Revised generic delimitations for Oxypolis and Ptilimnium (Apiaceae) based on leaf morphology, comparative fruit anatomy, and phylogenetic analysis of nuclear rDNA ITS and cpDNA trnQ -trnK intergenic spacer sequence data

Mary Ann Feist

Anatomical characters are very important to distinguish between closely related species and genera, especially in the family Apiaceae. Anatomy of Apiaceae fruits varies strongly, even among closely related species in the same genus. An ethnobotanical study was conducted between 2004 and 2005 in order to determine wild used plants in West part of Nigde-Aladaglar. According the data of this study’s, 3 plant species belong to Apiaceae family sold to different tea companies by local people. Anatomical characteristics of fruits of the Ferulago pachyloba, Prangos ferulacea, Prangos meliocarpoides have been studied. Ferulago pachyloba has winged dorsal ribs, equal to or shorter than seeds. Prangos ferulacea has five blocks of mesocarp, they are not separated by the exocarp and the vascular bundles usually surround each block. Prangos meliocarpoides has continuous mesocarp and not separated into blocks. Druse crystals do not exist in the endosperm of F. pachyloba, P. ferulacea, and P. melio...

Determination of phylogenetic affinity among the members of Apiaceae especially the subfamily Apioideae is much debatable. The morphological evidences on inflorescence, fruit and seed characters showed little concomitance with those derived from molecular parameters as far as groupings in clades are concerned. In the present study involving few selected Apiaceae members, a Dendrogram was computed and a putative evolutionary trend was outlined based on leaf margin insertion pattern and cotyledonary leaf features, general plant morphology as well as leaf surface micromorphology by SEM. The members of Apiaceae with broad-ovate pinnatifid to pinnatipartite leaves and ovate –lanceolate cotyledonary leaves were included in a distinct cluster but the members with pinnatisect – decompounds leaves and linear cotyledonary leaves were included in a separate cluster. Long Coriander (Eryngium foetidum) included in the subfamily Saniculoideae may represent the most primitive condition having simple entire leaf with spiny margin. The member like Long coriander might have served as progenitor stock from which Coriander, Celery, Parsley (all with ovate pinnatifid to pinnatipartite leaves) and Cumin, Fennel, Dill (all having pinnatisect to decompounds leaves with linear segments) might have originated as evidenced by gradual increase in marginal insertion. Similarly, gradual narrowing of ovate-lanceolate cotyledonary leaves might have given rise to narrow linear cotyledonary leaves. Leaf-margin insertion pattern and cotyledonary leaf features appeared to be promising in tracing phylogeny in Apiaceae especially in Apioideae. Simple entire leaf represents the primitive condition and evolution in Apiaceae supposed to have progressed towards gradual increase in marginal insertion leading to pinnately dissected, pinnatipartite and ultimately to decompound leaves with narrow linear segments. Dendrogram computed from morphological features showed concomitance with that phylogenetic trend as far as grouping and interrelationships of members are concerned. In the present study leaf margin insertion pattern and seedling morphology appeared to be instrumental in tracing the phylogeny in Apiaceae specially subfamily Apioideae while molecular systematics are not conclusive.

Feist & al. • Generic delimitations for Oxypolis and Ptilimnium TAXON 61 (2) • April 2012: 402–418 Revised generic delimitations for Oxypolis and Ptilimnium (Apiaceae) based on leaf morphology, comparative fruit anatomy, and phylogenetic analysis of nuclear rDNA ITS and cpDNA trnQ-trnK intergenic spacer sequence data Mary Ann E. Feist,1 Stephen R. Downie,2 Anthony R. Magee3,4 & Mei (Rebecca) Liu5 1 Illinois Natural History Survey, Prairie Research Institute, University of Illinois at UrbanaChampaign, Champaign, Illinois 61820, U.S.A. 2 Department of Plant Biology, University of Illinois at UrbanaChampaign, Urbana, Illinois 61801, U.S.A. 3 Compton Herbarium, South African National Biodiversity Institute, Private Bag X7, Claremont 7735, South Africa 4 Department of Botany and Plant Biotechnology, University of Johannesburg, P.O. Box 524, Auckland Park 2006, Johannesburg, South Africa 5 Department of Biology, Harbin Normal University, Hexing Road 50, Harbin 150080, People’s Republic of China Author for correspondence: Mary Ann E. Feist, mfeist@illinois.edu Abstract A phylogenetic study of Oxypolis and Ptilimnium, two small genera of tribe Oenantheae (Apiaceae: subfamily Apioideae), was carried out. Generic circumscriptions and infrageneric and infraspecific relationships were investigated through parsimony and Bayesian inference analyses of nuclear rDNA ITS and cpDNA trnQ5′rps16 and 3′rps165′trnK intergenic spacer sequences. Fruit anatomical characters were also examined and used in conjunction with leaf morphology to corroborate the results of the phylogenetic analyses. Each genus as currently delimited has both compound-leaved and rachis-leaved species. Results of the phylogenetic analyses show that neither Oxypolis nor Ptilimnium is monophyletic; each genus is split into two strongly supported clades that correspond to differences in leaf morphology within the groups. Fruit anatomical characters support these splits. The fruits of compound-leaved and rachis-leaved Oxypolis species differ in the number of commissural vittae per mericarp, the branching of the vittae, and the lignification of mericarp around the seed. The fruits of compoundleaved and rachis-leaved Ptilimnium species differ in the compression of the mericarps and the development of the marginal ribs. The fruits of rachis-leaved Oxypolis and rachis-leaved Ptilimnium species also differ in the compression of the mericarps and the development of the marginal ribs. Based on analyses of molecular data and corroboration with morphological and fruit anatomical data, new circumscriptions for the genera Oxypolis and Ptilimnium are formalized. Each of the two polyphyletic genera (Oxypolis and Ptilimnium) is split, two genera (Tiedemannia and Harperella) are resurrected, and three new combinations are made. Keywords Harperella; new combinations; Oenantheae; Oxypolis; Ptilimnium; rachis leaves, Tiedemannia INTRODUCTION Oxypolis Raf. and Ptilimnium Raf. are two small genera of tribe Oenantheae (Apiaceae: subfamily Apioideae). As currently circumscribed, the genus Oxypolis is comprised of seven species and the genus Ptilimnium of six species. These names with taxonomic authorities are presented in Table 1. Most species of Oxypolis and Ptilimnium are endemic to North America, but each genus has one species (i.e., O. filiformis, P. capillaceum) with a range that extends into the West Indies (Brace, 1929; Liogier & Martorell, 2000). Each genus also has one species (i.e., O. canbyi, P. nodosum) listed as federally endangered in the U.S.A. (U.S. Fish and Wildlife Service, 1986, 1988). The genera Oxypolis and Ptilimnium share several ecological and morphological traits including glabrous leaves and stems, fascicled roots, globose to broadly ovate fruits, and a preference for wet habitats. In addition, two very different leaf morphologies are found within each genus. While most species of Oxypolis and Ptilimnium have compound leaves, as is 402 typical in subfamily Apioideae, others share an unusual leaf morphology known as rachis leaves. Rachis leaves are linear, terete, hollow, and septate and are equivalent to the rachis of a pinnately compound leaf in which the pinnae are not fully expressed. Instead, the pinnae are highly reduced and transformed into nodal appendages that function as hydathodes (Kaplan, 1970). One other genus within tribe Oenantheae, Lilaeopsis Greene, also has species with rachis leaves, but this genus has no compound-leaved species. Cynosciadium DC. and Limno sciadium Mathias & Constance, also of tribe Oenantheae, have rachis-like basal leaves (linear and septate, but flattened and not hollow), but their cauline leaves are palmately and pinnately lobed, respectively. Rachis leaves are thought to be an adaptation to an aquatic or semi-aquatic habitat, and species of Oxypolis and Ptilimnium with rachis leaves (i.e., O. canbyi, O. filiformis, O. greenmanii, and P. nodosum) spend much of the growing season at least partially submerged. The compound-leaved species of Oxypolis have pinnately or ternately compound leaves with pinnae that are generally TAXON 61 (2) • April 2012: 402–418 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium somewhat broad (except for O. ternata in which they are long and narrow). Compound-leaved Ptilimnium species have finely dissected, pinnately decompound leaves with linear or filiform pinnae. Despite having radically different leaf morphologies from their compound-leaved congeners, the rachis-leaved Oxy polis and Ptilimnium species were placed in their respective genera based primarily on fruit morphology (Elliott, 1817; Mathias, 1936). Elliott (1817) placed the compound-leaved Sium rigidius L. (now O. rigidior) and the rachis-leaved S. tereti folium Elliott (now O. filiformis) in the same genus based on the shared features of their fruit, including their strong dorsal compression and broad marginal wings. Candolle (1829) created separate genera for the rachis-leaved and compound-leaved Oxypolis species (Tiedemannia and Archemora, respectively), but Coulter & Rose (1887, 1888) united them once again arguing that “no fruit character can be made to separate them, and the only distinction would have to be drawn from the leaves”. The rachis-leaved Ptilimnium (then divided into three species) were originally placed in their own genus, Harperella Rose, but Mathias (1936) moved them into the genus Ptilimnium based on morphological similarities of the fruit, including the corkythickened marginal ribs. Easterly (1957a), in his monograph of the genus Ptilimnium, agreed with Mathias stating that “the strikingly different vegetative character” of the leaves was not sufficient to warrant generic recognition for Harperella. Traditional higher-level classifications within Apiaceae have relied heavily on fruit characters. These classifications, however, have not always held up to molecular systematic studies (Downie & Katz-Downie, 1996; Downie & al., 1996, 1998, 2001; Plunkett & al., 1996; Calviño & Downie, 2007; Magee & al., 2010a). This is the result of a complex pattern of parallelisms and convergences of fruit characters within the family (Plunkett & al., 1996; Downie & al., 1998; Lee & al., 2001; Liu & al., 2009, Magee & al., 2009b). For example, characters such as dorsal flattening and wing formation have evolved in several independent lineages of Apioideae, most likely as a dispersal mechanism (Theobald, 1971; Downie & al., 2000c; Spalik & Downie, 2001; Winter & al., 2008), and therefore cannot necessarily be taken as evidence of close relationship. This has brought the taxonomic value of fruit characters into question (Downie & al., 2001; Spalik & Downie, 2001; Calviño & al., 2006; Liu & al., 2006). Liu & al. (2006), however, are against abandoning the use of these characters altogether. They argue that “fruit anatomy, if studied carefully, can provide an excellent source of characters to test, support, and supplement findings based on molecular evidence”. In the past decade, tribe Oenantheae has been the subject of several studies utilizing DNA sequence data to examine phylogenetic relationships (Downie & al., 2004, 2008; Hardway & al., 2004; Lee & Downie, 2006; Spalik & Downie, 2006; Feist & Downie, 2008; Spalik & al., 2009). Results of these studies have shown that as many as five genera within the tribe are not monophyletic. Two of these studies examined the relationship of the rachis-leaved Ptilimnium and Oxypolis species to their compound-leaved congeners (Downie & al., 2008; Feist & Downie, 2008). The Feist & Downie (2008) study was based solely on nrDNA internal transcribed spacer (ITS) sequences. The Downie & al. (2008) study incorporated both ITS and cpDNA psbI-5′trnK sequences but took a broader look at relationships within the whole tribe Oenantheae and did not thoroughly sample Oxypolis and Ptilimnium. The results of both studies agreed that Oxypolis and Ptilimnium as currently delimited are not monophyletic and should each be split into two genera. For each genus, the suggested split corresponds to the split between rachis-leaved and compound-leaved taxa. Although these two previous studies suggested the polyphyly of both Oxypolis and Ptilimnium, nomenclatural changes were postponed until confirmation from additional data could be obtained. In this paper we present additional sequences from the chloroplast genome and combine these with ITS and cpDNA sequences from these earlier studies. We also examine leaf morphology and an additional independent source of Table 1. Current circumscription of Oxypolis and Ptilimnium and proposed new combinations for taxa included in this study. Current classification New combinations and reinstatements Oxypolis canbyi (J.M. Coult. & Rose) Fernald Tiedemannia canbyi (J.M. Coult. & Rose) Feist & S.R. Downie Oxypolis fendleri (A. Gray) A. Heller Oxypolis filiformis (Walter) Britton Tiedemannia filiformis (Walter) Feist & S.R. Downie subsp. filiformis Oxypolis greenmanii Mathias & Constance Tiedemannia filiformis subsp. greenmanii (Mathias & Constance) Feist & S.R. Downie Oxypolis occidentalis J.M. Coult. & Rose Oxypolis rigidior (L.) Raf Oxypolis ternata (Nutt.) A. Heller Ptilimnium ahlesii Weakley & G.L. Nesom Ptilimnium capillaceum (Michx.) Raf. Ptilimnium costatum (Elliott) Raf. Ptilimnium nodosum (Rose) Mathias Harperella nodosa Rose Ptilimnium nuttallii (DC.) Britton Ptilimnium texense J.M. Coult. & Rose 403 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium evidence, fruit anatomical data, to determine if they corroborate the molecular results. We then provide new circumscriptions for the genera Oxypolis and Ptilimnium based on these multiple lines of evidence. Each of the two polyphyletic genera (Oxypolis and Ptilimnium) is split, two genera (Tiedemannia and Harperella) are resurrected, and three new combinations are made. A key to these four genera is included in the Taxonomic Treatment section below. Infrageneric and infraspecific relationships are also examined and our results are compared to former taxonomic treatments. MATERIALS AND METHODS Taxon sampling and outgroup selection. — In a study of tribe Oenantheae which included several species of Oxypolis and Ptilimnium, Downie & al. (2008) found that among the five noncoding plastid DNA loci they examined, the trnQ5′rps16 and 3′rps16-5′trnK intergenic spacers (hereafter, trnQ and trnK) were the most variable. Therefore these regions were used for this study. Complete sequences of the trnQ and trnK regions from 76 accessions and complete sequences of the nrDNA ITS region from 66 accessions were used in this study (Appendix 1). Sequences were obtained for all seven species of Oxypolis and six species of Ptilimnium. Sequence data for one accession of the purported hybrid Oxypolis filiformis × O. greenmanii were also included. All Oxypolis and Ptilimnium taxa were represented by multiple accessions to assess infraspecific variation. Two species of Perideridia Rchb. were chosen as outgroups (Appendix 1). Previous phylogenetic studies which have included the Oenanthe clade or tribe Oenantheae (Plunkett & al., 1996; Downie & al., 1996, 1998, 2000b) have shown the genus Perideridia to be sister group to a clade comprising all other members of the tribe. The genera Atrema DC., Cynosciadium, Daucosma Engelm. & A. Gray ex A. Gray, Lilaeopsis, Limnosciadium, Neogoezia Hemsl., and Trepocar pus Nutt. ex DC. have allied with Ptilimnium and Oxypolis in previous studies and together make up what has become known as the North American Endemics Clade (Hardway & al., 2004; Downie & al., 2008). Accessions of these genera were included in the phylogenetic analyses to show their placements relative to Oxypolis and Ptilimnium (Appendix 1). Nomenclature follows Kartesz (2010), except Oxypolis ternata which follows Feist (2009), Atrema americanum DC. (= Bifora americana Benth. & Hook. f. ex S. Watson) which follows Hardway & al. (2004), and Lilaeopsis which follows Affolter (1985). DNA extraction, purification, and sequencing. — For this study, trnQ, trnK, and ITS sequences were newly generated for 55, 44, and 9 accessions, respectively. The remaining sequences used had already been published (Downie & KatzDownie, 1996; Downie & al., 2000a, 2008; Hardway & al., 2004; Feist & Downie, 2008). The new sequences were generated according to the following methods. Leaf material was taken from either herbarium specimens or field-collected and silica-dried samples. DNA was isolated using a DNeasy Plant Mini Kit (Qiagen, Valencia, California, U.S.A.) according to the manufacturer’s instructions. The entire ITS region (ITS-1, 404 TAXON 61 (2) • April 2012: 402–418 5.8S rDNA, ITS-2) was amplified using primers 18S-ITS1-F (Spalik & Downie, 2006) and C26A (Wen & Zimmer, 1996) or primers 18S-for (Feist & Downie, 2008) and C26A. The PCR amplification methods used are described elsewhere (Downie & al., 2000a). PCR products (templates) were purified using either a QIAquick Gel Extraction Kit or a QIAquick PCR Purification Kit (Qiagen) following manufacturer’s instructions. For cpDNA, sequences were obtained for the trnQ(UUG)-rps16 5′ exon and rps16 3′ exon-trnK (UUU) intergenic spacers using the primers trnQ, rps16-1R, rps16-2, and trnK (Downie & KatzDownie, 1996; Lee & Downie, 2006; Downie & al., 2008). The “rpl16” program of Shaw & al. (2005) was used for the cpDNA regions because it is effective across a wide range of taxa and genomic regions (Shaw & al., 2007). PCR products were checked on 1% agarose gels and then purified according to the ExoSAP protocol of Werle & al. (1994) using 5 U of Exonuclease I (New England Biolabs, Ipswich, Massachusetts, U.S.A.) and 0.5 U of Shrimp Alkaline Phosphatase (Promega, Madison, Wisconsin, U.S.A.). Sequence reactions for all sequences were carried out using an ABI Prism Big Dye Terminator v.3.1 Ready Reaction Cycle Sequencing Kit (Applied Biosystems, Foster City, California, U.S.A.). Sequence reaction products were visualized using an ABI 3730XL high-throughput DNA capillary sequencer. All newly acquired sequences used in this study have been deposited in GenBank (www.ncbi.nlm.nih.gov/ genbank/). See Appendix 1 for GenBank accession numbers. Sequence alignment and phylogenetic analysis. — Sequences were aligned using Clustal X v.1.83 (Thompson & al., 1997) and manually adjusted as necessary using the alignment editor Bioedit v.7.0.9.0 (Hall, 1999). For the trnQ and trnK datasets only, informative gaps were scored as additional binary characters according to the “simple indel coding” method of Simmons & Ochoterena (2000). Indels were not scored for the ITS dataset, as they had been used in a previous study of the group (Feist & Downie, 2008) and had not been useful. Three matrices of sequence data were constructed. The first included the aligned nucleotide data from the trnQ and trnK regions (cpDNA), the second included these data and the binary-coded indels (cpDNA/indels), and the third included the aligned nucleotide data from the trnQ, trnK, and ITS regions (cpDNA/ITS). To facilitate analysis, identical sequences were represented by single terminals, except where identical sequences were from individuals from distinct geographic areas of interest (Table 2). The cpDNA and cpDNA/indels datasets included sequences of 76 accessions (65 terminals). The cpDNA/ITS dataset included sequences of 66 accessions (63 terminals). Most of the ITS sequences used represented a subset of a larger ITS matrix of 147 accessions used in Feist & Downie (2008). Sequence characteristics were obtained for the ITS and cpDNA (trnQ/ trnK) regions. Uncorrected pairwise nucleotide distances were calculated using the distance matrix option of PAUP* v.4.0b10 (Swofford, 2003). Before combining the ITS and cpDNA datasets, the incongruence length difference test of Farris & al. (1995) was performed using the partition-homogeneity test in PAUP* to evaluate the extent of conflict between them. This test was executed with 100 replicate analyses, using the heuristic search option, simple stepwise addition of taxa, and TBR TAXON 61 (2) • April 2012: 402–418 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium Table 2. Summary of identical sequences that were represented by a single terminal accession in the phylogenetic analyses, except when these sequences were from individuals from geographic areas of interest. Accession numbers in the second column are those identifying terminals in Fig. 1 and/or Fig. 2. Accession numbers in the third column have identical sequences to those presented in the second column for the regions indicated. Voucher information for all accessions is provided in Appendix 1. Taxon Accession number Accessions with identical sequences (regions for which it is identical) Daucosma laciniata 3411 3412 (trnQ/trnK) Oxypolis canbyi 2937 2938 (trnQ/trnK) Oxypolis fendleri 2350 2351, 2369 (trnQ/trnK) Oxypolis greenmanii 2941 2717 (trnQ/trnK/ITS) Oxypolis occidentalis 3464 2929 2928 2755 3465, 3466 (trnQ/trnK) 2937 (trnQ/trnK/ITS) 2927 (trnQ/trnK) 3442 (trnQ/trnK) Ptilimnium nodosum 2784 2635 (trnQ/trnK/ITS) branch swapping; MaxTrees was set to 20,000. The aligned data matrices are available in TreeBASE (http://purl.org/phylo/ treebase/phylows/study/TB2:S12295). Maximum parsimony (MP) analyses of each of the three data matrices were implemented in PAUP*. All characters were treated as unordered and all character transformations were weighted equally. Heuristic MP searches were replicated 10,000 times with random addition of taxa and the following options in effect: MULTREES, TBR branch swapping, and gaps treated as missing data. Bootstrap (BS) analyses were done on all datasets to assess clade support. For all datasets 1000 bootstrap replicates were performed with 100 random sequence addition replicates. All BS analyses were performed with the heuristic search option, TBR branch swapping, and MULTREES options in effect. For the cpDNA and cpDNA/ indels datasets MaxTrees was set to 5000 trees per replicate; for the cpDNA/ITS dataset it was set to 20,000 trees per replicate. MrBayes v.3.1.2 (Huelsenbeck & Ronquist, 2001) was used to conduct Bayesian inference (BI) analyses of the three datasets (cpDNA, cpDNA/indels, cpDNA/ITS). Modeltest v.3.7 (Posada & Crandall, 1998) was used to select the appropriate evolutionary models of nucleotide substitution for each of the three DNA regions. The models that best fit these data, as selected by the Akaike information criteria (AIC) estimator, were used in the BI analyses. A restriction site (binary) model was implemented for the indels partition in the cpDNA/indels dataset. For each dataset, two independent runs of four chains each were conducted for 2,000,000 generations with a sample frequency of 100. Plots of generation number vs. likelihood value were inspected and stationarity was determined to have been reached after 4000 trees were sampled. For each dataset, the 4000 trees sampled prior to stationarity (the burn-in) were discarded and a majority-rule consensus tree was constructed from the remaining trees to show the posterior probability values of all observed bipartitions. The MP and BI analyses were performed on cpDNA, cpDNA/indel, and combined cpDNA/ITS datasets that both included and excluded the single accession of the hybrid Oxypo lis filiformis × O. greenmanii. This was done to assess whether the inclusion of this hybrid had any effect on the placement of other taxa in the phylogenetic trees. Fruit anatomy. — Fruits from herbarium specimens were used in an anatomical study (Appendix 2). Each of the seven Oxypolis and six Ptilimnium species were represented. In addition, accessions of the closely related Cynosciadium digitatum, Limnosciadium pinnatum, and L. pumilum were included. For each species, fruits from between one and three accessions were examined. To make transverse sections, fruits were first rehydrated and then placed in FAA for a minimum of 24 h. These samples were subsequently treated according to a modification of the method of Feder & O’Brien (1968) for embedding in glycol methacrylate (GMA). Transverse sections of about 3 µm thick were made using a Porter-Blüm ultramicrotome and stained using the periodic acid Schiff/toluidine blue (PAS/ TB) method of Feder & O’Brien (1968). To study the threedimensional structure of the vittae, fruits were placed in boiling water and left to cool and soak for at least 24 h. Thereafter, the exocarp was removed while keeping the fruit submerged in water to prevent desiccation. RESULTS ITS dataset. — The ITS matrix had an aligned sequence length of 639 positions. One hundred and four positions were excluded from further analyses due to alignment ambiguities. The number of parsimony-informative positions was 242 and the number of autapomorphic positions was 30. The maximum pairwise sequence divergence value for the ITS region across all 63 terminals was 24.23%. Considering just Oxypolis s.l. and Ptilimnium s.l., the maximum sequence divergence was 21.86% between O. canbyi (a rachis-leaved species) and O. ternata (a compound-leaved species). Phylogenetic analyses were not carried out on this dataset alone since these analyses had been 405 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium performed on a larger ITS dataset previously (Feist & Downie, 2008). Modeltest selected the SYM + I + G model of evolution to be used for the ITS region in the BI analysis of combined data. cpDNA datasets. — The cpDNA dataset had an aligned sequence length of 2813 positions. Due to alignment ambiguities 409 positions were excluded from further analyses. Of the remaining 2404 aligned positions, 280 were parsimonyinformative and 88 were autapomorphic. The cpDNA/indel dataset with 63 binary-scored alignment gaps had a total of 343 parsimony-informative characters. The maximum pairwise sequence divergence value for the cpDNA region across all 65 terminals was 5.56%. Considering only Oxypolis s.l. and Ptilimnium s.l., the maximum sequence divergence was 3.82% between P. nodosum (rachis leaves) and O. ternata (compound leaves). Modeltest selected the GTR + G and GTR + I + G models for the trnQ and trnK regions, respectively, for use in the BI analyses. MP analysis of the cpDNA dataset resulted in 23,229 trees of 473 steps each (consistency index CI = 0.854 and 0.824, with and without uninformative characters, respectively; retention index RI = 0.964). MP analysis of the cpDNA/indel dataset resulted in 4698 trees of 554 steps each (CI = 0.843 and 0.811, with and without uninformative characters, respectively; RI = 0.963). The addition of the binary-scored indels into the cpDNA matrix had only a minimal effect on the resulting tree topologies and support values. The only significant change was that all accessions of P. costatum were resolved as monophyletic (BS = 79) in the cpDNA/indels MP strict consensus tree (Fig. 1) but not in the cpDNA MP strict consensus tree (not shown), and support for this group was increased from PP = 0.51 in the cpDNA BI majority-rule consensus tree (not shown) to PP = 0.89 in the cpDNA/indels BI majority-rule consensus tree (not shown). For this reason, only the results of cpDNA/indels analyses are discussed hereafter. The cpDNA/indels MP tree and cpDNA/indels BI tree were congruent except that in the BI tree, all accessions of Oxypolis rigidior formed a weakly supported monophyletic group (PP = 0.54), whereas in the cpDNA/indels MP tree, they did not. Other species not resolved as monophyletic in the cpDNA/indels MP and BI trees were P. ahlesii, P. capil laceum, O. filiformis, O. rigidior, and O. ternata. Ptilimnium ahlesii formed a strongly supported monophyletic group with P. capillaceum in the cpDNA/indels analyses (BS = 100, PP = 1.00), as did O. ternata with O. rigidior (BS = 96, PP = 1.00). One accession of O. filiformis came out as sister group to the clade which includes all accessions of O. canbyi (BS = 69, PP = 0.92), while the other accession of O. filiformis formed a weakly supported clade with O. greenmanii and their purported hybrid O. filiformis × O. greenmanii (BS = 64, PP = 0.79). These relationships did not change when O. filiformis × O. greenmanii was excluded from the MP and BI analyses. Oxypolis occidentalis is divided into two strongly to moderately supported clades (Fig. 1). The two major clades of O. occidentalis consist of populations that are separated geographically and can be referred to as the North clade and the South clade. The North clade (BS = 89, PP = 1.00) consists 406 TAXON 61 (2) • April 2012: 402–418 of individuals from populations from Haida Gwaii (formerly the Queen Charlotte Islands) in British Columbia, the Cascade Mountains in Oregon, and the Cascade Range Foothills in northern California. The South clade (BS = 71, PP = 0.98) consists of individuals from populations in the Sierra Nevada, San Bernardino, White, and Inyo mountain ranges in California. At the generic level, the MP and BI trees of the cpDNA/ indels data matrix were completely congruent. Ptilimnium s.l. and Oxypolis s.l. were each not monophyletic. Ptilimnium s.l. was split between two clades: Clade 1 (Ptilimnium) comprised of the compound-leaved species of Ptilimnium (i.e., P. ahlesii, P. capillaceum, P. costatum, P. nuttallii, and P. texense) and Clade 2 (Harperella) containing the single rachis-leaved species of Ptilimnium (i.e., P. nodosum). Oxypolis s.l. was also split between two clades: Clade 3 (Tiedemannia) comprised of the rachis-leaved species of Oxypolis (i.e., O. canbyi, O. fili formis, and O. greenmanii) and Clade 4 (Oxypolis) containing the compound-leaved species of Oxypolis (i.e., O. fendleri, O. occidentalis, O. rigidior, and O. ternata). Clade 3 (Tiede mannia) is sister to Clade 1 (Ptilimnium) and quite distant in the tree from Clade 4 (Oxypolis). Clades 1, 2, and 3 were each strongly supported, whereas Clade 4 (Oxypolis) had moderate support (BS = 83, PP = 0.99). In previous analyses based on ITS sequences (Feist & Downie, 2008), the evolutionary relationships among the following five groups were not resolved: Clade 1 (Ptilimnium), Clade 2 (Harperella), Clade 3 (Tiedemannia), Limnosciadium, and Cynosciadium. These five groups formed what was essentially a polytomy (fig. 1 in Feist & Downie, 2008). The results of the cpDNA/indels analyses presented herein, however, strongly support the sister relationship between Clade 1 (Ptilimnium) and Clade 3 (Tiedemannia) (BS = 98, PP = 1.00). Together, these two groups form a strongly supported sister relationship with Clade 2 (Harperella) (BS = 86, PP = 0.99), followed by the successively basal sister groups Limnosciadium (BS = 66, PP = 0.85) and Cynosciadium (BS = 100, PP = 1.00). Two new accessions of Daucosma laciniata Engelm. & A. Gray were used in this study. As a result, the placement of Daucosma differs from that found in previous studies where the accession D. laciniata 2397, a misidentified specimen of Lim nosciadium pinnatum (DC.) Mathias & Constance, was used (Hardway & al., 2004; Downie & al., 2008; Feist & Downie, 2008). In this study, the two accessions of D. laciniata are monophyletic and are sister to a clade formed by Atrema ameri canum and Trepocarpus aethusae Nutt. (BS = 87, PP = 1.00). Combined dataset (cpDNA/ITS). — The results of the partition homogeneity test for 63 terminals common to both the cpDNA and ITS datasets revealed that these loci do not yield significantly different phylogenetic estimates (ILD probability value = 0.21). Therefore, these datasets were combined for simultaneous analysis. The combined cpDNA (without binary-scored indels) and ITS data matrix (cpDNA/ITS) had an aligned sequence length of 3452 positions. Due to alignment ambiguities 513 positions were excluded from further analyses. Of the remaining 2939 aligned positions, 522 were parsimony-informative and 118 were autapomorphic. Adding the cpDNA trnQ and trnK sequences to the ITS matrix more TAXON 61 (2) • April 2012: 402–418 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium than doubled the number of parsimony-informative characters from the previous study by Feist & Downie (2008). The maximum pairwise sequence divergence value for the combined cpDNA/ITS dataset across all 63 terminals was 9.74% between Oxypolis canbyi and Daucosma laciniata. Considering just Oxypolis s.l. and Ptilimnium s.l., maximum sequence divergence was 7.62% between O. canbyi (a rachis-leaved species) and 64/0.99 100/1.00 63/0.87 87/1.00 79/0.89 100/1.00 62/0.99 100/1.00 98/1.00 98/1.00 100/1.00 75/0.97 69/0.92 86/0.99 100/1.00 64/0.79 65/0.96 66/0.85 100/1.00 100/1.00 64/0.89 100/1.00 80/0.99 100/1.00 99/1.00 100/1.00 100/1.00 99/1.00 100/1.00 87/1.00 100/1.00 100/1.00 63/0.99 100/1.00 89/1.00 98/1.00 71/0.98 63/0.76 83/0.99 96/1.00 100/1.00 100/1.00 O. ternata (a compound-leaved species). Maximum sequence divergence among compound-leaved Oxypolis species was 2.40%, whereas among rachis-leaved Oxypolis species it was 2.99%. Maximum sequence divergence among Ptilimnium s.l. species was 6.22%; however, it was just 1.43% among compound-leaved Ptilimnium species and just 0.19% among the rachis-leaved Ptili mnium. All species of Ptilimnium s.l. and Oxypolis s.l. showed Ptilimnium ahlesii 2648 Ptilimnium capillaceum 2701 Ptilimnium ahlesii 2969 Ptilimnium capillaceum 2703 Ptilimnium costatum 1646 Ptilimnium costatum 1970 Ptilimnium costatum 2402 Ptilimnium costatum 2707 Ptilimnium nuttallii 2165 Ptilimnium nuttallii 2617 Ptilimnium nuttallii 2623 Ptilimnium texense 1981 Ptilimnium texense 2905 Oxypolis canbyi 2744 Oxypolis canbyi 2747 Oxypolis canbyi 2751 Oxypolis canbyi 2937 (2) Oxypolis ﬁliformis 2713 Oxypolis ﬁliformis 2371 Oxypolis greenmanii 2941 (2) Oxypolis ﬁliformis × O. greenmanii 2714 Ptilimnium nodosum 2787 Ptilimnium nodosum 2900 Ptilimnium nodosum 2931 Ptilimnium nodosum 2934 Ptilimnium nodosum 2902 Ptilimnium nodosum 2930 Ptilimnium nodosum 2936 Limnosciadium pinnatum 1511 Limnosciadium pinnatum 2395 Limnosciadium pumilum 3742 Cynosciadium digitatum 1571 Cynosciadium digitatum 1986 Lilaeopsis carolinensis 2148 Lilaeopsis mauritiana 2150 Lilaeopsis novae-zelandiae 2152 Lilaeopsis occidentalis 1999 Daucosma laciniata 2912 Daucosma laciniata 3411 (2) Atrema americanum 1160 Trepocarpus aethusae 1660 Neogoezia gracilipes 2770 Neogoezia minor 2774 Oxypolis occidentalis 2755 (2) Oxypolis occidentalis 2756 Oxypolis occidentalis 2928 (2) Oxypolis occidentalis 3413 Oxypolis occidentalis 3417 North Oxypolis occidentalis 2899 clade Oxypolis occidentalis 3528 Oxypolis occidentalis 1142 Oxypolis occidentalis 1153 Oxypolis occidentalis 3376 South Oxypolis occidentalis 3464 (3) clade Oxypolis occidentalis 3532 Oxypolis ternata 2940 Oxypolis ternata 2738 Oxypolis ternata 2735 Oxypolis rigidior 2003 Oxypolis rigidior 1998 Oxypolis rigidior 1927 Oxypolis fendleri 2350 (3) Oxypolis fendleri 2368 Perideridia americana 2033 Perideridia kelloggii 778 Clade 1 Ptilimnium (compound leaves) Clade 3 Tiedemannia (rachis leaves) Clade 2 Harperella (rachis leaves) Clade 4 Oxypolis (compound leaves) Fig. 1. Strict consensus tree of 4698 minimal-length 554-step trees obtained from the MP analysis of the cpDNA/indels dataset (CI = 0.843 and 0.811, with and without uninformative characters, respectively; RI = 0.963). Numbers on branches represent bootstrap estimates and Bayesian posterior probability values, respectively. Numbers in parentheses following the name of a taxon indicate the number of accessions of that taxon having identical DNA sequences (Table 2). 407 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium intraspecific variation, except for P. ahlesii and P. capillaceum which had sequences that were identical to one another. MP analysis resulted in 32 trees of 1184 steps each (CI = 0.689 and 0.634, with and without uninformative characters, respectively; RI = 0.919). The BI majority-rule consensus tree is presented in Fig. 2 with branch lengths. The topology of the BI tree is consistent with that of the MP strict consensus tree TAXON 61 (2) • April 2012: 402–418 (not shown) except at the nodes where an asterisk (*) is given to indicate that the BS value was less than 50%. As in the cpDNA/indels analyses, Ptilimnium ahlesii and P. capillaceum are each not monophyletic, but together they form a strongly supported monophyletic group (BS = 100, PP = 1.00). Oxypolis filiformis is also not monophyletic, but it forms a monophyletic group with O. greenmanii and their 64/.99 Ptilimnium ahlesii 2648 Ptilimnium capillaceum 2701 Ptilimnium ahlesii 2969 Ptilimnium capillaceum 2703 */0.60 86/1.00 Ptilimnium costatum 1646 94/1.00 Ptilimnium costatum 1970 Ptilimnium costatum 2402 100/1.00 Ptilimnium texense 1981 100/1.00 Ptilimnium texense 2905 Ptilimnium nuttallii 2165 96/1.00 100/1.00 Ptilimnium nuttallii 2623 96/1.00 Oxypolis canbyi 2744 100/1.00 Oxypolis canbyi 2747 Oxypolis canbyi 2937 65/0.98 100/1.00 Oxypolis canbyi 2938 Oxypolis ﬁliformis 2371 100/1.00 Oxypolis ﬁliformis 2713 80/1.00 Oxypolis greenmanii 2941 (2) 74/0.99 Oxypolis ﬁliformis × O. greenmanii 2714 Ptilimnium nodosum 2784 (2) Ptilimnium nodosum 2787 */0.90 Ptilimnium nodosum 2934 */0.94 Ptilimnium nodosum 2936 */88 Ptilimnium nodosum 2900 100/1.00 Ptilimnium nodosum 2931 Ptilimnium nodosum 2902 Ptilimnium nodosum 2930 100/1.00 100/1.00 Limnosciadium pinnatum 1511 100/1.00 Limnosciadium pinnatum 2395 Limnosciadium pumilum 3742 100/1.00 100/1.00 Cynosciadium digitatum 1571 Cynosciadium digitatum 1986 99/1.00 Lilaeopsis carolinensis 2148 100/1.00 Lilaeopsis mauritiana 2150 100/1.00 Lilaeopsis novae-zelandiae 2152 100/1.00 Lilaeopsis occidentalis 1999 100/1.00 Daucosma laciniata 3411 96/1.00 Daucosma laciniata 3412 95/1.00 Atrema americanum 1160 Trepocarpus aethusae 1660 100/1.00 Neogoezia gracilipes 2770 Neogoezia minor 2774 Oxypolis occidentalis 2755 84/0.99 Oxypolis occidentalis 2756 */0.90 Oxypolis occidentalis 2928 100/1.00 Oxypolis occidentalis 2929 (2) North clade 100/1.00 Oxypolis occidentalis 3417 Oxypolis occidentalis 2899 Oxypolis occidentalis 3528 97/1.00 Oxypolis occidentalis 1153 Oxypolis occidentalis 3376 Clade 4 South clade 100/1.00 Oxypolis occidentalis 3466 Oxypolis 73/0.77 (compound Oxypolis occidentalis 3532 88/0.99 Oxypolis ternata 2735 100/1.00 Oxypolis ternata 2738 Oxypolis ternata 2940 99/1.00 Oxypolis rigidior 1927 98/1.00 99/1.00 Oxypolis rigidior 1998 Oxypolis fendleri 2350 100/1.00 Oxypolis fendleri 2351 Oxypolis fendleri 2369 Perideridia americana 2033 Perideridia kelloggii 778 100/1.00 Clade 1 Ptilimnium (compound leaves) Clade 3 Tiedemannia (rachis leaves) Clade 2 Harperella (rachis leaves) leaves) 0.3 Fig. 2. Majority-rule consensus tree obtained from the BI analysis of the cpDNA/ITS dataset. Numbers on branches represent bootstrap estimates and Bayesian posterior probability values, respectively; a bootstrap estimate of less than 50% is indicated with an asterisk (*). Numbers in parentheses following the name of a taxon indicate the number of accessions of that taxon having identical DNA sequences (Table 2). 408 TAXON 61 (2) • April 2012: 402–418 hybrid, O. filiformis × O. greenmanii (BS = 100, PP = 1.00). When O. filiformis × O. greenmanii is excluded from the MP and BI analyses, the relationships among the remaining taxa do not change. Both the North and South clades of O. occidentalis are again apparent and strongly supported (BS = 100, PP = 1.00, for both clades). Ptilimnium costatum is resolved as monophyletic (BS = 94, PP = 1.00) as it was in the cpDNA/indel strict consensus tree. In contrast to the results of the cpDNA/ indel analyses, O. ternata and O. rigidior are each resolved as monophyletic with strong support (BS = 100, PP = 1.00; BS = 99, PP = 1.00, respectively). At the generic level, Oxypolis s.l. and Ptilimnium s.l. are again each shown not to be monophyletic. The sister relationship of Clade 1 (Ptilimnium) and Clade 3 (Tiedemannia) continues to be strongly supported (BS = 96, PP = 1.00), while the relationships of Clade 2 (Harperella), Limnosciadium and Cynoscia dium to these groups and to each other are less well-supported. Clade 4 (Oxypolis) is strongly supported (BS = 99, PP = 1.00). As in the cpDNA/indels tree (Fig. 1), Daucosma laciniata is monophyletic and is sister to a clade formed by Atrema ameri canum and Trepocarpus aethusae (BS = 96, PP = 1.00). Fruit anatomy. — Based on fruit anatomy, the species previously recognized within Oxypolis s.l. and Ptilimnium s.l. can be separated into four groups corresponding to Clades 1–4 recovered in the molecular analyses. All species have homomericarpic fruits with a very broad commissure that extends over the full width of the mericarp. The mericarps are slightly to prominently dorsally compressed with narrow to broadly winged marginal ribs, except those of P. nodosum which are prominently isodiametric and with the marginal ribs not winged (Fig. 3A). The compound-leaved species of Ptilim nium (i.e., P. ahlesii, P. capillaceum, P. costatum, P. nuttallii, and P. texense) have fruits with slightly dorsally compressed mericarps and thick, narrowly winged marginal ribs that extend only slightly beyond the marginal vascular bundles (Fig. 3B–F). The fruits of the species of Ptilimnium s.l. (Fig. 3A–F) are distinguished from those of Oxypolis s.l. (Fig. 3G–L; O. ternata not shown), as well as Cynosciadium (Fig. 3M) and Limno sciadium (Fig. 3N–O), by the presence of prominent square or somewhat elongated cells external to the vittae (not shown), a character also reported for Dasispermum Raf. (Magee & al., 2009a, 2010b). The fruits of the species of Oxypolis s.l. are distinguished from the fruits of the species of Ptilimnium s.l., Cynosciadium, and Limnosciadium by their very broad, thin marginal wings and usually smaller, less lignified vascular bundles. While most of the species studied have a lignified layer of mesocarp cells surrounding the endocarp, this character is conspicuously absent in the compound-leaved species of Oxypolis (i.e., O. fendleri, O. occidentalis, O. rigidior, and O. ternata; Fig. 3J–L; O. ternata not shown). This latter group is distinguished furthermore from all other species studied by the presence of four to eight, often branching commissural vittae (Figs. 3J–L, 4A–C). The other species studied all have two commissural vittae which are never branching (Fig. 3A–I, M–O, 4D). The compound-leaved Oxypolis species are distinguished additionally from the rachis-leaved species of Oxypolis Feist & al. • Generic delimitations for Oxypolis and Ptilimnium (i.e., O. canbyi, O. filiformis, and O. greenmanii) in that the vittae are smaller than or equal in size to the vascular bundles (Fig. 3G–I). In the rachis-leaved Oxypolis species the vittae are distinctly larger than the vascular bundles (Fig. 3J–L). The fruits of the closely related genera Limnosciadium (Fig. 3N–O) and Cynosciadium (Fig. 3M) can be distinguished from both Oxypolis s.l. and Ptilimnium s.l. by the presence of a lignified commissural keel, and furthermore in Limnosciadium by the sclerification of the mesocarp between the vascular bundles so that they appear continuous. Oxypolis canbyi (a rachis-leaved species) has a unique wing type not found in any of the other winged species examined in this study. In O. canbyi, the marginal wing is formed through the expansion of the mesocarp between the vascular bundle and the endocarp so that the vascular bundle is located near the wing tip to form a pseudo-marginal wing (Fig. 3I). Also, as mentioned by Tucker & al. (1983), a sclerified band of mesocarp cells is formed between the vascular bundle and the endocarp. In the other winged species examined, the wing is formed through the expansion of the mesocarp beyond the vascular bundle so that the vascular bundle is located at the base of the wing (Fig. 3B–H, J–O) and a sclerified band of mesocarp cells is not present between the vascular bundle and the endocarp. In the compound-leaved Ptilimnium, Cynosciadium, and Limnosciadium species, the true marginal wing remains narrow and extends slightly beyond the marginal vascular bundle (Fig. 3B–F, M–O). In the compound-leaved Oxypolis, O. fili formis, and O. greenmanii, the true marginal wing extends significantly beyond the vascular bundle (Fig. 3G, H, J–M). DISCUSSION Ptilimnium/Harperella. — Morphological and molecular results from this study confirm what was proposed in previous studies (Downie & al., 2008; Feist & Downie, 2008), that the genus Ptilimnium is not monophyletic. Differences in leaf morphology, fruit anatomy, and DNA sequence data, as well as reproductive strategy and chromosome number, support removing P. nodosum from the genus. Ptilimnium nodosum has rachis leaves whereas the other members of the genus have pinnately decompound leaves that are finely dissected. The fruits of P. nodosum are isodiametric and prominently five-ribbed but not winged, whereas the fruits of the compound-leaved Ptilimnium species are slightly dorsally compressed and have thick, narrowly winged marginal ribs that extend beyond the vascular bundles. In addition, P. nodosum can proliferate extensively through vegetative reproduction. Vegetative shoots are produced at the nodes of decumbent flowering stems and develop into individual plantlets when the flowering stalks die back in winter (Marcinko & Randall, 2008). This method of reproduction has not been observed in any of the compoundleaved Ptilimnium species. Furthermore, the chromosome number for P. nodosum is n = 6, whereas chromosome numbers for the compound-leaved species are n = 16 (P. costatum), n = 7 (P. capillaceum, P. nuttallii), and n = 8 or 14 (P. capillaceum) (Easterly, 1957a; Constance & al., 1976; Weakley & Nesom, 409 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium TAXON 61 (2) • April 2012: 402–418 Fig. 3. Transverse sections of fruits of Ptilimnium s.l. (A–F), Oxypolis s.l. (G–L), Cynosciadium (M), and Limnosciadium (N–O). A, Ptilimnium nodosum (Feist & MolanoFlores 2967.1, ILLS); B, Ptilimnium ahlesii (Bozeman 6100, NCU); C, Ptilimnium capillaceum (Valentine s.n., BRITSMU); D, Ptilimnium costatum (Feist s.n., ILLS); E, Ptilimnium texense (Shinners 11830, BRIT-SMU); F, Ptilimnium nuttallii (Cory 53275, BRIT-SMU); G, Oxypolis filiformis (Feist & MolanoFlores 3197, ILLS); H, Oxypolis greenmanii (Godfrey 53756, NCSC); I, Oxypolis canbyi (Nelson 4269, USCH); J, Oxypolis fendleri (Sturges 205, RM); K, Oxypolis occidentalis (Feist & MolanoFlores 4106, ILLS); L, Oxypolis rigid ior (Webster & Webster 7206, DUKE); M, Cynosciadium digitatum (Sundell 15406, BRIT); N, Limnosciadium pumilum (Gentry 1996, BRIT); O, Limnosciadium pinnatum (Lundell 14012, LL). — Abbreviations: aod, additional oil duct; cv, commissural vitta; lck, lignified commissural keel; mr, marginal rib; pmw, pseudo-marginal wing; rod, rib oil duct; sm, sclerified mesocarp; vb, vascular bundle; vv, vallecular vitta. — Scale: A = 500 µm; B–O = 2 mm. — Additional voucher information is provided in Appendix 2. 410 TAXON 61 (2) • April 2012: 402–418 2004). Maximum pairwise sequence divergence is much greater between P. nodosum and the compound-leaved Ptili mnium species (6.22%) than it is among all compound-leaved species of Ptilimnium (1.43%), or among the nine accessions of P. nodosum (0.19%) included in the study. In addition, the compound-leaved Ptilimnium are more closely related to the rachis-leaved Oxypolis, with which they form a strongly supported sister group, than they are to P. nodosum. The type of the genus Ptilimnium is P. capillaceum. Ptili mnium ahlesii, P. costatum, P. nuttallii, and P. texense form a monophyletic group with P. capillaceum in the phylogenetic analyses and share a common fruit and leaf structure. These taxa should therefore remain together in the genus Ptilimnium. Prior to 1936 (Mathias, 1936), P. nodosum s.str. was recognized as belonging to the genus Harperella Rose. Three species of Harperella were recognized (H. nodosa Rose, H. fluvia tilis Rose, H. vivipara Rose). Easterly (1957b) synonymized P. viviparum (Rose) Mathias with P. fluviatile (Rose) Mathias based on their shared habitat type, phenology, and lack of clear morphological differences. Ptilimnium nodosum s.str. was still recognized as a separate species due to its unique habitat, phenology, and perceived larger size. Also, it was not thought to be able to proliferate through asexual reproduction, rooting at the nodes and producing individual plantlets, as P. fluviatile was known to do. Kral (1981) conducted a morphological study comparing P. fluviatile and P. nodosum s.str. Finding no consistent morphological differences he further grouped the taxa and included P. fluviatile within P. nodosum s.l. Furthermore, the ability to reproduce asexually, as described above, has been observed in populations of P. nodosum s.str. (Feist, pers. obs.) and so cannot be used to differentiate these taxa. Results from the ITS analyses of Feist & Downie (2008) showed some geographic separation of populations that conformed to the previous delimitations of P. nodosum as three species, however, the current analyses based on trnQ and trnK data and combined cpDNA and ITS data do not support these relationships. As we have found no consistent molecular or morphological evidence to suggest otherwise, the authors agree with the assessment of Kral (1981) that these three taxa should be recognized as one species. We propose the reinstatement of the genus Harperella Rose with the single species H. nodosa Rose, including H. flu viatilis and H. vivipara as taxonomic synonyms. The results of this study support the recent reinstatement of the species Ptilimnium texense (Feist, 2010). In the combined cpDNA/ITS analyses, P. texense and P. costatum are sister to one another but form separate strongly supported clades. The separation of P. ahlesii from P. capillaceum (Weakley & Nesom, 2004), however, is not supported. In this study and in the two previous studies where these two taxa have been included (Downie & al., 2008; Feist & Downie, 2008), P. ahlesii and P. capillaceum always form a strongly supported monophyletic group, but the two taxa are never separated. DNA sequences for both accessions of P. ahlesii used in this study were identical to sequences of accessions of P. capillaceum. This alone, however, is insufficient evidence for not recognizing P. ahlesii as a distinct species. While writing the treatment of the genus for the Flora of North America (Feist, unpub. data), Feist & al. • Generic delimitations for Oxypolis and Ptilimnium the morphological characters used to separate the two taxa were found to be inconsistent when specimens from across the geographic range of P. capillaceum were examined. Therefore, a morphological study is being conducted by the first author to determine the true status of P. ahlesii. Oxypolis/Tiedemannia. — Morphological and molecular results from this study confirm what was proposed in previous studies (Downie & al., 2008; Feist & Downie, 2008), that the genus Oxypolis is not monophyletic. As with Ptilimnium, differences in leaf morphology, fruit anatomy, and DNA sequence data support splitting the genus Oxypolis into two genera. Oxypolis canbyi, O. filiformis, and O. greenmanii have rachis leaves whereas the other members of the genus have pinnately or palmately compound leaves. Although the fruits of the compound-leaved and rachis-leaved Oxypolis species are Fig. 4. Three-dimensional structure of the commissural vittae in the fruit of Oxypolis s.l. A, Oxypolis rigidior (Oldham 6994, CAN); B, Oxypolis occidentalis (Feist & MolanoFlores 4106, ILLS); C, Oxy polis fendleri (Sturges 205, RM); D, Oxypolis filiformis (Feist & MolanoFlores 3197, ILLS). — Arrows indicate commissural vittae. Scale: A–D = 1.4 mm. 411 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium superficially similar in that they are all prominently dorsally compressed and have broadly winged marginal ribs, they can readily be distinguished anatomically. The mesocarp is lignified around the seed of the rachis-leaved Oxypolis species, but not of the compound-leaved species. The vittae are larger than the vascular bundles in the rachis-leaved taxa, but smaller than or equal to the vascular bundles in the compound-leaved species. The rachis-leaved Oxypolis taxa have four vallecular vittae and two commissural vittae, whereas the compoundleaved species have four vallecular vittae and four to eight commissural vittae, the latter often branching. Maximum pairwise sequence divergence is much greater between the rachisleaved and compound-leaved Oxypolis species (7.62%) than it is among all compound-leaved species of Oxypolis (2.40%), or among all rachis-leaved species of Oxypolis (2.99%). The clade formed by the rachis-leaved Oxypolis is quite distant from the compound-leaved Oxypolis clade. In fact, the rachisleaved Oxypolis species are more closely related to all the other genera within the North American Endemics Clade (i.e., Ptili mnium, Limnosciadium, Cynosciadium, Lilaeopsis, Atrema, Trepocarpus, Daucosma, and Neogoezia) than they are to the compound-leaved Oxypolis species. The chromosome numbers for the rachis-leaved Oxypolis taxa are all n = 14, whereas the chromosome numbers for the compound-leaved species are n = 16 or n = 18 (Bell & Constance, 1957, 1960; Crawford & Hartman, 1972; Tucker & al., 1983; Pimenov & al., 2003). The type of Oxypolis is O. rigidior. Oxypolis rigidior forms a monophyletic group with the other compound-leaved Oxy polis. We propose to split the genus Oxypolis into two genera conforming to the compound-leaved and rachis-leaved clades. In 1829 Candolle created the genus Tiedemannia (Candolle, 1829). He believed that the plant placed in the genus Oenanthe by Walter and Persoon (i.e., Oenanthe filiformis Walter, 1788; Oenanthe carolinensis Persoon, 1805) and Sium by Elliott (i.e., Sium teretifolium Elliott, 1817), all homotypic synonyms of O. filiformis, belonged in its own genus (i.e., Tiedemannia) based on the uniqueness of its fruit and its reduced leaf morphology. We agree with this assessment and propose the reinstatement of the genus Tiedemannia to accommodate the three rachis-leaved taxa, O. canbyi, O. filiformis, and O. greenmanii. Oxypolis filiformis is a widespread species, abundant in Florida and common in the Southeastern U.S.A. and occurs in a variety of different wetland habitats. Oxypolis greenmanii has a more restricted range, occurring in just a few counties in Florida, and seems to have a preference for Hypericum bogs, although it can be found in flatwoods, swamps, marshes, and roadside ditches as well. Although at first glance O. greenmanii appears to be strikingly different from O. filiformis, with its larger stature and maroon-colored flowers and fruits, Judd (1982) found that standard populations of O. greenmanii and O. filiformis are connected by an extensive series of intermediate populations that completely bridge the morphological gap between the two taxa. Judd (1982) hypothesized that these populations resulted from gradual geographic intergradation. In addition to these intermediate populations, populations that appear to be the result of recent hybridization also exist. These populations are not uniformly intermediate but are rather 412 TAXON 61 (2) • April 2012: 402–418 highly variable being composed of standard O. greenmanii and O. filiformis individuals, as well as a complete range of intermediate plants. Most plants within both the intermediate and highly variable populations were found to be highly fertile (ca. 90% to nearly all pollen grains stained with cotton-blue in lacto-phenol; Judd, 1982). Oxypolis filiformis and O. green manii also share the same chromosome number and flowering phenology (Judd, 1982). Furthermore, DNA sequence data provide no support for recognizing the two as distinct species. Maximum pairwise sequence divergence between O. filifor mis and O. canbyi is 2.99%, whereas maximum pairwise sequence divergence between O. filiformis and O. greenmanii is just 0.11%. Therefore, we accept the conclusion of Judd (1982) that Oxypolis greenmanii is a subspecies of O. filiformis (i.e., O. filiformis subsp. greenmanii (Mathias & Constance) Judd). Finally, the two major clades of Oxypolis occidentalis that were seen in previous analyses based on ITS sequence data (Feist & Downie, 2008) are again recovered in all analyses in this study. The cpDNA sequence data presented herein support the findings of the ITS data alone. A North clade and a South clade of O. occidentalis are recovered and strongly supported. Our results suggest that populations of O. occiden talis from Haida Gwaii, the Cascade Mountains of Oregon, and the Cascade Range Foothills in northern California are quite different from those of the Sierra Nevada and the other more southern mountain ranges of California (South clade). The maximum sequence divergence among the North clade is 0.24% and among the South clade 0.32%, whereas the maximum sequence divergence between the North clade and the South clade is 1.49%. This is greater than the maximum pairwise sequence divergence between O. rigidior and O. ternata (1.31%). The populations represented by the South clade are within a geographic region known as the California Floristic Province, which harbors more endemic plant and animal taxa and more identifiable subspecies than any other area of comparable size in North America (Calsbeek & al., 2003). Populations from this area could represent a new taxon and another example of a California Floristic Province endemic. There is a major disjunction between the populations of Oxypolis occidentalis that are represented in the North clade. The northernmost populations of O. occidentalis in Oregon are approximately 1400 km from the populations on Haida Gwaii (Cheney & Marr, 2007). Haida Gwaii, located approximately 80 km off the west coast of British Columbia, is an intriguing geographic area. It has been proposed that during the last glacial maximum (ca. 15,000 years ago), when glaciers covered much of present-day British Columbia and extended into present-day Washington State, much of Haida Gwaii remained ice-free (Heusser, 1960; Lacourse & al., 2005) and provided a refugium to plants and animals living there. The glacial refugium hypothesis may explain why O. occidentalis occurs on these islands and their highly disjunct geographic distribution. The distribution of this species might have once extended from Oregon to British Columbia with intervening populations being wiped out by glaciers. The morphology and phylogeography of O. occidentalis is currently being studied to confirm the taxonomic status of these plants. TAXON 61 (2) • April 2012: 402–418 TAXONOMIC TREATMENT Key to the genera 1. Leaves simple; blades reduced to the rachis (thus appearing linear, terete, hollow, and septate) ......................... 2 1. Leaves 1-pinnate, 1-ternate, or pinnately decompound; blades with well-developed lamina ....................... 3 2. Mericarps isodiametric, 1–2 × 1–2 mm; marginal ribs not winged; carpophore 2-fid at apex; plants without a rhizome or caudex ...................................... 1. Harperella 2. Mericarps strongly dorsally compressed, 4–9 × 3–5.5 mm; marginal ribs broadly winged; carpophore 2-cleft nearly to the base; plants with a rhizome or short caudex ....... ............................................... 2. Tiedemannia 3. Mericarps strongly dorsally compressed; marginal ribs broadly winged, thin, not corky; carpophore 2-cleft nearly to the base; leaves 1-pinnate or 1-ternate; ultimate leaf segments linear, lanceolate, ovate, or orbiculate; roots tuberous-thickened ............................. 3. Oxypolis 3. Mericarps slightly dorsally compressed; marginal ribs narrowly winged, thick, corky; carpophore 2-fid at the apex; leaves pinnately decompound; ultimate leaf segments filiform; roots fibrous........................... 4. Ptilimnium 1. Harperella Rose in Proc. Biol. Soc. Washington 19: 96. 1906 ≡ Harperia Rose in Proc. U.S. Natl. Mus. 29: 441. 1905, nom. illeg. non Fitzg. (1904) – Type: Harperella nodosa Rose. 1.1 Harperella nodosa Rose in Proc. Biol. Soc. Washington 19: 96. 1906 ≡ Harperia nodosa Rose in Proc. U.S. Natl. Mus. 29: 441. 1905, nom. illeg. ≡ Carum nodosum (Rose) Koso-Pol. in Bull. Soc. Imp. Naturalistes Moscou, n.s., 29: 199. 1916 ≡ Ptilimnium nodosum (Rose) Mathias in Brittonia 2: 244. 1936 – Type: U.S.A., Georgia, Schley County, 10 July 1902, Harper 1411 (US!, holotype; BM!, E!, MO, 2 sheets!, NY, 2 sheets!, US!, isotypes). = Harperella fluviatilis Rose in Contr. U.S. Natl. Herb. 13: 290. 1911 ≡ Ptilimnium fluviatile (Rose) Mathias in Brittonia 2: 244. 1936 – Type: U.S.A., Alabama, Dekalb County, 24 Nov. 1905, Harper 8 (US!, holotype; MO!, isotype). = Harperella vivipara Rose in Contr. U.S. Natl. Herb. 13: 290. 1911 ≡ Carum viviparum (Rose) Koso-Pol. in Bull. Soc. Imp. Naturalistes Moscou, n.s., 29: 199. 1916 ≡ Ptilimnium vivipa rum (Rose) Mathias in Brittonia 2: 244. 1936 ≡ Ptilimnium fluviatile var. viviparum (Rose) Reveal & C.R. Broome in Castanea 46(1): 67. 1981 – Type: U.S.A., Maryland, on the banks of the Potomac River near Hancock, 5 Oct. 1910, Rose s.n. (US!, holotype; NY, 2 sheets!, isotypes). The genus name Harperia Rose was found to be a later homonym for Harperia Fitzg. (1904) and, therefore, Rose replaced it with the name Harperella. Rose designated Har perella nodosa Rose as the type for this genus. Harperella nodosa, H. fluviatilis, and H. vivipara were transferred to the genus Ptilimnium by Mathias (1936). Easterly (1957a) synonmized P. viviparum under P. fluviatile and Kral (1981) later synonomized P. fluviatile under P. nodosum. Morphological Feist & al. • Generic delimitations for Oxypolis and Ptilimnium and molecular evidence shows that P. nodosum (Rose) Mathias does not belong in the genus Ptilimnium but rather should be placed in its own genus. Therefore the genus Harperella and the species name H. nodosa are reinstated and P. nodosum is synonymized under H. nodosa. 2. Tiedemannia DC., Coll. Mém. 5: 51. 1829 – Type: T. filifor mis (Walter) Feist & S.R. Downie Candolle (1829) created the genus Tiedemannia because he felt that the plant placed in Oenanthe by Walter and Persoon (i.e., Oenanthe filiformis Walter, 1788; Oenanthe carolinensis Persoon, 1805) and in Sium by Elliott (i.e., Sium teretifolium Elliott, 1817) did not belong in either of these genera based on the uniqueness of its fruit and reduced leaf morphology. The names Oenanthe carolinensis and Sium teretifolium are both homotypic synonyms of Oenanthe filiformis, and due to the rules of priority (Art. 11.2 of the International Code of Botani cal Nomenclature, McNeill & al., 2006) are illegitimate names. Candolle made the new combination Tiedemannia teretifolia based on an illegitimate name and, therefore, his name is also illegitimate. The correct combination is Tiedemannia filifor mis. Because Candolle indicated that the name Tiedemannia teretifolia is a synonym of Oenanthe filiformis, this latter name represents the type species of Tiedemannia. This is according to Art. 7.5 of the ICBN (McNeill & al., 2006) which states “a name that is illegitimate under Art. 52 is typified either by the type of the name that ought to have been adopted under the rules (automatic typification), or by a different type designated or definitely indicated by the author of the illegitimate name”. 2.1 Tiedemannia canbyi (J.M. Coult. & Rose) Feist & S.R. Downie, comb. nov. ≡ Oxypolis filiformis var. canbyi J.M. Coult. & Rose in Contr. U.S. Natl. Herb. 7: 193. 1900 ≡ Oxypolis canbyi (J.M. Coult. & Rose) Fernald in Rhodora 41: 139. 1939 – Type: U.S.A., Delaware, Ellendale, Aug. 1867, Canby s.n. (US!, holotype; E!, PH!, isotypes). 2.2 Tiedemannia filiformis (Walter) Feist & S.R. Downie, comb. nov. ≡ Oenanthe filiformis Walter, Fl. Carol.: 113. 1788 ≡ Oxypolis filiformis (Walter) Britton in Mem. Torrey Bot. Club 5: 239. 1894 ≡ Oenanthe carolinensis Pers., Syn. Pl. 1: 318. 1805, nom. illeg. ≡ Oxypolis caroliniana (Pers.) Raf. in Bull. Bot. (Geneva) 1: 218. 1830, nom. illeg. ≡ Oenanthe teretifolia Muhl., Cat. Pl. Amer. Sept.: 32. 1813, nom. illeg. ≡ Sium teretifolium (Muhl.) Elliott, Sketch Bot. S. Carolina 1: 354. 1817, nom. illeg. ≡ Tiedemannia teretifolia (Muhl.) DC., Coll. Mem. 5: 81. 1829, nom. illeg. ≡ Peucedanum teretifolium (Muhl.) Wood, Amer. Bot. Fl.: 136. 1870, nom. illeg. – Type: U.S.A., South Carolina, Berkeley County, 17 Sept. 1981, Porcher 1977a (BM!, neotype, designated by A.O. Tucker & al. in Syst. Bot. 8: 300. 1983; CITA, DOV, isoneotypes). = Tiedemannia bakeri H. Wolff ex Urb., Symb. Antill. 5: 452. 1908 ≡ Oxypolis bakeri (H. Wolff ex Urb.) Britton & P. Wilson ex Bracelin in Torreya 29: 16. 1929 – Type: Cuba, in Havana Province near Batabano, Oct. 1904, Baker & Wilson 2215 (location unknown). 413 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium 2.2.1 Tiedemannia filiformis (Walter) Feist & S.R. Downie subsp. filiformis 2.2.2 Tiedemannia filiformis subsp. greenmannii (Mathias & Constance) Feist & S.R. Downie, comb. nov. ≡ Oxypo lis greenmanii Mathias & Constance in Bull. Torrey Bot. Club 69: 152. 1942 ≡ Oxypolis filiformis subsp. greenmanii (Mathias & Constance) Judd in Rhodora 84: 277. 1982 – Type: U.S.A., Florida, Gulf County, Wewahitchka, Aug. 1896, Chapman s.n. (MO!, holotype). 3. Oxypolis Raf., Neogenyton: 2. 1825 – Type: Oxypolis ri gidior (L.) Raf. 3.1 Oxypolis rigidior (L.) Raf. in Bull. Bot. (Geneva) 1: 218. 1830 ≡ Sium rigidius L., Sp. Pl. 1: 251. 1753 ≡ Oenanthe rigidius (L.) Crantz, Cl. Umbell. Emend.: 85. 1767, ‘rigida’ ≡ Pastinaca rigidior (L.) Spreng. in Roemer & Schultes, Syst. Veg. 6: 586. 1820, ‘rigida’ ≡ Archemora rigidior (L.) DC., Prodr. 4: 188. 1830, ‘rigida’ ≡ Peucedanum ri gidius (L.) Wood, Amer. Bot. Fl.: 136. 1870, nom. illeg., non Bunge (1833), ‘rigidum’ ≡ Tiedemannia rigidior (L.) J.M. Coult. & Rose in Bot. Gaz. 12: 74. 1887, ‘rigida’ – Type: U.S.A. Virginia, Clayton 279 (BM!, lectotype, designated by J.L. Reveal in Taxon 55: 215. 2006). = Archemora serrata Raf., Herb. Raf.: 78. 1833 – Type: U.S.A., Kentucky & Tennessee (location unknown). = Archemora trifoliata Raf., Herb. Raf.: 78. 1833 – Type: U.S.A., Missouri (location unknown). = Oenanthe ambigua Nutt., Gen. N. Amer. Pl. 1: 189. 1818 ≡ Pastinaca ambigua (Nutt.) Torr., Fl. N. Middle United States: 315. 1824 ≡ Archemora ambigua (Nutt.) DC., Prodr. 4: 188. 1830 ≡ Archemora rigidior (L.) DC. var. ambigua (Nutt.) A. Gray, Manual: 158. 1848, ‘rigida’ ≡ Peucedanum rigidius (L.) Wood var. ambiguum (Nutt.) Wood, Amer. Bot. Fl.: 136. 1870, ‘rigidum’ ≡ Tiedemannia rigidior (L.) J.M. Coult. & Rose var. ambigua (Nutt.) J.M. Coult. & Rose, Rev. N. Amer. Umbell.: 47. 1888, ‘rigida’ ≡ Oxy polis rigidior (L.) Raf. var. ambigua (Nutt.) B.L. Rob. in Rhodora 10: 35. 1908 – Type: U.S.A., banks of the Delaware near Philadelphia (PH!). = Oxypolis turgida Small, Man. S.E. Fl.: 986. 1933 – Type: U.S.A., Virginia, Staunton Co., Staunton, 2 Oct. 1895, Murrill s.n. (lectotype, designated by M.A. Feist in J. Bot. Res. Inst. Texas 6: 662. 2009: NY!). = Sium denticulatum Baldwin in Elliott, Sketch Bot. S. Carolina 1: 354. 1817 ≡ Archemora denticulata (Baldwin) DC., Prodr. 4: 188. 1830 ≡ Oxypolis denticulata (Baldwin) Raf. in Bull. Bot. (Geneva) 1: 218. 1830 ≡ Pastinaca denticulata (Baldwin) D. Dietr., Syn. Pl. 2: 971. 1840 – Type: U.S.A., Georgia, 1817, Baldwin s.n. (lectotype, designated by Edmondson in Novon 15: 109. 2005: LINN-Smith, digital image!). = Sium longifolium Pursh, Fl. Amer. Sept.: 194. 1813 ≡ Oxypo lis rigidior (L.) Raf. var. longifolia (Pursh) Britton in Mem. Torrey Bot. Club 5: 239. 1894, ‘rigidus var. longifolius’ ≡ Oxypolis longifolia (Pursh) Small, Fl. S.E. U.S.: 875, 1336. 414 TAXON 61 (2) • April 2012: 402–418 1903 ≡ Oxypolis rigidior (L.) Raf. subsp. longifolia (Pursh) Stone, Pl. S. New Jersey 2: 600. 1911 – Type: U.S.A., New Jersey (lectotype, designated with reservations by J. Ewan in the introduction to the facsimile reprint of F.T. Pursh’s Flora Americae Septentrionalis, 1979: PH!). = Sium tricuspidatum Elliott, Sketch Bot. S. Carolina 1: 354. 1817 ≡ Archemora tricuspidata (Elliott) DC., Prodr. 4: 188. 1830 ≡ Oxypolis tricuspidata (Elliott) Raf. in Bull. Bot. (Geneva) 1: 218. 1830 ≡ Pastinaca tricuspidata (Elliott) D. Dietr., Syn. Pl. 2: 971. 1840 – Type: U.S.A., South Carolina (CHARL!, holotype). 3.2 Oxypolis ternata (Nutt.) A. Heller, Cat. N. Amer. Pl.: 5. 1898 ≡ Peucedanum ternatum Nutt., Gen. N. Amer. Pl. 1: 182. 1818 ≡ Sataria linearis Raf., New Fl. 4: 21. 1838, nom. illeg. ≡ Archemora ternata (Nutt.) Nutt. in Torr. & A. Gray, Fl. N. Amer. 1: 631. 1840 ≡ Tiedemannia ternata (Nutt.) J.M. Coult. & Rose in Bot. Gaz. 12: 74. 1887 – Type: U.S.A., North and South Carolina, Nuttall s.n. (BM! lectotype, designated herein). This taxon was neotypified by M.A. Feist in J. Bot. Res. Inst. Texas 3: 662. 2009. At that time a type could not be located. This type specimen has since been located, therefore the neotypification is revoked and this specimen is designated as the lectotype. = Sataria linearis Raf. var. longipes Raf., New Fl. 4: 21. 1838. Rafinesque did not cite a specific collection or give a locality (location unknown). = Neurophyllum longifolium Torr. & A. Gray, Fl. N. Amer. 1: 613. 1840 – Types: U.S.A., North Carolina, [Cravern Co.], “Swamps near Newbern, North Carolina, Mr. Croom! Dr. Loomis! Middle Florida, Mr. Croom! Sept.” (NY!, lectotype, designated by M.A. Feist in J. Bot. Res. Inst. Texas 3: 662. 2009). 3.3 Oxypolis fendleri (A. Gray) A. Heller in Bull. Torrey Bot. Club 24: 478. 1897 ≡ Archemora fendleri A. Gray in Mem. Amer. Acad. Arts, n.s., 4: 56. 1849 ≡ Tiedemannia fendleri (A. Gray) J.M. Coult. & Rose, Rev. N. Amer. Umbell.: 48. 1888 – Type: U.S.A., New Mexico, Santa Fe Creek, Fendler 272 (GH!, holotype; GH!, BM!, isotypes). 3.4 Oxypolis occidentalis J.M. Coult. & Rose in Contr. U.S. Natl. Herb. 7: 196. 1900 – Type: U.S.A., Oregon, west of Crater Lake, Leiberg 4413 (US!, holotype; ORE!, isotype). 4. Ptilimnium Raf. in Amer. Month. Mag. & Crit. Rev. 4: 192. 1819 (nom. nud.). Neogenyton: 2. 1825 – Type: Ptilimnium capillaceum (Michx.) Raf. 4.1 Ptilimnium capillaceum (Michx.) Raf. in Bull. Bot. (Geneva) 1: 217. 1830 ≡ Ammi capillaceum Michx., Fl. Bor.Amer. 1: 164. 1803 ≡ Sison capillaceum (Michx.) Spreng., Syst. Veg. 1: 897. 1825 ≡ Discopleura capillacea (Michx.) DC., Coll. Mém. 5: 38. 1829 – Type: U.S.A., South Carolina (P-MICH). = Ammi junceum Raf., Neogenyton: 2. 1825 (nom. nud.) ≡ Ptilim nium junceum (Raf.) Raf. in Bull. Bot. (Geneva) 1: 217. 1830 TAXON 61 (2) • April 2012: 402–418 ≡ Discopleura juncea (Raf.) Steud., Nomcl. Bot., ed. 2., 1: 520. 1840 – Type: U.S.A., Kentucky (location unknown). = Ammi majus Walter, Fl. Carol.: 113. 1788 non A. majus L. (1753) ≡ Discopleura major (Walter) Britton & al., Prelim. Cat.: 22. 1888 – Type: U.S.A., South Carolina (location unknown). = Ammi rubricaule Hornem., Hort. Bot. Hafn.: 272. 1813 ≡ Sison rubricaule (Hornem.) Eaton & Wright, Man. Bot., ed. 8 [= N. Amer. Bot]: 429. 1840 – Type: U.S.A., Maryland, Baltimore Co., near Baltimore (location unknown). 4.2 Ptilimnium ahlesii Weakley & G.L. Nesom in Sida 21: 744. 2004 – Type: U.S.A., North Carolina, Brunswick Co., tidal freshwater marsh of the Brunswick River, Weakley & LeBlond 7317 (NCU!, holotype; MO!, NY!, isotypes) 4.3 Ptilimnium costatum (Elliott) Raf. in Bull. Bot. (Geneva) 1: 217. 1830 ≡ Ammi costatum Elliott, Sketch Bot. S. Carolina 1: 350. 1817 ≡ Discopleura capillacea var. costata DC., Coll. Mem. 5: 39. 1829 ≡ Discopleura costata (Elliott) Steud., Nomencl. Bot., ed. 2, 1: 77, 520. 1840 – Type: U.S.A., Georgia, “In inundatis, Ogeechee” (GH, photo of type!). = Ptilimnium missouriense J.M. Coult. & Rose in Contr. U.S. Natl. Herb. 12: 444. 1909 – Type: U.S.A., Missouri, Allenton, 27 Aug. 1878, Letterman s.n. (US!, holotype). 4.4 Ptilimnium texense J.M. Coult. & Rose in Contr. U.S. Natl. Herb. 12: 445. 1909 – Type: U.S.A., Texas, Near Hockley, Sept. 1890, Thurow s.n. (US!, holotype). 4.5 Ptilimnium nuttallii (DC.) Britton in Mem. Torrey Bot. Club 5: 244. 1894 ≡ Discopleura nuttallii DC., Coll. Mem. 5: 39. 1829 ≡ Discopleura capillacea var. nuttallii (DC.) J.M. Coult. & Rose in Bot. Gaz. 12: 292. 1887 – Type: U.S.A., Red River, Nuttall s.n. (G-DC!, holotype; BM! p.p. the type is the specimen on the left side of the sheet, K! p.p. the type is the specimen on the right side of the sheet, PH!, NY!, isotypes). Peucedanum verticillatum Raf., Fl. Ludov.: 81. 1817, nom. dub. – Type: U.S.A., Louisiana (location unknown). This name has been listed as a possible synonym of Ptilim nium nuttallii (DC.) Britton. However, Rafinesque’s description associated with this name is vague and could be applied to a number of different species. It is unknown whether Rafinesque based his description on an actual specimen and no type could be located. Therefore, there is no way to determine for certain that this name is a synonym of P. nuttallii and the designation nomen dubium is applied. ACKNOWLEDGEMENTS The authors wish to thank the curators and collection managers of the following herbaria for providing specimens for this study: AUA, BAYLU, BRIT, CHICO, DAV, DOV, DUKE, EKY, F, FLAS, FSU, ILL, ILLS, JEPS, LAF, LSU, MO, NCSC, NCU, NO, NY, OS, OSC, PH, RM, Feist & al. • Generic delimitations for Oxypolis and Ptilimnium RSA-POM, TAMU, TENN, TEX, UARK, UC, US, USCH, USF, and WVA. Thanks to Brenda Molano-Flores for accompanying M.A. Feist on many collecting trips. Thanks to all the botanists, who are truly too many to mention, that helped M.A. Feist to locate populations of her study plants and even, in many cases, accompanied her into the field. This work was partially supported by grants to M.A. Feist, including the Graduate College Dissertation Research Travel Grant (University of Illinois), the Francis M. and Harlie M. Clark Research Support Grant (School of Integrative Biology), the Herbert H. Ross Memorial Award (Illinois Natural History Survey), the California Native Plant Society Research Award, the Catherine H. Beattie Fellowship (Center for Plant Conservation and the Garden Club of America), and the American Society of Plant Taxonomy Graduate Student Research Grant. This work was also supported by funding from the National Science Foundation (DEB 0089452) to S.R. Downie. LITERATURE CITED Affolter, J.M. 1985. A monograph of the genus Lilaeopsis (Umbelliferae). Syst. Bot. Monogr. 6: 1–140. Bell, C.R. & Constance, L. 1957. Chromosome numbers in Umbelliferae. Amer. J. Bot. 44: 565–572. Bell, C.R. & Constance, L. 1960. Chromosome numbers in Umbelliferae II. Amer. J. Bot. 47: 24–32. Brace, L.J.K. 1929. Notes on the occurrence of Oxypolis filiformis in the Bahamas. Torreya 1: 16–17. Calsbeek, R., Thompson, J.N. & Richardson, J.E. 2003. Patterns of molecular evolution and diversification in a biodiversity hotspot: The California Floristic Province. Molec. Ecol. 12: 1021–1029. Calviño, C.I. & Downie, S.R. 2007. Circumscription and phylogeny of Apiaceae subfamily Saniculoideae based on chloroplast DNA sequences. Molec. Phylogenet. Evol. 44: 175–191. Calviño, C.I., Tilney, P.M., Van Wyk, B.E. & Downie, S.R. 2006. A molecular phylogenetic study of southern African Apiaceae. Amer. J. Bot. 93: 1828–1847. Candolle, A.P. de 1829. Memoire sur la famille des Ombelliferes. Paris: Treuttel & Würtz. Cheney, M. & Marr, K.L. 2007. Cowbane, Oxypolis occidentalis, a new native vascular plant species for the Queen Charlotte Islands, British Columbia. Canad. FieldNaturalist 121: 421–422. Constance, L., Chuang, T.-I. & Bell, C.R. 1976. Chromosome numbers in Umbelliferae V. Amer. J. Bot. 63: 608–625. Coulter, J.M. & Rose, J.N. 1887. Notes on Umbelliferae of the E. United States III. Bot. Gaz. 12: 73–76 Coulter, J.M. & Rose, J.N. 1888. Revision of North American Umbel liferae. Crawfordsville, Indiana: Wabash College. Crawford, D.J. & Hartman, R.L. 1972. Chromosome numbers and taxonomic notes for Rocky Mountain Umbelliferae. Amer. J. Bot. 59: 386–392. Downie, S.R. & Katz-Downie, D.S. 1996. A molecular phylogeny of Apiaceae subfamily Apioideae: Evidence from nuclear ribosomal DNA internal transcribed spacer sequences. Amer. J. Bot. 83: 234–251. Downie, S.R., Katz-Downie, D.S. & Cho, K.-J. 1996. Phylogenetic analysis of Apiaceae subfamily Apioideae using nucleotide sequences from the chloroplast rpoC1 intron. Molec. Phylogenet. Evol. 6: 1–18. Downie, S.R., Katz-Downie, D.S. & Spalik, K. 2000a. A phylogeny of Apiaceae tribe Scandiceae: Evidence from nuclear ribosomal DNA internal transcribed spacer sequences. Amer. J. Bot. 87: 76–95. Downie, S.R., Katz-Downie, D.S., Sun, F.-J. & Lee, C.-S. 2008. Phylogeny and biogeography of Apiaceae tribe Oenantheae inferred from nuclear rDNA ITS and cpDNA psbI5′trnK(UUU) sequences, with emphasis on the North American Endemics clade. Botany 86: 1039–1064. 415 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium Downie, S.R., Katz-Downie, D.S. & Watson, M.F. 2000b. A phylogeny of the flowering plant family Apiaceae based on chloroplast DNA rpl16 and rpoC1 intron sequences: Towards a suprageneric classification of subfamily Apioideae. Amer. J. Bot. 87: 273–292. Downie, S.R., Plunkett, G.M., Watson, M.F., Spalik, K., KatzDownie, D.S., Valiejo-Roman, C.M., Terentieva, E.I., Troitsky, A.V., Lee, B.-Y., Lahham, J. & El-Oqlah, A. 2001. Tribes and clades within Apiaceae subfamily Apioideae: The contribution of molecular data. Edinburgh J. Bot. 58: 301–330. Downie, S.R., Ramanath, S., Katz-Downie, D.S. & Llanas, E. 1998. Molecular systematics of Apiaceae subfamily Apioideae: Phylogenetic analyses of nuclear ribosomal DNA internal transcribed spacer and plastid rpoC1 intron sequences. Amer. J. Bot. 85: 563–591. Downie, S.R., Sun, F.-J., Katz-Downie, D.S. & Colletti, G.J. 2004. A phylogenetic study of Perideridia (Apiaceae) based on nuclear ribosomal DNA ITS sequences. Syst. Bot. 29: 737–751. Downie, S.R., Watson, M.F., Spalik, K. & Katz-Downie, D.S. 2000c. Molecular systematics of Old World Apioideae (Apiaceae): Relationships among some members of tribe Peucedaneae sensu lato, the placement of several island-endemic species, and resolution within the apioid superclade. Canad. J. Bot. 78: 506–528. Easterly, N.W. 1957a. A morphological study of the genus Ptilimnium. Dissertation, University of West Virginia, Morgantown. Easterly, N.W. 1957b. A morphological study of Ptilimnium. Brittonia 9: 136–145. Elliott, S. 1817. A sketch of the botany of South Carolina and Georgia. Charleston: J.R. Schenck. Farris, J.S., Källersjö, M., Kluge, A.G. & Bult, C. 1995. Constructing a significance test for incongruence. Syst. Biol. 44: 570–572. Feder, N. & O’Brien, T.P. 1968. Plant microtechnique: Some principles and new methods. Amer. J. Bot. 55: 123–42. Feist, M.A.E. 2009. Clarifications concerning the nomenclature and taxonomy of Oxypolis ternata (Apiaceae). J. Bot. Res. Inst. Texas 3: 661–666. Feist, M.A.E. 2010. The reinstatement of Ptilimnium texense (Apiaceae) and a new key to the genus. J. Bot. Res. Inst. Texas 4: 641–651. Feist, M.A.E. & Downie, S.R. 2008. A phylogenetic study of Oxypolis and Ptilimnium based on nuclear rDNA ITS sequences. Syst. Bot. 33: 447–458. Hall, T.A. 1999. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids Sym. Ser. 41: 95–98. Hardway, T.M., Spalik, K., Watson, M.F., Katz-Downie, D.S. & Downie, S.R. 2004. Circumscription of Apiaceae tribe Oenantheae. S. African J. Bot. 70: 393–406. Heusser, C.J. 1960. Late Pleistocene environments of North Pacific North America. Special Publication, vol. 35. New York: American Geographical Society. Huelsenbeck, J.P. & Ronquist, F. 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755. Judd, W.S. 1982. The taxonomic status of Oxypolis greenmanii (Apiaceae). Rhodora 84: 265–279. Kaplan, D.R. 1970. Comparative development and morphological interpretation of ‘rachis-leaves’ in Umbelliferae. Bot. J. Linn. Soc. 63(Suppl. 1): 101–126. Kartesz, J.T. 2010. A synonymized checklist and atlas with biological attributes for the vascular flora of the United States, Canada, and Greenland. In: Kartesz, J.T. & Meacham, C.A. (eds.), Synthesis of the North American flora, version 1.0. Chapel Hill: North Carolina Botanical Garden. CD-ROM. Kral, R. 1981. Notes on some “quill-leaved” umbellifers. Sida 9: 124–134. Lacourse, T., Mathewes, R.W. & Fedje, D.W. 2005. Late-glacial vegetation dynamics of the Queen Charlotte Islands and the adjacent continental shelf, British Columbia, Canada. Palaeogeogr. Pal aeoclimatol. Palaeoecol. 226: 36–57. Lee, B.-Y., Levin, G.A. & Downie, S.R. 2001. Relationships within the spiny-fruited Umbellifers (Scandiceae subtribes Daucinae and 416 TAXON 61 (2) • April 2012: 402–418 Torilidinae) as assessed by phylogenetic analysis of morphological characters. Syst. Bot. 26: 622–642. Lee, C.-S. & Downie, S.R. 2006. Phylogenetic relationships within Cicuta (Apiaceae tribe Oenantheae) inferred from nuclear rDNA and cpDNA sequence data. Canad. J. Bot. 84: 453–468. Liogier, H.A. & Martorell, L.F. 2000. Flora of Puerto Rico and ad jacent islands: A systematic synopsis, 2nd ed. San Juan, Puerto Rico: Editorial de la Universidad de Puerto Rico. Liu, M., Plunkett, G.M., Lowry, P.P., Van Wyk, B.-E. & Tilney, P.M. 2006. The taxonomic value of fruit wing types in the order Apiales. Amer. J. Bot. 93: 1357–1368. Liu, M., Van Wyk, B.-E., Tilney, P.M., Plunkett, G.M. & Lowry, P.P. 2009. Evidence from fruit structure supports in general the circumscription of Apiaceae subfamily Azorelloideae. Pl. Syst. Evol. 280: 1–13. Magee, A.R., Calviño, C.I., Liu, M., Downie, S.R., Tilney, P.M. & Van Wyk, B.-E. 2010a. New tribal delimitations for the early diverging lineages of Apiaceae subfamily Apioideae. Taxon 59: 567–580. Magee, A.R., Van Wyk, B.-E., Tilney, P.M. & Downie, S.R. 2009a. Generic delimitations and relationships of the Cape genera Capno phyllum, Dasispermum and Sonderina, the North African genera Krubera and Stoibrax, and a new monotypic genus of the subfamily Apioideae (Apiaceae). Syst. Bot. 34: 580–594. Magee, A.R., Van Wyk, B.-E., Tilney, P.M. & Downie, S.R. 2010b. A taxonomic revision of the South African endemic genus Dasi spermum (Apiaceae, Apioideae) S. African J. Bot. 76: 308–323 Magee, A.R., Van Wyk, B.-E., Tilney, P.M., Sales, F., Hedge, I. & Downie, S.R. 2009b. Billburttia, a new genus of Apiaceae (tribe Apieae) endemic to Madagascar. Pl. Syst. Evol. 283: 237–245. Marcinko S.E. & Randall J.L. 2008. Protandry, mating systems, and sex expression in the federally endangered Ptilimnium nodosum (Apiaceae). J. Torrey Bot. Soc. 135: 178–188. Mathias, M. 1936. Studies in the Umbelliferae V. Brittonia 3: 239–245. McNeill, J., Barrie, F.R., Burdet, H.M., Demoulin, V., Hawksworth, D.L., Marhold, K., Nicolson, D.H., Prado, J., Silva,P.C., Skog, J.E., Wiersema, J.H. & Turland, N.J. (eds.) 2006. International code of botanical nomenclature (Vienna Code): Adopted by the Seventeenth International Botanical Congress Vienna, Austria, July 2005. Regnum Vegetabile 146. Ruggell: Gantner. Persoon, C.H. 1805. Synopsis plantarum seu enchiridium botanicum, vol. 1. Tübingen, Germany: J.G. Cotta. Pimenov, M.G., Vasil’eva, M.G., Leonov, M.V. & Daushkevich, J.V. 2003. Karyotaxonomical analysis in the Umbelliferae. Enfield: Science Publishers. Plunkett, G.M., Soltis, D.E. & Soltis, P.S. 1996. Evolutionary patterns in Apiaceae: Inferences based on matK sequence data. Syst. Bot. 21: 477–495. Posada, D. & Crandall, K.A. 1998. Modeltest: Testing the model of DNA substitution. Bioinformatics 14: 817–818. Shaw, J., Lickey, E.B., Beck, J.T., Farmer, S.B., Liu, W., Miller, J., Siripun, K.C., Winder, C.T., Schilling, E.E. & Small, R.L. 2005. The tortoise and the hare II: Relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. Amer. J. Bot. 92: 142–166. Shaw, J., Lickey, E.B., Schilling, E.E. & Small, R.L. 2007. Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: The tortoise and the hare III. Amer. J. Bot. 94: 275–288 Simmons, M.P. & Ochoterena, H. 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49: 369–381. Spalik, K. & Downie, S.R. 2001. The utility of morphological characters for inferring phylogeny in Scandiceae subtribe Scandicinae (Apiaceae). Ann. Missouri Bot. Garden 88: 270–301. Spalik, K. & Downie, S.R. 2006. The evolutionary history of Sium sensu lato (Apiaceae): Dispersal, vicariance, and domestication as inferred from ITS rDNA phylogeny. Amer. J. Bot. 93: 76–95. Spalik, K., Downie, S.R. & Watson, M.F. 2009. Generic delimitation TAXON 61 (2) • April 2012: 402–418 within the Sium alliance (Apiaceae tribe Oenantheae) inferred from cpDNA rps165′trnK(UUU) and nrDNA ITS sequences. Taxon 58: 735–748. Swofford, D.L. 2003. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4. Sunderland, Massachusetts: Sinauer. Theobald, W.L. 1971. Comparative anatomical and developmental studies in the Umbelliferae. Pp. 177–197 in: Heywood, V.H. (ed.), The biology and chemistry of the Umbelliferae. Botanical Journal of the Linnean Society 64 Supplement 1. London: Academic Press for the Linnean Society London. Thompson, J.D., Gibson, T.J., Plewniak, F., Jeanmougin, F. & Higgins, D.G. 1997. The Clustal X windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl. Acids Res. 24: 4876–4882. Tucker, A.O., Dill, N.H., Pizzolato, T.D. & Kral, R.D. 1983. Nomenclature, distribution, chromosome numbers, and fruit morphology of Oxypolis canbyi and O. filiformis (Apiaceae). Syst. Bot. 8: 299–304. Walter, T. 1788. Flora Caroliniana. London: J. Fraser. Feist & al. • Generic delimitations for Oxypolis and Ptilimnium Weakley, A.S. & Nesom, G.L. 2004. A new species of Ptilimnium (Apiaceae) from the Atlantic coast. Sida 21: 743–752. Wen, J. & Zimmer, E.A. 1996. Phylogeny and biogeography of Panax L. (the ginseng genus, Araliaceae): Inferences from ITS sequences of nuclear ribosomal DNA. Molec. Phylogenet. Evol. 6: 167–177. Werle, E., Schneider, C., Renner, M., Volker, M. & Fiehnet, W. 1994. Convenient single-step, one tube purification of PCR products for direct sequencing. Nucl. Acids Res. 22: 4354–4355 Winter, P.J.D., Magee, A.R., Phephu, N., Tilney, P.M., Downie, S.R. & Van Wyk, B.-E. 2008. A new generic classification for African peucedanoid species (Apiaceae). Taxon 57: 347–364. U.S. Fish and Wildlife Service. 1986. Determination of Oxypolis can byi (Canby’s dropwort) to be an endangered species. Fed. Reg. 37: 6690–6693. U.S. Fish and Wildlife Service. 1988. Determination of Ptilimnium nodosum (harperella) to be an endangered species. Fed. Reg. 53: 37978–37982. Appendix 1. Accessions from which nuclear rDNA ITS and cpDNA sequence data were obtained, with corresponding DNA accession and GenBank reference numbers and voucher information. Taxon name — DNA accession number; voucher information; cpDNA GenBank accession number(s), ITS GenBank accession number. An asterisk (*) indicates sequences newly generated for this study; a number sign (#) indicates the trnQ and trnK sequences used were part of a longer sequence of the entire psbItrnK region deposited to GenBank; a paragraph sign (§) indicates the trnK sequence used was part of a longer sequence of the rps16 intron-trnK region deposited to GenBank. Atrema americanum Benth. & Hook. f. ex S. Watson — 1160; U.S.A., Texas, Williamson Co., jnt. Hwys. 183 and 29, Barclay & Perdue 785 (UC 184750); EF185206#, EF177699. Cynosciadium digitatum DC. — 1571; U.S.A., Louisiana, Madison Parish, 1 mi E of Indian Lake, 28 May 1973, Jones 215 (ILL); EF185219#, EF177704. 1986; U.S.A., Arkansas, Lafayette Co., 4 mi E of Red River, Hwy. 82, 24 May 1993, Sundell & al. 10,500 (ILL); EF185221#, EF177706. Daucosma laciniata Engelm. & A. Gray — 2912; U.S.A., Texas, Uvalde Co., Garner State Park, 21 June 1958, Sullivan & Turner 33 (GA 114044); JQ652503*, JQ652549*. 3411; U.S.A., Texas, Bexar Co., San Antonio, 31 July 1921, Schulz 594 (US 1087113); JQ652500*, JQ652550*, JQ652463*. 3412; U.S.A., Texas, Hays Co., Wimberly, 5 July 1942, Fisher s.n. (F 1501788); JQ652501*, JQ652551*, JQ652464*. Lilaeopsis carolinensis J.M. Coult. & Rose — 2148; U.S.A., cultivated, origin unknown, 1985, Bogner s.n., material sent from Petersen GPL41 (C); EF185225#, AF466276. Lilaeopsis mauritiana G. Petersen & Affolter — 2150; Republic of Mauritius, Le Val Nature Park, 3 May 1992, Windelov s.n., material sent from Petersen GPL81 (C); EF185226#, AF466277. Lilaeopsis novae-zelandiae (Gand.) A.W. Hill — 2152; New Zealand. cultivated, origin unknown, material sent from Petersen GPL9 (C); EF185227#, AF466278. Lilaeopsis occidentalis J.M. Coult. & Rose — 1999; U.S.A., Oregon, Douglas Co., East Gardiner, Hill & Dutton 32982 (ILLS 203634); EF185228#, AY360242. Limnosciadium pinnatum (DC.) Mathias & Constance — 1511; U.S.A., Louisiana, Ouachita Parish, Ouachita W.M.A., 20 May 1987, Thomas & al. 99586I (MO 3680921); EF185229#, EF177717. 2395; U.S.A., Missouri, Stoddard Co., Otter Slough Conservation Area, 31 May 2000, Brant & al. 4380 (MO 5186226); EF185230#, EF177720. Limnosciadium pumilum (Engelm. & A. Gray) Mathias & Constance — 3742; U.S.A., Texas, San Ptricio Co., US 181 NW of Sinton, 5 Apr. 1984, Ertter 5263 (NY); JQ652499*, JQ652548*, JQ652470*. Neogoezia gracilipes (Hemsl.) Hemsl. — 2270; Mexico, Oaxaca, Nochixtlan, N of La Joya, 2 Oct. 1993, Panero 3614 (UC 1611523); EF185232#, EF177726. Neogoezia minor Hemsl. — 2274; Mexico, Oaxaca, Cerro San Felipe summit, Breedlove & Almeda 59951 (UC 1518420); EF185236#, EF177730. Oxypolis canbyi (J.M. Coult. & Rose) Fern. — 2744; U.S.A., South Carolina, Richland Co., Carolina Bay on N side of Vero Road and ca. 0.3 mi E of Sec. Hwy. 2206, ca. 2 air mi NW of downtown Gadsden, 7 Sept. 1984, Nelson 3687 (NCU 537890); JQ652489*, JQ652540*, EF647756. 2747; U.S.A., North Carolina, Scotland Co., McIntosh Carolina Bay, US 401 NE of Laurinburg, 13 Sept. 1992, Sorrie 6946 (NCU 562048); JQ652490*, JQ652541*, EF647757. 2751; U.S.A., South Carolina, Lee Co., just NE of Mt Pleasant Church, W of Lynchburg, 10 Sept. 1985, Nelson 4271 (NCU 537512); JQ652491*, JQ652542*. 2937; U.S.A., South Carolina, Bamberg Co., Bamberg Bay Preserve, 28 Aug. 2005, Feist, MolanoFlores & Glitzenstein 3193 (ILLS); JQ652492*, JQ652543*, EF647759. 2938; U.S.A., South Carolina, Bamberg Co., Oxypolis Bay Preserve, 28 Aug. 2005, Feist, MolanoFlores & Glitzenstein 3194 (ILLS); JQ652493*, JQ652544*, EF647760. Oxypolis fendleri (A. Gray) A. Heller — 2350; U.S.A., Colorado, Boulder Co., Forth of July Canyon, 10 July 1962, Jones 34084 (ILL); JQ652502*, JQ652552*, EF647767. 2351; U.S.A., Colorado, Boulder Co., along Boulder Creek, 24 June 1962, Jones 34450 (ILL); JQ652504*, JQ652553*, EF647768. 2368; U.S.A., New Mexico, Rio Arriba Co., Ortega Mountains, 17 Aug. 1984, Hill 15181 (UC 1508862); JQ652505*, JQ652554*. 2369; U.S.A., Colorado, Chafee Co., CO 306, 14 mi W of Buena Vista, 2 Aug. 1973, Haber & Given 2049 (CAN 370800); JQ652506*, EF185239§, EF177734. Oxypolis filiformis (Walter) Britton — 2371; U.S.A., Louisiana, Vernon Parish, E of Drake’s Creek, ca. 2 mi E of Johnsville Church and LA 10, ca. 7 mi E of Pickering, Kisatchie National Forest, 7 Sept. 1987, Thomas 101486 (DAO 574521); JQ652494*, EF185240§, EF177736. 2713; U.S.A., Florida, Alachua Co., Gainesville, N side of NE 39th Ave. N. just E of Main St., 9 Sept. 1987, Alcorn 155 (FLAS 166610); JQ652495*, JQ652545*, EF177737. Oxypolis filiformis × O. greenmanii — 2714; U.S.A., Florida, Bay Co., along US 231, 1.8 mi N of the junction with FL Rt. 388, N of Youngstown, 29 Aug. 1980, Judd & Perkins 2714 (FLAS 174297); JQ652498*, EF185242§, EF177739. Oxypolis greenmanii Mathias & Constance — 2717; U.S.A., Florida, Bay Co., Tyndall Airforce Base, 15 Sept. 1979, Judd & Perkins 2439 (FLAS 174274); JQ652496*, JQ652546*, EF177738. 2941; U.S.A., Florida, Gulf Co., just E of Wetappo Creek and 3.6 mi S of FL 22, 2 Sept. 2005, Feist & MolanoFlores 3244 (ILLS); JQ652497*, JQ652547*, EF647780. Oxypolis occidentalis J.M. Coult. & Rose — 1142; U.S.A., California, El Dorado County, Osgood Swamp, Follette s.n. (JEPS 82187); EF185243#, AY360254. 1153; U.S.A., California: Fresno Co., Wishon Reservoir Dam, Call 2455 (UC 282880); EF185244#, EF177740. 2755; U.S.A., Oregon, Douglas and Jackson Co., Abbott Creek Research Natural Area, ca. 20 mi W of Crater Lake near Abbott Butte, 29 July 1972, Mitchell 348 (USFS 406185), JQ652507*, JQ652555*, EF647784. 2756; Canada. British Columbia: Queen Charlotte Islands, Graham Island, 2003, Cheney s.n. (ILLS), JQ652508*, JQ652556*, EF647786. 2899; U.S.A., California: Tehama Co., Forest Rd. 26N09 at Cascade Creek, NW of Chico Meadows, 1.1 mi SE of Hwy. 32, 9 Sept. 1997, Oswald & Ahart 8863 (JEPS 94369), JQ652509*, JQ652557*, EF647789. 2927; U.S.A., California: Sierra Co., ca. 1.25 mi N of Scales, ca. 2 mi (air) SE of Poverty Hill, 28 Sept. 2001, Ahart 9295 (JEPS 102455), JQ652510*, JQ652558*, EF647790. 2928; U.S.A., Oregon, Lane Co., Quaking Aspen Swamp, 6 mi W of S end of Cougar Reservoir, 28 July 1979, Wagner 2318 (ORE 103195); JQ652511*, JQ652559*, EF647791. 2929; U.S.A., Oregon, Klamath Co., near W boundary of Crater Lake Park, 24–29 Aug., Wynd 1745 (ORE 64782); JQ652512*, JQ652560*, EF647785. 3376; U.S.A., California, Kern Co., French Meadow, 3 Sept. 2007, Feist & Molano Flores FM11; JQ652513*, JQ652561*, JQ652465*. 3413; Canada, Haida Guai, British Columbia, Feist & MolanoFlores Site 42; JQ652514*, JQ652562*. 3417; U.S.A., California, Butte Co., 9 Sept. 2007, Feist & MolanoFlores BTC13; JQ652515*, JQ652563*, JQ652466*. 3442; U.S.A., Oregon, Lane Co., Quaking Aspen Swamp, 13 Sept. 2007, Feist & MolanoFlores QAS25; JQ652516*, JQ652564*. 3464; U.S.A., California, San Bernardino Co., Lemon Lily Springs, 2 Sept. 2007, 417 Feist & al. • Generic delimitations for Oxypolis and Ptilimnium TAXON 61 (2) • April 2012: 402–418 Appendix 1. Continued. Feist & MolanoFlores LLS24; JQ652517*, JQ652565*. 3465; U.S.A., California, San Bernardino Co., Lemon Lily Springs, 2 Sept. 2007, Feist & MolanoFlores LLS16; JQ652518*, JQ652566*. 3466; U.S.A., California, San Bernardino Co., Lemon Lily Springs, 2 Sept. 2007, Feist & MolanoFlores LLS10; JQ652519*, JQ652567*, JQ652467*. 3528; U.S.A., California, Sierra Co., Scales, 25 Oct. 2007, Feist & MolanoFlores S11; JQ652520*, JQ652568*, JQ652468*. 3532; U.S.A., California, Tulare Co., Nelson Trail, 4 Sept. 2007, Feist & MolanoFlores NT16; JQ652521*, JQ652569*, JQ652469*. Oxypolis rigidior (L.) Raf. — 1927; U.S.A., Illinois, Vermilion Co., Windfall Hill Prairie Nature Preserve, Windfall Prairie Seep, 17 July 1991, Phillippe, Morris, & Simon 19411 (ILLS 177487); JQ652522*, EF185245§, AY360255. 1998; U.S.A., Louisiana, Winn Parish, along LA 126, 1.2 mi E of Jct. LA 1233, Kisatchie National Forest, 20 Sept. 1981, Kessler 1877 (ILL); EF185246#, EF177743. 2003; U.S.A., Illinois, Lake Co., SE corner of Tri-state Tollway and Buckley Rd. by RR, 13 Aug. 1981, Robertson 2640 (ILLS 166045); EF185247#, EF177744. Oxypolis ternata (Nutt.) A. Heller — 2735; U.S.A., South Carolina, Horry Co., 3.8 mi S. of Socastee and ca. 1 mi W. on dirt road, 25 Oct. 1970, Massey & Thomas 3480 (NCU 422851); JQ652523*, EF185248§, EF177745. 2738; U.S.A., North Carolina, Pender Co., Holly Shelter Game Land, 3 Oct. 1997, Horn & Dirig 362 (DUKE 363865); JQ652524*, EF185249§, EF177746. 2940; U.S.A., Florida, Wakulla Co., Saint Mark’s National Wildlife Refuge, Panacea Unit Longterm Burn Plot (P13), ca. 2 km SE of Sopchoppy, 1 Sept. 2005, Feist, MolanoFlores & Glitzenstein 3222 (ILLS); JQ652525*, JQ652570*, EF647809. Perideridia americana (Nutt. ex DC.) Reichenb. — 2033; U.S.A., Illinois, Shelby Co., NE of Assumption, 2 June 1981, Shildneck 12868 (ILL); JQ652526*, JQ652527*, JQ652471*. Perideridia kelloggii (A. Gray) Mathias — 778; U.S.A., California, Sonoma Co., King Ridge Rd., 5 mi N. of Cazadero, 6 Aug. 1993, Ornduff & al. s.n. (UC), cult. University of California Botanical Garden, Berkeley (no. 81.0521); EF185251#, U78373. Ptilimnium ahlesii Weakley & G.L. Nesom — 2648; U.S.A., South Carolina, Berkeley Co., Cooper River at the mouth of Durham Creek, 7 June 1990, McAninch 23 (NCU 557199); JQ652472*, EF185252§, EF177747. 2969; U.S.A., North Carolina, Brunswick Co., just E of Brunswick River and just N of the 74–76 causeway, ca. 2 mi W of Wilmington, 10 June 2004, Weakley & LeBlond 7317 (sheet 2 of 2) (NCU); JQ652473*, JQ652528*, EF647814. Ptilimnium capillaceum (Michx.) Raf. — 2701; U.S.A., Virginia, Lancaster Co., Bellwood Marsh, S of Rt. 3 bridge, W of Lancaster, 22 July 1994, Weldy 849 (BRIT); JQ652474*, JQ652529*, EF177748. 2703; U.S.A., Florida, Nassau Co., White Oak Plantation in the wedge formed by the junction of the Little St. Mary’s River and St. Mary’s River about 8–10 mi NW of Yulee, 19 June 1997, Wilbur 67597 (BRIT); JQ652475*, JQ652530*, EF647822. Ptilimnium costatum (Elliott) Raf. — 1646; U.S.A., Illinois, Jackson Co., Shawnee National Forest, 20 Sept. 1989, Stritch 2159 (ILLS 172136); EF185253#, EF177749. 1970; U.S.A., Illinois: Jackson Co., Shawnee National Forest, 11 Sept. 1989, Stritch 2124 (ILLS 172160); EF185254#, EF177750. 2402; U.S.A., Missouri, Wayne Co., Hattie’s Ford Fen Area, 12 Oct. 2001, Brant 4857 (MO 5573699); EF185256#, EF177752. 2707; U.S.A., Kentucky, Calloway Co., right 0.7 miles on KY 121S from KY 614, 9 Oct. 1972, Athey 2197 (NCU 473641); JQ652476*, JQ652531*. Ptilimnium nodosum (Rose) Mathias — 2635; U.S.A., South Carolina; Aiken Co., Monetta, Windmill High Pond, Carolina Bay Road, 20 July 1992, Hill 23921 (USF 206922); JQ652480*, JQ652533*, EF647843. 2784; U.S.A., South Carolina, Aiken Co., Maddox SC74; JQ652481*, EF185257§, EF177753. 2787; U.S.A., Maryland, Mad dox MG4, JQ652482*, EF185258§, EF177754. 2900; U.S.A., South Carolina, Saluda Co., near Hibernia, Saluda Highpond, 11 May 2005, Feist & MolanoFlores 3287 (ILLS); JQ652483*, JQ652534*, EF647841. 2902; U.S.A., Georgia, Greene Co., Siloam Outcrop, 12 May 2005, Feist & MolanoFlores Siloam5; JQ652484*, JQ652535*, EF177754. 2930; U.S.A., Arkansas, Yell Co., Ouachita Mountains, below the Hwy. 27 bridge over Irons Fork, 16 Oct. 1990, Bates 10558 (UARK); JQ652485*, JQ652536*, EF647851. 2931; U.S.A., Alabama, DeKalb Co., Little River near AL Hwy. 35 bridge, DeSoto State Park, 15 July 1987, Freeman s.n. (AUA 46749); JQ652486*, JQ652537*, EF647853. 2934; U.S.A., West Virginia, Berkeley Co., along Back Creek, 25 Aug. 2005, Feist, Harmon & O’Malley 3285 (ILLS); JQ652487*, JQ652538*, EF647854. 2936; U.S.A., North Carolina, Granville County, Tar River, 26 Aug. 2005, Feist 3286 (ILLS); JQ652488*, JQ652539*, EF647856. Ptilimnium nuttallii (DC.) Britton — 2165; U.S.A. Oklahoma, Rogers Co., Claremore, along railroad in moist ground, 12 June 1974, Jones 3030 (ILL); EF185259#, AY360256. 2617; U.S.A. Arkansas, Ashley Co., SE of Hamburg and NE of Ark. 52, 20 June 1986, Thomas 97154 (WVA 114836); JQ652478*, EF185260§, EF177758. 2623; U.S.A., Illinois, Randolph Co., W of Sparta, 16 July 2003, Feist 2510 (ILLS); JQ652477*, EF185261§, EF177759. Ptilimnium texense J.M. Coult. & Rose — 1981; U.S.A., Louisiana, Natchitoches Parish, moist seepage area beside LA 479 at Strange Rd., W of Goldonna in Kisatchie National Forest, 14 Aug. 1989, Thomas & Bell 112081 (ILL); EF185255#, EF177751. 2905; U.S.A., Texas, Anderson Co., Gus Engeling Wildlife Management Area, NW of Palestine, Lake 2 bog area, pasture 2, 16 Oct. 1993, Dubrule Reed 1354 (TAMU 24011); JQ652479*, JQ652532*, EF647825. Trepocarpus aethusae Nutt. — 1660; U.S.A., Illinois, Saline Co., US Rt. 45, E of Harrisburg levee, 7 July 1999, Hill 31876 (ILLS 201642); EF185279#, EF177761. Appendix 2. Accessions from which fruit anatomical observations were made with voucher information. Taxon name — voucher information. Cynosciadium digitatum DC. — U.S.A., Arkansas, Monroe Co., Ark. Hwy. 1, ca. 2 mi NE of Cross Roads at Branch Missionary Baptist Church, Sundell 15406 (BRIT). — U.S.A., Louisiana, Morehouse Pa., ½ mi E of Jones, 2 July, 1968, J. Thieret s.n. (SMU 37382). Limnosciadium pinnatum (DC.) Mathias & Constance — U.S.A., Texas, Red River Co., N of Clarksville, 28 June 1945, Lundell 14012 (LL). — U.S.A., Missouri, Stoddard Co., Otter Slough Conservation Area, 31 May 2000, Brant & al. 4380 (MO 5186226). Limnosciadium pumilum (Engelm. & A. Gray) Mathias & Constance — U.S.A., Texas, Calhoun Co., vacant lot in Port O’Connel, 13 Apr. 1952, Gentry 1996 (BRIT). — U.S.A., Texas, San Patricio Co., US 181 NW of Sinton, 5 Apr. 1984, Ertter 5263 (NYBG). Oxypolis canbyi (J.M. Coult. & Rose) Fern. — U.S.A., South Carolina, Lee Co., W of Lynchburg, 10 Sept. 1985, J. Nelson 4269 (USCH 032054). — U.S.A., Georgia, Lee Co., NE Leesberg, 22 Aug. 1948, Muenscher s.n. (NCU 65120). Oxypolis fendleri (A. Gray) A. Heller — U.S.A., Wyoming, Albany, Medicine Bow Forest, Elk Creek Study Bog, Sturges 205 (RM 272453). — U.S.A., Wyoming, La Plata Mines, E. Nelson s.n. (RM 12350). — U.S.A., Colorado, Boulder Co., Eldora, 22 July 1953, G.N. Jones 20071 (ILL). Oxypolis filiformis (Walter) Britton — U.S.A., South Carolina, Charleston Co., Francis Marion N.F., 29 Aug. 2009, Feist & MolanoFlores 3197 (ILLS). — U.S.A., Florida, Calhoun Co., 29 Aug. 1980, Judd & Perkins 2729 (FLAS 174320). Oxypolis greenmanii Mathias & Constance — U.S.A., Florida, Gulf Co., 17 mi N of Port St. Joe, 7 Sept. 1955, Godfrey 53756 (NCSC 52879). — U.S.A., Florida, Bay Co., Tyndall Airforce Base, 15 Sept. 1979, Judd & Perkins 2439 (FLAS 174274). Oxypolis occidentalis J.M. Coult. & Rose — U.S.A., California, Lemon Lily Springs, 2 Sept. 2007, Feist & MolanoFlores 4106 (ILLS). — U.S.A., Oregon, Lane Co., Quaking Aspen Swamp, 17 Sept. 2007, Feist & MolanoFlores 4131 (ILLS). — U.S.A., California, Kern Co. Portuguese Meadow, 27 Sept. 1936, L. Benson 8007 (POM 287809). Oxypolis rigidior (L.) Raf. — Canada, Ontario, Essex Co., 5.3 km E & 5.7 km N of Leamington, Oldham 6994 (CAN 522163). — U.S.A., Indiana, Tippecanoe Co., Ross Biological Reserve, 2 Oct. 1958, Webster & Webster 7206 (DUKE 147518). — U.S.A., Tennessee, Munroe Co., Cherokee N.F., 21 Sept. 1938, M. Shaver s.n. (UGA 79326). Oxypolis ternata (Nutt.) A. Heller — U.S.A., Florida, Franklin Co., near Wright Lake, Apalachicola N.F., 12 Nov. 1969, Godfrey 69254 (USF 88592). — U.S.A., North Carolina, Lee Co., 0.3 mi E of jct. of 1176 and 1179 on 1176, 17 Oct. 1967, Bozeman 11639 (NCU 303709). Ptilimnium ahlesii Weakley & G.L. Nesom — U.S.A., Georgia, Chatham Co., Savannah National Wildlife Refuge, along Hwy 17, 14 July 1966, Bozeman 6100 (NCU 339364). — U.S.A., South Carolina, Beaufort Co., Trichinham Plantation, 27 June 1956, Bell 3767 (NCU 97910). Ptilimnium capillaceum (Michx.) Raf. — U.S.A., Louisiana, Vermillion Pa., Pecan Island, 28 June 1963, Valentine s.n. (SMU). — U.S.A., South Carolina, Clarendon Co., at St. Paul, J. Nelson 13280 (USCH 58378). Ptilimnium costatum (Elliott) Raf. — U.S.A., Illinois, Washington Co., 25 Sept. 2001, Feist s.n. (ILLS). U.S.A., Oklahoma, Pushtamaha Co., Antlers, 23 Oct. 1915, Palmer 8989 (MO 793414). Ptilimnium nodosum (Rose) Mathias — U.S.A., South Carolina, Aiken Co., Janet Harrison High Pond, 11 May 2005, Feist & MolanoFlores 2967.1 (ILLS). — U.S.A., West Virginia, Morgan Co., along Sleepy Creek, 23 Aug. 2005, Feist 3188 (ILLS) — U.S.A., Alabama, Tuscaloosa Co., North River, 15 mi N of Tuscaloosa, Easterly Ala. 146 (WVA114892). Ptilimnium nuttallii (DC.) Britton — U.S.A., Texas, Kaufman Co., 1.75 mi E of Terrell, 14 June 1946, Cory 53275 (SMU). — U.S.A., Kansas, Labette Co., Mound Valley, 19 July 1995, Freeman 7322 (WVU 114956). Ptilimnium texense J.M. Coult. & Rose — U.S.A., Texas, Freestone Co., 14.5 mi S. of Fairfield, 2 Oct. 1949, Shinners 11830 (SMU). — U.S.A., Texas, Tyler Co., ca. 2 mi E of Warren, 5 Oct. 1969, Correll 3813 (LAF 46206). 418

RELATED PAPERS

RELATED TOPICS

Log In

Revised generic delimitations for Oxypolis and Ptilimnium (Apiaceae) based on leaf morphology, comparative fruit anatomy, and phylogenetic analysis of nuclear rDNA ITS and cpDNA trnQ -trnK intergenic spacer sequence data

Revised generic delimitations for Oxypolis and Ptilimnium (Apiaceae) based on leaf morphology, comparative fruit anatomy, and phylogenetic analysis of nuclear rDNA ITS and cpDNA trnQ -trnK intergenic spacer sequence data

Related Papers

RELATED PAPERS

RELATED TOPICS