistics 10: 315–319.[CrossRef][Web of Science]
Ferguson, I. K., AND J. J. Skvarla. 1991. Pollen morphology of the tribe Swartzieae (subfamily Papilionoideae: Leguminosae). The genera Aldina Endlicher and Swartzia Schreber and systematic conclusions. Review of Palaeobotany and Palynology 67: 153–177.[CrossRef][Web of Science]
Goldblatt, P. 1981. Cytology and the phylogeny of Leguminosae. In R. M. Polhill, and P. H. Raven [eds.], Advances in legume systematics, vol. 2, 427–463. Royal Botanic Gardens, Kew, UK.
Herendeen, P. S. 1994. Phylogenetic relationships of the tribe Swartzieae. In M. D. Crisp, and J. J. Doyle [eds.], Advances in legume systematics, vol. 7, 123–132. Royal Botanic Gardens, Kew, UK.
Hurr, K. A., P. J. Lockhart, P. J. Heenan, AND D. Penny. 1999. Evidence for the recent dispersal of Sophora (Leguminosae) around the southern oceans: Molecular data. Journal of Biogeography 26: 565–577.[CrossRef][Web of Science]
Hutchinson, J. 1964. The genera of flowering plants, Clarendon Press, Oxford, UK.
Ireland, H. E. 2001. The taxonomy and systematics of Ateleia and Cyathostegia (Leguminosae-Swartzieae). Ph.D. dissertation, University of Reading, Reading, UK.
Ireland, H. E. 2005. Swartzieae. In G. Lewis, B. Schrire, B. MacKinder, and M. Lock [eds.], Legumes of the world, 214–225, Royal Botanic Gardens, Kew, UK.
Ireland, H. E. 2007. Taxonomic changes in the South American genus Bocoa (Leguminosae–Swartzieae): Reinstatement of the name Trischidium, and a synopsis of both genera. Kew Bulletin 62: 333–350.
Ireland, H. E., R. T. Pennington, AND J. Preston. 2000. Molecular systematics of the tribe Swartzieae. In P. S. Herendeen, and A. Bruneau [eds.], Advances in legume systematics, vol. 9, 217–231. Royal Botanic Gardens, Kew, UK.
Jakob, S. S., AND F. R. Blattner. 2006. A chloroplast genealogy of Hordeum (Poaceae): Long-term persisting alleles, incomplete lineage sorting, regional extinction, and the consequences for phylogenetic inference. Molecular Biology and Evolution 23: 1602–1612.
Janzen, D. H. 1989. Natural history of a wind-pollinated Central American dry forest legume tree (Ateleia herbert-smithii Pittier). In C. H. Stirton, and J. L. Zarucchi [eds.], Advances in legume biology. Monographs in Systematic Botany from the Missouri Botanical Garden 29: 293–376.
Kirkbride, J. H., AND J. H. Wiersema. 1997. Bobgunnia, a new African genus of the tribe Swartzieae (Fabaceae, Faboideae). Brittonia 49: 1–23.[CrossRef][Web of Science]
Lewis, G. P., J. T. Knudsen, B. B. Klitgaard, AND R. T. Pennington. 2003. The floral scent of Cyathostegia mathewsii (Leguminosae, Papilionoideae) and preliminary observations on reproductive biology. Biochemical Systematics and Ecology 31: 951–962.[CrossRef][Web of Science]
Lewis, G. P., B. Schrire, B. MacKinder, AND M. Lock. [eds.]. 2005. Legumes of the world. Royal Botanic Gardens, Kew, UK.
Mesquita, M. R., AND M. J. G. Hopkins. 1999. Leguminosae: Papilionoideae. In J. E. L. S. Ribeiro, M. J. G. Hopkins, A. Vicentini, C. A. Sothers, M. A. da Costa, J. M. Brito, M. A. D. de Souza, L. H. P. Martins, L. G. Lohmann, P. A. C. L. Assunção, E. da Pereira, C. F. da Silva, M. R. Mesquita, and L. C. Procópio [eds.], Flora da Reserva Ducke: Guia de identificação de uma floresta de terra-firme na Amazônia Central, INPA/DIFD, Manaus, Brazil.
Nylander, J. A. A. 2004. MrModeltest v2. Computer program distributed by the author. Evolutionary Biology Centre, Uppsala University, Sweden. Available at http://people.scs.fsu.edu/
nylander/ [accessed 15 November 2007].
Pennington, R. T., M. Lavin, H. Ireland, B. B. Klitgaard, AND J. Preston. 2001. Phylogenetic relationships of basal papilionoid legumes based upon sequences of the chloroplast trnL intron. Systematic Botany 26: 537–566.[Web of Science]
Polhill, R. M. 1994. Classification of the Leguminosae and complete synopsis of legume genera. In F. A. Bisby, J. Buckingham, and J. B. Harborne [eds.], Phytochemical dictionary of the Leguminosae, vol. 1, Plants and their constituents. Chapman and Hall, London, UK.
Rambaut, A. 1996. Se-Al: Sequence alignment editor. Computer program available at http://tree.bio.ed.ac.uk/software/seal/ [accessed 15 November 2007].
Richardson, J. E., R. T. Pennington, T. D. Pennington, AND P. M. Hollingsworth. 2001. Rapid diversification of a species-rich genus of neotropical rain forest trees. Science 293: 2242–2245.
Rieseberg, L. H., AND D. E. Soltis. 1991. Phylogenetic consequences of cytoplasmic gene flow in plants. Evolutionary Trends in Plants 5: 65–84.[Web of Science]
Ronquist, F., AND J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (Oxford, England) 19: 1572–1574.[CrossRef]
Ronquist, F., AND J. P. Huelsenbeck. 2005. Bayesian analysis of molecular evolution using MrBayes. In R. Nielsen [ed.], Statistical methods in molecular evolution. Springer, New York, New York, USA.
Ronquist, F., J. P. Huelsenbeck, AND P. van der Mark. 2005. MrBayes 3.1 manual. Available at http://mrbayes.scs.fsu.edu/ [accessed 15 November 2007].
Simmons, M. P., AND H. Ochoterena. 2000. Gaps as characters in sequence-based phylogenetic analysis. Systematic Biology 49: 369–381.[CrossRef][Web of Science][Medline]
Strand, A. E., J. Leebens-Mack, AND B. G. Milligan. 1997. Nuclear DNA-based markers for plant evolutionary biology. Molecular Ecology 6: 113–118.[CrossRef][Medline]
Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4 0b10. Sinauer, Sunderland, Massachusetts, USA.
Taberlet, P., L. Gielly, G. Pautou, AND J. Bouvet. 1991. Universal primers for the amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109.[CrossRef][Web of Science][Medline]
Torke, B. M. 2004. Two new species of Swartzia (Leguminosae) from the Amazon basin of Brazil, with notes on the genus and a key to the unifoliolate species. Systematic Botany 29: 358–365.[CrossRef][Web of Science]
Torke, B. M. 2006. Systematics and evolutionary diversification in the species-rich neotropical tree genus Swartzia (Leguminosae–Papilionoideae). Ph.D. dissertation, Washington University, St. Louis, Missouri, USA.
Torke, B. M. 2007a. New combinations and species-level synonyms in Swartzia (Leguminosae: Papilionoideae). Novon 17: 110–119.[CrossRef][Web of Science]
Torke, B. M. 2007b. Three new species of Swartzia from northern South America. Botanical Journal of the Linnean Society 153: 343–355.[CrossRef][Web of Science]
Tucker, S. C. 1987. Floral initiation and development in legumes. In C. H. Stirton [ed.], Advances in legume systematics, vol. 3, 183–239. Royal Botanic Gardens, Kew, UK.
Tucker, S. C. 1990. Loss of floral organs in Ateleia (Leguminosae: Papilionoideae: Sophoreae). American Journal of Botany 77: 750–761.[CrossRef][Web of Science]
Tucker, S. C. 2003. Floral ontogeny in Swartzia (Leguminosae: Papilionoideae: Swartzieae): Distribution and role of the ring meristem. American Journal of Botany 90: 1271–1292.
Tucker, S. C., AND A. W. Douglas. 1994. Ontogenetic evidence and phylogenetic relationships among basal taxa of legumes. In I. K. Ferguson, and S. C. Tucker [eds.], Advances in legume systematics, vol. 6, 11–32. Royal Botanic Gardens, Kew, UK.
Vásquez-Martínez, R. 1997. Flórula de las reservas biológicas de Iquitos, Perú. Missouri Botanical Garden Press, St. Louis, Missouri, USA.
White, T. J., T. D. Bruns, S. B. Lee, AND J. W. Taylor. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White [eds.], PCR protocols: A guide to methods and applications. Academic Press, San Diego, California, USA.
Wojciechowski, M. F., M. Lavin, AND M. J. Sanderson. 2004. A phylogeny of legumes (Leguminosae) based on analysis of the plastid matK gene resolves many well-supported subclades within the family. American Journal of Botany 91: 1845–1861.
Wojciechowski, M. F., M. J. Hu, AND J.-M. Sanderson. 1999. Evidence on the monophyly of Astragalus (Fabaceae) and its major subgroups based on nuclear ribosomal DNA ITS and chloroplast DNA trnL intron data. Systematic Botany 24: 409–437.[CrossRef][Web of Science]
Young, N. D., AND J. Healy. 2003. Gapcoder automates the use of indel characters in phylogenetic analysis. BMC Bioinformatics 4: 6.[CrossRef][Medline]
This article has been cited by other articles:
|
|
 |
|
  C. Saslis-Lagoudakis, M. W. Chase, D. N. Robinson, S. J. Russell, and B. B. Klitgaard Phylogenetics of neotropical Platymiscium (Leguminosae: Dalbergieae): systematics, divergence times, and biogeography inferred from nuclear ribosomal and plastid DNA sequence data Am. J. Botany, October 1, 2008; 95(10): 1270 - 1286. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
80% MBP and 95% PP) placements that conflict with their placements in aspartate aminotransferase (AAT1) and ITS consensus trees (
