Molecular Systematics and Biogeography of Descurainia (Brassicaceae) Based on Nuclear ITS and Non-Coding Chloroplast DNA Author(s): Barbara E. Goodson, Sumaiyah K. Rehman, and Robert K. Jansen Source: Systematic Botany, 36(4):957-980. 2011. Published By: The American Society of Plant Taxonomists URL: http://www.bioone.org/doi/full/10.1600/036364411X604976 BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder. BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. Systematic Botany (2011), 36(4): pp. 957–980 © Copyright 2011 by the American Society of Plant Taxonomists DOI 10.1600/036364411X604976 Molecular Systematics and Biogeography of Descurainia (Brassicaceae) based on Nuclear ITS and Non-Coding Chloroplast DNA Barbara E. Goodson,1,4 Sumaiyah K. Rehman,1,2 and Robert K. Jansen1,3 Section of Integrative Biology and Institute of Cellular and Molecular Biology, University of Texas, Austin, Texas 78712 U. S. A. Department of Molecular and Cellular Oncology, University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030 U. S. A. 3 Genomics and Biotechnology Section, King Abdulaziz University, Jeddah, 21589 Saudi Arabia 4 Author for correspondence (begoodson@gmail.com) 1 2 Communicating Editor: Kenneth M. Cameron Abstract—Descurainia is a genus in the Brassicaceae distributed throughout portions of the temperate Old and New World. The genus is most diverse in western North America and western South America, with a smaller center of distribution in the Canary Islands and three additional Old World species. Descurainia is well known for its taxonomic complexity, especially within New World species, on account of its numerous intergrading forms coupled with circumscriptions dependent upon inconsistent and overlapping characters. A molecular-based analysis of Descurainia was conducted using DNA sequences from nuclear ribosomal ITS and non-coding cpDNA regions. Descurainia and related genera form the monophyletic tribe Descurainieae, which is divided into two lineages: 1) Hornungia and Tropidocarpum and 2) Descurainia (including Hugueninia), Ianhedgea, and Robeschia. The genus is strongly-supported as monophyletic, and appears to be of Old World origin with recent diversification within the Canary Islands and the New World. Within the New World, three major well-supported lineages can be identified, with South American taxa distributed into four clades that correlate well with fruit dehiscence and orientation. A phylogeny recovered from combined ITS and cpDNA data is not well-resolved with respect to relationships between some major New World lineages, but suggests that multiple independent dispersals of Descurainia have taken place between North and South America. Substantial incongruence between ITS and cpDNA phylogenies, as well as the presence of mixed ITS sequences, point to a complex evolutionary history involving extensive gene flow and hybridization for North American Descurainia. Keywords—Brassicaceae, hybridization, incongruence, long-distance dispersal, molecular dating, phylogeny. Descurainia has traditionally been divided into two sections (Schulz 1924). Section Sisymbriodendron (Christ) O. E. Schulz comprises seven species endemic to the Canary Islands (Bramwell 1977). These species are self-incompatible suffrutescent perennials, possessing relatively large flowers, conspicuous nectaries, and slightly winged seeds. A recent molecular study (Goodson et al. 2006) suggests that the island taxa are recently derived from a single colonization event, most likely from southwestern Europe. Section Descurainia (= sect. Seriphium O. E. Schulz), consisting of small-flowered herbaceous annuals, biennials, or rarely perennials with inconspicuous nectaries and unwinged seeds, comprises the majority of species in the genus. In contrast to sect. Sisymbriodendron, taxa in sect. Descurainia are generally weedy and wide-ranging. The reproductive biology of this section has not been studied thoroughly, but several species are known to be autogamous (Rollins and Rüdenberg 1971; Best 1977; Bramwell 1977; Wiens 1984; Wiens et al. 1987). Excluding D. sophioides, which extends westward from northern Canada into Siberia, only three members of this section (D. kochii, D. sophia, and D. tanacetifolia) are found in the Old World. (D. tanacetifolia formerly constituted the unispecific genus Hugueninia, but molecular evidence [e.g. Goodson et al. 2006] clearly supports its inclusion within Descurainia). North American species of Descurainia are distributed in the western U. S. A., with only two species (D. incana and D. pinnata) extending east of the Mississippi. Three to four species are endemic to Mexico. Various authors (e.g. Schulz 1924; Detling 1939; Rollins 1993a) have recognized a conflicting number of species, differing primarily in the classification of taxa associated with D. incana, D. incisa, and the wide-ranging and morphologically variable D. pinnata. While sampling was limited, a molecular and morphological study to assess the taxonomic status of D. torulosa (Bricker et al. 2000) uncovered evidence suggesting that D. pinnata and D. incana may not be monophyletic. The number of South American Descurainia taxa is poorly understood. Roughly 20–25 species are generally recognized, Descurainia Webb & Berthel., as traditionally circumscribed, includes approximately 45 species with centers of diversity in the Canary Islands (seven species), North America (ca. 12 species), and western South America (ca. 21 species) (Fig. 1; Appendix 1). Only four species are not wholly confined to these areas, instead either occupying portions of Eurasia or southwestern Europe, or extending from arctic North America into northern Siberia. Members of the genus are characterized by minute dendritic trichomes, pinnate to tripinnate leaves, small yellow or whitish flowers with spathulate petals, filiform fruiting pedicels, and narrow siliques containing seeds that are mucilaginous when wet (Schulz 1924; Al-Shehbaz 1988; Rollins 1993a). Many species also possess unicellular clavate glands, a feature not found elsewhere in the Brassicaceae (Al-Shehbaz et al. 2006). Although one of the principal genera in the Brassicaceae (Mabberley 1997) on the basis of species diversity, there have been no comprehensive molecular systematic studies of Descurainia. This paucity of critical studies is largely due to the taxonomic complexity of Descurainia, especially within New World species that comprise the majority of the genus. Extensive morphological variation exists within numerous species, many wide-ranging and widely-overlapping taxa appear to intergrade endlessly, and descriptions are frequently based on inconsistent and overlapping diagnostic characters which depend on mature fruiting material. Although the extent of hybridization in the genus is unknown, gene flow between populations has undoubtedly been facilitated by the ready dispersability of the small mucilaginous seeds, range contractions and expansions during previous glacial cycles, and the effects of human activity and disturbance over past centuries. Polyploidy (x = 7) is widespread, at least within North American species, with tetraploid (2n = 14) and hexaploid (2n = 42) populations reported for many taxa (Appendix 1). This ploidal level variation, coupled with the presence of morphologically intergrading forms, is suggestive of frequent hybridization. 957 958 SYSTEMATIC BOTANY [Volume 36 ern Argentina possess fruit strongly appressed to the rachis that dehisce from the apex to the base. Species in the second group, distributed primarily throughout low- to mid-elevation portions of Argentina and Chile, feature fruit that is erect, spreading, or reflexed, and valves that dehisce from the base. North American species of Descurainia are separated from their congeners in South America by a distance of approximately 2,200 km. This well-known pattern of disjunction is postulated to have arisen via long-distance dispersal most likely by migrating birds (Wen and Ickert-Bond 2009). Although molecular phylogenetic studies focusing on genera with many species on one continent and one or two on the other have appeared, only a few have produced broadlysampled and well-resolved phylogenies of genera that are species rich on both continents (reviewed in Wen and IckertBond 2009). Descurainia, like the genera in these latter studies, is of moderate size with centers of diversity on both continents, and thus well suited for studying the origins of New World temperate zone disjunctions. Although Descurainia has historically been included in Sisymbrieae subtribe Descurainiinae (Schulz 1924, 1936), molecular evidence (Warwick et al. 2004b; R. A. Price, unpublished data cited in Al-Shehbaz 2003; Beilstein et al. 2006, 2008) does not support the monophyly of either Sisymbrieae or Descurainiinae. Al-Shehbaz et al. (2006) have proposed new tribal classifications for the Brassicaceae, one of which is Descurainieae Al-Shehbaz, Beilstein and E. A. Kellogg, comprising Descurainia, Hornungia Rchb., Ianhedgea Al-Shehbaz and O’Kane, Robeschia, Trichotolinum, and Tropidocarpum Hook. The goals of this study were two-fold. The first was to establish generic boundaries and confirm the taxonomic position of Descurainia by a thorough sampling of the genus with respect to the Descurainieae, Descurainiinae and other related taxa. To accomplish this goal, DNA sequences from the ITS region of the nuclear ribosomal DNA repeat (ITS1, 5.8S rRNA, ITS2; Kim and Jansen 1994) and from the cpDNA trnL intron (Taberlet et al. 1991) were utilized. Because these markers have been widely used within the Brassicaceae for investigating generic level relationships, published sequences were available for incorporation along with data generated by this study. The second, and major, focus of this study was the construction of a molecular-based phylogeny for Descurainia to assess infrageneric relationships and investigate New World biogeography, including the timing, number, and direction of dispersal events. To accomplish this goal, the genus was widely sampled and seven noncoding cpDNA regions and the nuclear ITS region were employed as phylogenetic markers. Materials and Methods Fig. 1. World-wide distribution of Descurainia. Darker areas correspond to regions of greatest species diversity. The primary range of widely-introduced D. sophia, based on Anderberg and Anderberg (1997), is shown in black outline. although the actual number is probably much smaller (I. Al-Shehbaz, pers. comm.). On the basis of geography and morphology, these species can be readily separated into two groups. Species distributed along the Andes from Columbia to north- Sampling—To assess the monophyly of Descurainia, seven representative Descurainia species, six other species of Descurainieae (i.e. Hornungia alpina, H. petraea, H. procumbens, Ianhedgea minutiflora, Robeschia schimperi, and Tropidocarpum gracile), and a selection of representatives from other tribes, primarily Brassicaceae “lineage I” (Al-Shehbaz et al. 2006; Bailey et al. 2006), were included in the analysis (Appendix 2). Species of Descurainia were selected to represent major lineages identified from preliminary phylogenetic analysis of a broader data set. Arabis alpina, Brassica rapa, and Sisymbrium altissimum were used as outgroups. For the study of relationships within Descurainia, DNA was isolated from 135 Descurainia samples (Appendix 3), representing all ten Old World species (25 accessions), ten of 13 named North American species (71 accessions), and 12 of 21 named South American species (39 accessions), and from putative congeners Ianhedgea minutiflora (one accession) and Robeschia schimperi (two accessions). On the basis of preliminary analyses of our data and published studies (Warwick et al. 2004b; Al-Shehbaz 2011] GOODSON ET AL.: DESCURAINIA et al. 2006; Bailey et al. 2006) Arabidopsis thaliana, Sisymbrium altissimum and Smelowskia americana were included as outgroups (Appendix 3). Leaf material was field-collected and dried over silica, or harvested from cultivated plants grown from seed in the greenhouse at the University of Texas at Austin. Total DNA was extracted using the CTAB method of Doyle and Doyle (1987) followed by purification using cesium chloride and ethidium bromide gradients. Additionally, DNA was isolated from herbarium specimens following the protocol in Loockerman and Jansen (1996). PCR Amplification and DNA Sequencing—To ascertain the phylogenetic position of Descurainia, cpDNA trnL intron and nuclear ITS regions were amplified for six Descurainia taxa and three close relatives. Published and unpublished ITS and trnL sequences of other closely related taxa and a selection of representatives from other tribes (Al-Shehbaz et al. 2006; Bailey et al. 2006) were also incorporated into the data set (Appendix 2). Although no trnL sequence was available for Tropidocarpum gracile, the ITS sequence was provided by Robert A. Price and included in the data set. Seven non-coding cpDNA regions and the nuclear ITS region were utilized as phylogenetic markers in the broader study (Table 1). DNA regions were amplified via PCR in 50 µL volumes containing 5 µL of 10 × buffer, 4 µL of 25 mM MgCl2, 4 µL of 0.25 µM dNTPS, 0.5 µL of a 100 µM solution of each primer, 0.5 µL of Taq polymerase and 1 µL of unquantified DNA template. For a few samples extracted from herbarium material, PCR amplifications were accomplished in 25 µL volumes containing 0.25 µL of a 100 µM solution of each primer, 0.25 µL of Taq polymerase, 0.5–1 µL of unquantified DNA template, and 12.5 µL of FailSafe PCR 2 × Premix A, D, E or H (Epicentre, Madison, Wisconsin). For ITS amplifications, reaction conditions were as follows: one cycle of denaturation at 96°C for 3 min, annealing at 50°C for 1 min, and extension at 72°C for 1 min, followed by 35 cycles of 95°C for 1 min, 50°C for 1 min, and 72°C for 45 sec, with a final extension step of 72°C for 7 min. Chloroplast regions were amplified using the following conditions: denaturation at 96°C for 3 min, followed by 35 cycles of 94°C for 35 sec, 50°C for 45 sec, and 72°C for 1 min, with a final extension of 72°C for 12 min. Following amplification, PCR products were cleaned with Qiagen (Valencia, California) spin columns following the manufacturer’s protocols. Sequencing reactions were carried out using Big Dye Terminator chemistry. The sequencing products were cleaned with Centri-cep columns (Applied BioSystems, Foster City, California) and sequenced on either an MJ Research (Waltham, Massachusetts) BaseStation or ABI Prism 3730 automated sequencer. Sequences were deposited in GenBank (Appendices 2 and 3) and data matrices are archived in TreeBASE (study number S11210). Direct sequencing of the ITS region of 10 Descurainia accessions generated traces exhibiting double peaks at multiple nucleotide positions. These samples were cloned with a TOPO TA kit (Invitrogen with vector pCR 2.1-TOPO, Carlsbad, California) using 1/3 the recommended reaction volumes. For each cloning reaction, 5–10 positively transformed colonies were amplified and products sequenced. PCR amplification of clones containing ITS inserts were carried out as follows: 94°C for 10 min, followed by addition of Taq polymerase while the reactions were held at 72°C for 5–10 min, and then 37 cycles of 94°C for 1 min, 48°C for 1 min, and 72°C 959 for 1 min, followed by a final extension step of 72°C for 12 min. For each cloned sample, sequencing of transformed colonies generated two distinct sequence types, i.e. clones for a given type either formed a monophyletic group or, if not forming a monophyletic group of clones, were identical or differed by a single autapomorphic nucleotide substitution from other clones of that type. For each sample, two representative sequences (one for each putative parental type) were selected for inclusion. There were also five sequences (from three accessions), which upon visual inspection were clearly chimeric and probably an artifact of PCR-mediated recombination (Bradley and Hillis 1997; Cronn et al. 2002) or the result of incomplete concerted evolution following in vivo recombination (Álvarez and Wendel 2003; Buckler et al. 1997). These were excluded along with the redundant cloned sequences. Phylogenetic Analyses—Sequences were edited with Sequencher 4.1.2 (Gene Codes Corp., Ann Arbor, Michigan) and aligned with ClustalX (Thompson et al. 1997) followed by manual adjustments. Indels that were potentially phylogenetically informative were coded as binary (presence/absence) characters following Simmons and Ochoterena (2000) and appended to the alignment. Parsimony analyses were performed on each data set with PAUP* 4.0b10 (Swofford 2002). For the ITS, trnL, and combined ITS-trnL Brassicaceae data sets, heuristic searches were conducted using 10,000 random addition sequence replicates, holding 10 trees at each step, and with tree-bisection-reconnection (TBR) branch swapping, characters equally weighted, and gaps treated as missing. For the larger ITS, cpDNA, and combined ITS-cpDNA Descurainia data sets, 20 independent parsimony ratchet (Nixon 1999) runs of 200 iterations each were carried out in PAUP* using batch files generated by PAUPRat (Sikes and Lewis 2001). Support for internal nodes was assessed using bootstrap analysis (Felsenstein 1985). For the Brassicaceae data sets, 500 bootstrap replicates were conducted with 100 random additions per replicate, holding 10 trees at each step; for the Descurainia data sets, this entailed 100 bootstrap replicates of 10 random additions each, holding one tree at each step and saving no more than 500 trees of length greater than or equal to 200 steps in each replicate. Bootstrap support was categorized as strong (> 85%), moderate (70–85%), weak (50–69%), or unsupported (< 50%). Bayesian analyses were carried out separately on individual and combined data sets using MrBayes 3.1 (Ronquist and Huelsenbeck 2003). Bestfit models of evolution for each data set were selected in MrModeltest 2.2 (Nylander 2004) based on the Akaike information criterion (Akaike 1974; Posada and Buckley 2004). For those data sets containing both nucleotide and coded indels, separate evolutionary models were applied to the data partitions with all parameters unlinked except for topology and branch length; the model(s) selected by MrModeltest were applied to each nucleotide partition and the BINARY model (with coding bias set to variable) was applied to the coded indels. Two independent analyses were performed on each data set. Each analysis was run for two to six million generations with four Markov chains (three heated and one cold) and trees saved every 100 generations. Trees were checked for stationarity by plotting log likelihood values vs. generation, and trees from the burn-in period were discarded. A 50% majority-rule consensus tree was Table 1. Sequence characteristics of DNA regions used for the 150 sample set. *Including outgroups and indels; ** including only ingroups and indels; *** excluding uninformative characters; †including one inversion; ††data missing for ndhF-rpl32 (D. antarctica D52, D. incisa D21, and D. incisa subsp. filipes B195) and rps16 (D. antarctica D52 and D. pinnata F11). Sources for primers are: rps16 intron (Oxelman et al. 1997); trnDguc-trnEuuc, trnEuuc-trnTggu and psbZtrnfMcau (Demesure et al. 1995); trnCgca-petN and petN-psbM (Shaw et al. 2005). Primers for ndhF-rpl32 (ndhF-F: 5′-ACTGGAAGTGGAATGAAAGG-3′; rpl32-R: 5′-GCTTTCAACGATGTCCAATA-3′; internal sequencing primers ndhF-iF: 5′-CGTGTAAATCTTTGTTCTAT-3′; rpl32-iR: 5′-ATAGAACAAA GATTTACACG-3′) were designed based on the Arabidopsis thaliana chloroplast genome (NC_000932). trnC-petN Seq. length (bp) # taxa Alignment length % missing/% gaps No. of nonautapomorphic indels No. informative characters (%)* No.informative characters (%)** No. of MPTs Length of MPTs Consistency index *** Retention index petN-psbM trnD-trnE trnE-trnT psbZ-trnfM ndhF-rpl32 rps16 intron Combined chloroplast ITS 519–585 137 613 0/7.5 9 579–619 137 647 0/7.8 5 517–541 137 573 0/8.0 3 534–775 135 820 0/26.1 1 647–731 136 788 0.2/15.5 9 656–940 133 1,044 0.3/15.0 9 787–828 134 866 0.2/6.3 3 4,423–4,857 135†† 5,351 0.8/12.8 39† 596–614 150 627 0.1/2.5 0 73 (11.7%) 62 (9.5%) 62 (10.7%) 75 (9.1%) 88 (11.0%) 134 (12.7%) 87 (10.0%) 581 (10.8%) 127 (20.3%) 68 (10.9%) 45 (6.9%) 47 (8.2%) 51 (6.2%) 62 (7.8%) 89 (8.5%) 59 (6.8%) 505 (9.4%) 103 (16.4%) – – 0.835 0.962 – – 0.753 0.948 – – 0.745 0.948 – – 0.740 0.935 – – 0.653 0.886 – – 0.765 0.918 (3419) 1538 0.713 0.918 (4020) 338 0.630 0.925 – – 0.754 0.921 960 SYSTEMATIC BOTANY constructed in PAUP* from the remaining trees. Branches with posterior probabilities ≥ 95% were considered to be strongly-supported, with posterior probabilities < 95% indicating weak support. To explore alternative hypotheses regarding some New World relationships, a 95% credible set of trees (Huelsenbeck and Rannala 2004) was constructed from the phylogenies recovered by Bayesian analysis of a combined ITS-cpDNA data set. The “filter constraints” command in PAUP* was subsequently used to search this 95% credible set for topologies consistent with alternative hypotheses of interest. Maximum likelihood (ML) analyses were conducted on selected data sets, excluding coded gap characters, using PAUP* or GARLI 0.95 (Zwickl 2006). For ML analyses using GARLI, the tree with the best likelihood score was chosen from the results of 10–15 independent runs. ML bootstrapping was conducted in GARLI using 300 replicates. Substitution models for both PAUP* and GARLI ML searches were chosen based on the Akaike information criterion calculated in Modeltest 3.7 (Posada and Crandall 1998). Shimodaira-Hasegawa (SH) tests (Shimodaira and Hasegawa 1999; Goldman et al. 2000) of alternative hypotheses were conducted on constrained and unconstrained ML trees as implemented in PAUP* with 1,000 RELL bootstrap replicates (one-tailed test). The incongruence length difference (ILD) test as implemented in PAUP* (Farris et al. 1994) was used to assess global topological incongruence. Each test consisted of 100 replicates, with 10 random additions per replicate, holding 20 trees per step. For ITS vs. combined cpDNA data sets, if the ILD test indicated significant data heterogeneity, conflicting clades were identified by means of visual inspection and degree of partitioned branch (Bremer) support (Baker and DeSalle 1997; Baker et al. 1998). Taxa which appeared to contribute to the observed incongruence were removed until the ILD test indicated no significant conflict. Partitioned branch support indices were calculated using the program TreeRot.v2 (Sorenson 1999). To explore possible New World dispersal patterns in Descurainia, continent of distribution was mapped onto topologies representing phylogenetic relationships between major New World lineages recovered from phylogenetic analysis of the combined ITS-cpDNA data set. This was accomplished using MacClade 4.0 (Maddison and Maddison 2000) with Fitch parsimony optimization (unordered characters and unweighted character state changes). Estimates of Divergence Times–—Absolute divergence times were calculated from the ITS and cpDNA data. To eliminate zero- or near zero-length terminal branches (Sanderson 2004) and expedite computation, redundant taxa were removed by pruning identical or nearly-identical sequences from the full data sets. To eliminate arbitrary zero-length branches at the root of the tree in PAUP*, Aethionema grandiflorum (GenBank accession DQ452067), representing an early diverging genus in the Brassicaceae (Zunk et al. 1999; Koch et al. 2003a; Beilstein et al. 2006), was also incorporated as an extra outgroup. The best ML tree was generated from the resulting 27-taxon ITS data set under the SYM + I + Γ model of evolution using GARLI 0.95. Rate heterogeneity was assessed using a likelihood ratio test (Felsenstein 1981; Huelsenbeck and Rannala 1997) in PAUP* to compare ML trees generated with and without enforcing a molecular clock. For the ITS data a molecular clock could be rejected at the p = 0.01 level but not at p = 0.05. For this data set, divergence times were consequently estimated under both the assumption of a molecular clock and using a relaxed molecular clock model. As a calibration point, the age of the node joining Arabidopsis to Brassica was fixed at 43.2 mya (Beilstein et al. 2010), and the program r8s 1.71 (Sanderson 2004) was used to estimate divergence times on the reduced ITS ML tree. The distant outgroup Aethionema was pruned from the tree, and divergence times were first calculated in r8s under the assumption of a molecular clock using the Langley-Fitch (LF) method (Langley and Fitch 1974) with the truncated Newton (TN) algorithm. Divergence times were also estimated using penalized likelihood rate smoothing (Sanderson 2002). The cross-validation procedure in r8s was used to determine an optimal smoothing level of 1,000, and divergence times of selected nodes were calculated with this smoothing factor using the PL method with the TN algorithm. All solutions were evaluated for correctness using the checkGradient option. Confidence intervals on divergence dates were generated using a parametric bootstrapping approach as recommended by Sanderson (2004). SG Runner, a graphical user interface to Seq-Gen 1.3.2 (Rambaut and Grassly 1997), was used to generate 100 bootstrapped data sets based on the reduced ML tree, its branch lengths, and the model parameters selected by Modeltest for the original data set. The ML trees with different branch lengths but the same topology as the original tree were then generated in PAUP* from these data sets. These trees were imported into r8s, divergence times were estimated for nodes of interest, and the [Volume 36 “profile” command was used to summarize rate and time information for each node across the collection of trees. Divergence times were also estimated using cpDNA data with the same calibration point employed in the ITS-based dating. A ML tree was generated using PAUP* under a GTR + Γ model from a reduced cpDNA data set of 25 taxa into which sequence data from Cardamine diphylla, Rorippa palustris and Aethionema grandiflorum had been incorporated. A likelihood ratio test to assess rate constancy strongly rejected a molecular clock ( p << 0.001). Divergence times were calculated in r8s using the PL method with the TN algorithm and a smoothing factor of 100, and confidence intervals were estimated as described for the ITS data. Results Analysis of ITS and trnL Data to Assess the Monophyly of Descurainia—The ITS data set for 36 taxa had 647 characters, including 644 nucleotide positions and three indels, with 5.9% gaps and 0.5% missing characters. Three hundred four characters (50.0%) were variable and 212 (32.8%) were parsimony informative. Parsimony analysis of the ITS data set recovered 553 trees of 953 steps (CI [excluding informative characters] = 0.44; RI = 0.60) (Supplemental Fig. 1). Bayesian analysis (SYM + I + Γ, four million generations) resulted in a strict consensus tree (Supplemental Fig. 1) that was essentially identical to the parsimony topology. In the strict parsimony and Bayesian consensus trees, taxa classified by Al-Shehbaz et al. (2006) as constituting the tribe Descurainieae (except Trichotolinum, which was not sampled in this study) are well supported as a distinct group (bootstrap support [BS] = 91%, posterior probability [PP] = 100%). The Descurainieae are partitioned into two lineages: 1) Hornungia and Tropidocarpum (BS = 70%, PP = 96%) and 2) Descurainia, Ianhedgea, and Robeschia (BS = 86%, PP = 100%). Within the latter clade, Ianhedgea is sister to a polytomy (BS = 71%, PP = 99%) composed of D. kochii + D. sophia (BS = 84%, PP = 100%), Robeschia, and a stronglysupported (BS = 100%, PP = 100%) New World-Canary Island-D. tanacetifolia clade. The Bayesian tree places Robeschia at the base of this polytomy, but support is weak (PP = 66%). The aligned trnL intron data set was 554 base pairs (bp) in length, including gaps (15.1%) and missing (0.84%) data. In addition, eight autapomorphic indels were appended to the data set. Of the 562 characters in the resulting data set, 134 (23.8%) were variable and 60 (10.7%) were parsimonyinformative. The same taxa were included in the trnL data set as in the ITS data set, excluding Tropidocarpum gracile, for which a trnL intron sequence was not available. Parsimony analysis of the trnL data set for 35 taxa generated 8,734 most parsimonious trees (length = 185 steps, CI = 0.717, RI = 0.843) (Supplemental Fig. 2). Compared to the ITS tree, the strict consensus tree from parsimony and Bayesian (GTR + I, three million generations) analyses is not well-resolved, although it is generally congruent. Like the ITS results, a strongly-supported (BS = 97%, PP = 100%) Descurainia New World-Canary Island-D. tanacetifolia clade is present, but its relationship to other members of the Descurainieae, and most of the other included taxa, is unresolved. In contrast to the ITS tree, where D. sophia is sister to D. kochii, the trnL tree joins D. sophia with Ianhedgea. Support for this relationship, however, is weak (BS = 71%, PP = 90%). As in the ITS results, Hornungia alpina is strongly-supported (BS and PP = 100%) as sister to H. petraea, but additional relationships within Descurainieae and between Descurainieae and other tribes is largely unresolved. The Bayesian phylogeny weakly supports (PP = 89%) a polytomy comprising various members of the Descurainieae and a strongly-supported (BS = 94%, PP = 99%) Smelowskieae. 2011] GOODSON ET AL.: DESCURAINIA 961 Fig. 2. One of 41 most parsimonious trees derived from combined ITS-trnL data to assess the monophyly of Descurainia. Branch lengths are indicated above branches; bootstrap values > 50%/Bayesian posterior probabilities are below branches. Dashed lines indicate branches that collapse in the strict consensus tree. Species tagged with an * belong to the subtribe Descurainiinae as circumscribed by Schulz (1924); tribal classifications on the right-hand side are those proposed by Al-Shehbaz et al. 2006. The ITS and trnL data were combined into a single data set comprising 1,198 bp (1.9% gaps and 9.9% missing) and 11 indels. The high percentage of missing data was due to the absence of the trnL sequence for Tropidocarpum gracile. Because the exclusion of T. gracile did not appreciably affect the outcome of preliminary phylogenetic analyses, this taxon was retained in the combined data set. Of the 1,209 charac- ters in the combined data set, 438 (36.2%) were variable and 272 (22.5%) were parsimony-informative. Visual comparison of trees from the separate data sets, as well as results from the larger-scale study to be described later, suggested incongruence due to the varying placement of D. sophia. The ILD test (p = 0.18) indicated that the two partitions were not heterogeneous, however, and parsimony analysis of the combined 962 SYSTEMATIC BOTANY data set with and without D. sophia gave essentially identical results. Parsimony analysis of the combined data set (with D. sophia included) generated 41 most parsimonious trees of 1,149 steps (CI = 0.470, RI = 0.627) (Fig. 2). The Descurainieae are strongly-supported (BS = 96%) as a distinct group comprised of Hornungia and Tropidocarpum (BS = 77%) as a sister lineage to Descurainia, Robeschia, and Ianhedgea (BS = 89%). Within the latter clade, New World and Canary Island Descurainia, along with D. tanacetifolia, form a stronglysupported clade (BS = 100%), but relationships among D. sophia, D. kochii, R. schimperi, and I. minutiflora are largely unresolved (BS ranging from < 50–55%). The Descurainieae are sister to the Smelowskieae, but the support is weak (BS = 53%). Bayesian analysis (SYM + I + Γ for the ITS partition, GTR + I for trnL, and BINARY for trnL indels, six million generations) of the combined data set generated a tree of similar topology to that of parsimony, but the presence or absence of D. sophia affected which taxon diverges first within the Descurainia/Ianhedgea/Robeschia lineage. When D. sophia is included, Robeschia is placed sister to the clade (PP = 90%) and D. kochii is strongly-supported (PP = 99%) as sister to D. sophia + Ianhedgea (PP = 100%). When D. sophia is excluded from the analysis, Ianhedgea is placed sister to the clade (PP = 99%) with Robeschia next (PP = 99%) followed by D. kochii (PP = 88%). Regardless of whether D. sophia is included or not, Bayesian analyses join the Descurainieae and Smelowskieae as sister tribes (PP = 96–97%). Analysis of ITS Data to Assess Relationships within Descurainia—The ITS data set for 150 accessions and representative cloned samples was readily alignable, comprising 627 nucleotide positions including gaps (2.5%) and missing (0.1%) characters. Two hundred twenty-eight characters (36.4%) were variable and 127 (20.3%) were parsimony informative (Table 1). Parsimony analysis of the ITS data generated 4,020 most parsimonious trees of 388 steps (CI = 0.63; RI = 0.93) (Fig. 3). Bayesian analysis (SYM + Γ, three million generations) produced a consensus tree (Fig. 3) that recovered the same major clades as parsimony. In contrast to the analysis to assess the monophyly of Descurainia, these results place Robeschia as sister to Descurainia with moderate to strong support (BS = 73%, PP = 100%). This support is sensitive to the presence of Hornungia and Tropidocarpum; when those taxa were added to the analysis, bootstrap support for the branch uniting Robeschia with Descurainia dropped from 73% to 55%. Within Descurainia, D. sophia, and D. kochii form a clade (BS = 91%, PP = 100%) that is sister to the remainder of the genus. This well-supported clade (BS = 100%, PP = 100%) comprises a polytomy with four distinct lineages. Lineage “A” (BS = 86%, PP = 100%) is exclusively North American, and lineage “B” (BS = 81%, PP = 100%) is North American except for the presence of half of the South American D. antarctica clones. The third main lineage, clade “C”, is strongly-supported (BS = 99%, PP = 100%). It includes North American taxa and all South American species sampled (except two D. antarctica clones). Of the four lineages, C had the greatest sequence divergences, ranging up to 2.46%. There is little resolution within lineage C, but a few weakly supported clades consistent with geography and morphology are evident. One of these (clade C-I) encompasses D. pinnata and the Mexican endemic D. virletii (BS = 57%, PP = 99%); another (C-II) comprises all South American taxa characterized by fruit spreading away from the rachis (BS = 61%, PP = 93%). Every accession of D. pinnata [Volume 36 exhibiting sequence polymorphism yielded clones in both lineage B and lineage C. The fourth lineage, “D”, is weakly supported (BS = 53%, PP = 93%), but includes all the species from the Canary Islands, and is sister (BS = 55%, PP = 99%) to D. tanacetifolia subsp. suffruticosa. Surprisingly, clade D + D. tanacetifolia is joined in a trichotomy with not only another accession of D. tanacetifolia (subsp. tanacetifolia) but also with New World clade B. Bootstrap support for the branch leading to this trichotomy is weak, although it is stronglysupported in the Bayesian topology (BS = 57%, PP = 100%). To further evaluate support for this relationship, a SH test was conducted comparing this topology to one where New World Descurainia species (clades A, B and C) were constrained to monophyly. The outcome of the SH test ( p = 0.12) indicated that a tree with New World taxa monophyletic could not be rejected. Chloroplast Data—To assess incongruence, the ILD test was applied to all possible pairings of the seven cpDNA data sets. All data combinations were supported as homogeneous at the p = 0.01 level, but were rejected at the p = 0.05 level for the combinations trnD-trnE vs. rps16 and psbZ-trnfM vs. ndhFrpl32 ( p = 0.04 and 0.02, respectively). Since the topologies generated from the individual cpDNA partitions (not shown) do not appear to seriously conflict, and the chloroplast genome is inherited uniparentally as a single unit and does not usually undergo recombination, this borderline significant heterogeneity is assumed to be a type I error (i.e. inference of incongruence where none exists) to which the ILD test has been shown to be highly susceptible (Dolphin et al. 2000; Barker and Lutzoni 2002; Darlu and Lecointre 2002; Dowton and Austin 2002). The data from the seven non-cpDNA coding regions were consequently combined into a single data set. Sequence characteristics for the individual cpDNA regions and combined data set are found in Table 1. The combined cpDNA data set for 135 accessions contained 5,351 nucleotide positions including gaps (12.8%) and missing (0.8%) characters. Thirty-eight indels, ranging in length from four to 278 bp, and one five-bp inversion, were binary-coded and appended to the data set. The resulting data set comprised 5,390 characters, of which 1,107 (20.5%) were variable and 581 (10.8%) were parsimony informative. Parsimony analysis of the combined cpDNA data set for 135 accessions yielded 3,419 most parsimonious trees from the 4,020 trees produced using the parsimony ratchet (length = 1,538, CI = 0.713, RI = 0.918) (Fig. 4). The cpDNA data place Robeschia schimperi sister to Descurainia with moderate support (BS = 81%). In contrast to the ITS tree where D. sophia and D. kochii form a separate clade sister to the remainder of the genus, the cpDNA tree strongly supports D. kochii as sister to the rest of the genus (BS = 96%) and D. sophia as sister to the remaining taxa (BS = 100%). The cpDNA phylogeny is generally consistent with the results obtained from the ITS data set with respect to lineages A-D. The Canary Island taxa (lineage D) are strongly-supported as monophyletic (BS = 100%) and sister (BS = 100%) to European D. tanacetifolia subsp. suffruticosa C6. This clade in turn is sister to the other subspecies of D. tanacetifolia, subsp. tanacetifolia B111 (BS = 100%). The Canarian/European lineage is sister with strong support (BS = 100%) to New World Descurainia. Within the New World clade, lineages A and B are still present (BS = 100% and 82%, respectively) and form a strongly-supported clade (BS = 91%) along with one accession of D. incisa which is found in clade A in the ITS phylogeny. The relationship of the remainder of 2011] GOODSON ET AL.: DESCURAINIA 963 Fig. 3A. Strict consensus of 4,020 most parsimonious trees derived from ITS data to assess relationships within Descurainia using the parsimony ratchet. Bootstrap values > 50%/Bayesian posterior probabilities are indicated below branches. Generic names are abbreviated as follows: A. = Arabidopsis, D. = Descurainia, I. = Ianhedgea, R. = Robeschia, and S. = Sisymbrium (altissimum) or Smelowskia (americana). The designations A, B, C, C-I, C-II, and D refer to clades described in the text. Distributions are abbreviated as follows: NA = North America, SA = South America, CI = Canary Islands, and EU = Europe, Eurasia and/or Middle East. Additional clades recovered from parsimony bootstrap or Bayesian analysis which do not appear in the strict consensus tree are marked by: * (PP = 81%); ** (all members of clade C except these accessions) (PP = 72%); *** (BS = 63%; PP = 81%). Bayesian analysis also groups D. tanacetifolia B111 as sister to clade D (55%). Accessions that exhibit incongruence with respect to the cpDNA topology are indicated as follows: solid square- mixed ITS types; slashed square - between-clade; open square- within-clade. 964 SYSTEMATIC BOTANY [Volume 36 Fig. 3B. Strict consensus of 4,020 most parsimonious trees derived from ITS data to assess relationships within Descurainia using the parsimony ratchet. See caption for Fig. 3A for more information. 2011] GOODSON ET AL.: DESCURAINIA 965 Fig. 4A. Strict consensus of 3,419 most parsimonious trees derived from combined cpDNA data to assess relationships within Descurainia using the parsimony ratchet. Bootstrap values > 50%/Bayesian posterior probabilities are indicated below branches. Generic names are abbreviated as follows: A. = Arabidopsis, D. = Descurainia, I. = Ianhedgea, R. = Robeschia, and S. = Sisymbrium (altissimum) or Smelowskia (americana). The designations A, B, C, C-I, C-II, C-III, C-IV, C-V and D refer to clades described in the text. Distributions are abbreviated as follows: NA = North America, SA = South America, CI = Canary Islands, and EU = Europe, Eurasia and/or Middle East. Two additional branches within clade B recovered from Bayesian analysis with weak support (PP < 72%) are not shown. 966 SYSTEMATIC BOTANY [Volume 36 Fig. 4B. Strict consensus of 3,419 most parsimonious trees derived from combined cpDNA data to assess relationships within Descurainia using the parsimony ratchet. See caption for Fig. 4A for more information. the New World taxa (which are primarily found in clade C in the ITS tree) to the A+B clade is unresolved, but provides further support for clades C-I (D. pinnata + D. virletii, BS = 99%) and C-II (South American spreading fruit, BS = 100%). Clade C-II is grouped with D. sophioides, D. obtusa subsp. obtusa, D. obtusa subsp. adenophora, and one sample of D. californica, but support for this clade is weak (BS = 55%). In addition, all South American taxa with fruit appressed to the rachis are 2011] GOODSON ET AL.: DESCURAINIA strongly-supported in two lineages (henceforth designated as C-III and C-IV, BS = 100% in both). The 50% majority-rule consensus tree generated from Bayesian analysis (GTR + I + Γ, four million generations) is nearly identical to the parsimony tree, with support values for most of the branches recovered under parsimony generally above 98%. The only significant topological difference between the trees recovered from the two methods is that all unresolved “clade C” taxa are placed by Bayesian inference in a strongly-supported (PP = 100%) clade that also includes C-II (PP = 100%). Sequencing of ITS clones revealed that 10 accessions possessed both clade B and clade C (either C-I or C-II) types. In the cpDNA phylogeny, all but one of these accessions are placed in clade B. The exception, D. pinnata subsp. intermedia C19, is found in clade C-I in the cpDNA tree. While there are many similarities between the cpDNA and ITS trees, there are a number of obvious incongruences, especially within North American taxa (Fig. 3). Outside of North America, the only major incongruence is that of D. sophia; in the ITS phylogeny it is sister to D. kochii, whereas in the cpDNA phylogeny it is sister to D. tanacetifolia and species of the New World and Canary Islands. Tests for Incongruence—Before combining ITS and cpDNA data, incongruent and redundant taxa were identified and removed as described under Materials and Methods. This process was straightforward with two exceptions. First, the considerable incongruence within lineage A could be resolved by removal of varying sets of taxa, so that the choice of accessions remaining in that clade in the final combined data set represents only one alternative among several. Fortunately, which set of taxa was chosen did not affect the relationship of major lineages in the combined topology. Secondly, the ILD test detected significant incongruence when clade D was included (p = 0.05 with compared to p = 0.17 without). Clade D (the Canary Island/D. tanacetifolia clade) is stronglysupported as monophyletic and sister to all New World taxa in the cpDNA phylogeny. While the topology of the most parsimonious ITS tree conflicts with such a relationship, the SH test discussed previously does not reject it. With respect to clade D and New World species, the phylogeny based on the combined data set is consistent with morphology and geography. Based on these observations, and noting that phylogenetic accuracy does not always depend on congruent data sets (Hipp et al. 2004), clade D was accordingly retained in the combined ITS-cpDNA data set. Analysis of Combined Data—The combined ITS and cpDNA data set for 74 accessions included 5,894 bp, with 0.32% missing characters and 10.9% gaps. Twenty-three indels were coded as binary characters and appended to the combined data set. The resulting data set comprised 5,917 characters, of which 1,159 (19.6%) were variable and 441 (7.5%) were parsimony informative. The ITS partition contributed 20.6% (91) and the cpDNA partition 79.4% (350) of the parsimony informative characters. Parsimony analysis of the combined data set was carried out with Arabidopis thaliana, Sisymbrium altissimum, and Smelowskia americana as outgroups. For the Bayesian analysis, a mixed model analysis was conducted (two million generations) with the SYM + Γ, GTR + I + Γ, and BINARY models applied to the ITS, cpDNA, and indel partitions, respectively. Parsimony analysis recovered 4,020 trees of 1,534 steps (CI = 0.709, RI = 0.907) (Fig. 5). The strict consensus Bayesian tree 967 was identical to the parsimony tree except for one weakly supported branch described below. The combined ITS-cpDNA tree provides strong support for the major clades previously observed in the trees from the separate data sets. Robeschia schimperi is placed sister (BS = 92%, PP = 100%) to Descurainia, with D. kochii, in the absence of D. sophia, sister to the rest of the genus (BS = 100%, PP = 100%). Canary Island taxa and D. tanacetifolia (BS = 100%, PP = 100%) are sister (BS = 100%, PP = 100%) to the New World species. New World Descurainia are strongly-supported (BS = 100%, PP = 100%) as a monophyletic group composed of clades A, B and C (all with BS and PP = 100%). Clade C is sister to a lineage (BS = 89%, PP = 100%) comprising clades A and B. Within clade C are the North and South American sub-lineages C-I (North American D. pinnata and D. virletii), C-II (South American spreading fruit + North American D. obtusa), C-III (South American appressed fruit), C-IV (South American appressed fruit), and C-V (South American D. cumingiana var. tenuissima), all of which have bootstrap support values ranging from 97–100% and Bayesian posterior probabilities of 100%. Relationships among these lineages are largely unresolved: parsimony analysis weakly (BS = 64%) joins clades C-I, C-II, and C-III in a polytomy and places C-IV and C-V in a sister relationship (BS = 58%). Bayesian analysis (PP = 94–95%) supports these same relationships, and weakly (PP = 87%) implies that clades C-I and C-III are most closely related. Most of the remaining sampled Bayesian trees that differ from this topology place C-II, rather than C-III, as sister to C-I. Optimization of New World Distribution on Phylogenies— An examination of the 4,020 most parsimonious trees from the combined ITS-cpDNA data set revealed only two distinct resolutions of major New World lineages (Fig. 6). These two topologies were also present in a 95% credible set of trees from the Bayesian analysis. The topologies differ in the placement of North American clade C-I and South American C-III with respect to South American C-II + North American D. obtusa. The first topology (Fig. 6a), representing 35% of most parsimonious trees and 84% of the set of 95% credible Bayesian trees, places clade C-III sister to C-II + D. obtusa. The second topology (Fig. 6b), representing 65% of most parsimonious trees and 10% of 95% credible Bayesian trees, groups clade C-I with C-II + D. obtusa. When New World continental distribution was traced onto simplified trees representing these two topologies, five most parsimonious reconstructions were recovered for the first topology and seven reconstructions were generated for the second. All reconstructions are consistent with four separate dispersals of Descurainia between North and South America. Nine of these reconstructions are illustrated in Fig. 6. The remaining three reconstructions are not shown because they are inconsistent with a North American origin for one parent of D. antarctica strongly suggested by the molecular data. Divergence Time Estimates—Divergence times from the ITS data were virtually identical regardless of which dating method (LF or PL) was employed (Fig. 7). The PL algorithm dated the origin of the Descurainieae to 23.18 +/− 2.2 mya (Hornungia-Ianhedgea split) with the last common ancestor of Canary Island taxa arising 1.57 +/− 0.57 mya. The rate of sequence evolution, 3.3 +/− 0.5 × 10−9 substitutions/site/ year, is similar to rates reported for the ITS region in other annual and perennial herbs including crucifers (Richardson et al. 2001; Koch et al. 2006). Divergence times estimated from 968 SYSTEMATIC BOTANY [Volume 36 Fig. 5. One of 4,020 most parsimonious trees recovered from combined ITS-cpDNA data to assess relationships within Descurainia using the parsimony ratchet. Bootstrap values > 50%/Bayesian posterior probabilities are indicated below branches. Dashed lines indicate branches that collapse in the strict consensus tree. Generic names are abbreviated as follows: A. = Arabidopsis, D. = Descurainia, I. = Ianhedgea, R. = Robeschia, and S. = Sisymbrium (altissimum) or Smelowskia (americana). The designations A, B, C, C-I, C-II, C-III, C-IV, C-V and D refer to clades described in the text. Distributions are abbreviated as follows: NA = North America, SA = South America, CI = Canary Islands, and EU = Europe, Eurasia and/or Middle East. the cpDNA data (Fig. 8) give the last common ancestor of Robeschia-Descurainia, New World Descurainia, and Canary Island Descurainia as 22.45 +/− 1.41 mya, 5.20 +/− 0.50 mya, and 2.40 +/− 0.45 mya, respectively. The overall sequence evolution rate for these cpDNA regions was calculated to be 8.4 +/− 0.4 × 10−8 substitutions/site/year. This rate is comparable to typical rates of evolution for other noncoding cpDNA regions (Richardson et al. 2001). 2011] GOODSON ET AL.: DESCURAINIA 969 Fig. 6. Most parsimonious reconstructions from optimization of New World Descurainia continental distribution (North America [NA] or South America [SA]) on topologies from phylogenetic analysis of the combined ITS-cpDNA data set. Outgroups and Old World taxa are not shown because they have no effect on the outcome of the reconstructions. A. Optimization on the topology found in 35% of most parsimonious trees and 84% of the Bayesian 95% credible set of trees. One additional reconstruction, which is inconsistent with the recent introduction of D. antarctica to SA seen in molecular data, is not illustrated. B. Optimization on the topology found in 65% of most parsimonious trees and 10% of the Bayesian 95% credible set of trees from phylogenetic analysis of the combined ITS-cpDNA data set. Two additional reconstructions, inconsistent with the recent introduction of D. antarctica to SA seen in molecular data, are not illustrated. Discussion Taxonomic Position and Monophyly of Descurainia—This study provides support for the monophyly of the recently designated tribe Descurainieae (Al-Shehbaz et al. 2006). With the exception of Trichotolinum, all putative members of Descurainieae were included and form a monophyletic group. Another recent molecular-based study (Beilstein et al. 2008) has also addressed the monophyly of Descurainieae. They report that combined cpDNA ndhF and nuclear PHYA data generated a phylogeny in which the position of Hornungia with respect to other Descurainieae sampled was equivocal. While sampling across the Brassicaceae was much more extensive in their paper, they only included four putative Descurainieae, i.e. Descurainia sophia, Hornungia procumbens, Ianhedgea, and Robeschia. Our data indicate that Descurainieae is divided into two distinct lineages (Fig. 2), one composed of Hornungia and Tropidocarpum and the other comprising Descurainia, Ianhedgea, and Robeschia. Fruit morphology varies widely in the Hornungia/Tropidocarpum clade, although in crosssection the fruits are all angustiseptate. This morphological feature distinguishes the Hornungia/Tropidocarpum clade from the second Descurainieae lineage, in which the fruits are terete, quadrangular or, in the case of D. sophioides, slightly latiseptate. Within the second lineage, the taxonomic position of monotypic genera Ianhedgea and Robeschia with respect to Descurainia is unclear; their placement is affected by the inclusion or exclusion of D. sophia and Hornungia in the analyses. When D. sophia or Hornungia are excluded, Ianhedgea is placed at the base of this clade and Robeschia is sister to the remaining taxa. Some treatments (e.g. Appel and Al-Shehbaz 2003) have placed Robeschia within Descurainia. Approximately two-thirds of recognized genera in the Brassicaceae consist of one to three species (Koch and Kiefer 2006), and Al-Shehbaz et al. (2006) suggest that the vast majority of these should be united with larger genera. Regardless of the exact position of Ianhedgea and Robeschia, the molecular data presented here would support the inclusion of these two genera within Descurainia. Species Concepts and Relationships within Descurainia— ITS and cpDNA phylogenies offer mixed support for Schulz’s sectional classifications. Section Sisymbriodendron, which comprises the Canary Island taxa, is shown to form a single lineage, but sect. Descurainia, which was considered to include all the non-Canary Island species, is polyphyletic: New World species are clearly separated in the tree from D. kochii and D. sophia, the Old World members of this section. Descurainia kochii and D. sophia are sister to the remainder of the genus (Figs. 3, 4). While D. kochii has a relatively 970 SYSTEMATIC BOTANY [Volume 36 Fig. 7. The maximum likelihood tree (lnL = -3,845.573) derived from a reduced ITS data set used for divergence time estimates in r8s. Calculated divergence times for labeled nodes are in million years before present (mya). The calibration node (MRCA of Arabidopsis and Brassica, 43.2 mya) is marked with a circle. 2011] GOODSON ET AL.: DESCURAINIA 971 Fig. 8. The maximum likelihood tree (lnL = -19,564.939) based on a reduced cpDNA data set used for divergence time estimates in r8s. Calculated divergence times for labeled nodes are in million years before present (mya). The calibration node (MRCA of Arabidopsis and Brassica, 43.2 mya) is marked with a circle. Generic names are abbreviated as follows: A. = Arabidopsis, D. = Descurainia, I. = Ianhedgea, R. = Robeschia, and S. = Sisymbrium (altissimum) or Smelowskia (americana). narrow distribution (Turkey and Caucasia), D. sophia is wideranging throughout most of Europe and temperate Asia and is an introduced weed in other temperate areas of the world. The successful colonization, vigor and weediness of D. sophia compared to its Old World congeners is consistent with a hybrid origin (Grant 1981; Doyle et al. 1999; Rieseberg et al. 2007), which can be deduced for D. sophia from its tetraploid chromosome number (2n = 28) and differing placements in the ITS and cpDNA phylogenies. Descurainia kochii appears to be the maternal parent of D. sophia since cpDNA is inherited maternally in most angiosperms, including crucifers such as Brassica (Johannessen et al. 2005) and Arabidopsis (Martínez 972 SYSTEMATIC BOTANY et al. 1997). The paternal ancestor is presumably extinct, because there are no other described Descurainia species occupying the same phylogenetic position as D. sophia in the ITS tree. Species of Descurainia in the Canary Islands comprise a monophyletic lineage (Fig. 5), suggesting that these woody perennials are descended from a single colonization. Relationships within the island taxa have been discussed previously (Goodson et al. 2006). The Canary Island species are most closely related to their nearest continental neighbor, D. tanacetifolia. Based on the limited sampling in this study, D. tanacetifolia subsp. suffruticosa is more closely related to the Canarian species than to subsp. tanacetifolia. A range disjunction exists between these two subspecies of D. tanacetifolia, with subsp. suffruticosa restricted to the Pyrenees and mountains of northern Spain and subsp. tanacetifolia distributed in the Italian and Swiss Alps. Morphologically, the two subspecies of D. tanacetifolia are similar, differing only in minor details such as degree of pubescence, number of leaf lobes or teeth, fruiting pedicel length, and degree of woodiness at the base of the stem (Schulz 1924; Ball 1964; Ortiz 1993). The genetic differentiation and present-day distribution of D. tanacetifolia most likely reflect range disruption during Pleistocene glaciation and subsequent evolution in isolation in Iberian and Italian glacial refugia (Hewitt 1996; Taberlet et al. 1998). Although ITS sequence data is equivocal regarding the monophyly of New World Descurainia, cpDNA data strongly support New World Descurainia as monophyletic and sister to D. tanacetifolia and Canary Island species (Fig. 4). Within the New World, there are three major, well supported groups: clade A is exclusively North American, clade B is North American except for the maternal type of one South American species, and clade C contains a mixture of North and South American taxa. Despite much incongruence between ITS and cpDNA phylogenies, four major North American lineages are evident (Figs. 3–5). The first lineage (clade A) is distributed along the Rocky Mountains and Sierra Madre Oriental, and includes D. incana, D. incisa (except subsp. filipes), D. streptocarpa, and D. obtusa subsp. brevisiliqua. In addition, D. sophioides, D. californica and D. impatiens are placed by either ITS or cpDNA data (but not both) in clade A. The second lineage, clade B, has its center of distribution in the Great Basin region of the western U. S. A., and extends from southern California to Wyoming, with hybrids between this clade and clade C found north to Montana and east to Minnesota. When putative hybrids with other clades are removed, this group includes D. paradisa, D. incisa subsp. filipes, and some subspecies of D. pinnata (mainly subsp. menziesii, nelsonii, and halictorum). The maternal ancestor of the Patagonian species D. antarctica is also found in this clade (Fig. 4). The final two North American lineages are part of Clade C (Figs. 3–5). One is composed of a single species, D. obtusa, which is distributed in the mountains and plateau regions of New Mexico, Arizona, northern Baja California and northern Chihuahua. It is well-supported as sister to a group of South American taxa distributed in Argentina and Chile. The other lineage (clade C-I) comprises the Mexican endemic D. virletii and a southern subspecies complex of D. pinnata, particularly subsp. pinnata, glabra, ochroleuca, and halictorum, ranging from the coastal plains of the southeastern U. S. A. into Arizona, New Mexico, and northern Mexico. (Additional sampling indicates northern subsp. brachycarpa also belongs in this clade [B. Goodson, unpubl. data]). [Volume 36 With respect to North American Descurainia, taxonomic implications of this study include the following. First, D. pinnata, as currently circumscribed, is polyphyletic. Descurainia pinnata subsp. pinnata, subsp. glabra and subsp. ochroleuca, and D. virletii form a species complex which, in the western part of its range, undergoes extensive hybridization with other taxa. Other subspecies traditionally placed in D. pinnata, subsp. nelsonii, intermedia, and menziesii, do not belong there, and are more closely related to D. incisa subsp. filipes. This latter species, in turn, appears to be distinct from D. incisa and may be better referred to by the earlier designation D. longipedicellata O. E. Schulz. Some subspecies, especially D. pinnata subsp. halictorum, probably represent hybrid populations of polyphyletic origin. Secondly, D. obtusa subsp. brevisiliqua does not belong in D. obtusa, but is more closely allied with D. incisa; Goodson and Al-Shehbaz (2007) have proposed the new combination D. brevisiliqua for this taxon. With the exception of clade B, which is morphologically heterogeneous, each of the North American lineages identified in this study correlate strongly with morphology. A more detailed discussion of North American Descurainia species concepts and morphology, which expands the number of native North American species to 17 (based primarily on this study), may be found in Goodson and Al-Shehbaz (2010). Based on morphology and geography, two major divisions of South American Descurainia are evident: 1) high Andean species with appressed fruit that range from Colombia to northern Argentina and northern Chile, and 2) species with spreading fruit that occupy mid-level elevations primarily in Argentina and Chile. Unlike their North American congeners, there is little geographic overlap between these two divisions. ITS and cpDNA molecular data resolve South American Descurainia into four strongly-supported lineages that generally correlate well with these major morphological divisions (Figs. 3, 4, 5). The monophyly of the first South American lineage (clade C-II) is supported by both ITS and cpDNA data. This clade includes all sampled South American spreading-fruit species except D. cumingiana var. tenuissima. There is little phylogenetic structure within the group, and overall sequence divergence is low (maximum of 0.66% and 0.15% divergence for ITS and cpDNA, respectively, with most ITS sequences identical or only differing by one bp). At least some populations of D. antarctica, which is distributed throughout Patagonia, appear to be a product of hybridization between a member of this South American clade C-II and a presumed dispersant from North American clade B. All four accessions of D. antarctica form a monophyletic group strongly-supported as part of clade B in the cpDNA phylogeny (Fig. 4). Mixed ITS types were detected for three of these four samples; when two of them were cloned, the resulting sequences were placed in both clade B and clade C-II. Inspection of the additive ITS sequence for the third polymorphic sample revealed polymorphisms consistent with the same two clades. The other sampled South American species of Descurainia with spreading fruit is D. cumingiana. This species is morphologically distinct from the species in clade C-II, being easily distinguished from the latter by tripinnatisect leaves and long narrow siliques. The four accessions of D. cumingiana (all var. tenuissima from Chile and Argentina) are well-supported by molecular data as forming a distinct lineage (C-V) that is not part of clade C-II. The species is distributed throughout 2011] GOODSON ET AL.: DESCURAINIA central Chile and also scattered across central Patagonian regions of Argentina, where it has been reported to hybridize with D. antarctica (Romanczuk 1984). Reports of D. cumingiana var. cumingiana in Mendoza and northern Neuquén provinces of Argentina (Romanczuk 1984; Zuloaga and Morrone 1999) appear to be based on an overly-broad species concept for D. cumingiana by those authors. Consistent with their ITS position, such specimens more closely resemble taxa such as D. antarctica and D. pimpinellifolia with respect to leaf morphology and seed arrangement. Morphologically, the high Andean species constituting the other two South American lineages (denoted as C-III and C-IV) are united by distinct characters such as fruit appressed to the rachis and valves of the fruit dehiscing from the apex to the base. These characters are absent from other South American Descurainia species. Hybridization within North American Descurainia— Although ITS and cpDNA trees are in general agreement regarding major lineages within Descurainia, many cases of phylogenetic conflict are evident throughout the trees. The degree of incongruence with respect to North American lineages (Figs. 3, 4) is particularly striking (Fig. 3), with nearly two-thirds of sampled North American accessions exhibiting some degree of incongruence. Of the 71 sampled North American accessions, 22 (31%) either differ between the two trees in major clade placement or possess mixed ITS types for different major clades. Discordance between nuclear and plastid phylogenies is often seen as evidence of past hybridization events, although other processes, such as lineage sorting (especially in recentlydiversified groups) can also give rise to conflicting topologies (Wendel and Doyle 1998). While both processes may have contributed to the observed conflict, we believe that much of the incongruence is due to hybridization. Little is known about the reproductive biology of North American Descurainia, but widely noted infraspecific morphological variation and confusing taxonomic boundaries (e.g. Detling 1939; Rollins 1993a,b; Welsh et al. 1993; de Rzedowski and Rzedowski 2001; Holmgren et al. 2005), overlapping ranges, and the occurrence of possible intermediate forms (Detling 1939; pers. obs.) suggest that interpopulational and interspecific gene flow is occurring. Polyploidy is also relatively common, with 15 out of 29 North American chromosome counts (excluding D. sophioides) reported as tetraploid or higher (Appendix 1). Hybridization is an extremely common phenomenon in the Brassicaceae (Marhold and Lihová 2006), and hybrid polyploid complexes have been extensively characterized and studied in genera such as Boechera (e.g. Koch et al. 2003b; Schranz et al. 2005; Sharbel et al. 2005), Brassica (Osborn 2004 and references therein), and Draba (Brochmann 1992; Koch and Al-Shehbaz 2002). The eight North American accessions with mixed ITS types (seven clade C × clade B and one vice versa) are presumably allopolyploids arising from relatively recent hybridization events: with one exception, they belong to taxa which have known tetraploid populations and whose ranges occur in areas where members of the two clades are sympatric. The detection of ITS additive sequences is often considered strong evidence for a recent hybrid origin, especially when hybridization has previously been suspected for the taxa under investigation (e.g. Kim and Jansen 1994; Whittall et al. 2000; Tate and Simpson 2003). No ITS additivity was observed for the 14 remaining strongly incongruent accessions. In contrast to the samples 973 with mixed ITS sequences, for all but two accessions the incongruence exhibited in this category is between clades A and either C or B. Most involve taxa for which only diploid chromosome counts have been reported or none are known. In the absence of additional molecular or cytological information, it is difficult to distinguish between potential processes responsible for the incongruence observed in these cases. Some, such as D. pinnata subsp. nelsonii C47, might be recently derived polyploids like the accessions possessing mixed types, but this history has been obscured by either complete concerted evolution leading to fixation of the paternal type or by preferential PCR amplification of one parental type. For the species exhibiting incongruence between clades A and C, the ITS phylogeny is much more consonant with morphology than the cpDNA phylogeny, a pattern which has been observed in other groups with similar phylogenetic discordance and often attributed to cytoplasmic introgression (e.g. Hardig et al. 2000; Ferguson and Jansen 2002). In addition to between-clade incongruence, 21 other accessions (30%) exhibit conflicting placements within a given clade but possess few if any polymorphic loci. In many of these cases branch lengths are short and terminal clade support in a given tree is weak. Such conflicts could equally well reflect within-clade gene flow, the effects of lineage sorting, or rapid or recent diversification (Wendel and Doyle 1998). Origins and Biogeography—Divergence time estimates support a Miocene origin (14–23 mya) for Descurainia and related species. Ianhedgea (central and southwest Asia), Robeschia (Middle East), and D. kochii (Turkey and Caucasia) are in a basal position with respect to the remainder of the genus. Assuming that present-day distributions reflect ancestral areas, this suggests that the genus arose in the IranoTuranian region posited by Hedge (1976) as a likely center of origin for the Brassicaceae. Geographic expansion of Descurainia out of ancestral areas appears to have begun in the late Miocene with diversification accelerating during the late Pliocene or early Pleistocene. This latter period was marked by dramatic climate changes which opened up new niches for speciation, and by the final uplift of Eurasian and American mountain systems which could serve as corridors for migration (Agakhanjanz and Breckle 1995; Hewitt 1996, 2000). The genus spread into Europe giving rise to D. tanacetifolia as well as an unknown (or now extinct) taxon, which hybridized with D. kochii to form D. sophia. Pleistocene glacial cycles have had a profound effect on the composition and distribution of species in Europe, and one of the early glacial cycles may have abetted the extirpation of the maternal parent of D. sophia as well as contributed to the expansion of D. sophia throughout Eurasia. Such an event has been reported, for example, in Paeonia L. (Paeoniaceae), where European populations of present-day Asian species appear to have been completely replaced by their hybrids during the Pleistocene (Sang et al. 1997). Descurainia species in the Canary Islands are closely related to D. tanacetifolia of the Pyrenees and northern Spain. The ancestral species arrived in the Canary Islands 1–2.8 mya, probably from the Iberian peninsula, a refugial area during Pleistocene glacial maxima (Hewitt 1996; Taberlet et al. 1998). New World species of Descurainia are of Pliocene/earlyPleistocene origin, with molecular clock calculations from cpDNA data estimating a date of 4.7–5.7 mya for the last common ancestor of all New World taxa, and ITS data suggesting an origin of approximately two mya for each of the 974 SYSTEMATIC BOTANY three major New World clades. Biogeographic reconstructions with MacClade (Fig. 6) are equivocal regarding whether Descurainia was first introduced into North or South America, although most of the reconstructions support initial introduction into North America. Although approximately twice as many species have been recognized in South America compared to North America, maximum ITS sequence divergence within North American taxa (5.1%), and even among many western North American taxa alone (e.g. clades A+B, 3.0%), is much greater than for all of South America (1.3%); cpDNA data reveal a similar trend (Table 2). This greater genetic diversity within North America relative to South America argues for North America as the continent of initial establishment, and is consistent with the general distribution pattern observed in New World Brassicaceae. Assuming a North American origin for Descurainia in the New World, the ancestor of North American Descurainia could have conceivably migrated either westward from Europe or eastward from Eurasia. The close relationship between New World Descurainia species and those of the Canary Islands and Europe (i.e. D. tanacetifolia) would be congruent with dispersal from the European continent. Any introduction from Europe via a North Atlantic land bridge, however, can be ruled out; such a land connection is believed to have been broken by the early Eocene (Tiffney 1985; Tiffney and Manchester 2001), which considerably predates the origin of the genus. More recent long-distance dispersal from Europe to North America is a possibility though; examples of such trans-Atlantic dispersals, while not common, include South American Hypochaeris arriving from northwest Africa during the Pliocene or Pleistocene (Tremetsberger et al. 2005) and various amphi-Atlantic arctic species whose North American populations are of late Quaternary origin (Brochmann et al. 2003). Eastern North America is not the likely ancestral area for the genus in North America, however, because the area of greatest species and sequence diversity for North American Descurainia is the Great Basin region of the western U. S. A. While disjunctions between southwest North America and the Mediterranean/European flora are known (cited in Coleman et al. 2003), their origins date from times much earlier than the Pliocene. Migration from Eurasia, rather than Europe, Table 2. Maximum sequence divergences (based on uncorrected “p” distances) within major Descurainia lineages. †Not present as distinct clade in ITS; members unresolved as part of clade C. Clade ITS % cpDNA % New World - A New World - B New World - A+B New World - C C-I (NA D. pinnata + D. virletii) C-II (SA spreading-fruit) C-III (SA appressed fruit) C-IV (SA appressed fruit) C-V (D. cumingiana) D. obtusa subsp. obtusa (NA) New World - A+B+C All North American accessions All South American accessions (excluding D. antarctica type 2) D (Canary Is. + D. tanacetifolia) Canary Is. A+B+C+D 1.16 1.63 2.95 2.46 1.80 0.66 0.67† 0.33† 0.17 0.49† 5.08 5.08 1.31 0.44 0.43 1.02 1.15 0.49 0.16 0.45 0.09 0.00 0.11 1.18 1.18 0.83 – 1.16 5.08 1.22 0.54 1.94 [Volume 36 into western North America is therefore the most plausible route of introduction of Descurainia into the Americas. Immigration from central Asia to western North America via the Bering land bridge, which served as a glacial refugium and corridor for migration of temperate taxa during the late Tertiary and Quaternary (Colinvaux 1996; Hewitt 2000), has been invoked to explain the distribution of a number of genera in the Brassicaceae, such as Braya Sternb. and Hoppe, Eutrema R. Br., and Parrya R. Br. (Rollins 1982). Moreover, many recent molecular studies have uncovered evidence of dispersal from Asia (especially southwest Asia and Eurasia) to North America during the Pliocene/Pleistocene, either by long-distance dispersal or via Beringia. Examples from the Brassicaceae include Braya (Warwick et al. 2004a), Draba L. (Koch and Al-Shehbaz 2002), Lepidium (Mummenhoff et al. 2001) and Smelowskia C. A. Mey (Warwick et al. 2004b); representatives from other families are Androsace L. (Primulaceae; Schneeweiss et al. 2004), Gentianella Moench (Gentianaceae; von Hagen and Kadereit 2001), Halenia Borkh. (Gentianaceae; von Hagen and Kadereit 2003), Hordeum (Poaceae; Blattner 2006), and Senecio mohavensis (Asteraceae; Coleman et al. 2003). A Eurasian–North American link within Descurainia is exemplified in the present day by arctic/subarctic D. sophioides, which is distributed from western Canada (with outlier populations around Hudson Bay) across Alaska and northern Siberia westward almost to the Ural Mountains. This species occupies a derived phylogenetic position with respect to New World Descurainia, however, and its current range may represent expansion from Beringia after the last glacial maximum rather than a relictual connection between Asia and western North America. The most recent common ancestor of European and New World Descurainia could have been eliminated from the mountains of northern Asia during a period of rapid glaciation in the Pleistocene, as has been considered for Androsace (Schneeweiss et al. 2004). Regardless of whether or not one assumes a North American origin for New World Descurainia, it is clear that there have been several independent dispersals between North and South America. The exact number, and direction, of all dispersal events is difficult to infer, however, because the major North/South American clade (clade C) is not resolved with respect to some lineages in the parsimony tree (Fig. 5). While the topology (Fig. 5) recovered by Bayesian inference is wellresolved, the posterior probabilities joining some of the clade C lineages are low, suggesting uncertainty in the placement of those branches. The inability to obtain a well-supported resolution for these lineages suggests a period of rapid diversification with some dispersals occurring nearly simultaneously. Figure 6 illustrates most parsimonious reconstructions of New World continental distribution (North or South America) traced onto simplified trees representing the two topologies recovered from parsimony analysis and 94% of the 95% credible set of trees from Bayesian analysis of the combined ITS-cpDNA data set. All of these optimizations suggest multiple dispersals of Descurainia between North and South America. The majority of reconstructions support a New World origin in North America followed by three or four independent dispersals to South America. The general trend of relatively-recent colonization of South America from North America is consistent with that seen in many genera (e.g. Chrysosplenium [Saxifragaceae; Soltis et al. 2001], Draba [Koch and Al-Shehbaz 2002], Gentianella [von Hagen and Kadereit 2001], Gilia [Polemoniaceae; Morell et al. 2000], Lepidium 2011] GOODSON ET AL.: DESCURAINIA [Mummenhoff et al. 2001], and others reviewed in Wen and Ickert-Bond 2009). Studies have uncovered evidence of multiple independent dispersals from North to South America in Halenia (Gentianaceae; von Hagen and Kadereit 2003) and several other genera (Wen and Ickert-Bond 2009). Although multiple dispersals are also suggested in Lycium L. (Solanaceae; Fukuda et al. 2001; Levin and Miller 2005), the direction of dispersal is uncertain. Hoffmannseggia (Simpson et al. 2005) also illustrates a history of multiple dispersals, but the dispersal direction has been from South to North America. Several of the scenarios in Fig. 6 that support a North American origin for Descurainia also indicate re-dispersal from South America to North America has occurred. They suggest that D. obtusa, which is distributed in the mountains and plateaus of the southwest U. S. A. and northern Mexico, arose from a common ancestor of clade C-II, which comprises all of the species (except D. cumingiana) with spreading fruit ranging primarily throughout Argentina and parts of Chile. Such a situation would not be without precedent; Blattner (2006) has detected evidence that after introduction into South America from the north ca. two mya, Hordeum redispersed to North America on two separate occasions. It is equally or more likely, however, that South American clade C-II arose from long-distance dispersal of D. obtusa or a close relative or ancestor from North America to South America. It is not possible to determine which of these two scenarios is correct from the current data. With the exception of the branch joining South American clade C-II with North American D. obtusa, the branches resolving relationships between major lineages in clade C, on which the above scenarios are based, are not well-supported (Fig. 5). Generation of trees with alternative arrangements of these weakly-supported branches (not shown) requires only one or two additional steps compared to the shortest trees recovered from the parsimony analysis. A few of these alternate topologies are also found in the 95% set of credible trees generated from Bayesian analysis. Nonetheless, optimization of continental distribution on any of these alternate topologies also yields most parsimonious reconstructions (not shown) supporting three or four dispersals between North and South America similar to those illustrated in Fig. 6. As first proposed by authors such as Cruden (1966) and Carlquist (1983), adhesion of seeds or fruits to (or ingestion by) migrating birds has been suggested as the mechanism for long-distance dispersal in many studies (e.g. Ballard and Sytsma 2000; Morrell et al. 2000; Mummenhoff et al. 2001; Levin and Miller 2005; Blattner 2006; Wen and IckertBond 2009). Because Descurainia has seeds that are mucilaginous when wet, it is easy to envision such a bird-mediated transport over tropical areas into temperate regions of South America (and vice versa). Mummenhoff et al. (1992) cite several reports of Lepidium and Capsella seeds, both crucifers with mucilaginous seeds, found attached to birds. Long-distance transport of seed from North to South America, whether by birds or otherwise, would be consistent with the origin of D. antarctica in Patagonia, since maternal alleles would only be dispersed by seed. Acknowledgments. The authors wish to thank Romey Haberle, Heidi Meudt, and Jennifer Tate for reviewing the dissertation on which this manuscript is based, Ihsan Al-Shehbaz for helpful advice, Jan Barber for providing lab facilities for some of this research, and Andy Alverson, Mary Guisinger, Leah Larkin, and Andrea Weeks for answering software questions. César Gomez-Campo, Jay McKendrick, T. Chumley, R. Haberle, 975 José Panero, Bonnie Crozier, Ali Dönmez, and Arnoldo Santos-Guerra provided seeds or collected plant material, Mark Beilstein donated DNA of Ianhedgea, Robeschia, and D. sophia, and Robert Price shared unpublished ITS sequences of Tropidocarpum. For assistance with field work in South America, BEG thanks Lázaro J. Novara, Maximo Liberman, Franz Guzmán, Mónica Stronati, and the students and staff of LPB, BAA, and BAB, especially Carolina Garcia, Stephan Beck, Renee Fortunato, Rosa Isela Meneses, Matías Morales, and Silvia Perfetti, as well as the curators of BAA, GH, LPB, MO, NY, UNM, OSC, SI, TEX, BAB, LIL, MCNS, MERL, RM, and SGO. Funding from a National Science Foundation IGERT fellowship and University of Texas botany program research and travel support, including Lorraine I. Stengl endowment and Linda Escobar memorial funds, is gratefully acknowledged. Literature Cited Agakhanjanz, O. and S.-W. Breckle. 1995. Origin and evolution of the mountain flora of Middle Asia and neighbouring mountain regions. Pp. 63–80 in Arctic and alpine biodiversity: Patterns, causes and ecosystem consequences, eds. F. S. Chapin III and C. Körner. Berlin: Springer-Verlag. Akaike, H. 1974. A new look at the statistical model identification. IEEE Transactions on Automatic Control 19: 716–723. Al-Shehbaz, I. A. 1988. The genera of Sisymbrieae (Cruciferae; Brassicaceae) in the southeastern United States. Journal of the Arnold Arboretum 69: 213–237. Al-Shehbaz, I. A. 2003. A synopsis of Tropidocarpum. Novon 13: 392–395. Al-Shehbaz, I. A., M. A. Beilstein, and E. A. Kellogg. 2006. Systematics and phylogeny of the Brassicaceae (Cruciferae): an overview. Plant Systematics and Evolution 259: 89–120. Álvarez, I. and J. F. Wendel. 2003. Ribosomal ITS sequences and plant phylogenetic inference. Molecular Phylogenetics and Evolution 29: 417–434. Anderberg, A. A. and A.-L. Anderberg. 1997. Den virtuella floran. Naturhistoriska riksmuseet: http://linnaeus.nrm.se/flora/di/Brassica/descu/ descsop.html. Accessed 22 March 2007. Appel, O. and I. A. Al-Shehbaz. 2003. Cruciferae. Pp. 75–174 in The families and genera of vascular plants, vol. V. flowering plants. Dicotyledons: Malvales, Capparales, and non-betalain Caryolphyllales, eds. K. Kubitzki and C. Bayer. Berlin: Springer-Verlag. Bailey, C. D., M. A. Koch, M. Mayer, K. Mummenhoff, S. L. O’Kane Jr., S. I. Warwick, M. D. Windham, and I. A. Al-Shehbaz. 2006. Toward a global phylogeny of the Brassicaceae. Molecular Biology and Evolution 23: 2142–2160. Baker, R. H. and R. DeSalle. 1997. Multiple sources of character information and the phylogeny of Hawaiian Drosophilids. Systematic Biology 46: 654–673. Baker, R. H., X. B. Yu, and R. DeSalle. 1998. Assessing the relative contribution of molecular and morphological characters in simultaneous analysis trees. Molecular Phylogenetics and Evolution 9: 427–436. Ball, P. W. 1964. Hugueninia Reichenb. In Flora Europaea, Vol. 1 Lycopodiaceae to Platanaceae, eds. T. G. Tutin, V. H. Heywood, N. A. Burges, D. H. Valentine, S. M. Walters, and D. A. Webb. Cambridge: Cambridge University Press. Ballard, H. E. and K. J. Sytsma. 2000. Evolution and biogeography of the woody Hawaiian violets (Viola, Violaceae): arctic origins, herbaceous ancestry and bird dispersal. Evolution 54: 1521–1532. Barker, F. K. and F. M. Lutzoni. 2002. The utility of the incongruence length difference test. Systematic Biology 51: 625–637. Beilstein, M. A., I. A. Al-Shehbaz, and E. A. Kellogg. 2006. Brassicaceae phylogeny and trichome evolution. American Journal of Botany 93: 607–619. Beilstein, M. A., I. A. Al-Shehbaz, S. Mathews, and E. A. Kellogg. 2008. Brassicaceae phylogeny inferred from phytochrome A and ndhF sequence data: tribes and trichomes revisited. American Journal of Botany 95: 1307–1327. Beilstein, M. A., N. S. Nagalingum, M. D. Clements, S. R. Manchester, and S. Mathews. 2010. Dated molecular phylogenies indicate a Miocene origin for Arabidopsis thaliana. Proceedings of the National Academy of Sciences USA 107: 18724–18728. Best, K. F. 1977. The biology of Canadian weeds. 22. Descurainia sophia (L.) Webb. Canadian Journal of Plant Science 57: 499–507. Blattner, F. R. 2006. Multiple interncontinental dispersals shaped the distribution area of Hordeum (Poaceae). The New Phytologist 169: 603–614. Bradley, R. D. and D. M. Hillis. 1997. Recombinant DNA sequences generated by PCR amplification. Molecular Biology and Evolution 14: 592–593. Bramwell, D. 1977. A revision of Descurainia Webb. & Berth. section Sisymbriodendron (Christ) O. E. Schulz in the Canary Islands. Botanica Macaronesica 4: 35–53. 976 SYSTEMATIC BOTANY Bricker, J. S., G. K. Brown, and T. L. P. Lewis. 2000. Status of Descurainia torulosa (Brassicaceae). Western North American Naturalist 60: 426–432. Brochmann, C. 1992. Polyploid evolution in arctic-alpine Draba (Brassicaceae). Sommerfeltia 4(Supplement): 1–34. Brochmann, C., T. M. Gabrielsen, I. Nordal, J. Y. Landvik, and E. Reidar. 2003. Glacial survival or tabula rasa? The history of North Atlantic biota revisited. Taxon 52: 417–450. Buckler, E. S., A. Ippolito, and T. P. Holtsford. 1997. The evolution of ribosomal DNA: divergent paralogues and phylogenetic implications. Genetics 145: 821–832. Carlquist, S. C. 1983. Intercontinental dispersal. Pp. 37–47 in Dispersal and distribution: an international symposium. Sonderbände des Naturwissenschaftlichen Vereins in Hamburg 7, ed. K. Kubitzki. Hamburg: Paul Parey. Coleman, M., A. Liston, J. W. Kadereit, and R. J. Abbott. 2003. Repeat intercontinental dispersal and Pleistocene speciation in disjunct Mediterranean and desert Senecio (Asteraceae). American Journal of Botany 90: 1446–1454. Colinvaux, P. A. 1996. Low-down on a land bridge. Nature 382: 21–22. Cronn, R., M. Cedroni, T. Haselkorm, C. Grover, and J. F. Wendel. 2002. PCR-mediated recombination in amplification products derived from polyploid cotton. Theoretical and Applied Genetics 104: 482–489. Cruden, R. W. 1966. Birds as agents of long-distance dispersal for disjunct plant groups of the temperate Western Hemisphere. Evolution 20: 517–532. Darlu, P. and G. Lecointre. 2002. When does the incongruence length difference test fail? Molecular Biology and Evolution 19: 432–437. Demesure, B., N. Sodzi, and R. J. Petit. 1995. A set of universal primers for amplification of polymorphic non-coding regions of mitochondrial and chloroplast DNA in plants. Molecular Ecology 4: 129–131. Detling, L. E. 1939. A revision of the North American species of Descurainia. American Midland Naturalist 22: 481–520. Dolphin, K., R. Belshaw, D. L. C. Orme, and D. L. J. Quicke. 2000. Noise and incongruence: interpreting results of the incongruence length difference test. Molecular Phylogenetics and Evolution 17: 401–406. Dowton, M. and A. D. Austin. 2002. Increased congruence does not necessarily indicate phylogenetic accuracy: the behavior of the incongruence length difference test in mixed-model analyses. Systematic Biology 51: 19–31. 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, and A. H. D. Brown. 1999. Origins, colonization, and lineage recombination in a widespread perennial soybean polyploid complex. Proceedings of the National Academy of Sciences USA 96: 10741–10745. Farris, J. S., M. Källersjö, A. G. Kluge, and C. Bult. 1994. Testing significance of incongruence. Cladistics 10: 315–319. Felsenstein, J. 1981. Evolutionary trees from DNA sequences: a maximum likelihood approach. Journal of Molecular Evolution 17: 368–376. Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. Ferguson, C. J. and R. K. Jansen. 2002. A chloroplast DNA phylogeny of eastern Phlox (Polemoniaceae): implications of congruence and incongruence with the ITS phylogeny. American Journal of Botany 89: 1324–1335. Fukuda, T., J. Yokoyama, and H. Ohashi. 2001. Phylogeny and biogeography of the genus Lycium (Solanaceae): inferences from chloroplast DNA sequences. Molecular Phylogenetics and Evolution 19: 246–258. Goldman, N., J. P. Anderson, and A. G. Rodrigo. 2000. Likelihood-based tests of topologies in phylogenetics. Systematic Biology 49: 652–670. Goodson, B. E. and I. A. Al-Shehbaz. 2007. Two new combinations in North American Descurainia (Brassicaceae). Harvard Papers in Botany 12: 421–423. Goodson, B. E. and I. A. Al-Shehbaz. 2010. Descurainia. Pp. 518–529 in Flora of North America north of Mexico, vol. 7: Magnoliophyta: Dilleniidae, pt. 1, eds. Flora of North America Editorial Committee. New York: Oxford University Press. Goodson, B. E., A. Santos-Guerra, and R. K. Jansen. 2006. Molecular systematics of Descurainia (Brassicaceae) in the Canary Islands: biogeographic and taxonomic implications. Taxon 55: 671–682. Grant, V. 1981. Plant speciation. New York: Columbia University Press. von Hagen, K. B. and J. W. Kadereit. 2001. The phylogeny of Gentianella (Gentianaceae) and its colonization of the southern hemisphere as revealed by nuclear and chloroplast DNA sequence variation. Organisms, Diversity & Evolution 1: 61–79. von Hagen, K. B. and J. W. Kadereit. 2003. The diversification of Halenia (Gentianaceae): ecological opportunity versus key innovation. Evolution 57: 2507–2518. [Volume 36 Hardig, T. M., P. S. Soltis, and D. E. Soltis. 2000. Diversification of the North American shrub genus Ceanothus (Rhamnaceae): conflicting phylogenies from nuclear ribosomal DNA and chloroplast DNA. American Journal of Botany 87: 108–123. Hedge, I. C. 1976. A systematic and geographical survey of the Old World Cruciferae. Pp. 1–45 in The biology and chemistry of the Cruciferae, eds. J. G. Vaughn, A. J. MacLeod, and B. M. G. Jones. London: Academic Press. Hewitt, G. M. 1996. Some genetic consequences of ice ages, and their role in divergence and speciation. Botanical Journal of the Linnean Society 58: 247–276. Hewitt, G. 2000. The genetic legacy of the Quaternary ice ages. Nature 405: 907–913. Hipp, A. L., J. C. Hall, and K. J. Sytsma. 2004. Congruence versus phylogenetic accuracy: revisiting the incongruence length difference test. Systematic Biology 53: 81–89. Holmgren, N. H., P. K. Holmgren, and A. Cronquist. 2005. Intermountain flora, vol. 2B: subclass Dilleniidae. New York: New York Botanical Garden Press. Huelsenbeck, J. P. and B. Rannala. 1997. Phylogenetic methods come of age: testing hypotheses in an evolutionary context. Science 276: 227–232. Huelsenbeck, J. P. and B. Rannala. 2004. Frequentist properties of Bayesian posterior probabilities under simple and complex substitution models. Systematic Biology 53: 904–913. Johannessen, M. M., B. A. Andersen, C. Damgaard, and R. B. Jørgensen. 2005. Maternal inheritance of chloroplasts between Brassica rapa and F1-hybrids demonstrated by cpDNA markers specific to oilseed rape and B. rapa. Molecular Breeding 16: 271–278. Kim, K.-J. and R. K. Jansen. 1994. Comparisons of phylogenetic hypotheses among different data sets in dwarf dandelions (Krigia, Asteraceae): additional information from internal transcribed spacer sequences of nuclear ribosomal DNA. Plant Systematics and Evolution 190: 157–185. Koch, M. and I. A. Al-Shehbaz. 2002. Molecular data indicate complex intra- and intercontinental differentiation of American Draba (Brassicaceae). Annals of the Missouri Botanical Garden 89: 88–109. Koch, M. A. and C. Kiefer. 2006. Molecules and migration: biogeographical studies in cruciferous plants. Plant Systematics and Evolution 259: 121–142. Koch, M., I. A. Al-Shehbaz, and K. Mummenhoff. 2003a. Molecular systematics, evolution, and population biology in the mustard family (Brassicaceae). Annals of the Missouri Botanical Garden 90: 151–171. Koch, M. A., C. Dobeš, and T. Mitchell-Olds. 2003b. Multiple hybrid formation in natural populations: concerted evolution of the internal transcriber spacer of nuclear ribosomal DNA (ITS) in North American Arabis divaricarpa (Brassicaceae). Molecular Biology and Evolution 20: 338–350. Koch, M. A., C. Kiefer, D. Ehrich, J. Vogel, C. Brochmann, and K. Mummenhoff. 2006. Three times out of Asia Minor: the phylogeography of Arabis alpina L. (Brassicaceae). Molecular Ecology 15: 825–839. Langley, C. H. and W. Fitch. 1974. An estimation of the constancy of the rate of molecular evolution. Journal of Molecular Evolution 3: 161–177. Levin, R. A. and J. S. Miller. 2005. Relationships within tribe Lycieae (Solanaceae): paraphyly of Lycium and multiple origins of gender dimorphism. American Journal of Botany 92: 2044–2053. Loockerman, D. and R. K. Jansen. 1996. The use of herbarium material for molecular systematic studies. Pp. 205–220 in Sampling the green world, eds. T. F. Stuessy and S. Sohmer. New York: Columbia University Press. Mabberley, D. J. 1997. The Plant-book: A portable dictionary of the vascular plants. Cambridge: Cambridge University Press. Maddison, D. R. and W. P. Maddison. 2000. MacClade 4: Analysis of phylogeny and character evolution. Sunderland: Sinauer Associates. Marhold, K. and J. Lihová. 2006. Polyploidy, hybridization and reticulate evolution: lessons from the Brassicaceae. Plant Systematics and Evolution 259: 143–174. Martínez, P., C. López, M. Roldán, B. Sabater, and M. Martín. 1997. Plastid DNA of five ecotypes of Arabidopsis thaliana: sequence of ndhG gene and maternal inheritance. Plant Science 123: 113–122. Morrell, P. L., J. M. Porter, and E. A. Friar. 2000. Intercontinental dispersal: the origin of the widespread South American plant species Gilia laciniata (Polemoniaceae) from a rare California and Oregon coastal endemic. Plant Systematics and Evolution 224: 13–32. Mummenhoff, K., H. Hurka, and H.-J. Bandelt. 1992. Systematics of Australian Lepidium species (Brassicaceae) and implications for the origin: evidence from IEF analysis of Rubisco. Plant Systematics and Evolution 183: 99–112. Mummenhoff, K., H. Brüggemann, and J. L. Bowman. 2001. Chloroplast DNA phylogeny and biogeography of Lepidium (Brassicaceae). American Journal of Botany 88: 2051–2063. 2011] GOODSON ET AL.: DESCURAINIA Nixon, K. C. 1999. The Parsimony Ratchet, a new method for rapid parsimony analysis. Cladistics 15: 407–414. Nylander, J. A. A. 2004. MrModeltest v2. Program distributed by the author. Uppsala: Evolutionary Biology Centre, Uppsala University. Ortiz, S. 1993. Hugueninia Rchb. Pp. 36–37 in Flora Iberica: plantas vasculares de la Península Ibérica e Islas Baleares. Vol. IV Cruciferae-Monotropaceae, eds. S. Castroviejo, C. Aedo, C. Gómez Campo, M. Laínz, P. Montserrat, R. Morales, F. Muñoz Garmendia, G. Nieto Feliner, E. Rico, S. Talavera, and L. Villar. Madrid: Real Jardín Botáncio, C. S. I. C. Osborn, T. C. 2004. The contribution of polyploidy to variation in Brassica species. Physiologia Plantarum 121: 531–536. Oxelman, B., M. Lidén, and D. Berglund. 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Plant Systematics and Evolution 206: 393–410. Posada, D. and T. R. Buckley. 2004. Model selection and model averaging in phylogenetics: advantages of the AIC and Bayesian approaches over likelihood ratio tests. Systematic Biology 53: 793–808. Posada, D. and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817–818. Rambaut, A. and N. C. Grassly. 1997. Seq-Gen: an application for the Monte Carlo simulation of DNA sequence evolution along phylogenetic trees. Computer Applications in the Biosciences 13: 235–238. 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., S.-C. Kim, R. A. Randell, K. D. Whitney, B. L. Gross, C. Lexer, and K. Clay. 2007. Hybridization and the colonization of novel habitats by annual sunflowers. Genetica 129: 149–165. Rollins, R. C. 1982. A new species of the Asiatic genus Stroganowia (Cruciferae) from North America and its biogeographic implications. Systematic Botany 7: 214–220. Rollins, R. C. 1993a. The Cruciferae of continental North America: systematics of the mustard family from the Arctic to Panama. Stanford: Stanford University Press. Rollins, R. C. 1993b. Descurainia. Pp. 413–416 in The Jepson manual: Higher plants of California, ed. J. C. Hickman. Berkeley and Los Angeles: University of California Press. Rollins, R. C. and L. Rüdenberg. 1971. Chromosome numbers of Cruciferae II. Contributions of the Gray Herbarium 201: 117–133. Romanczuk, M. C. 1984. XVI. Descurainia Webb et Berthelot. Pp. 414–430 in Flora Patagonica, Part IVa, ed. M. N. Correa. Buenos Aires: Colleción cientifica del INTA. Ronquist, F. and J. P. Huelsenbeck. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574. de Rzedowski, G. C. and J. Rzedowski. (eds.). 2001. Flora fanerogámica del Valle de México. Pátzcuaro, Michoacán: Instituto de Ecología, A. C. Y Comisión Nacional para el Conocimiento y Uso de la Biodiversidad. Sanderson, M. J. 2002. Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach. Molecular Biology and Evolution 19: 101–109. Sanderson, M. J. 2004. r8s, version 1.70, user’s manual and software. Distributed by the author at http://ginger.ucdavis.edu/r8s/. Sang, T., D. J. Crawford, and T. F. Stuessy. 1997. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). American Journal of Botany 84: 1120–1136. Schneeweiss, G. M., P. Schönswetter, S. Kelso, and H. Niklfeld. 2004. Complex biogeographical patterns in Androsace (Primulaceae) and related genera: evidence from phylogenetic analyses of nuclear internal transcribed spacer and plastid trnL-F sequences. Systematic Biology 53: 856–876. Schranz, M. E., C. H. Dobeš, M. A. Koch, and T. Mitchell-Olds. 2005. Sexual reproduction, hybridization, apomixis, and polyploidization in the genus Boechera (Brassicaceae). American Journal of Botany 92: 1797–1810. Schulz, O. E. 1924. Cruciferae-Sisymbrieae. Pp. 1–388 in Pflanzenreich IV. 105 (Heft 86), ed. A. Engler. Leipzig: Verlag Wilhelm Engelmann. Schulz, O. E. 1936. Cruciferae. Pp. 227–658 in Die Natürlichen Pflanzenfamilien, vol. 17B, ed. A. Engler and K. Prantl. Leipzig: Verlag Wilhelm Engelmann. Sharbel, T. F., T. Mitchell-Olds, C. Dobeš, L. Kantama, and H. de Jong. 2005. Biogeographic distribution of polyploidy and B chromosomes in the apomictic Boechera holboellii complex. Cytogenetic and Genome Research 109: 283–292. Shaw, J., E. B. Lickey, J. T. Beck, S. S. Farmer, W. Liu, J. Miller, K. C. Siripun, C. T. Winder, E. E. Schilling, and R. L. Small. 2005. The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. American Journal of Botany 92: 142–166. 977 Shimodaira, H. and M. Hasegawa. 1999. Multiple comparisons of loglikelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16: 1114–1116. Sikes, D. S. and P. O. Lewis. 2001. Beta software, version 1. PAUPRat: PAUP* implementation of the parsimony ratchet. Distributed by the authors. Storrs: Department of Ecology and Evolutionary Biology, University of Connecticut. Simmons, M. P. and H. Ochoterena. 2000. Gaps as characters in sequencebased phylogenetic analyses. Systematic Botany 49: 369–381. Simpson, B. B., J. A. Tate, and A. Weeks. 2005. The biogeography of Hoffmannseggia (Leguminosae, Caesalpinioideae, Caesalpinieae): a tale of many travels. Journal of Biogeography 32: 15–27. Soltis, D. E., M. Tago-Nakazawa, Q.-Y. Xiang, S. Kawano, J. Murata, M. Wakabayashi, and C. Hibsch-Jetter. 2001. Phylogenetic relationships and evolution in Chrysosplenium (Saxifragaceae) based on matK sequence data. American Journal of Botany 88: 883–893. Sorenson, M. D. 1999. TreeRot, version 2. Boston: Boston University. Suda, J., T. Kyncl, and R. Freiová. 2003. Nuclear DNA amounts in Macaronesian angiosperms. Annals of Botany 92: 153–164. Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4. Sunderland: Sinauer Associates. Taberlet, P., L. Fumagalli, A.-G. Wust-Saucy, and J.-F. Cosson. 1998. Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology 7: 453–464. Taberlet, P., L. Gielly, G. Patou, and J. Bouvet. 1991. Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109. Tate, J. A. and B. B. Simpson. 2003. Paraphyly of Tarasa (Malvaceae) and diverse origins of the polyploid species. Systematic Botany 28: 723–737. Thompson, J. D., T. J. Gibson, F. Plewniak, F. Jeanmougin, and D. G. Higgins. 1997. The Clustal X windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 24: 4876–4882. Tiffney, B. H. 1985. The Eocene North Atlantic Land Bridge: its importance in Tertiary and modern phytogeography of the Northern Hemisphere. Journal of the Arnold Arboretum 66: 243–273. Tiffney, B. H. and S. R. Manchester. 2001. The use of geological and paleontological evidence in evaluating plant phylogeographic hypotheses in the Northern Hemisphere Tertiary. International Journal of Plant Sciences 162: S3–S17. Tremetsberger, K., H. Weiss-Schneeweiss, T. Stuessy, R. Samuel, G. Kadlec, M.Á. Ortiz, and S. Talavera. 2005. Nuclear ribosomal DNA and karyotypes indicate a NW African origin of South American Hypochaeris (Asteraceae, Cichorieae). Molecular Phylogenetics and Evolution 35: 102–116. Warwick, S. I. and I. A. Al-Shehbaz. 2006. Brassicaceae: chromosome number index and database on CD-ROM. Plant Systematics and Evolution 259: 237–248. Warwick, S. I., I. A. Al-Shehbaz, C. Sauder, J. G. Harris, and M. Koch. 2004a. Phylogeny of Braya and Neotorularia (Brassicaceae) based on nuclear ribosomal internal transcribed spacer and chloroplast trnL intron sequences. Canadian Journal of Botany 82: 376–392. Warwick, S. I., I. A. Al-Shehbaz, C. A. Sauder, D. F. Murray, and K. Mummenhoff. 2004b. Phylogeny of Smelowskia and related genera (Brassicaceae) based on nuclear ITS DNA and chloroplast trnL intron DNA sequences. Annals of the Missouri Botanical Garden 91: 99–123. Warwick, S. I., A. Francis, and I. A. Al-Shehbaz. 2006. Brassicaceae: species checklist and database on CD-ROM. Plant Systematics and Evolution 259: 249–258. Welsh, S. L., N. D. Atwood, S. Goodrich, and L. C. Higgins, eds. 1993. A Utah flora, second edition. Provo: Brigham Young University. Wen, J. and S. M. Ickert-Bond. 2009. Evolution of the Madrean-Tethyan disjunctions and the North and South American amphitropical disjunctions in plants. Journal of Systematics and Evolution 47: 331–348. Wendel, J. F. and J. J. Doyle. 1998. Phylogenetic incongruence: window into genome history and molecular evolution. Pp. 265–296 in Molecular systematics of plants II: DNA sequencing, eds. D. E. Soltis, P. S. Soltis, and J. J. Doyle. Boston: Kluwer Academic Press. Whittall, J., A. Liston, S. Gisler, and R. J. Meinke. 2000. Detecting nucleotide additivity from direct sequences is a SNAP: an example from Sidalcea (Malvaceae). Plant Biology 2: 211–217. Wiens, D. 1984. Ovule survivorship, brood size, life history, breeding systems, and reproductive success in plants. Oecologia 64: 47–53. Wiens, D., C. L. Calvin, C. A. Wilson, C. I. Davern, D. Frank, and S. R. Seavey. 1987. Reproductive success, spontaneous embryo abortion, and genetic load in flowering plants. Oecologia 71: 501–509. 978 SYSTEMATIC BOTANY Zuloaga, F. O. and O. Morrone, eds. 1999. Catálogo de las plantas vasculares de la Repœblica Argentina II. Acanthaceae – Euphorbiaceae (Dicotyledoneae). St. Louis: Missouri Botanical Garden Press. Zunk, K., K. Mummenhoff, and H. Hurka. 1999. Phylogenetic relationships in tribe Lepidieae (Brassicaceae) based on chloroplast DNA restriction site variation. Canadian Journal of Botany 77: 1504–1512. Zwickl, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph. D. dissertation. Austin: The University of Texas. Appendix 1. Traditionally recognized species concepts for Descurainia and related taxa (primarily based on the Brassicaceae checklist of Warwick et al. [2006]). Taxa sampled in this study are marked with an * (see Appendix 3). More questionable taxa are marked by a † and explained as follows: 1most likely mislabelled South American collections (cf. Rollins 1993a); 2known only from the type locality (Saguache Co., Colorado) and probably an intermediate form between D. pinnata and D. incana; 3studies by Bricker and al. (2000) suggest that D. torulosa is not sufficiently distinct from D. incana to merit taxonomic recognition; and 4the species description (Schulz 1924) is apparently based on an immature non-fruiting specimen and herbarium material labelled as D. argentea is nearly non-existent. Citations for chromosome numbers appear in Warwick and Al-Shehbaz (2006) except 5Suda et al. 2003 and 6Ortiz 1993. 7Appears as D. incisa in Warwick and Al-Shehbaz (2006), but the specimen on which it is based is actually D. obtusa subsp. brevisiliqua. Order is taxon, distribution, Reported chromosome numbers n =, 2n =. Section Sisymbriodendron (Christ) O. E. Shulz (Canary Islands): D. artemisioides Svent.*, Gran Canaria,–, 14; D. bourgaeana Webb. ex. O. E. Schulz*, Tenerife, 7, 14; D. gilva Svent.*, La Palma,–, 14;5 D. gonzalezi Svent.*, Tenerife, 7, 14, 14, 21; D. lemsii Bramwell*, Tenerife, 7, 14; D. millefolia (Jacq.) Webb & Berthel.*, Tenerife, La Palma, La Gomera,–, 14; D. preauxiana (Webb) Webb ex O. E. Schulz*, Gran Canaria,–, 14. Section Descurainia: Eurasia: D. kochii (Petri) O. E. Schulz*, Turkey, Caucasus region,–,–; D. sophia (L.) Webb*, Europe, Asia (except SE), N Africa (& New World), 14, 28; D. tanacetifolia (L.) Prantl, 7, 14-16; subsp. suffruticosa, Pyrenees, N Spain,–, 14;6 subsp. tanacetifolia*, SW Alps,–,–. North America: D. californica (A. Gray) O. E. Schulz*, W U. S. A., 7, 14; D. hartwegiana (E. Fourn.) Britton1†, Central Mexico?,–,–; D. impatiens (Cham. & Schltdl.) O. E. Schulz*, Central and S Mexico, 7,–; D. incana (Bernh. ex Fisch. & C. A. Mey.) Dorn*, Canada, N U. S. A. incl. Alaska, 7, 14, 28; D. incisa (Engelm. ex A. Gray) Britton, subsp. filipes (A. Gray) Rollins*, W Canada, W U. S. A.,–, 14, 28, 42; subsp. incisa*, W Canada, W U. S. A.,–,–; subsp. paysonii (Detling) Rollins*, W U. S. A.,–,–; subsp. viscosa (Rydb.) Rollins*, W U. S. A., 7, 14; D. obtusa (Greene) O. E. Schulz, subsp. adenophora (Wooton & Standley)*, SW U. S. A., N Mexico, 21,–; subsp. brevisiliqua Detling*, N Arizona, N New Mexico,–, 42;7 subsp. obtusa*, Arizona, New Mexico, N Mexico,–, 14; D. paradisa (A. Nels. & Kenn.) O. E. Schulz, subsp. nevadensis Rollins*, Nevada, Oregon, California,–,–; subsp. paradisa*, Nevada,–,–; D. pinnata (Walter) Britton, subsp. brachycarpa (Richardson) Detling*, Canada, N & central U. S. A., 7, 28; subsp. glabra (Wooton & Standl.) Detling*, SW U. S. A., N Mexico,–, 28; subsp. halictorum (Cockerell) Detling*, Central & W U. S. A., N Mexico, 7, 14,28, 42; subsp. intermedia (Rydb.) Detling*, W Canada, W U. S. A.,–, 28; subsp. menziesii (DC.) Detling*, S California, Baja California,–, 28; subsp. nelsonii (Rydb.) Detling*, W Canada, W U. S. A.,–, 14; subsp. ochroleuca (Wooton) Detling*, SW U. S. A., N Mexico, 14,–; subsp. pinnata*, SE U. S. A., 7,–; D. ramosissima Rollins2†, Saguache Co., Colorado,–,–; D. sophioides (Fischer ex Hook.) O. E. Schulz*, N Canada, Alaska, N Siberia,–, 14; D. streptocarpa (E. Fourn.) O. E. Schulz*, Central Mexico, Guatemala, 14,–; D. torulosa Rollins3†, W Wyoming,–,–; D. virletii (E. Fourn.) O. E. Schulz*, Central Mexico, 14,–. South America: D. adpressa Boelcke, Prov. San Juan, Argentina,–,–; D. altoandina Romanczuk, Prov. Neuquén, Argentina,–,–; D. antarctica (E. Fourn.) O. E. Schulz, var. antarctica, S Argentina, S Chile,–,–; var. bonarelli O. E. Schulz*, S Argentina,–,–; var. patagonica (Speg.) O. E. Schulz*, S Argentina, S Chile,–,–; D. appendiculata (Griseb.) O. E. Schulz*, Central & NW Argentina, Uruguay, S Bolivia,–,–; D. argentea O. E. Schulz4†, Prov. Chubut, Argentina,–,–; D. argentina O. E. Schulz*, Central and NW Argentina,–,–; D. athrocarpa (A. Gray) O. E. Schulz* (includes D. gilgiana Muschl., D. macbridei O. E. Schulz, and D. urbaniana Muschl.), Peru, Bolivia, N Chile,–,–; D. brevifructa Boelcke ex Mart.-Laborde, Prov. San Juan, Argentina,–,–; D. cumingiana (Fisch. & C. A. Mey.) Prantl, 7, 14; var. cumingiana, Chile, W Argentina,–,–; var. glabrescens (Speg.) Speg., Chile, S Argentina,–,–; var. tenuissima (Phil.) Reiche*, Chile, S Argentina,–,–; D. depressa (Phil.) Reiche*, Bolivia, Peru, N Chile, NW Argentina, D. erodiifolia (Phil.) Prantl ex Reiche*, Chile,–,–; D. glaucescens (Phil.) Prantl ex Reiche*, Chile, W Argentina,–,–; D. heterotricha [Volume 36 Speg.*, Argentina,–,–; D. latisiliqua (Muschl.) O. E. Schulz, S Bolivia,–,–; D. leptoclada Muschl. ex O. E. Schulz*, Bolivia, Peru, NW Argentina, N Chile,–, 14; D. myriophylla (Willd. ex DC.) R. E. Fr.* (includes D. perkinsiana Muschl. and D. pulcherrima Muschl.), Colombia, Ecuador, Peru, Bolivia, N Chile, NW Argentina, 7, 14 (28 in some cells); D. nana Romanczuk, Prov. Santa Cruz, Argentina,–,–; D. nuttallii (Colla) O. E. Schulz, Chile,–,–; D. pimpinellifolia (Barnéoud) O. E. Schulz*, Chile, W Argentina,–,–; D. stricta (Phil.) Reiche*, N Chile,–,–; var. minutiflora (Phil.) O. E. Schulz, var. rubescens (Phil.) O. E. Schulz, var. stricta, var. florida (Phil.) O. E. Schulz, D. titicacensis (Walpers) Lillo, Peru, Bolivia, NW Argentina,–,–. Related taxa: Ianhedgea minutiflora (Hook. f & Thoms.) Al-Shehbaz & O’Kane*, Central and SW Asia,–, 28; Robeschia schimperi (Boiss.) O. E. Schulz*, Middle East, 8,–. Appendix 2. Plant material and GenBank accession numbers (ITS, trnL) used to assess the monophyly of Descurainia and its relationship to other genera. Collection information - location, DNA voucher (herbarium) - is given before accession numbers if applicable. Seed source for cultivated plants designated as follows: [ETSIA] = Escuela Técnica Superior de Ingenieros Agrónomos de Madrid crucifer seedbank, Universidad Politécnica de Madrid, Spain; [B&T] = B&T World Seeds, Paguignan, France. Arabidopsis thaliana (L.) Heynh.: NC_000932; Arabis alpina L.: AF137559, AY034180; Boechera holboellii (Hornem.) Á. Löve & D. Löve: AY457932, DQ013055; Brassica rapa L.: AF531563, AY236217; Cardamine amara L.: AY260584, AF266633; Descurainia argentina O. E. Schulz var. brachysiliqua (Chodat & Wilczek) O. E. Schulz: Cultivated, seed [ETSIA 240-5886-81] from Argentina (TEX), HQ896510, JF298528; D. gilva Svent.: Spain: Canary Islands, A. Santos s. n. (ORT), DQ418714, JF298529; D. incisa (Engelm. ex A. Gray) Britton subsp. incisa: U. S. A.: Colorado, Goodson 1502 (TEX), DQ418717, JF298530; D. kochii (Petri) O. E. Schulz: Turkey: Çankiri, A. Dönmez 11789 (TEX), HQ896547, JF298531; D. pinnata (Walter) Britton subsp. glabra (Wooton & Standley) Detling: U. S. A.: Arizona, R. C. Haberle 177 (TEX), HQ896585, JF298532; D. sophia (L.) Webb ex Prantl: U. S. A.: New Mexico, Beilstein 01-19 (MO), HQ896614, JF298533; D. tanacetifolia (L.) Prantl subsp. suffruticosa: Cultivated from seeds [B&T] (TEX), HQ896624, JF298534; Halimolobos elatus (Rollins) Al-Shehbaz & C. D. Bailey: DQ336388, DQ336387; Hornungia alpina (L.) O. Appel: DQ310527, DQ310515; H. petraea (L.) Reichenbach: AJ440308, AY015905; H. procumbens (L.) Hayek: AJ440309, AY015903; Ianhedgea minutiflora (Hook. f. & Thoms.) Al-Shehbaz & O’Kane: Tajikistan: Badakhson, Solomon et al. 21646 (MO), HQ896625, JF298535; Lepidium campestre (L.) R. Br.: AF055197, AY015845; L. virginicum L.: AY662280, AY015902; Mancoa bracteata (S. Wats.) Rollins: AF307633, AF307556; Nasturtium officinale R. Br.: AY254531, AY122457; Nerisyrenia linearifolia (S. Wats.) Greene: AF055200, AF055267; Nevada holmgrenii (Rollins) N. H. Holmgren: AY230589, AY230555; Physaria fendleri (A. Gray) O’Kane & Al-Shehbaz: AF055199, AF055266; Polyctenium fremontii (S. Wats) Greene: AY230614, AY230614; Robeschia schimperi (Boiss.) O. E. Schulz: Iran: Prov. Esfahan, American-Iranian Botanical Delegation 33719 (TUH), HQ896627, JF298536; Sisymbrium altissimum L.: U. S. A.: Colorado, Goodson 1460 (TEX), HQ896628, JF298537; Smelowskia americana (Regel & Herder) Rydb.: U. S. A.: Colorado, Goodson 1462 (TEX), DQ418729, JF298538; S. bartholomewii (Al-Shehbaz) Al-Shehbaz & S. I. Warwick: AY230609, AY230550; S. calycina (Stephen) C. A. Mey: AY230576, AY230523; S. jacutica (Botsch. & Karav.) Al-Shehbaz & S. I. Warwick: AY230606, AY230548; S. parryoides (Cham.) Polunin: AY230625, AY230540; S. sisymbrioides (Regel & Herder) Lipsky ex Paulsen: Tajikistan: Pil’doni-Poyen, Chukavina 352 (GH), JF298540, JF29839; S. sophiifolia (Cham. & Schltdl.) Al-Shehbaz & S. I. Warwick: AY230608, AY230542; S. tibetica (Thomson) Lipsky: AY230627, AY230551; Tropidocarpum gracile Hook.: unpublished ITS seq. from R. A. Price. Appendix 3. Plant material used to examine phylogenetic relationships within Descurainia. Seed source for cultivated plants designated as follows: [ETSIA] = Escuela Técnica Superior de Ingenieros Agrónomos de Madrid crucifer seedbank, Universidad Politécnica de Madrid, Spain; [B&T] = B&T World Seeds, Paguignan, France. Following each taxon name is given, for each sample: DNA sample number, collection location, collector, DNA voucher (herbarium), and ITS, rps16, trnD-trnE, trnE-trnT, psbZ-trnfM, trnC-petN, petN-psbM, and ndhF-rpl32 GenBank accession numbers. Arabidopsis thaliana (L.) Heynh.: GenBank; NC_000932; Descurainia antarctica (Fourn.) O. E. Schulz: var. bonarelli O. E. Schulz – D37: Argentina: Prov. Santa Cruz, O. Boelke et al. 16806 (BAA); HQ896501/ HQ896502, HQ913409, HQ913185, HQ913297, JF298282, JF298394, JF298650, JF298541; var. patagonica (Speg.) O. E. Schulz – D52: Argentina: Prov. Chubut, O. Boelcke 16038 (BAA); HQ896503, -, HQ913186, HQ913298, 2011] GOODSON ET AL.: DESCURAINIA JF298283, JF298395, JF298651, -; E2: Cultivated from seed collected by B. Goodson, Prov. Chubut, Argentina, B. Goodson 1629 (TEX); HQ896504/ HQ896505, HQ913410, HQ913187, HQ913299, JF298284, JF298396, JF298652, JF298542; F15: Cultivated from seed collected by B. Goodson, Prov. Chubut, Argentina, B. Goodson 1630 (TEX); HQ896506, HQ913411, HQ913188, HQ913300, JF298285, JF298397, JF298653, JF298543; D. appendiculata (Griseb.) O. E. Schulz: B126: Cultivated from seed collected by B. Goodson, Prov. Salta, Argentina, B. Goodson 1627 (TEX); HQ896507, HQ913412, HQ913189, HQ913301, JF298286, JF298398, JF298654, JF298544; C25: Cultivated from seed collected by B. Goodson, Prov. Tucumán, Argentina, B. Goodson 1621 (TEX); HQ896508, HQ913413, HQ913190, HQ913302, JF298287, JF298399, JF298655, JF298545; D47: Argentina: Prov. Mendoza, Paladina s. n. (BAA); HQ896509, HQ913414, HQ913191, HQ913303, JF298288, JF298400, JF298656, JF298546; D. argentina O. E. Schulz: var. brachysiliqua (Chodat & Wilczek) O. E. Schulz – B37: Cultivated from seed [ETSIA 240-5886-81] collected from Prov. La Pampa, Argentina, B. Goodson 1622 (TEX); HQ896510, HQ913415, HQ913192, HQ913304, JF298289, JF298401, JF298657, JF298547; var. indet. – B96: Cultivated from seed [ETSIA 239-5908-81] collected from Prov. Mendoza, Argentina, B. Goodson 1623 (TEX); HQ896511, HQ913416, HQ913193, HQ913305, JF298290, JF298402, JF298658, JF298548; D. artemisioides Svent.: B36: Cultivated from seed collected [ETSIA 241-4201-76] from Gran Canaria, Canary Is., Spain, B. Goodson 1632 (TEX); DQ418708, DQ418686, DQ418554, DQ418576, DQ418598, JF298403, JF298659, DQ418620; D. athrocarpa (A. Gray) O. E. Schulz: B94: Peru: Dept. Ancash, D. N. Smith et al. 9450 (MO); HQ896512, HQ913417, HQ913194, HQ913306, JF298291, JF298404, JF298660, JF298549; C27: Bolivia: Dept. La Paz, B. Goodson 1506 (TEX); HQ896513, HQ913418, HQ913195, HQ913307, JF298292, JF298405, JF298661, JF298550; D. bourgaeana Webb ex O. E. Schulz: B14: Spain: Tenerife, Canary Is., A. Santos s. n. (ORT); DQ418709, DQ418665, DQ418555, DQ418577, DQ418599, JF298406, JF298662, DQ418621; B171: Cultivated, seed [ETSIA 242-1629-68], Tenerife, Canary Is., Spain (TEX); DQ418710, DQ418688, DQ418556, DQ418578, DQ418600, JF298407, JF298663, DQ418622; D7: Spain: La Palma, Canary Is., A. Santos s. n. (ORT); DQ418711, DQ418689, DQ418557, DQ418579, DQ418601, JF298408, JF298664, DQ418623; D. californica (A. Gray) O. E. Schulz: C9: U. S. A.: Nevada, B. Goodson 1493 (TEX); HQ896514, HQ913419, HQ913196, HQ913308, JF298293, JF298409, JF298665, JF298551; D12: U. S. A.: Utah, B. Goodson 1466 (TEX); HQ896515, HQ913420, HQ913197, HQ913309, JF298294, JF298410, JF298666, JF298552; D. cumingiana (Fisch. & C. A. Mey) Prantl: var. cumingiana – D34: Argentina: Prov. Neuquén, AbadieSpeck 7 (BAA); HQ896517, -, -, -, -, -, -, -; D39: Argentina: Prov. Neuquén, M. N. Correa et al. 8699 (BAA); HQ896518, HQ913422, HQ913199, HQ913311, JF298296, JF298412, JF298668, JF298554; var. tenuissima (Phil.) Reiche – B103: Chile: Atacama (Region III), M. Muñoz et al. 2930 (MO); HQ896519, -, -, -, -, -, -, -; D38: Argentina: Prov. Chubut, Irisarri 180 (BAA); HQ896520, -, -, -, -, -, -, -; D43: Argentina: Prov. Rio Negro, Abadie-Vallerini 1020 (BAA); HQ896521, HQ913423, HQ913200, HQ913312, JF298297, JF298413, JF298669, JF298555; D49: Chile: Valparaíso (Region V), A. Garaventa 8072 (BAA); HQ896522, HQ913424, HQ913201, HQ913313, JF298298, JF298414, JF298670, JF298556; D. depressa (Phil.) Reiche: C26: Bolivia: Dept. La Paz, B. Goodson 1505 (TEX); DQ418712, DQ418690, DQ418558, DQ418580, DQ418602, JF298415, JF298671, DQ418646; C37: Bolivia: Dept. La Paz, B. Goodson 1520 (TEX); HQ896523, HQ913425, HQ913202, HQ913314, JF298299, JF298416, JF298672, JF298557; D17: Bolivia: Dept. La Paz, B. Goodson 1510 (TEX); HQ896524, HQ913426, HQ913203, HQ913315, JF298300, JF298417, JF298673, JF298558; D31: Argentina: Prov. Jujuy, L. Giusti et al. 558 (BAA); HQ896525, HQ913427, HQ913204, HQ913316, JF298301, JF298418, JF298674, JF298559; D. cf. erodiifolia (Phil.) Reiche (Chile): D50: Argentina: Prov. Mendoza, Roig 10766 (BAA); HQ896516, HQ913421, HQ913198, HQ913310, JF298295, JF298411, JF298667, JF298553; D. gilva Svent: B22: Spain: La Palma, Canary Is., S. Santos s. n. (ORT); DQ418714, DQ418692, DQ418560, DQ418582, DQ418604, JF298420, JF298676, DQ418626; B163: Cultivated, seed collected [ETSIA 243-4055-76] from La Palma, Canary Is., Spain (TEX); DQ418713, DQ418691, DQ418559, DQ418581, DQ418603, JF298419, JF298675, DQ418625; D. glaucescens (Phil.) Prantl ex Reiche: D20: Chile: Region III (Atacama), E. Werdermann 971 (MO); HQ896526, HQ913428, HQ913205, HQ913317, JF298302, JF298421, JF298677, JF298560; D. gonzalezi Svent.: B19: Spain: Tenerife, Canary Is., A. Santos s. n. (ORT); DQ418562, DQ418694, DQ418584, DQ418606, DQ418716, JF298423, JF298679, DQ418628; B160: Cultivated, seed [ETSIA 244-3172-74] from Tenerife, Canary Is., Spain (TEX); DQ418561, DQ418693, DQ418583, DQ418605, DQ418715, JF298422, JF298678, DQ418627; D. heterotricha Speg.: B124: Cultivated from seed collected by B. Goodson, Prov. Mendoza, Argentina, B. Goodson 1628 (TEX); HQ896527, HQ913429, HQ913206, HQ913318, JF298303, JF298424, JF298680, JF298561; D. impatiens (Cham. & Schlecht.) O. E. Schulz: C40: 979 Mexico: Oaxaca, G. B. Hinton et al. 26690 (TEX); HQ896528, HQ913430, HQ913207, HQ913319, JF298304, JF298425, JF298681, JF298562; C42: Mexico: Veracruz, F. Ventura A. 1338 (TEX); HQ896529, HQ913431, HQ913208, HQ913320, JF298305, JF298426, JF298682, JF298563; D. incana (Bernh. ex Fischer & C. A. Meyer) Dorn: B109: U. S. A.: Montana, P. P. Lowrey 2693 (MO); HQ896530, HQ913432, HQ913209, HQ913321, JF298306, JF298427, JF298683, JF298564; C2: U. S. A.: Montana, R. C. & K. W. Rollins 83308 (GH); HQ896531, HQ913433, HQ913210, HQ913322, JF298307, JF298428, JF298684, JF298565; D29: U. S. A.: Idaho, R. C. & K. W. Rollins 86118 (TEX); HQ896532, HQ913434, HQ913211, HQ913323, JF298308, JF298429, JF298685, JF298566; D. incisa (Engelm. ex A. Gray) Britton: subsp. filipes (A. Gray) Rollins – B195: U. S. A.: Nevada, A. Tiehm 11104 (GH); HQ896533, HQ913435, HQ913212, HQ913324, JF2983309, JF298430, JF298686, -; C21: U. S. A.: Utah, B. Goodson 1499 (TEX); HQ896534, HQ913436, HQ913213, HQ913325, JF298310, JF298431, JF298687, JF298567; C45: U. S. A.: Nevada, A. Tiehm 13845 (TEX); HQ896535, HQ913437, HQ913214, HQ913326, JF298311, JF298432, JF298688, JF298568; D14: U. S. A.: Utah, B. Goodson 1500 (TEX); HQ896536/ HQ896537, HQ913438, HQ913215, HQ913327, JF298312, JF298433, JF298689, JF298569; subsp. incisa – C24: U. S. A.: Colorado, B. Goodson 1502 (TEX); DQ418717, DQ418695, DQ418563, DQ418585, DQ418607, JF298434, JF298690, DQ418629; D6: U. S. A.: Utah, B. Goodson 1528 (TEX); HQ896541, HQ913442, HQ913219, HQ913331, JF298316, JF298438, JF298694, JF298573; D25: U. S. A.: Idaho, R. C. & K. W. Rollins 86101 (TEX); HQ896538, HQ913439, HQ913216, HQ913328, JF2983313, JF298435, JF298691, JF298570; D56: U. S. A.: New Mexico, J. McGrath 157 (UNM); HQ896539, HQ913440, HQ913217, HQ913329, JF298314, JF298436, JF298692, JF298571; D57: U. S. A.: New Mexico, T. Dunbar 609 (UNM); HQ896540, HQ913441, HQ913218, HQ913330, JF298315, JF298437, JF298693, JF298572; subsp. paysonii (Detling) Rollins – D28: U. S. A.: Colorado, W. A. Weber & P. Salamun 12649 (TEX); HQ896542, HQ913443, HQ913220, HQ913332, JF298317, JF298439, JF298695, JF298574; D73: U. S. A.: New Mexico, W. L. Wagner 1932 (UNM); HQ896543, HQ913444, HQ913221, HQ913333, JF298318, JF298440, JF298696, JF298575; subsp. viscosa (Rydb.) Rollins – D21: U. S. A.: Wyoming, Porter & Porter 10187 (TEX); HQ896544, HQ913445, HQ913222, HQ913334, JF298319, JF298441, JF298697, -; D24: U. S. A.: Arizona, A. R. & H. N. Moldenke 27885 (LL); HQ896545, HQ913446, HQ913223, HQ913335, JF298320, JF298442, JF298698, JF298576; D. kochii (Petri) O. E. Schulz: D2: Turkey: Çankiri, A. Dönmez 11789 (TEX); HQ896547, HQ913448, HQ913225, HQ913337, JF298322, JF298444, JF298700, JF298578; D3: Turkey: Kastamonu, A. Dönmez 11793 (TEX); DQ418718, DQ418696, DQ418564, DQ418586, DQ418608, JF298445, JF298701, DQ418630; D18: Turkey: Kirikkale, A. Dönmez 11928 (TEX); HQ896546, HQ913447, HQ913224, HQ913336, JF298321, JF298443, JF298699, JF298577; D. lemsii Bramwell: B23: Spain: Tenerife, Canary Is., A. Santos s. n. (ORT); DQ418720, DQ418698, DQ418566, DQ418588, DQ418610, JF298447, JF298703, DQ418632; B170: Cultivated, seed [ETSIA 245-3094-74] collected from Tenerife, Canary Is., Spain (TEX); DQ418719, DQ418697, DQ418565, DQ418587, DQ418609, JF298446, JF298702, DQ418631; D. leptoclada Muschl.: C33: Bolivia: Dept. Oruro, B. Goodson 1514 (TEX); HQ896548, HQ913449, HQ913226, HQ913338, JF298323, JF298448, JF298704, JF298579; C34: Bolivia: Dept. Oruro, B. Goodson 1515 (TEX); HQ896549, HQ913450, HQ913227, HQ913339, JF298324, JF298449, JF298705, JF298580; C36: Bolivia: Dept. Oruro, B. Goodson 1516 (TEX); HQ896550, HQ913451, HQ913228, HQ913340, JF298325, JF298450, JF298706, JF298581; D46: Argentina: Prov. Jujuy, J. H. Hunziker et al. 10531 (BAA); DQ896621, HQ913511, HQ913289, HQ913401, JF298386, JF298518, JF298774, JF298642; D. millefolia (Jacq.) Webb & Berthel.: B24: Spain: La Palma, Canary Is., A. Santos s. n. (ORT); DQ418721, DQ418699, DQ418567, DQ418589, DQ418611, JF298451, JF298707, DQ418633; B38: Cultivated from seed [ETSIA 246-1073-67] collected from Tenerife, Canary Is., Spain, B. Goodson 1633 (TEX); DQ418722, DQ418700, DQ418568, DQ418590, DQ418612, JF298452, JF298708, DQ418634; D1: Spain: Tenerife, Canary Is., A. Santos & J. Francisco-Ortega 6987 (TEX); DQ418723, DQ418701, DQ418569, DQ418591, DQ418613, JF298453, JF298709, DQ418635; D5: Spain: Tenerife, Canary Is., A. Santos s. n. (ORT); DQ418724, DQ418702, DQ418570, DQ418592, DQ418614, JF298454, JF298710, DQ418636; F1: Spain: La Gomera, Canary Is., leg. ign. AAU71-7259 (MO); HQ896552, -, -, -, -, -, -, -; F2: Spain: La Gomera, Canary Is., leg. ign. AAU71-7533 (MO); DQ418725, DQ418703, DQ418571, DQ418593, DQ418615, JF298455, JF298711, DQ418637; D. myriophylla (Willdenow ex DC.) R. E. Fries: C29: Bolivia: Dept. La Paz, B. Goodson 1508 (TEX); HQ896553, HQ913452, HQ913229, HQ913341, JF298326, JF298456, JF298712, JF298582; C52: Peru: Cuzco, Tupayachi 1065 (MO); HQ896554, HQ913453, HQ913230, HQ913342, JF298327, JF298457, JF298713, JF298583; D9: Bolivia: Dept. La Paz, B. Goodson 1507 (TEX); HQ896557, HQ913456, HQ913233, HQ913345, JF298330, JF298460, JF298716, JF298586; D13: Bolivia: Dept. La Paz, 13 980 SYSTEMATIC BOTANY March 2004, B. Goodson 1511 (TEX); HQ896555, HQ913454, HQ913231, HQ913343, JF298328, JF298458, JF298714, JF298584; D16: Bolivia: Dept. La Paz, B. Goodson 1512 (TEX); HQ896556, HQ913455, HQ913232, HQ913344, JF298329, JF298459, JF298715, JF298585; D. obtusa (E. L. Greene) O. E. Schulz: subsp. adenophora (Wooton & Standley) – D61: U. S. A.: New Mexico, C. A. Huff & D. Stevens 2310 (UNM); HQ896559, HQ913458, HQ913235, HQ913347, JF298332, JF298462, JF298718, JF298588; D62: U. S. A.: Arizona, W. N. Anderson A201 (UNM); HQ896560, HQ913459, HQ913236, HQ913348, JF298333, JF298463, JF298719, JF298589; subsp. brevisiliqua Detling – D58: U. S. A.: New Mexico, B. Hutchins 4450 (UNM); HQ896561, HQ913460, HQ913237, HQ913349, JF298334, JF298464, JF298720, JF298590; D59: U. S. A.: New Mexico, Fletcher 823 (UNM); HQ896562, HQ913461, HQ913238, HQ913350, JF298335, JF298465, JF298721, JF298591; D72: U. S. A.: New Mexico, B. Hutchins 5099 (UNM); HQ896563, HQ913462, HQ913239, HQ913351, JF298336, JF298466, JF298722, JF298592; D4 (cf. subsp. brevisiliqua): U. S. A.: New Mexico, B. Goodson 1527 (TEX); HQ896582, HQ913476, HQ913254, HQ913366, JF298351, JF298481, JF298737, JF298607; subsp. obtusa – B26: U. S. A.: New Mexico, T. Chumley 7359 (TEX); HQ896564, HQ913463, HQ913240, HQ913352, JF298337, JF298467, JF298723, JF2985593; D63: U. S. A.: New Mexico, O. Baca 262 (UNM); HQ896565, HQ913464, HQ913241, HQ913353, JF298338, JF298468, JF298724, JF298594; D64: U. S. A.: New Mexico, W. Wagner 1157 (UNM); HQ896566, HQ913465, HQ913242, HQ913354, JF298339, JF298469, JF298725, JF298595; D65: U. S. A.: New Mexico, B. Hutchins 10245 (UNM); HQ896567, HQ913466, HQ913243, HQ913355, JF298340, JF298470, JF298726, JF298596; D. paradisa (A. Nels. & Kenn.) O. E. Schulz: subsp. nevadensis Rollins – C8: U. S. A.: Nevada, B. Goodson 1492 (TEX); HQ896569/ HQ896570, HQ913468, HQ913245, HQ913357, JF298342, JF298472, JF298728, JF298598; C48: U. S. A.: Nevada, A. Tiehm 11582 (TEX); HQ896568, HQ913467, HQ913244, HQ913356, JF298341, JF298471, JF298727, JF298597; subsp. paradisa – C7: U. S. A.: Nevada, B. Goodson 1490 (TEX); HQ896572, HQ913470, HQ913247, HQ913359, JF298344, JF298474, JF298730, JF298600; C46: U. S. A.: Nevada, A. Tiehm 13794 (TEX); HQ896571, HQ913469, HQ913246, HQ913358, JF298343, JF298473, JF298729, JF298599; D. pimpinellifolia (Barnéoud) O. E. Schulz: D11: Argentina: Prov. Mendoza, B. Goodson 1475 (TEX); HQ896573, HQ913471, HQ913248, HQ913360, JF298345, JF298475, JF298731, JF298601; D42: Argentina: Prov. San Juan, Nicora 8466 (BAA); HQ896574, HQ913472, HQ913249, HQ913361, JF298346, JF298476, JF298732, JF298602; D51: Argentina: Prov. Mendoza, Roig 7409 (BAA); HQ896575, HQ913473, HQ913250, HQ913362, JF298347, JF298477, JF298733, JF298603; E3: Argentina: Prov. San Juan, R. Kiesling 8760 (SI); HQ896576, HQ913474, HQ913251, HQ913363, JF298348, JF298478, JF298734, JF298604; D. pinnata (Walter) Britton: subsp. brachycarpa (Richardson) Detling – F11: U. S. A.: Minnesota, S. D. Swanson 490 (MO); HQ896577/ HQ896578, -, HQ913252, HQ913364, JF298349, JF298479, JF298735, JF298605; F12: U. S. A.: Kentucky, R. Athey 536 (MO); HQ896579, -, -, -, -, -, -, -; subsp. glabra (Wooton & Standley) Detling – B144: U. S. A.: Arizona, R. C. Haberle 177 (TEX); HQ896585, HQ913479, HQ913257, HQ913369, JF298354, JF298484, JF298740, JF298610; D27: Mexico: Sonora, R. S. Felger 94-88 (TEX); HQ896588, HQ913481, HQ913257, HQ913371, JF298356, JF298486, JF298742, JF298612; C10: U. S. A.: California, T. Chumley 7434 (TEX); HQ896586/ HQ896587, HQ913480, HQ913258, HQ913370, JF298355, JF298485, JF298741, JF298611; subsp. halictorum (Wooton) Detling – C12: U. S. A.: Nevada, T. Chumley 7437 (TEX); HQ896589, HQ913482, HQ913260, HQ913372, JF298357, JF298487, JF298743, JF298613; C14: U. S. A.: Nevada, T. Chumley 7440 (TEX); HQ896590/ HQ896591, HQ913483, HQ913261, HQ913373, JF298358, JF298488, JF298744, JF298614; D10: U. S. A.: Texas, B. Goodson 1521 (TEX); HQ896593, HQ913485, HQ913263, HQ913375, JF298360, JF298490, JF298746, JF298616; D19: U. S. A.: Texas, B. Goodson 1523 (TEX); HQ896594, HQ913486, HQ913264, HQ913376, JF298361, JF298491, JF298747, JF298617; D67: U. S. A.: Arizona, M. Kurzius 84-5 (UNM); HQ896595, HQ913487, HQ913265, HQ913377, JF298362, JF298492, JF298748, JF298618; D69: U. S. A.: New Mexico, A. C. Cully & M. Medrano s. n. (UNM); HQ896596, HQ913488, HQ913266, HQ913378, JF298363, JF298493, JF298749, JF298619; D71: U. S. A.: New Mexico, G. Tierney A84575 (UNM); HQ896597, HQ913489, HQ913267, HQ913379, JF298364, JF298494, JF298750, JF298620; subsp. intermedia (Rydb.) Detling – C19: U. S. A.: Utah, B. Goodson 1498 (TEX); HQ896598, HQ913490, HQ913268, HQ913380, JF298365, JF298495, JF298751, JF298621; subsp. menziesii (DC.) Detling – B35: Cultivated from seed [ETSIA 248-1725-69] collected from California, U. S. A., B. Goodson 1635 (TEX); HQ896599, HQ913491, HQ913269, HQ913381, JF298366, JF298496, JF298752, JF298622; C3: Mexico: Baja California Norte, T. Chumley 7429 (TEX); HQ896600/ HQ896601, HQ913492, HQ913270, HQ913382, JF298367, JF298497, JF298753, JF298623; D53: U. S. A.: California, E. LaRue s. n. (TEX); HQ896602, HQ913493, [Volume 36 HQ913271, HQ913383, JF298368, JF298498, JF298754, JF298624; D55: U. S. A.: California, A. L. & H. N. Moldenke 30653 (TEX); HQ896603, HQ913494, HQ913272, HQ913384, JF298369, JF298499, JF298755, JF298625; subsp. nelsonii (Rydb.) Detling – C17: U. S. A.: Wyoming, B. Goodson 1495 (TEX); HQ896604, HQ913495, HQ913273, HQ913385, JF298370, JF298500, JF298756, JF298626; C47: U. S. A.: Nevada, A. Tiehm 13911 (TEX); HQ896605, HQ913496, HQ913274, HQ913386, JF298371, JF298501, JF298757, JF298627; D23: U. S. A.: Montana, R. C. & K. W. Rollins 86185 (TEX); HQ896606, HQ913497, HQ913275, HQ913387, JF298372, JF298502, JF298758, JF298628; subsp. ochroleuca (Wooton) Detling – D8: U. S. A.: Texas, B. Goodson 1524 (TEX); HQ896607, HQ913498, HQ913276, HQ913388, JF298373, JF298503, JF298759, JF298629; D26: Mexico: Chihuahua, C. W. Pennington 325 (TEX); HQ896558, HQ913457, HQ913234, HQ913346, JF298331, JF298461, JF298717, JF298587; subsp. pinnata – B12a: U. S. A.: Texas, B. Goodson 1457 (TEX); HQ896608, HQ913499, HQ913277, HQ913389, JF298374, JF298504, JF298760, JF298630; D15: U. S. A.: Texas, B. Goodson 1522 (TEX); HQ896609, HQ913500, HQ913278, HQ913390, JF298375, JF298505, JF298761, JF298631; F5: U. S. A.: Georgia, S. B. Jones 24758 (MO); HQ89661, HQ913280, HQ913392, JF298377, JF298507, JF298763, JF298633; F6: U. S. A.: Florida, M. R. Crosby 4844 (MO); HQ896612, -, -, -, -, -, -, -; F17: U. S. A.: Florida, B. Goodson 1616 (TEX); HQ896610, HQ913517, HQ913279, HQ913391, JF298376, JF298506, JF298762, JF298632; subsp. undet. – C4: U. S. A.: Arizona, T. Chumley 7427 (TEX); HQ896580/ HQ896581, HQ913475, HQ913253, HQ913365, JF298350, JF298480, JF298736, JF298606; C15: U. S. A.: Nevada, T. Chumley 7439 (TEX); HQ896592, HQ913484, HQ913262, HQ913374, JF298359, JF298489, JF298745, JF298615; D68: U. S. A.: New Mexico, T. Maddux & S. Loftin 12 (UNM); HQ896583, HQ913477, HQ913255, HQ913367, JF298352, JF298482, JF298738, JF298608; D70: U. S. A.: New Mexico, T. Maddux 327 (UNM); HQ896584, HQ913478, HQ913256, HQ913368, JF298353, JF298483, JF298739, JF298609; D. preauxiana (Webb) Webb ex O. E. Schulz: B117: Cultivated from seed [ETSIA 249-4135-76] collected from Gran Canaria, Canary Is., Spain, B. Goodson 1631 (TEX); DQ418726, DQ418704, DQ418572, DQ418594, DQ418616, JF298508, JF298764, DQ418638; D. sophia (L.) Webb ex Prantl: B20: U. S. A.: Colorado, B. Goodson 1461 (TEX); DQ418727, DQ418705, DQ418573, DQ418595, DQ418617, JF298509, JF298765, DQ418639; E6: Argentina: Prov. Chubut, B. Goodson 1560 (TEX); HQ896613, HQ913503, HQ913281, HQ913393, JF298378, JF298510, JF298766, JF298634; MB3: U. S. A.: New Mexico, Beilstein 01-19 (MO); HQ896614, HQ913504, HQ913282, HQ913394, JF298379, JF298511, JF298767, JF298635; D. sophioides (Fischer) O. E. Schulz: B112: Canada: Yukon Territory, Cooper 715 (NY); HQ896615, HQ913505, HQ913283, HQ913395, JF298380, JF298512, JF298768, JF298636; F13: Cultivated, seed collected from North Slope Co., Alaska, U. S. A. (TEX); HQ896616, HQ913506, HQ913284 HQ913396, JF298381, JF298513, JF298769, JF298637; D. streptocarpa (Fourn.) O. E. Schulz: B33: Mexico: Veracruz, B. F. Hansen & M. Nee 7702 (MO); HQ896617, HQ913507, HQ913285, HQ913397, JF298382, JF298514, JF298770, JF298638; C44 (D. cf. streptocarpa): Mexico: Chihuahua, M. H. Mayfield et al. 206 (TEX); HQ896618, HQ913508, HQ913286, HQ913398, JF298383, JF298515, JF298771, JF298639; D. stricta (Phil.) Reiche: var. undet. – C38: Chile: Tarapacá (Region I), J. L. Panero & B. S. Crozier 8435 (TEX); HQ896619, HQ913509, HQ913287, HQ913399, JF298384, JF298516, JF298772, JF298640; D45: Chile: Tarapacá (Region I), C. Villagrán 2457 (BAA); HQ896620, HQ913510, HQ913288, HQ913400, JF298385, JF298517, JF298773, JF298641; D. tanacetifolia (L.) Prantl: subsp. suffruticosa – C6: Cultivated from seeds [B&T] (TEX); HQ896624, HQ913514, HQ913292, HQ913404, JF298389, JF298522, JF298778, JF298645; subsp. tanacetifolia (SW Alps) – B111: Italy: Piemonte, Pistarino 2027 (NY); DQ418728, DQ418706, DQ418574, DQ418596, DQ418618, JF298521, JF298777, DQ418640; D. virletii (Fourn.) O. E. Schulz: B108: Mexico: Mexico, I. Piña E. 100 (MO); HQ896622, HQ913512, HQ913290, HQ913402, JF298387, JF298519, JF298775, JF298643; C39: Mexico: Chiapas, J. Pérez 266 (TEX); HQ896623, HQ913513, HQ913291, HQ913403, JF298388, JF298520, JF298776, JF298644; Ianhedgea minutiflora (Hook. f. & Thoms.) Al-Shehbaz & O’Kane: MB2: Tajikistan: Badakhson, Solomon et al. 21646 (MO); HQ896625, HQ913515, HQ913293, HQ913405, JF298390, JF298523, JF298779, JF298646; Robeschia schimperi (Boiss.) O. E. Schulz: B106: Iran: Prov. Kerman, K. H. & F. Rechinger 3076 (MO); HQ896626, HQ913516, HQ913294, HQ913406, JF298391, JF298524, JF298780, JF298647; MB1: Iran: Prov. Esfahan, American-Iranian Botanical Delegation 33719 (TUH); HQ896627, HQ913517, HQ913295, HQ913407, JF298392, JF298525, JF298781, JF298648; Sisymbrium altissimum L.: B21: U. S. A.: Colorado, B. Goodson 1460 (TEX); HQ896628, HQ913518, HQ913296, HQ913408, JF298393, JF298526, JF298782, JF298649; Smelowskia americana (Regel & Herder) Rydb.: B146: U. S. A.: Colorado, B. Goodson 1462 (TEX); DQ418729, DQ418707, DQ418575, DQ418597, DQ418619, JF298527, JF298783, DQ418641.