HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (21) Citing Articles via Google Scholar Google Scholar Articles by Kelly, LawrenceM. Search for Related Content PubMed Articles by Kelly, LawrenceM. Agricola Articles by Kelly, LawrenceM.

Phylogenetic relationships in Asarum(Aristolochiaceae) based on morphology and ITSsequences¹

^a L. H. BaileyHortorium, 462 Mann Library, Cornell University, Ithaca, New York14853

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
A cladistic analysis of Asarum was conducted using data frommorphology and nuclear ribosomal internal transcribed spacer sequences(ITS) to examine the relationships among several groups of taxa thathave often been treated as segregate genera. Morphological andmolecular data were analyzed simultaneously for a set of 36 terminalsrepresenting the taxonomic, morphological, and geographic diversity ofthe genus. The two data sources are generally congruent and togetherprovide more resolution than either by itself. The data support twomain clades in the genus: an Asarum sensu stricto clade and anAsiasarum + Hexastylis +Heterotropa clade. The former consists of ~17 species andis distributed through North America, Europe, and Asia. Within thisgroup, the North American species are monophyletic and derived fromwithin the paraphyletic Asiatic species group. The Asiasarum+ Hexastylis + Heterotropa clade consists oftwo Asiatic segregates and the North American segregate,Hexastylis. Resolution within this group supports both themonophyly of Heterotropa and a sister group relationship ofAsiasarum to Hexastylis + Heterotropa. Hexastylis is paraphyletic and occurs as two separate clades onthe tree; one of these is sister to Heterotropa and the otheris sister to the latter plus Heterotropa. The phylogeneticdata provide several clues about the biogeographic history ofAsarum and suggest that: (1) the genus likely originated inAsia and underwent substantial diversification prior to colonizing NorthAmerica; and (2) Asarum in North America represents at leasttwo historically distinct groups that likely achieved their NorthAmerican distributions at different times. The results of theseanalyses support recognition of two subgenera, Asarum andHeterotropa, each with two sections: Asarum andGeotaenium of the former, and Asiasarum andHeterotropa of the latter. Asarum sect.Ceratasarum (Hexastylis) is here treated as a synonymof Asarum sect. Heterotropa. A taxonomic conspectusof the genus is provided, and the combination Asarum sect.Geotaenium (F. Maek.) L. Kelly is made in accordance with therevised circumscription oftaxa.

Key Words: Aristolochiaceae • Asarum • Asiasarum • cladistics • Geotaenium • Heterotropa • Hexastylis • ITSsequence data • phylogeny

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Asarum consists of ~85 species of low-growing,rhizomatous herbs. The genus is exclusively north temperate indistribution and is most diverse in Asia (~65 species), but also hasrepresentatives in eastern and western North America (~15 species),and a single species in Europe. Asarum is well supported asmonophyletic and is distinguished from other genera of Aristolochiaceaeby several characters including determinate seasonal growth, simple leafbases, cataphylls, preformed flowers and leaves, and fleshy, irregularlydehiscent fruits (Kelly, 1997). WithinAristolochiaceae, Asarum has traditionally been aligned withSaruma in subfamily Asaroideae based on a host of sharedcharacters, including herbaceous habit, similar seed morphology, andactinomorphic flowers that are solitary and terminal (e.g., Schmidt, 1935; Duchartre, 1864; Gregory, 1956; Huber,1993). However, more recent analyses have suggested that manyof these characters are plesiomorphic and that Asarum is moreappropriately placed as sister to the Aristolochioideae(Aristolochia and Thottea; Loconte and Stevenson, 1991).

The taxonomic history of Asarum is one of disagreement onthe definition of the genus. While some floristic and revisionaltreatments recognized a single genus, other authors have defined generamore restrictively and admitted as many as six segregates. Of thesesegregates, Heterotropa (~50 species), Asiasarum,and Geotaenium (each with 3–4 species) are exclusivelyAsiatic; Hexastylis (8–9 species) is entirely NorthAmerican; and Asarum sensu stricto (s.s.) (~17 species) isthe most widely distributed, with species in North America, Europe, andAsia (primarily China). Characters that have been used to distinguishthese genera include chromosome number (Maekawa,1933, 1963; Maekawa and Ono,1965; Tanaka, 1935; Yuasa and Maekawa, 1976; Sugawara, 1981) and several aspects of floralmorphology, such as degree of fusion of the sepals, ovary position, andstamen, style, and stigma morphology (Maekawa,1978; Sugawara, 1987).

The majority of authors have recognized Asarum as a singlegenus. This system was adopted in the most comprehensive revisions ofthe genus (Araki, 1937, 1953), in recent treatments of Chinese andTaiwanese species (Cheng and Yang, 1983;Huang, Hsieh, and Huang, 1995), and forthe Flora of Japan (Ohwi, 1965). Incontrast, recognition of the segregate genera has been advocated in Asiaprimarily by a single author (Maekawa,1933, 1953, 1978), and in the United States, two genera,Asarum and Hexastylis, have been recognized followingBlomquist's influential (1957)treatment. Unfortunately, Blomquist did not study variation in any ofthe Asiatic members of Asarum and his recognition ofHexastylis was based on its distinctness from North Americanspecies of Asarum s.s. Recently, Barringer (1993) recommended against therecognition of Hexastylis, but the genus was neverthelessadopted for the Flora of North America (Whittemore and Gaddy, 1997). Questions onthe status of Hexastylis and the other segregates, and thepersistence of these segregates in the literature (e.g., Gaddy, 1987; Yamaki et al.,1996), highlight the need for a broad phylogenetic study ofAsarum.

In a recent paper (Kelly, 1997), Iinitiated such a study and incorporated morphological data into aphylogenetic analysis of Asarum to examine morphologicalsupport for relationships of the segregates. The primary conclusions ofthis work were that Asarum is monophyletic and that the genus,in the broad sense, consists of two clades: an Asarum s.s.clade (including Geotaenium), which is characterized by connatestyles and inferior ovaries, and an Asiasarum +Hexastylis + Heterotropa clade, which ischaracterized by ridges on the inner sepal surface, dorsal stigmas, andbifid style extensions. Resolution within the Asiasarum +Hexastylis + Heterotropa clade further supportedboth a sister group relationship between Asiasarum andHexastylis + Heterotropa and the monophyly ofHeterotropa. However, despite an abundance of charactersupport for these larger groups in the genus, morphological data yieldedvery little reliable resolution within these groups. This lack ofresolution resulted from the small number of qualitative morphologicalcharacters that separate closely related species. Given theselimitations in the morphological data, the best approach was deemed theintegration of a new source of data, with the goal of providing a betterresolved, more robust phylogenetic hypothesis for Asarum.

The internal transcribed spacer region (ITS) of nuclear ribosomal DNAis a good candidate for extending phylogenetic studies ofAsarum. ITS has been the most widely sequenced gene in studiesof relationships within and among closely related genera of floweringplants, and the gene has contributed much toward our understanding ofangiosperm systematics. In some cases, ITS has provided improvedresolution within genera where morphology and/or other DNA datafailed to yield informative variation (e.g., Wojciechowski et al., 1993; Baldwin and Robichaux, 1995). Thecontributions of ITS, as well as the overall properties and potentialpitfalls of the gene, are reviewed by Baldwin etal. (1995). In this study, I integrate ITS sequence data withthe morphological data to shed further light on phylogeneticrelationships in Asarum.

This combined ITS and morphological analysis of Asarum aimsto address questions that remain unanswered because of limitedresolution in the morphological data (Kelly,1997). A primary focus of this analysis is on therelationships of Hexastylis. The fundamental questionunaddressed by the morphological data is whether this North Americansegregate is monophyletic, which is obviously central to the issue ofHexastylis as a distinct genus. A second question involves therelationships within Asarum s.s., which were also poorlyresolved by the morphological data. Asarum s.s. is the onlysegregate to span the northern continents, and the relationships amongthe North American, the Asiatic, and the single European species of thisgroup are still unknown. Tepliakova(1985), Cheng and Yang (1983),and Araki (1953) all recognized sectionsthat imply transcontinental species relationships in Asarums.s., which is not unexpected given Asian–North American speciespairs in a large number of both woody and herbaceous genera (e.g.,Li, 1952). However, in Asarumthese taxa were based primarily on sepal morphology, and given thatthere is limited morphological variation within Asarum s.s.(Kelly, 1997), these hypotheses must beconsidered tentative.

Any further resolution of relationships within Asarum willhelp provide a clearer picture of the biogeographic history of thegenus. Relatively few cladistic analyses are available for northtemperate plant genera, and Asarum is of potential interest asan herbaceous taxon with limited dispersal ability, representative of apotentially ancient group. Considering herbaceous angiosperms of northtemperate regions, it is most likely rhizomatous perennials of theforest floor association that were among the original components of theboreotropical flora and thus were present when the northern continentswere well connected via the North Atlantic and Bering land bridges(Tiffney, 1985a). A well-resolvedphylogenetic hypothesis for Asarum can be extended to addressthe geographic origin of the genus, to assess patterns ofinterconnections and floristic interchanges between the northerncontinents, and to examine concordance in the data with what is known ofthe geological history of Laurasia.

Finally, as an additional source of data, ITS sequences are valuableas potential corroboration for the conclusions supported by themorphological data. This analysis is thus a further test of themonophyly of the Asarum s.s. and Asiasarum +Hexastylis + Heterotropa clades and of theresolution (Asiasarum(Hexastylis Heterotropa)). Anyadditional support from the ITS data will serve as an indication ofgreater robustness of these clades.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Taxon sampling
Sampling of taxa for this analysis was designed for complete overlapbetween morphological and ITS data matrices. Representation of speciesis thus nearly identical to the morphological analysis, which includeddiverse representation of all segregates, with multiple species of thelarger and geographically diverse Asarum, Hexastylis,and Heterotropa. However, there are a few minor differences insampling due to the incorporation of ITS data. First, to test morerigorously the monophyly of Hexastylis, an additional speciesfrom that group, Asarum minor, has been added to both datamatrices. Second, to examine the degree and/or effects ofinfraspecific ITS polymorphism, this analysis includes multipleaccessions from four of the more variable and widespread species:Asarum caulescens, A. sieboldii, A.minor, and A. arifolium. Finally, onespecies that was included in the morphological analysis, A.petelotii, was not included in this combined analysis becauseITS was not successfully sequenced. Outgroup relevance is also affectedby the inclusion of ITS data. Among the outgroup taxa included in themorphological analysis, only Saruma had ITS sequences that werereadily alignable with those of Asarum, whereas the othergenera were either highly divergent from Asarum(Aristolochia and Pararistolochia) or due to technicaldifficulties or practical limitations were not available for molecularstudy (Thottea). As a result, ITS data are included in thematrix only for Saruma of the outgroup; however, the otheroutgroup genera are included in the simultaneous analysis because ofpotential morphological relevance. A complete list of the speciessampled, along with collection data and voucher information, is providedin Table1.

DNAisolation, amplification, and sequencing
Leaf material was collected fresh and either used immediately orpreserved on silica gel or Drierite desiccant. Total DNA was isolatedusing the CTAB method (Doyle and Doyle,1987). Double-stranded PCR (polymerase chain reaction)amplifications of the entire ITS region were done using ITS5 and ITS4primers (White et al., 1990). Thetemperature profile for amplification consisted of initial denaturationat 97°C (1 min), followed by 39 cycles of 97°C denaturation (1min), 48°C annealing (1 min), and 72°C extension (3 min), with afinal extension for 7 min at 72°C. PCR products were run out in a1% agarose TAE gel and purified using Wizard(TM) preps(Promega, Madison, WI). Direct sequencing of gel purifieddouble-stranded PCR products was accomplished using the snap-chillmodification of the dideoxy chain termination method (Nickrent, 1994) with the use of the Sequenaseversion 2.0 kit (US Biochemical, Cleveland, OH). Sequencing primerswere ITS5, ITS4, ITS2, and ITS3 (White et al.,1990). All ITS sequences included in this study have beendeposited in Genbank under accession numbers GBANAF061499-GBANAF061502and GBANAF061534-GBANAF061553.

Data analysis
Sequence alignments were performed using the multiple alignmentcomputer program Malign (Wheeler and Gladstein,1994), which uses parsimony as an optimality criterion toselect among possible alignments. Sequence distance matrices and GCcontent were calculated for the ITS data using PAUP (Swofford, 1991) and DNASTAR (DNASTAR, Inc.,Madison, WI), respectively. Data matrices were assembled and editedusing Dada (Nixon, 1996), and parsimonyanalyses were conducted using Nona version 2.1 (Goloboff, 1994). Nona runs were executed usingthe mult*100, hold/20 commands, which generate 100 random taxonentry sequences, with TBR swapping on up to 20 starting trees periteration. The resulting trees were then subjected to TBR swapping tocompletion using the max* command. This process—100 multreplications followed by max*—was repeated 25 times. Allanalyses were run using the default option amb-, which collapses all butunambiguously supported dichotomies. Character distributions on treeswere studied using ClaDOS (Nixon,1996).

Data from ITS and morphology were combined and analyzedsimultaneously, allowing parsimony to maximize congruence over bothsources of data (see Nixon and Carpenter,1996, for discussion). In addition, several subanalyses wereconducted to examine stability in the data and sensitivity of theresults to different treatments of the characters. Morphological andITS data were analyzed separately to examine the topology supported byeach data set independently. For the ITS data, the effects ofalignments and different treatments of gaps were examined for theirinfluence on the resulting phylogenetic hypothesis. Alignmentparameters were manipulated such that the ratio of gap cost to changecost was varied through a broad range. Different ways of coding gaps inthe data matrix were explored, including gaps as missing values, gaps asa fifth character state, and gaps condensed to single characters andappended to the nucleotide change matrix. The examination of theseresults provides an indication of which phylogenetic hypotheses dependentirely on a very restricted interpretation (or subset) of thecharacter data as opposed to those that are supported more broadly andunder a wide range of alternative means of analysis. This knowledge isuseful to dictate the amount of confidence warranted in extending theresults of the phylogenetic analysis to study the biogeography,evolution, and taxonomy of the genus.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
ITSlength, variation, and GC content
Aligned ITS sequences of Asarum are shown in Fig. 1. Sequences from adjacent codingregions, including the 5.8S, were found to be invariable; thus they arenot included in Fig. 1 and arenot considered in the following results and discussion. ITS sequencesof Asarum vary in length from 254 to 258 bp in ITS 1 and from219 to 228 bp in ITS 2. Pairwise distances in the genus range from 0 to7.7% in ITS 1 and from 0 to 9.6% in ITS 2. ITS sequencesof Saruma are readily comparable with those of Asarum,with pairwise divergence (uncorrected) between the two genera varyingfrom 7.9 to 9.6% for ITS 1 and from 7.4 to 13.7% for ITS2. GC content of the species ranges from 46.67 to 53.33% in ITS1 and from 46.49 to 56.62% in ITS 2. Lengths of the alignedsequences are 270 for ITS 1 and 241 for ITS 2 and the number of sitesthat are potentially phylogenetically informative are 34 (14%)for ITS 1 and 42 (17.4%) for ITS 2. The lengths, relativelengths, and divergence values observed for Asarum ITSsequences fall well within the range reported for other angiospermgroups (Baldwin et al.,1995).

View larger version (54K): [in this window] [in a new window]   Fig. 1. Aligned ITS sequences of Asarum and Saruma . Gaps are indicated by a "-".

View larger version (40K): [in this window] [in a new window]   Fig. 2. One of two equally most parsimonious trees resulting from analysis of the ITS data alone. Length = 175 steps, CI = 0.73, and RI = 0.89. For this analysis, gaps were assigned missing values in the data matrix. Solid hashmarks represent unique origins of states; gray hashmarks represent states not uniquely derived under the optimization presented. The North American taxa are indicated. The two nodes indicated by open arrows do not occur in the strict consensus tree.

The topology of Fig. 2(strict consensus) is stable across a broad range of alignmentparameters, indicating that alignment ambiguity of Asarum ITSsequences is of little concern. This result is hardlysurprising—inspection of the ITS sequences indicates that there isrelatively little length variation across taxa (Fig. 1). This is borne out by morerigorous analyses of the data in which change-cost to gap-cost ratioswere varied through a range of 1:1 to 1:8: all parameter values withinthis range yield nearly identical alignments, with roughly the samenumber and distribution of gaps, and identical topological support asindicated by comparing the strict consensus trees of each analysis (datanot shown).

Different coding of gaps in the phylogenetic analysis, however, doesyield minor differences in topology. These differences affect only theplacement of Asiasarum and Geotaenium. Asiasarum is supported as the sister group to theHexastylis + Heterotropa clade in every analysisin which gaps are included in the data matrix, no matter how they arecoded; Asiasarum is unresolved if gaps are excluded. Primarysupport for Asiasarum as sister to Hexastylis andHeterotropa is derived from a 9-bp deletion in ITS 2 (Fig. 1; positions 296–304):inclusion of this gap as a single binary character in the data matrixresolves the Asiasarum + Hexastylis +Heterotropa relationship. The position of Geotaeniumis more uncertain and variable, depending on the treatment of gaps. When gaps are excluded, Geotaenium is placed equallyparsimoniously as sister to Asarum s.s. or as sister to allother species of Asarum s.l. Most of the analyses in whichgaps are included support the latter placement ofGeotaenium.

In summary, a thorough series of analyses indicates that therelationships supported by ITS are not affected by alignment ambiguityand that the inclusion of gaps only slightly affects the topology byproviding more resolution than when gaps are excluded. The results thatare consistently supported by ITS and unaffected by the coding of gapsare the following: (1) Asarum s.s. (excludingGeotaenium) is monophyletic and the resolution within thatgroup, as described above, does not change; (2) Heterotropa ismonophyletic; (3) Hexastylis is paraphyletic; and (4)Hexastylis + Heterotropa is monophyletic. Therelationships of Asiasarum and Geotaenium are moreuncertain based on the results of all of theseanalyses.

Simultaneous analysis:ITS and morphology
Simultaneous analysis of ITS and morphological data resulted in 288trees of 249 steps, CI = 0.69, and RI = 0.90. One ofthese trees is shown in Fig. 3,and the strict consensus of all 288 trees is shown in Fig. 4. The strict consensus isgenerally similar to the tree supported by ITS data alone (Fig. 2), especially in its support forboth Asarum s.s. and Hexastylis + Heterotropa. Furthermore, the resolution withinAsarum s.s. and of the relationships (i.e., paraphyly) ofHexastylis is as supported by ITS alone. However, simultaneousanalysis provides better overall resolution than ITS alone, especiallyof the relationships of Asiasarum and Geotaenium. Asiasarum is supported as sister to Hexastylis +Heterotropa, and Geotaenium is supported as sister toAsarum s.s. Both of these relationships were supported insome, but not all, analyses of the ITS data alone. Table 2 summarizes the results overall of the analyses.

View larger version (42K): [in this window] [in a new window]   Fig. 3. One of 288 equally most parsimonious trees from the simultaneous analysis of ITS and morphological data. Length = 249 steps, CI = 0.69, and RI = 0.90. For this analysis, gaps were assigned missing values in the data matrix. Hashmarks are as in Fig. 2 .

View larger version (43K): [in this window] [in a new window]   Fig. 4. The strict consensus tree of the 288 equally most parsimonious trees from the simultaneous analysis of ITS and morphological data. Taxon areas are indicated to the right. The 9-bp deletion is mapped on the tree. Sepal morphology is mapped for species of Asarum s.s.; the characters shown were not included in the morphological analysis because of continuous variation. Mapping indicates that the extremes of variation have been independently derived in North America and in Asia.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In addition to concordance among data sets and stability to differenttreatments of the data, other aspects of the results of this analysisprovide some reassurance of robustness. First, the genus is wellsampled and none of its major geographic or taxonomic units have beenomitted from this study. Second, geographically distinct accessions ofthe same species in all cases but one come out as sister terminals. Theexception, A. minor, is paraphyletic based on a singlehomoplastic character. Thus infraspecific polymorphism of ITS sequencesin Asarum does not appear to be a problem. Finally, of all themajor groups supported by the simultaneous analysis, even those that aresupported by only a few characters, none is seriously contradicted bythe available data, morphological or molecular. Thus while theavailability of only a few characters may limit our confidence in somegroups, these groups nevertheless represent the best hypotheses givenall of the data. Following is a discussion of the major groupssupported by all of the data (i.e., in the simultaneous analysis), withan indication of the conclusions that are warranted given the samplingfor this analysis.

Asarum s.s.
This group consists of 15 species from North America, Europe, andAsia (primarily China). In addition to five nucleotide substitutions,the group is supported by several morphological synapomorphies,including inferior ovaries, connate styles, pubescent inner calyxsurface, and well-developed anther connectives (Kelly, 1997). Although a strongly supportedclade, most of the morphological variation within Asarum s.s.is continuous, and the paucity of potential cladistic characters hasmade it difficult to infer relationships in the group. Araki (1953), Cheng and Yang(1983), and Tepliakova(1985) all divided Asarum s.s. into sections basedprimarily on sepal characters and distinguished species withlong-caudate sepals from those with short, reflexed sepals, whichimplies transcontinental species relationships. My previous analysis ofmorphological data resulted in similar groupings, but support for thesegroupings was so weak (one character) that one could not assume thisresolution was reliable (Kelly, 1997). In contrast, the ITS and combined analyses unambiguously support NorthAmerican species as monophyletic and derived from within theparaphyletic, Asiatic species group. ITS sequences of the NorthAmerican species are nearly identical (0–0.6% sequencedivergence), whereas sequence divergence between North American andAsiatic species and among the Asiatic species is considerably higher(1.4–4.4% and 0.8–4.6%, respectively). Overall, this resolution within Asarum s.s. indicates thatsimilar morphological features (e.g., small flowers with reflexed sepallobes, large flowers with caudate sepal lobes, variegated leaves,evergreen leaves) have arisen independently on two continents (Fig. 4).

The topology supported by the simultaneous analysis also suggests areappraisal of the homology of synsepaly in Asarum s.s. Basedon morphology alone, all synsepalous species of Asarum s.s.were weakly supported as basal within the clade, and sharing theplesiomorphic condition of synsepaly. However, the integration of ITSdata provides a strong indication of the secondary derivation ofsynsepaly in the clade consisting of Asarum debile andA. caudigerellum and suggests that these species arenot closely related to A. epigynum (the othersynsepalous species of Asarum s.s.). Nevertheless, the resultsof the simultaneous analysis do not change the conclusion thatchorisepaly is derived from synsepaly in Asarum, nor theconclusion that synsepaly evolved before epigyny in the genus (Kelly, 1997).

Geotaenium
This Asiatic segregate consists of three species (Sugawara, Ogisu, and Cheng, 1990), and one ofthese, Asarum epigynum, was sampled in this analysis. WhileITS data are somewhat ambiguous as to the placement ofGeotaenium, the results of the simultaneous analysis supportthis taxon as sister to the Asarum s.s. clade. Morphologicalsynapomorphies that unite Geotaenium with Asarum s.s.include adaxial sepal trichomes, inferior ovaries, connate styles, andwell-developed anther connective extensions (Kelly,1997). Geotaenium has generally been considered tobe most closely related to Asarum s.s. regardless of whether itwas recognized as a distinct genus or not (e.g., Cheng and Yang, 1983; Sugawara, 1987). All three species ofGeotaenium are very much alike morphologically, with small,slightly zygomorphic, synsepalous, pubescent calyces, and all would becoded identically for all characters in the morphological data matrix. This group is putatively monophyletic based on a chromosome number of2n = 12, which is unique in Asarum (and withinthe Aristolochiaceae).

Asiasarum
This segregate comprises four species (Yamakiet al., 1996), one of which was sampled in this analysis. Inthe past, Asiasarum has been allied with Asarum s.s.(Schmidt, 1935) or withHexastylis and Heterotropa (Araki, 1937, 1953), and the segregate combines themorphological characteristics of the two groups. The simultaneousanalysis supports placement of Asiasarum as sister to theHexastylis + Heterotropa clade. Furthermore,outgroup comparison and optimization of the morphological charactersresolves the apparent conflict in morphology: Asiasarum hassynapomorphies of the Hexastylis + Heterotropagroup (e.g., bifurcate styles, dorsal stigmas, glabrous abaxial calyxsurface, and raised ridges of the calyx) and much of its similarity toAsarum s.s. is plesiomorphic (e.g., chromosome number2n = 26, long stamen filaments, and paired leaves)(Kelly, 1997). The shared possession ofa 9-bp deletion in ITS 2 (Fig.4) is here taken as strong support for an Asiasarum+ Hexastylis + Heterotropa relationship. All four species of Asiasarum are alike in their morphology andwould be coded identically in the morphological data matrix. Thespecies of this group all have pronounced longitudinal ridges on theinner calyx surface and completely lack transverse ridges; based on thischaracter, Asiasarum is potentiallymonophyletic.

Hexastylis
This segregate consists of nine species (Blomquist, 1957), five of which were sampledfor this analysis. Members of the group possess no identifiablemorphological synapomorphies (Kelly,1997), and the genus has been recognized solely on the basisof its distinctness from North American Asarum s.s. Morphological cladistic analysis strongly indicates that thecharacteristics that "unite" species of Hexastylisare symplesiomorphic for the group; however, the morphological data donot resolve whether Hexastylis is, in fact, positivelyparaphyletic (Kelly, 1997). Hexastylis species occur as two separate clades in all analysesthat include the ITS data. In the context of Blomquist's (1957) three species groups,one clade represents the Virginica group and the other corresponds tothe union of the Arifolia and Speciosa groups. ITS sequences withineach of these two clades are nearly identical (0–0.8%sequence divergence), whereas differences between the clades are greater(2.9–3.8% sequence divergence). Simultaneous analysis ofthe data supports the following resolution of these clades:(Speciosa/Arifolia (Virginica + Heterotropa)). Theobvious implication is that some species of Hexastylis are moreclosely related to Asiatic Heterotropa than they are to otherHexastylis.

Heterotropa
This is a large and taxonomically complex group, with ~55 speciesof exclusively Asiatic distribution. The status of Heterotropawithin the Asarum s.l. complex has been uncertain, with someauthors regarding the group as basal (e.g., Tepliakova, 1985), while others considered itderived (e.g., Maekawa, 1953). Heterotropa is well nested within the Asiasarum +Hexastylis + Heterotropa clade and stronglysupported as monophyletic. Morphological synapomorphies ofHeterotropa include glandular trichomes on the inner surface ofthe calyx, the orifice ring on the calyx, and a chromosome number2n = 24 (Kelly, 1997). In addition, Heterotropa is supported by a large number ofsynapomorphic nucleotidesubstitutions.

Biogeography
Asarum exhibits a classic Laurasian distribution, as a northtemperate genus with species in eastern and western North America,Europe, and Asia. The geological history of the northern continents isreasonably well known (reviewed by Wolfe,1975; McKenna, 1983; Tiffney, 1985a, b), and the timing of some of the mostimportant geologic events coincided with the rise and spread ofangiosperms. Many modern angiosperm taxa appeared in the lateCretaceous, and through the early Eocene spread across the northerncontinents as components of a boreotropical flora (Tiffney, 1985a). It was during this timeperiod that many taxa achieved their current distributions, and severalcritical factors influenced these distributions. First, although theopening of the North Atlantic and North Pacific Oceans and the breakupof Laurasia were initiated long before the origin of the angiosperms,the potential for floristic exchange among the northern continents wassignificant through much of the Tertiary. This is due to the presenceof a direct North Atlantic land connection until the early Eocene, andto the Bering land connection, which was likely a viableintercontinental route in the later Tertiary (Wolfe, 1975). The relative importance of theseland connections is uncertain, but there is little doubt that theyaccount for much of the floristic similarity among the northerncontinents. Second, in the late Cretaceous, the northern continentswere divided into two floristic provinces: Europe + eastern NorthAmerica, and Asia + western North America. The boundaries of theseprovinces coincided with the presence of the Turgai straits in Eurasiaand the Transcontinental Sea in North America. The primary results ofthese geologic events are complex patterns of area relationships alongwith considerable potential for intercontinental floristic exchange. For example, the history of North America has included subdivision ofthe continent and both eastern and western links to Eurasia. As aresult, North America does not necessarily represent a unifiedbiogeographic entity, and different regions may be expected to havedifferent historical influences.

Asarum was likely present as an early Tertiary component ofa boreotropical flora, and may have been affected by any of theaforementioned events. The following lines of evidence support an earlyorigin for Asarum. (1) Aristolochia, a derived memberof the sister group of Asarum, is recorded in the fossil recordfrom the middle Eocene (MacGinitie,1969). (2) Of all north temperate herbs, it is most likelyrhizomatous members of the forest floor association, particularly those(such as Asarum) adapted to low light, stable environments,that would be among the herbaceous taxa most expected to have been anoriginal component of the boreotropical flora (Tiffney, 1985a). (3) Asarum has poorlong-distance dispersal capabilities, with ant-dispersed seeds that arehighly prone to desiccation.

Optimization of areas onto the cladogram of Asarum (Fig. 4) supports the followingconclusions on the biogeographic history of the genus. (1)Asarum originated in Asia and subsequently expanded its rangeinto other areas. This was suggested previously by others, includingMaekawa (1963), based on Asia as theapparent center of diversity of the genus. Two lines of evidence,interpreted under the assumption of the progression rule (Hennig, 1966), coincide to support an Asiaticorigin for Asarum. First, optimization of areas results in theunambiguous assignment of Asia to the basal ingroup node, and the basalmembers of both major clades of Asarum are Asiatic (Fig. 4). Second, the basal members ofboth of the immediate outgroups to Asarum (Saruma andThottea), are Asiatic (Fig.4). (2) A vicariance explanation for the distribution ofAsarum would demand that the genus had diversified, at least interms of the divergence of the two major clades, prior to thefragmentation of the broad ancestral range of the genus. This makes itlikely that Asarum had not only originated, but had undergonesubstantial diversification, prior to the early Eocene. (3) NorthAmerican species of Asarum are not monophyletic. The genusconsists of at least two geographic groups in North America. Thesegroups have separate histories and likely achieved their North Americandistributions at different times. The first of these groups isHexastylis, of eastern North America, which itself is notmonophyletic. The relationships of eastern North America in this partof the cladogram are with Asia. Furthermore, the paraphyly of easternNorth America with respect to Asia is a possible indication of anextended connection between the two areas, one that was sustainedthrough the divergence of lineages of Asarum. This is alsoconsistent with this area relationship being relatively earlier in thehistory of the genus. Another North American clade is derived fromwithin Asarum s.s. (Fig.4). This monophyletic group shows its relationships withEuropean and Asiatic areas. All but one species of this clade arewestern, but the single eastern species, A. canadense,is potentially basal in the group. This, along with a close (butunresolved) relationship of North American, European, and two westernChinese species, is suggestive of a North Atlantic crossing withinAsarum s.s., although a Beringian connection cannot be ruledout. The maximum divergence values within the North AmericanAsarum s.s. clade are low in comparison to those ofHexastylis (0.6 vs. 3.8%), which suggests either thatITS is not evolving in a clock-like fashion in Asarum or thatAsarum s.s. colonized North America more recently thanHexastylis.

In summary, the biogeographic history of Asarum parallelsthe complex relationships of the northern continents, and nonmonophylyof North American species of Asarum is a further reflection ofthis complexity. Some of the most important implications of thebiogeography of Asarum are taxonomic. The question of thevalidity of Hexastylis and its distinctness fromAsarum has focused primarily on differences between these twospecies groups in North America, without recognition that these groupsmay individually have closer ties with Asiatic taxa (see Araki [1953] and Barringer [1993] for exceptions). Once it is recognized that North American Asarum representsdifferent historical groups, the question of their relationships mustimmediately incorporate a broader context.

Taxonomic conclusions
Here and in my previous paper, I treat Asarum in the broadsense, as a single genus. This system is followed as the mostconservative approach and that which is most consistent with previoustaxonomic work. The alternative of recognizing more than one genuswould require many new nomenclatural combinations. For example, onecould recognize each of the two major clades as a genus and apply thename Hexastylis to one of these (as would be required bypriority); this alone would create the need for ~50 newcombinations. In contrast, most named species are valid combinations inAsarum. Furthermore, recognizing Asarum as a singlegenus highlights the morphological unity of the group withinAristolochiaceae.

The Asarum s.s. clade (including Geotaenium) wasrecognized in my previous paper as Asarum L. subgenusAsarum. The new data further support the recognition of twosections within this subgenus, Asarum sectionGeotaenium (F. Maek.) L. Kelly and Asarum sectionAsarum. The former consists of the three species heretoforereferred to Geotaenium: Asarum epigynum, A.geophyllum, and A. yunnanense, and the latterconsists of the remainder of Asarum s.s. TheAsiasarum + Hexastylis +Heterotropa clade corresponds to Asarum subgenusHeterotropa (Morr. & Decne.) O. C. Schmidt. This subgenusconsists of two sections: Asarum section Asiasarum (F.Maek.) Araki and Asarum section Heterotropa (Morr.& Decne) A. Braun. Circumscription of the latter is here broadenedto include species of Hexastylis, thus Asarum sect.Ceratasarum A. Braun (Hexastylis) should be treated asa synonym of Asarum section Heterotropa. See theAppendix for a conspectusof the taxonomy of Asarum including the placement of allspecies within the taxonomic system.

	FOOTNOTES

 
¹ The author thanks Melissa A. Luckow for consistently providing advice and encouragement, and for reading drafts of the manuscript; Tony Hall and Toby Pennington for generous assistance in obtaining plant material; Ihsan Al-Shehbaz for help with planning fieldwork in China; Tang Ya, Sun Hang, and especially Rob Soreng for invaluable assistance in the field; Jeff Doyle for advice and for reading a draft of the manuscript; Jane Doyle and Helga Ochoterena-Booth for technical assistance; Jim Liebherr, Michael Mesler, and Kerry Barringer for helpful reviews of the manuscript; and Kevin Nixon for help with implementing NONA analyses. This research was supported by NSF grants DEB-9411644 to Melissa Luckow and LMK and DEB-9420215 to Jeff Doyle, and by grants from the Mellon Foundation, the Cornell Graduate School, and the American Society of Plant Taxonomists. This work is part of a doctoral dissertation submitted to the Graduate School of Cornell University.

View larger version (54K): [in this window] [in a new window]   Fig. 1. Continued.

View larger version (37K): [in this window] [in a new window]   Fig. 1. Continued.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

———. 1953. Systemageneris Asari. Acta Phytotaxonomica et Geobotanica XV:33–36.

Baldwin, B. G., M. J. Sanderson, J. M.Porter, M. F. Wojciechowski, C. S. Campbell, and M. J.Donoghue. 1995. The ITS region of nuclear ribosomalDNA: a valuable source of evidence on angiosperm phylogeny. Annalsof the Missouri Botanical Garden 82: 247–277. [CrossRef][ISI]

Baldwin, B. G., and R. H.Robichaux. 1995. Historical biogeography and ecology ofthe Hawaiian silversword alliance: new molecular phylogeneticperspectives. In V. A. Funk and W. L. Wagner [eds.],Hawaiian biogeography: evolution on a hot spot archipelago,259–287. Smithsonian Institution Press, Washington,DC.

Barringer, K. 1993. Newcombinations in North American Asarum (Aristolochiaceae). Novon 23: 225–227.

Blomquist, H. L. 1957. Arevision of Hexastylis of North America. Brittonia 8:255–281.[CrossRef]

Cheng, C.-Y., and C.-S.Yang. 1983. A synopsis of the Chinese species ofAsarum (Aristolochiaceae). Journal of the ArnoldArboretum 64: 565–597.

Doyle, J. J., and J. L.Doyle. 1987. A rapid DNA isolation method for smallquantities of fresh tissues. Phytochemical Bulletin 19:11–15.

Duchartre,P. 1864. Aristolochiaceae. In A. DeCandolle[ed.], Prodromus 15: 421–498.

Gaddy, L. L. 1987. A reviewof the taxonomy and biogeography of Hexastylis(Aristolochiaceae). Castanea 52: 186–196. [ISI]

Goloboff, P. 1994. Nona,version 2.1. Program and documentation. Distributed by the author,Buenos Aires.

Gregory, M. P. 1956. Aphyletic rearrangement in the Aristolochiaceae. American Journal ofBotany 43: 110–122. [CrossRef][ISI]

Hennig,W. 1966. Phylogenetic systematics. University of IllinoisPress, Urbana, IL.

Huang, S.-F., T.-H. Hsieh, and T.-C.Huang. 1995. Notes on the Flora of Taiwan (21)—genus Asarum L. (Aristolochiaceae). Taiwania 40:91–120.

Huber,H. 1993. Aristolochiaceae. In K. Kubitzki, J. G.Rohwer, and V. Bittrich [eds.], The families and genera ofvascular plants, vol. 2, 129–137. Springer-Verlag,Berlin.

Kelly, L. M. 1997. Acladistic analysis of Asarum (Aristolochiaceae) andimplications for the evolution of herkogamy. American Journal ofBotany 84: 1752–1765. [Abstract]

Li, H.-L. 1952. Floristicrelationships between eastern Asia and eastern North America. Transactions of the American Philosophical Society 42:371–429.

Loconte, H., and D. W.Stevenson. 1991. Cladistics of the Magnoliidae. Cladistics 7: 267–296.

Maekawa,F. 1933. Journal of Japanese Botany 9: 39–49,96–103, 174–179, 241–246, 281–285,364–370, 505–512.

———. 1953. Geotaenium. VII Pacific Science Congress, Procedings 5: 217.

———. 1963. Reduction ofchromosomes and major polyploidy: their bearing on plant evolution. Journal of the Faculty of Science, University of Tokyo, Section III:Botany 8: 377–398.

———. 1978. Notes onAsarum and Heterotropa from Taiwan. Journal ofJapanese Botany 53: 289–299.

———, and M.Ono. 1965. Karyotype analysis in the genusHexastylis (Aristolochiaceae). Journal of the Faculty ofScience, University of Tokyo, Section III: Botany 9:151–159.

MacGinitie, H.D. 1969. The Eocene Green River flora of northwesternColorado and northeastern Utah. University of CaliforniaPublications in Geological Sciences 83: 140 pp.

McKenna, M.C. 1983. Cenozoic paleogeography of North Atlantic landbridges. In M. H. P. Bott, S. Saxov, M. Talwani, and J. Thiede[eds.], Structure and development of the Greenland-Scotlandridge, 351–399. Plenum Press, New York, NY.

Nickrent, D. L. 1994. Fromfield to film: rapid sequencing methods for field-collected plantspecies. Biotechniques 16: 470–475. [ISI][Medline]

Nixon, K. C. 1996. Dada andClados program and documentation. Distributed by the author, Ithaca,NY.

———, and J.M.Carpenter. 1996. On simultaneous analysis. Cladistics 12: 221–242. [CrossRef][ISI]

Ohwi, J. 1965. Flora of Japan. Smithsonian Institution Press. Washington, DC.

Schmidt, O.C. 1935. Aristolochiaceae. In A. Engler andK. Prantl [eds.], Die Naturlichen Pflanzenfamilien, 16b,204–242. Verlag Von Wilhelm Engelmann, Leipzig.

Sugawara,T. 1981. Taxonomic studies in Asarum sensulato I. Karyotype and C-banding pattern in Asarum s. str.,Asiasarum and Heterotropa. The BotanicalMagazine, Tokyo 94: 225–238.

———. 1987. Taxonomic studiesof Asarum sensu lato III. Comparative floral anatomy. Botanical Magazine, Tokyo 100: 335–348. [CrossRef][ISI]

Sugawara, T., M. Ogisu, and C.-Y.Cheng. 1990. Asarum yunnanense, a new speciesof Asarum (Aristolochiaceae) from southwestern China. ActaPhytotaxonomica et Geobotanica 41: 7–13.

Swofford, D. 1991. PAUP:phylogenetic analysis using parsimony, version 3.0s. Illinois NaturalHistory Survey, Champaign, IL.

Tanaka, N. 1935. Cytologicalstudies in Asaraceae. Botanical Magazine, Tokyo 49:709–716.

Tepliakova, T.E. 1985. The genus Asarum and its position in thefamily Aristolochiaceae. Botanicheskii Zhurnal 70:1051–1057.

Tiffney, B.H. 1985a. Perspectives on the origin of the floristicsimilarity between eastern Asia and eastern North America. Journalof the Arnold Arboretum 66: 73–94.

———. 1985b. TheEocene North Atlantic land bridge: its importance in Tertiary and modernphytogeography of the Northern Hemisphere. Journal of the ArnoldArboretum 66: 243–273.

Wheeler, W. C., and D. L.Gladstein. 1994. MALIGN, version 1.93. American Museumof Natural History, New York, NY.

White, T. J., T. Bruns, S. Lee, and J.Taylor. 1990. Amplification and direct sequencing offungal ribosomal RNA genes for phylogenetics. In M. Innes, D.Gelfand, J. Sninsky, and T. White [eds.], PCR protocols: aguide to methods and applications, 315–322. Academic Press, SanDiego, CA.

Whittemore, A., and L. L.Gaddy. 1997. Hexastylis. In Flora ofNorth America Editorial Committee [ed.], Flora of NorthAmerica, vol. 3, 54–58. Oxford University Press, New York,NY.

Wojciechowski, M. F., M. J. Sanderson,B. G. Baldwin, and M. J. Donoghue. 1993. Monophyly ofaneuploid Astragalus (Fabaceae): evidence from nuclearribosomal DNA internal transcribed spacer sequences. AmericanJournal of Botany 80: 711–722. [CrossRef][ISI]

Wolfe, J. A. 1975. Someaspects of plant geography of the northern hemisphere during the lateCretaceous and Tertiary. Annals of the Missouri BotanicalGarden 62: 264–279. [CrossRef][ISI]

Yamaki, K., S. Terabayashi, M. Okada, andJ.-H. Pak. 1996. A new species and a new variety ofAsiasarum (Aristolochiaceae) from Korea. Journal ofJapanese Botany 71: 1–10.

Yuasa, H., and F.Maekawa. 1976. Chromosomes of Asarum andHeterotropa (Aristolochiaceae) in the Ryukyu Islands. LaKromosomo II-1: 8–18.

This article has been cited by other articles:

				A. Schnabel, P. E. McDonel, and J. F. Wendel Phylogenetic relationships in Gleditsia (Leguminosae) based on ITS sequences Am. J. Botany, February 1, 2003; 90(2): 310 - 320. [Abstract] [Full Text] [PDF]

				G. D. Weiblen Phylogenetic relationships of functionally dioecious FICUS (Moraceae) based on ribosomal DNA sequences and morphology Am. J. Botany, September 1, 2000; 87(9): 1342 - 1357. [Abstract] [Full Text]

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (21) Citing Articles via Google Scholar Google Scholar Articles by Kelly, LawrenceM. Search for Related Content PubMed Articles by Kelly, LawrenceM. Agricola Articles by Kelly, LawrenceM.