Äîęóěĺíň âç˙ň čç ęýřŕ ďîčńęîâîé ěŕřčíű. Ŕäđĺń îđčăčíŕëüíîăî äîęóěĺíňŕ : http://herba.msu.ru/shipunov/author/pillon2007.pdf Äŕňŕ čçěĺíĺíč˙: Wed Dec 5 14:03:34 2007 Äŕňŕ číäĺęńčđîâŕíč˙: Mon Oct 1 22:05:03 2012 Ęîäčđîâęŕ: Ďîčńęîâűĺ ńëîâŕ: missouri

TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

Evolution and temporal diversif ication of western European polyploid species complexes in Dactylorhiza (Orchidaceae)
Yohan Pi l lon1,4, Michael F. Fay1, Mikael HedrČn2, R ichard M. Bateman1, Dion S. Devey1, Alexey B. Shipunov1, 5, Michel le van der Bank1,4 & Mark W. Chase1
1

2

3

4

5

Jodrell Laborator y, Royal Botanic Gardens, Ke w, Richmond, Sur re y TW9 3DS , U.K. m.chase@ke w.org (author for cor respondence) Depart ment of Ecolog y, Plant Ecolog y and Systematics, Universit y of Lund, SĆlvegatan 37, 223 62 Lund, Sweden Depart ment of Botany and Plant Biotechnolog y, Universit y of Johannesburg, APK Campus, PO BOX 524, Auckland Park , 2006, South Af rica Cur rent address: Laboratoire de botanique, Institut de Recherche pour DČveloppement, UniversitČ de la Nouvelle CalČdonie, BP A5, 98848 NoumČa Cedex, Ne w Caledonia Cur rent address: Depart ment of Forest Resources, Universit y of Idaho, Moscow, Idaho 83844, U.S .A.

Pat ter ns of poly ploid evolution in the t axonomically cont roversial Dact ylorhiza incarnata/maculata groups were infer red genetically by analyzing 399 individuals f rom 177 localities for (1) fou r poly mor phic plastid regions yielding agg regate haplot y pes and (2) nuclear r ibosomal ITS allele f requencies. Concordance bet ween pat ter ns obser ved in dist r ibutions of plastid haplot y pes and ITS alleles renders ancest ral poly mor phism an u nli kely cause of genetic var iation in diploids and allopoly ploids. Combining the deg ree of concer ted evolution in ITS alleles (thought to ref lect gene conversion) with infer red parent age provides suppor t for a quad r ipar tite classif ication of wester n Eu ropean allopoly ploid dact ylorchids according to thei r respective parent age and relative dates of or igin. T he older allotet raploids that generally ex hibit only one parent al ITS allele can be divided into those der ived via hybr idization bet ween the divergent complexes we now call D. incarnata s.l. and D. fuchsii (e.g., D. majalis) and those der ived via hybr idization bet ween D. incarnata s.l. and D. maculata (e.g., D. elata). Similarly, the you nger allotet raploids that maint ain evidence of both parent al ITS alleles can be divided into those der ived f rom hybr idization bet ween D. incarnata s.l. and D. fuchsii, or perhaps in some cases a diploid species resembling D. saccifera (e.g., D. praeter missa, D. purpurella, D. t raunsteineri s.l., D. baltica), and those der ived f rom hybr idization bet ween the D. incarnata s.l. and D. maculata g roups (e.g., D. occidentalis, D. sphag nicola). Older allotet raploids are infer red to have passed th rough glacially induced mig ration bot tlenecks in souther n Eu rasia, whereas at least some you nger allotet raploids now occupying nor ther n Eu rope are infer red to have or iginated post-glacially and remain sy mpat r ic with thei r parents, a scenar io that is largely in ag reement with the mor pholog y and ecolog y of these allotet raploids. ITS conversion is in most cases biased toward the mater nal parent, event ually obscu r ing evidence of the or iginal allopoly ploidization event because plastid haplot y pes also ref lect the mater nal cont r ibution. Gene f low appears u nexpectedly low among allotet raploids relative to diploids, whereas several mechanisms may assist the gene f low obser ved across ploidy levels. T here is good concordance bet ween (1) the genetically delimited species that are required to accu rately represent the infer red evolutionar y events and processes and (2) mor phologically based species recog nized in cer tain moderately conser vative mor phological classif ications previously proposed for the genus. Fu r ther research will seek to improve sampling, especially in easter n Eu rasia, and to develop more sensitive markers for disting uishing different lineages within (1) the remarkably genetically u nifor m D. incarnata group (diploids) and (2) locally differentiated populations of (in some cases u n named) allotet raploids.

KEY WORDS: allopoly ploid, autopoly ploid, concer ted evolution, Dact ylorhiza, ecological differentiation, ITS r ibosomal DNA, phylogeny, plastid microsatellites, speciation

INTRODUCTION
Dact ylorhiza is a ta xonomically problematic, evolutionarily complex genus. -- Dact ylorhiza Necker ex Nevsk y (1937) is a genus of ter rest r ial orchids with a circumboreal to war m-temperate dist r ibution and

cent res of diversit y in Eu rope and the Near East. Taxonomy of these dact ylorchids is widely considered to have been complicated by relatively g reat mor phological var iabilit y within species and high f requency of hybr idization bet ween species. Recent opinions expressed on the total number of species occu r r ing in Eu rope, Nor th
1185


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

Af r ica / Macaronesia and the Near East range f rom six (Su nder man n, 1980) to 61 (Delforge, 2005), whereas Aver yanov (1990) estimated the number of species at 75 world-wide. The mean number of species recog nized in published st udies has increased prog ressively th rough time (reviewed by Pedersen, 1998) and is epitomized by the jump f rom 49 to 61 species bet ween the f irst and third editions of Delforge (1993, 2005). Ref ining, with justif ication, the taxonomy of the genus has become increasingly impor tant because many putative Dact ylorhiza species are declining and others have probably always been endangered nar row endemics. Reconciling mor phologically and genetically circumscribed entities (recognized at whatever taxonomic level) is a necessar y pre-requisite for a meaningf ul taxonomic hierarchy, which in t ur n is needed to accurately characterize their biogeography, ecology and conser vation stat us. It is therefore cause for concer n that Bateman & al. (2003: 22) concluded that "For the present (and despite considerable research effor t), Dact ylorhiza remains perhaps the most tantalizing of the dominantly European clades of Orchidinae, its phylogenetic histor y obscured par tly by a combination of iterative hybridization and chromosomal instability, and par tly by suboptimal species delimitation and misidentif ications of chosen st udy organisms." Many of the European dact ylorchids belong to the D. incarnata/maculata poly ploid complex, as def ined by HedrČn (2001a, 2002), which is best viewed as consisting of three groups of species: D. incarnata s.l., D. maculata s.l., and allotetraploids for med by crosses between species of the f irst two complexes (Table 1). Dact ylorhiza incarnata s.l. is an aggregate of diploid taxa (eight or more named taxa, depending on the author) that is mor phologically variable but genetically homogenous according to data

from isozymes (HedrČn, 1996) and amplif ied fragment length poly mor phisms (A FLPs; Hed rČn & al., 2001). Dact ylorhiza euxina, endemic to the Near East, is a close relative of D. incarnata, albeit clearly distinct (Bateman, 2001; HedrČn, 2001b; Bateman & al., 2003), which is also involved in allopoly ploidization (HedrČn, 2001b). These species are hereafter ter med the D. incarnata group. Dact ylorhiza maculata s.l. (hereafter ter med the D. maculata group) is a heterogeneous set of diploid (D. fuchsii, D. saccifera) and tetraploid (D. maculata) species that are more readily distinguished in peripheral por tions of their respective ranges (e.g., Heslop-Har rison, 1951; Duf rÉne & al., 1991; Bateman & Den holm, 2003). For example, in the British Isles, D. fuchsii and D. maculata are easily separated using f loral and vegetative characters and have distinct ecological preferences: the for mer grows on alkaline to neutral soils that var y from unusually dr y habitats to marshland, whereas the latter is an acid-heath specialist. In contrast, in Ger many, Austria and easter n France some taxonomists wholly reject the distinction bet ween D. fuchsii and D. maculata (e.g., Bauman n & KÝn kele, 1988), whereas others identify as D. maculata plants that grow in habitats that in the British Isles would be occupied strictly by D. fuchsii. Hybridization, resulting in the reputedly near-sterile t riploid D. âtransiens, occurs occasionally, especially where the two taxa have been brought into unusually close proximit y, of ten by anthropogenic habitat dist urbance (Bateman & Haggar, in press). The importance of polyploidy. -- Dact ylorhiza maculata has long been viewed as an autotetraploid derivative of the diploid D. fuchsii (e.g., Heslop-Har rison, 1951, 1954), a view more recently given credence by their similar allozyme prof iles (HedrČn, 1996). Nonetheless, nuclear

Table 1. General ta xonomy and distribution of Western European species of Dact ylorhiza discussed in this paper.

Ploidy Diploid Diploid Diploid Diploid Diploid Diploid Diploid Autotet raploid Allotet raploid
a

Sp e cie s D. foliosa D. fuchsii D. saccifera D. incarnata D. euxina D. sambucina D. (Coeloglossum) viridis D. maculata D. majalis s.l. (including alpest ris, elata, occidentalis, praetermissa, purpurella, sphagnicola, traunsteineri)

Distribut ion Madeira Wester n Eu rope, Nor th Af r ica and Wester n and Cent ral Asia; in the east replaced by D. saccifera Italy, Greece, the Bal kans Wester n Eu rope, Nor th Af r ica and Wester n and Cent ral Asia Near East Sweden to souther n France, east to Greece and Easter n Eu rope Nor th Temperate Zone Wester n Eu rope, but diff icult to separate f rom D. fuchsii in Cent ral and Easter n Eu rope and rare in souther n Eu rope Broadly dist r ibuted in Eu rope and Asia, with isolated occu r rences in Nor th A mer ica and Iceland; some taxa with localized dist r ibution (e.g., occidentalis conf ined to Ireland)

a

Thought to have or iginated as crosses bet ween the Dact ylorhiza incarnata g roup and the D. maculata g roup, but exact parentage highly speculative and the pr imar y focus of this st udy.

1186


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

ribosomal spacer DNA sequences (ITS nrDNA: Pridgeon & al., 1997; Bateman & al., 2003), A FLPs (Hed rČn & al., 2001) and nuclear chalcone sy nthetase (Inda & al., submitted) clearly disting uish bet ween these t wo taxa, indicating that they are better viewed as independent evolutionar y units. Dact ylorhiza maculata is more closely related to other, more clearly distinct diploids, such as D. foliosa (from Madeira) than it is to D. fuchsii (HedrČn & al., 2001; Bateman & al., 2003; Inda & al., submitted). Moreover, this relationship is reinforced by genetic data obtained in this st udy, which allowed us to distinguish easily between allopoly ploids parented by the diploid D. fuchsii from rarer allopoly ploids such as D. sphagnicola (HedrČn, 2003) and cer tain populations in nor ther n Russia (Shipunov & al., 2004) that were parented by the tetraploid D. maculata s.str. Hence, D. maculata s.str. and D. fuchsii are here treated as separate species, although ran k is a matter of choice and some of us (MH) would prefer to recognise these as subspecies because of the evidence found in this and other st udies that they often hybridize to such an extent that distinguishing them becomes diff icult. Allotet raploid taxa possess a mixt u re of characters derived f rom members of the D. maculata and D. incarnata g roups and are t y pically refer red to as the D. majalis aggregate (sensu latissimo). Although allozymes (McLeod, 1995; HedrČn, 1996, 2001b) and AFLPs (Hed rČn & al., 2001) conf ir med their hybrid origin(s), neither technique identif ied precisely the parental lineages involved in poly ploid for mation. Multiple origins of allotetraploids have long been suspected (Heslop-Har rison, 1954, 1968). More recently, they have been demonstrated for some taxa in Sweden using allozymes (HedrČn, 1996) and PCR-R FLPs (HedrČn, 1996, 2003; Devos & al., 2003) in European Russia by combining ITS sequences with plastid and nuclear microsatellites (Shipunov & al., 2004), and by conventional and landmark-based mor phometrics (Shipunov & Bateman, 2005). However, to obtain a more complete pict ure, these intentionally parochial integrated st udies need to be expanded to a Europe-wide scale. Relevance of plastid microsatellite and ITS sequences. -- Repeating u nits of shor t DNA motifs ter med microsatellites are abundant in the plastid genome of higher plants (e.g., Powell & al., 1995). Their variability makes them usef ul markers to st udy patter ns of diversity within and bet ween closely related species (Powell & al., 1995; Provan & al., 2001), par ticularly with respect to biogeography. They have already demonstrated their usef ulness in genetic st udies of orchids (Fay & Cowan, 2001; Cozzolino & al., 2003a, b; For rest & al., 2004; Shipunov & al., 2004). Fur ther more, they are easy to develop and can be used on degraded DNA, such as that t y pically ext racted f rom herbarium specimens (Fay & Cowan, 2001). In orchids and most other angiosper ms, the plastid genome is exclusively mater nally in herited

(Cor r iveau & Coleman, 1988), so these markers have the potential to identify the mater nal parents of hybrids and thus to indicate the direction of the crosses between the D. incarnata and D. maculata groups that under pin allopoly ploid events. They can also provide infor mation on identity and geographical origin of mater nal parents and number of times allotet raploid lineages of similar parentage have succeeded in becoming established. The inter nal transcribed spacers (ITS) of nuclear ribosomal DNA have been widely used to reconstr uct phylogenetic relationships because of their variability and ease of amplif ication with nearly universal primers (cf. Baldwin & al., 1995). They have also proved useful in the detection of hybrids because for a period following the hybridization event hybrids are likely to display both parental alleles; examples include Paeonia (Paeoniaceae: Sang & al., 1995), Miscanthus (Poaceae: Hodkinson & al., 2002) and Anacamptis s.l. (Orchidaceae: Bateman & Hollingswor th, 2004). However, this region can undergo concer ted evolution (Hillis & Dixon, 1991), resulting in loss of one parental allele from taxa of hybrid origin (Wendel & al., 1995; Alvarez & Wendel, 2003; Chase & al., 2003). Previous molecular phylogenetic studies indicated that this phenomenon occurs in Dact ylorhiza ; only a single ITS allele was recovered by PCR from taxa known to be allotetraploids (Pridgeon & al., 1997; Bateman & al., 2003). Considering that ITS sequences from D. incarnata, D. fuchsii, D. saccifera and D. maculata differ by both substitution and length polymor phisms (Bateman & al., 2003), more detailed study could distinguish relative contributions of putative parents to at least some allotetraploids. Impor tant cont rasts have been obser ved among groups of f lowering plants in patter ns of ITS evolution. For instance, loss of one parental allele occur red in polyploids estimated to have for med about 100 years ago in Cardamine (Brassicaceae: Fran zke & Mu m men hoff, 1999), whereas both parental ty pes have been maintained in older, putatively Plio-Pleistocene allotet raploids of Amelanchier (Rosaceae: Campbell & al., 1997) and Paeonia (Paeoniaceae: Sang & al., 1995). Concer ted evolution of ITS is consistently biased towards one parent in some cases (e.g., in Cardamine: Franzke & Mummenhoff, 1999), whereas in others it can conver t in opposite directions in different allotet raploids for med within the same genus (e.g., Gossypium : Wendel & al., 1995; Nicotiana : Chase & al., 2003; Clarkson & al., 2004, 2005). For t unately, comparison of patter ns derived from ITS sequences with those derived from plastid microsatellites can reveal such biases. In this st udy, we use these t wo categories of marker to st udy a large number of Dact ylorhiza accessions sampled across a wide geographical area. Sampling was designed to deter mine whether we can discriminate among putative species in the D. maculata group, explore the extent to which they hybridize and
1187


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

identify which of them has hybridized with a member of the D. incarnata group during for mation of each identif iable allotetraploid. Additional goals were to investigate possible multiple origins of par ticular allotetraploid taxa and deter mine which of the two species was the mater nal parent of each allotet raploid lineage. We also explored patter ns of gene conversion in ITS n rDNA and assessed the degree to which it can provide a relative time frame for dating for mation of these allotetraploid lineages.

MATERIALS AND METHODS
Sampling and DNA extraction. -- A total of 399 accessions was analyzed, representing 177 localities and 44 named taxa, together sampled widely across the range of genus Dact ylorhiza (Electronic supplement). Many of the populations were analyzed for more than one accession, par ticularly if suspected hybrids were obser ved in the f ield. Vouchers for many of the accessions consist of pickled f lowers (most deposited in the Herbarium at the Royal Botanic Gardens, Kew). Samples of Gymnadenia s.l. (including Nigritella), which is the undoubted sister genus of Dact ylorhiza, and Pseudorchis, which is a member of the Platanthera clade that is sister to Dact ylorhiza plus Gymnadenia (Pridgeon & al., 1997; Bateman & al., 2003, 2006), constit uted outgroups in the phylogenetic analysis of ITS sequences. DNA was extracted from leaves or, more often, f lowers using a 2â CTAB extraction protocol (Doyle & Doyle, 1987), but with some modif ications. Although most DNA extractions were taken from either fresh or silica gel-dried materials (Chase & Hills, 1991), a few herbarium specimens up to 100 years old were also used. Most DNAs were f ur ther cleaned on a caesium chloride/ethidium bromide gradient (1.55 g . ml­1) or with QIAquick columns (Qiagen Ltd, Crawley, West Sussex, U.K., following the manufact urer's protocol for PCR reactions), although some were simply precipitated with ethanol and resuspended in 0.25â TE buffer without f ur ther cleaning. Plastid microsatellites. -- We f irst examined seven plastid regions in search of length variation (e.g., microsatellites or larger inser tions/deletions) by sequencing these f rom a caref ully selected reference set of species that included D. fuchsii, D. maculata and D. incarnata. This per mitted us to select for f ur ther st udy four lengthvariable sites in th ree regions: the trnL int ron, the intergenic spacer (IGS) bet ween trnL and trnF, and the spacer between trnS and trnG. We then developed new primers that closely f lan ked the length-variable regions producing f ragments less than 250 base-pairs (bp) in length. For sequencing, the PCR mix included 45 µL of 1.5 mM MgCl2 Reddy PCR Master Mix 1.1X (ABgene Ltd, Epsom, Sur rey, U.K.), 2 µL of 25 mM MgCl2, 1 µL of
1188

0.4% bovine ser um albumin (BSA), 0.5 µL of each primer (100 ng/µL), and 2 µL of template DNA. The trnL intron and the trnL-trnF IGS were amplif ied using the primers c/d and e/f of Taberlet & al. (1991), whereas the trnS-trnG IGS was amplif ied using the primers of Hamilton (1999). The following PCR program was used: 4 min at 94°C, 28 cycles of 1 min at 94°C, 1 min at 48°C and 1 min at 72°C, with a f inal extension of 5 min at 72°C. PCR products were purif ied using QIAquick columns (Qiagen Ltd) following the manufacturer's protocols. Both st rands were sequenced using Big Dye Ter minator 3.0 (Applied Biosystems Inc., ABI, War rington, Cheshire, U.K.), and cycle sequencing products were r un on an ABI 3100 Prism genetic analyzer, all following the manufact u rer's protocols. Sequences were edited in Sequence Navigator and assembled in Autoassembler (both ABI). Alignments were perfor med manually in PAUP* 4.01b10 (Swofford, 2001), following the recommended procedures of Kelchner (2000). The four length-variable plastid regions are either shor t, tandem, mixed-base repeats (two) or microsatellites (t wo with homopolymer repeats). More closely spaced primers were then designed to amplify these four regions: Dact Ms1, Dact Ms2 (both in the trnS-trnG IGS), Orch1 (trnL-F IGS) and Dact Msf (trnL intron; all in Table 2). One of each pair of primers was labelled with a f luorescent dye. The four length-variable plastid fragments were then amplif ied cheaply and eff iciently in a single 10 µL PCR reaction; fragments were separated on an ABI 3100 genetic analyzer, and the length of each amplif ied fragment was deter mined with Genescan 3.1 and Genoty per version 2.0 (ABI). Each reaction contained 9.2 µL 2.5 mM MgCl2 PCR Master Mix (ABgene), 0.2 µL of 0.4% BSA, 0.1 µL of each of the eight primers (100 ng/µL), and 0.4 µL of template DNA. The program used was 4 min at 94°C, 26 cycles of 30 s at 94°C, 1 min at 55°C, and 1 min at 72°C, with a f inal elongation of 10 min at 72°C. The four target microsatellites were then combined to def ine a number of plastid haplot y pes. A minimum span ning tree was drawn by hand to summarize the relationships between the haploty pes found in the D. maculata group; no variation was discovered in the D. incarnata group, but this haploty pe is so divergent from the others that it was excluded from the minimum span ning tree. nrDNA markers. -- First, the entire ITS region (ITS1 spacer plus 5.8S rDNA gene plus ITS2 spacer) was amplif ied using the primers 17SE and 26SE (Sun & al., 1994). Each 50 µL volume PCR reaction comprised 45 µL of 1.5 mM MgCl2 Reddy PCR Master Mix 1.1X (ABgene), 1 µL of 0.4% BSA, 1 µL of dimethylsulphoxide (DMSO), 0.8 µL of ddH2O, 0.6 µL of each primer (100 ng/µL), and 1 µL of template DNA. The program used was: 2 min at 94°C, 26 cycles of 1 min at 94°C, 1 min at 52°C, and 1 min 30 s at 72°C, with a f inal extension of 5 min at


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

Table 2. List of the primers used in the study. All primer sequences read 53.

Fragment ampl if ied Dact Ms1 D a c t M s2 Orch1 Dact Msf D f u ch D mac

P r i me r t r nS Dact Ms1 D a c t M s2 t r nG Orch1 F Orch1 R Dact Msf trnL f ITS.dact.f uch F ITS.dact.f uch R ITS.dact.mac F ITS.dact.mac R

Sequence or reference Hamilton (1999) CGT TGG A AC A A A AGA AGT AC GAG TA A TAG TGT TTC TA A GAG Hamilton (1999) Fay & Cowan (2001) Fay & Cowan (2001) CTA AGA A AT TA A GGG GGC TA Taberlet & al. (1991) ATT GA A TCG CTC CAT A AG AC ACC GCA TGA CGG GCC ATT CT TGT GCC A AG GTA A AT ATG CA TAG GAG CA A ACA ACT CCA CA

72°C. Sequencing procedures and sequence analysis were identical to those applied to the plastid regions, except for the addition of DMSO to the for mer (to reduce the effects of paired-stem for mation on st rand extension). Cloning using standard recombinant DNA tech niques was required for ITS whenever direct sequencing revealed heterogeneity ascribed to the presence of multiple alleles. Although several thousand copies of n rDNA ITS are present in each individual dactylorchid, we will nonetheless employ the ter m allele in its broadest sense to describe each characteristic, collective n rDNA sequence (i.e., individual repeats are most unlikely to be identical, so this ter m is used to refer to the consensus sequence). To establish the relationships of each allele, phylogenetic analysis was perfor med in PAUP*4.01b10 using maximum parsimony; heuristic searches employed 200 replicates of random taxon entr y order with tree bisection-reconnection (TBR) swapping, and no tree limit per replicate. These complete ITS sequences were submit ted to Gen Ban k (DQ022863 to DQ022894). Length var iation was obser ved in the alig n ment; moreover, some of the underlying inser tions/deletions (indels) clearly disting uished among D. incarnata, D. fuchsii, D. saccifera and D. maculata. In our second phase of analysis, we therefore desig ned pr imers to amplif y two shor t, length-variable fragments that taken together would be diagnostic of the alleles present in each accession analyzed (Table 3). These markers are expected to be codominant (unless gene conversion is complete) and thus are usef ul for deter mining the parental taxa involved in hybridization. The two polymor phic fragments in the ITS regions were amplif ied in a single t ube. The PCR reaction contained 18 µL of 1.5 mM MgCl2 PCR Master Mix (ABgene), 0.4 µL of 0.4% BSA, 0.4 µL of DMSO, 0.32 µL of H2O, 0.24 µL of each of the four primers (100 ng/µL), and 0.4 µL template DNA. The PCR program followed that for the plastid regions but with an annealing

temperat ure of 52°C. As in the plastid analysis, ITS fragments were r un on a 3100 genetic analyzer, and fragment lengths for each accession were deter mined using Genescan 3.1 and Genoty per 2.0. W hen multiple alleles were found in a single accession, their relative propor tions were estimated using the signal strength (peak height) on the original electropherograms. Although we recognize that these propor tions are not rigorously def ined, exact ratios are highly inf luenced by gene conversion and so are not relevant to our conclusions. However, the fact that the ratios obser ved are not consistent with expected simple ratios demonstrates that conversion has occur red.

RESULTS
Plastid haplotypes. -- The four plastid regions were successf ully amplif ied in all samples, including several herbarium samples collected up to 100 years ago. Only the two longer (approximately 220 bp) microsatellites, Dact Ms1 and Dact Ms2, failed to amplify from most DNAs ext racted f rom herbarium specimens. Full data for the plastid microsatellites and ITS markers are presented in the Electronic supplement. W hen analyzed in combination, the four plastid fragments def ined 34 haploty pes. All species of the D. maculata s.l. group are characterized by a unique deletion in the trnL-trnF IGS; consequently, the Dact Msf fragment is 4 bp shor ter than that of any other Dact ylorhiza species. Haploty pes recorded in the D. incarnata group differ so much from those of other Dact ylorhiza species that it is diff icult to assess how they are related to those recovered from the D. maculata group. The D. incarnata group has been shown to be distantly related within Dact ylorhiza to the D. maculata group in analyses of both ITS and plastid DNA data (Pridgeon & al., 1997; Pillon & al., 2006). Hence, they are omitted from the minimum spanning tree
1189


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

Table 3. Summar y of the main plastid haplot ypes and ITS alleles found in the sampled ta xa.

Plast id haplot y pe Diploid and autotetraploid A B D A, C, G, W E K, Y S1, S2 R1, R 2, R 3 F J V1, V2, V3, V4, V5, V6

I TS a l l e l e III b-V I I VI-III b-V Xa Xb IIIc IX V I I Ia XI VIII b, I X

Taxa fuchsii (incl. okellyi, cornubiensis) maculata (incl. ericetor um, islandica) foliosa saccifera incarnata (incl. cr uenta, pulchella, coccinea, ochroleuca, borealis) euxina sambucina romana aristata iberica viridis (for merly Coeloglossum)

A l lotetraploid incarnata â maculata group A C B O III b-V III b-V I III b majalis s.st r., praetermissa (incl. junialis), t raunsteineri (incl. lapponica), baltica, purpurella (inc. cambrensis) majalis s.st r., praetermissa (incl. junialis), t raunsteineri (incl. lapponica), alpest ris, nieschalkior um elata (Eu rope), occidentalis (incl. kerr yensis), sphagnicola elata ( Nor th Af r ica)

A l lotetraploid eu xina â maculata group C III b ur villeana A l lotetraploid eu xina â incarnata E Xa armeniaca Note: For clar it y, only the more com mon combinations of markers are show n, and rare individuals that are genetically at y pical of their mor phologically circu mscr ibed species are omit ted. In the case of allotet raploids the pater nal ITS allele is excluded, as it was missing f rom many accessions.

that shows relationships infer red among the haploty pes; instead, this focuses on the D. maculata group (Fig. 1), which provides nearly all of the infor mation that allows genetic differentiation of allopoly ploid taxa. W hen we were assigning individual plants to a par ticular haplotype, we did not initially characterize them according to which species they had originally been assigned. Nonetheless, it soon became clear that most haploty pes could readily be ascribed to a par ticular species epithet. Diploids and autopolyploids. Haploty pe A occur red in most accessions of D. fuchsii throughout its range, including the anthocyanin-def icient D. fuchsii var. okellyi from the wester n seaboard of Ireland and the anthocyaninrich D. fuchsii var. cornubiensis from Cor nwall (southwester n England; cf. Bateman & Den holm, 1989). In contrast, our sole accession of a similar anthocyanin-rich for m from the wester n seaboard of Scotland, D. fuchsii var. hebridensis, contained the B haploty pe. This is the most common haploty pe in D. maculata s.str., including
1190

putative varieties ericetorum from the British Isles and islandica from Iceland. Haploty pes of the B group all differ from the A haploty pe by a 4 bp inser tion in the Dact Ms1 fragment. Several other haploty pes (F, M, N, T) derived from the B haplotype were also found in D. maculata. The N haplotype is of par ticular interest as it was also found in D. fuchsii in many par ts of its range (occur ring in 14 out of the 108 samples of D. fuchsii examined). The N haploty pe differs from the B haploty pe by only a single change and from the A haploty pe by two changes. Among rarer haploty pes, only the two samples of D. maculata from Ireland had the M haploty pe; only our single accession from Por t ugal, D. maculata var. caramulensis, had the P haploty pe, and just one accession of D. maculata from Sweden yielded the X haploty pe (Fig. 1). In a few cases, plants initially identif ied as D. fuchsii on mor phological criteria were shown to contain not the characteristic A haploty pe but rather the B haploty pe, more ty pical of D. maculata, and vice versa.


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

Dactylorhiza saccifera, a diploid long considered on mor phological grounds to be closely related to D. fuchsii (cf. Landwehr, 1977), frequently contained the A haplotype typical of D. fuchsii. This was particularly tr ue of samples from Croatia, where the two species are known to form inter mediates, and Turkey. However, many accessions of D. saccifera contained one of the rarer haplotypes, G and W, which were not found in any other diploids and thus are probably more characteristic of this species. One accession of D. saccifera from Croatia had the E haplotype characteristic of D. incarnata (see below). Although the C haplotype differs only by a single base in one of the microsatellite markers (homopolymer; Dact Ms2) from the A haplotype, it was never found in D. fuchsii. However, the C haplotype was found in a single Greek individual of D. saccifera, the only putatively diploid accession to exhibit this haplotype. Otherwise, the C haplotype was found only in allotetraploids, predominating in some populations (see below and Electronic supplement). In contrast with the Dact ylorhiza maculata group, the diploid D. incarnata group maintained its already wellestablished record for genetic homogeneity (cf. HedrČn, 1996; Bateman, 2001; HedrČn & al., 2001; Bateman & al., 2003). Although our 55 accessions of the D. incarnata group encompassed six putative taxa and a wide geographical range, almost all yielded the characteristic E haploty pe. The notable exception was D. euxina from the

U G C W Z X A D Q L N B T M F
Fig. 1. Minimum spanning tree showing the relationships of the plastid haplot ypes of the Dact ylorhiza maculata g ro u p. A l l l i n e s i n d i c a t e si n g l e - s t e p t r a n si t i o n s . A i s t h e most common haplot ype in the diploid species D. fuchsii, whereas B is most common in the autotetraploid D. maculata. Haplot ype C was found only once in a diploid (a single accession of D. saccifera), but it is common in several allotetraploids.

O

P

Near East, which had haploty pes K and Y, most similar to haploty pe E of D. incarnata. Diploid species phylogenetically inter polated between the D. incarnata and D. maculata groups (Bateman & al., 2003), such as D. (formerly Coeloglossum) viridis, D. aristata, D. sambucina and D. romana, maintain distinct (and, in some cases, diverse) haploty pes. Allopolyploids . Allotet raploids yielded ten haploty pes, of which the most common, A, B and C, occupy central positions in the minimum span ning tree of haploty pes derived from the maculata group (Fig. 1). These haplotypes allow division of allotetraploids into two major categories that largely correspond to groups of mor phologically def ined taxa: (1) The Dact ylorhiza majalis group has predominantly the A ( fuchsii-derived) and C haploty pes. It includes accessions f rom several mor phologically def ined species groups--the majalis group (also including D. alpestris and probably D. pratermissa), the traunsteineri group (also including lapponica), the purpurella group (also including D. cambrensis) --as well as the more geographically isolated D. baltica (southeaster n Baltic region and nor thwester n Russia) and D. nieschalkiorum (Turkey). Except for one accession of D. saccifera from Greece, the C haploty pe was found only in some of the taxa in this categor y of allotetraploids: it was present in the majority of accessions of D. majalis (16 of 25 samples), D. alpestris (4 of 6) and D. traunsteineri (20 of 32). The C haploty pe was rarely found in D. praetermissa (only 4 of 19) and never in D. purpurella (20). A geographical split was evident within D. lapponica ; the single British sample had the C haploty pe, whereas the three Swedish accessions had the A haploty pe ty pical of D. fuchsii. The N haploty pe, which occur red in a minority of populations of both D. maculata and D. fuchsii, was also found in three accessions of D. majalis from two populations in France, located close to populations of D. fuchsii that also contained this unusual haploty pe. (2) The Dact ylorhiza elata group predominantly exhibits the maculata B or similar haploty pes. The group is geographically disparate, including not only the widespread wester n Mediter ranean D. elata but also the Irish endemic D. occidentalis (incor porating D. kerr yensis) and the nor thwest European D. sphagnicola. Wester n European D. elata reliably has the B haploty pe, whereas most Nor th African accessions yielded either the O or Z haplot y pes. The similar X haploty pe was found in one un named allotetraploid putatively locally synthesized in Sweden (M. HedrČn, unpublished data); other wise it was found only in a single accession of D. maculata, also from Sweden. The plastid haploty pes were suff iciently discriminating to allow "forensic hor ticult ure" in the Dact ylorhiza elata group. Apparently clonal clusters of plants labelled
1191


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

D. elata that have long been cultivated at the Royal Botanic Garden Edinburgh (U.K.) and the National Botanical Garden Glasnevin (Ireland) have the D haploty pe ty pical of D. foliosa, a commonly cultivated endemic from the isolated island of Madeira. These plants are most likely hybrids between D. elata and D. foliosa that were created in cultivation. Several other samples of un k nown origin collected in gardens, such as the misnamed D. "fuchsii cv. Bressingham Bonus", also presented evidence of hybridization, and thus ultimately proved to be of little value in this st udy. The E haplotype characteristic of the D. incarnata group was rarely found in allotetraploids. Unsurprisingly, it occurred in the only analyzed sample

of D. armeniaca, an allotetraploid derived from D. euxina â incarnata (HedrČn, 2003), although it was also found in some accessions (4 of 20) of the nor thwester n European D. purpurella. Figure 2 shows the geographical distribution of haplotypes found in the allotetraploids. ITS alleles. -- ITS fragments were successf ully amplif ied for each accession, including most targeted herbarium samples. Sequencing of the complete ITS region was only under taken if a potentially new allele was expected based on a novel fragment length because we were able to use length to distinguish among the ITS alleles of all parental diploids and D. maculata s.str. Relationships among the eleven ITS alleles detected in Dact ylorhiza

A

B

C

E

G

N

O

T

X

Z

Tibet

Fig. 2. Distribution of plastid haplot ypes found in the allotetraploid ta xa of Dact ylorhiza. Each population is represented by at least one dot , although polymorphic populations are represented by as many dots as the number of haplot ypes they contained. To aid presentation, populations from Anglesey (nor thern Wales) and Gotland (southeastern Sweden) are each represented as single populations. Symbols for haplot ypes are explained on top of the f igure.

1192


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

Fig. 3. Phylogenetic tree showing the relationships among the dif ferent ITS alleles found in diploid and autotetraploid Dactylorhiza based on DNA sequences of the entire ITS region. All allotetraploid taxa were excluded from this tree. Alleles occurring at dif ferent positions in the tree but represented by the same Roman numeral exhibit the same indel patterns and so these produced fragments of the same size dif fering in base substitutions. Numbers above and below branches are branch lengths and bootstrap percentages, respectively. Arrows indicate clades that collapse in the strict consensus tree.

are summarized in Fig. 3 (cf. Bateman & al., 2003), and examples of t race f iles for one ITS f ragment obtained from f ive contrasting species are shown in Fig. 4. Diploids and autopolyploids. Many putatively diploid samples yielded two, or even three, ITS alleles. Although presence of three alleles in a diploid appears unint uitive,

theoretically it could occu r if conversion had not yet reached completion when a plant crossed with another plant that possessed yet another allele. Less sur prisingly, some of the allotetraploids were able to maintain three or occasionally four ITS alleles. We assigned an approximate ratio (1 : 1, 1 : 2 or 1 : 3) to accessions maintaining two or
1193


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

more high-frequency alleles (Fig. 4). In addition, we noted cases when only a trace of an allele was present, def ining a minor allele as one at least th ree times less f requent than the most abundant major allele present. No PCR bias was obser ved when we perfor med amplif ications of the ITS f ragments on samples of k nown ratios (i.e., when we used a mixt ure of the entire ITS regions previously cloned as template DNA; data not shown). We estimate from the above mixing experiments that we can begin to detect the presence of a minor allele at a frequency somewhere between 5% and 10%. However, because the alleles overlap in their respective f ragment-length patter ns, in some cases it was not possible to exclude the presence of a small amount of one additive ty pe. For instance, when the 72 + 75 bp and 70 + 80 bp fragment pairs were both found in a single accession, we could not wholly reject the potential presence of a small propor tion of the mut ually overlapping 75 + 80 bp patter n. Most accessions of D. fuchsii had either allele V or IIIb or both, whereas most accessions of D. maculata had the I allele (Fig. 3). In most D. maculata var. islandica accessions examined (4 out of 5), we found evidence for fuchsii ITS alleles as well as the ty pical maculata allele I. Allele VI was found in most samples of D. saccifera, but in many it was mixed with allele V or allele III b, indicating a close relationship with D. fuchsii. The Xa allele was found in all samples of D. incarnata s.l., rarely associated with the VIIIc allele, whereas the similar Xb allele was found in the closely related D. euxina.

Fig. 4. Examples of traces obtained with one ITS fragment (Dmac). The autotetraploid Dact ylorhiza maculata generally displays a 72- bp long fragment and the diploid D. incarnata an 80 - bp long fragment . Dact ylorhiza occiden talis (Ireland), D. sphagnicola (nor thwestern Europe) and D. elata (southwestern Europe and nor thwestern Africa) are all allotetraploids formed by hybridization bet ween D. maculata (the maternal parent) and D. incarnata. Both parental alleles are present in both D. occidentalis and D. sphagnicola, but the maternal allele is dominant in the former and the paternal allele is dominant in the lat ter. In contrast , the paternal allele has been lost from D. elata.

Allotetraploids. In allotetraploids, several patter ns of ITS alleles were obser ved. Overall, they either possessed between one and three alleles of the D. maculata group plus that of the D. incarnata group or they maintained alleles of only one of these two parental groups; in most cases it was the D. incarnata group that was not represented. Fig. 5 shows the geographical dist ribution of allotet raploids displaying ITS t y pes f rom either one or both parental lineages. This demonstrates that complete loss of one parental allele is more common in Nor th Africa and souther n Europe, par ticularly in the characteristic allotetraploids of these regions, D. elata and D. majalis s.str. Alleles I, III and V of the D. maculata group were all frequent in the allotetraploids, whereas allele VI characteristic of D. saccifera was rare. A mong allotet raploids, ITS alleles mir rored the plastid haplot y pes in revealing st r uct ured patter ns that cor responded well with groups of named taxa: (1) Members of the Dact ylorhiza majalis group (characterized by fuchsii plastid haplotypes) had predominantly the fuchsii alleles V or IIIb, but degrees of evidence of the presumed incarnata parent varied both between and within species. Some D. alpestris possessed only the two fuchsii alleles, V and IIIb, whereas in others these were balanced by equal copy f requencies of incarnata allele Xa. The single sample of D. baltica had only the fuchsii allele V. Most of the 25 D. majalis samples had fuchsii V and IIIb alleles, only one exhibiting a trace of the incarnata Xa allele. The four specimens of D. lapponica had predominantly the fuchsii allele V, but the three accessions from Sweden also exhibited a trace of the incarnata Xa allele. The closely related D. traunsteineri was especially heterogeneous. A few individuals had predominantly incarnata Xa alleles (some also possessed the incarnata haploty pe E), and only f ive accessions lacked any trace of incarnata Xa; nonetheless, in most individuals the fuchsii alleles V and IIIb were dominant. Another variable tetraploid, D. praetermissa, was distinguished mainly by the presence in half the accessions of the saccifera allele VI, which ranged in frequency from dominance in two accessions, to presence as just a trace in four. Other wise, this species contained a mix of the t wo fuchsii alleles (V and III b) and occasionally the incarnata allele Xa, although this was rarely equal or dominant. Dact ylorhiza purpurella/cambrensis combined the incarnata Xa allele with the fuchsii V allele, the for mer always with at least comparable frequency with the latter. Shifting the geographic focus to Turkey, our limited samples of both D. ur villeana and D. nieschalkiorum exhibited only the fuchsii IIIb allele. (2) Members of the Dact ylorhiza elata group (characterized by maculata B or related plastid haplot y pes) had dominantly maculata alleles (I, or more rarely its variant, IV). The notable exception was D. sphagnicola,

1194


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

which dominantly had the incarnata allele Xa, most often occur ring in a 3 : 1 ratio with the maculata I allele. The single analyzed sample of D. kerr yensis had no detectable copies of the incarnata allele Xa, whereas the mor phologically similar D. occidentalis s.str. showed a 3 : 1 ratio of maculata to incarnata alleles. As with plastid haploty pes, a clear distinction was evident between accessions of D. elata from Nor th Africa versus those from southwester n Europe. The nine samples from Nor th Africa exhibited only the maculata IIIa allele (a fragment equal in length to one of the common fuchsii alleles but showing a different set of substit utions) or the maculata IV allele, with no t race of the incarnata X allele. In cont rast, the six samples of D. elata from Europe had the maculata allele I, occasionally supplemented with a trace of the incarnata Xa allele.

Correlation bet ween patterns in ITS alleles and plastid haplot ypes. -- T here was a sig n if icant cor relation bet ween plastid haplot y pes and ITS t y pes fou nd i n the allotet raploids ( p < 0.0 01; 2 test). Most not ably, haplot y pes A and C were most f requently associated with ITS alleles III b, V and V I ( fuchsii and saccifera markers, respectively), whereas the B haplo t y pe was most f requently associated with ITS allele I (maculata markers). The Dact ylorhiza maculata group proved far more genetically diverse than D. incarnata s.l., containing four common ITS alleles and 16 haplot y pes. Many samples displayed both D. fuchsii and D. maculata alleles. The A haploty pe ty pical of D. fuchsii was found in a few accessions that had been designated at the time of collection as D. maculata, and the B haploty pe ty pical of D. maculata

Fig. 5. Concer ted evolution of ITS in allotetraploids across Europe and adjacent regions. Populations of allotetraploids containing ITS alleles of both parental groups, D. incarnata s.l. and D. maculata s.l., are represented by blue squares. Populations from which one parental allele has been lost are represented by red circles. To aid presentation, populations from Anglesey (nor thern Wales) and Gotland (southeastern Sweden) are each represented as single populations.

1195


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

was similarly found in a few samples initially identif ied as D. fuchsii. The frequency of individuals that combined fuchsii and maculata ITS alleles showed a geographical t rend, being f requent in Aust ria, occasional in France and rare in the British Isles (Fig. 6). Most D. saccifera exhibited the A haploty pe shared with D. fuchsii and had a mixt ure of ITS alleles characteristic of both D. fuchsii and itself (i.e., alleles III and V versus VI). However, some Greek samples contained only the t y pical D. saccifera allele VI, combined with plastid haplot y pes G and W that were not found in any other diploid species. The D. saccifera allele VI was also found in a few D. fuchsii f rom Croatia and the British Isles. In Britain, allele VI was detected only in D. fuchsii individuals cohabiting with allotetraploid D. praetermissa, many of which also exhibited the VI allele. This indicated introgression between the two species, which differ in ploidy.

DISCUSSION
Comparison with previous studies of Dact ylo rhiza that used similar molecular markers. -- Our st udy of plastid haploty pes shares several accessions with a plastid PCR-R FLP st udy by HedrČn (2003), which in t ur n overlapped taxonomically with the PCR-R FLP st udy by Devos & al. (2003). Not sur prisingly, all three plastid investigations yielded broadly similar results. Our ITS work built on previous molecular phylogenetic st udies by Pridgeon & al. (1997) and Bateman & al. (2003), but here we have sampled far more extensively within the target species. The main advances in our st udy are dense sampling of individuals and amalgamation of the t wo previously separate lines of evidence f rom mater nally in her ited plastid regions and a biparentally in her ited nuclear region.

Fig. 6. Mixing of Dact ylorhiza maculata s.str. and D. fuchsii ITS alleles across Europe. Populations of the D. maculata group that contained ITS alleles characteristic of D. fuchsii and those characteristic of D. maculata are represented by green squares, whereas populations that contained alleles of only one of the t wo species are represented by purple circles.

1196


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

In all accessions examined in our st udy, a strong positive cor relation was obser ved between par ticular plastid haploty pes and par ticular ITS alleles. This allows us to address HedrČn's (2003) concer n that, on the basis of plastid data alone, he could not distinguish between ancestral polymor phisms and those due to recent hybridization. Our inter pretation is predicated on the assumption that if ancestral polymor phisms were segregating then no strong cor relation would be expected bet ween markers f rom plastid and nuclear genomes (cf. Ramsey & Schemske, 1998; Morjan & Rieseberg, 2004). Thus, for the parental taxa D. incarnata, D. fuchsii and D. maculata, most populations in which more than one marker was detected are assumed to result from hybridization rather than retention of ancestral polymor phism. The positive cor relation bet ween ITS alleles and plastid haploty pes constit utes f ur ther evidence that the diploid D. fuchsii and autotetraploid D. maculata should be regarded as distinct species, even though there is clearly cur rent gene f low across the divide of their different ploidy levels in many par ts of their shared range (Fig. 6: cf. Devos & al., 2003; HedrČn, 2003; Shipunov & al., 2005). These species undoubtedly experienced a period of isolation in the past that was suff icient to allow them to develop their distinct mor phologies and ecological preferences; some of their secondar y contact has resulted from post-glacial migration patter ns and also recent human dist urbance of the landscape. Evidence that hybridization between these two taxa is a recent phenomenon comes from the fact that the mixing of haploty pes and ITS alleles now obser ved in these species rarely occurs in allotetraploids. If local populations of D. fuchsii show evidence of introgression with D. maculata, then locally for med allotet raploids should show similar mixt ures of markers, cont rar y to our obser vations. Locally for med allotet raploids could have for med f rom single hybridizations, so it might be expected that such allotet raploid populations would be genetically consistent, but it would be highly unlikely that all parental D. fuchsii/maculata parents of such hybrids would themselves never be introgressed individuals. Most authors are also happy to treat D. foliosa (from Madeira) as a distinct species. This species is more closely related to D. maculata than the latter is to D. fuchsii, so if D. foliosa is recognized as a distinct species, then so must D. fuchsii and D. maculata (if DNA data are considered to have a bearing on which taxa are to be recognized). Molecular markers developed for this st udy were subsequently employed in the same laborator y as par t of a more geographically constrained st udy that focused on dactylorchids of European Russia (Shipunov & al., 2004, 2005; Shipunov & Bateman, 2005). This demonst rated that in Russia the N haplotype, which is most similar to the B haplotype of D. maculata but is also commonly found in D. fuchsii, occur red only in D. maculata. This obser vation

suppor ts our inference that the N haploty pe originated in D. maculata and subsequently became introgressed into D. fuchsii. In European Russia and Georgia, populations of the D. incarnata group contain not only the ty pical E haplotype but also the similar H haplotype. Allotetraploids occur ring in nor ther n Russia, most notably D. baltica, exhibited a mixt ure of fuchsii and maculata markers, but lacked the C haploty pe commonly found in souther n Europe. Overall, the patter ns obser ved in European Russian Dact ylorhiza revealed few new haploty pes and are highly congr uent with those repor ted here for wester n European Dact ylorhiza. Employing ITS alleles as nuclear markers. -- In this st udy we took advantage of indels in ITS to amplify length-variable f ragments that are usually codominant, unless alleles are lost through concer ted evolution /gene conversion. Our technique allowed us to screen a large number of samples, including herbarium specimens that ty pically yield highly degraded DNA. We amplif ied both fragments in the same reaction in small volumes, thereby establishing a highly eff icient screening method that did not require cloning when two or more alleles were present. A nalysis of elect ropherograms f rom direct sequencing of PCR products containing t wo or more alleles was made diff icult due to this length variation. However, we found that coincidence of these indels with alleles discovered earlier th rough sequencing (and in some cases cloning) was suff icient to identify each allele when the two fragments were considered together. Thus, we made the conser vative assumption that neither fragment alone was diagnostic. No apparent bias was detected when evaluating mixtures of DNA of known ratio, as previously obser ved by Rauscher & al. (2002). However, we obser ved only infrequently the simple allelic ratios expected from hybrids, indicating that in many cases copies of some alleles were being eliminated or back-crosses were occur ring; both processes "reinforce" one allele relative to others. For example, although one accession of D. fuchsii from Sorvilier, Switzerland, had the expected diploid chromosome number of 2n = 40 (L. Hanson, M. Fay & M. Chase, unpublished data), it also had the D. fuchsii IIIb and V alleles in a 1 : 2 ratio, which would have been most parsimoniously inter preted as indicating that this plant was triploid rather than diploid. It also proved diff icult to analyze patter ns from individuals that yielded more than two alleles because one par ticular fragment length can sometimes be attributed to more than one allele (Fig. 3). Nonetheless, we could routinely detect alleles representing only 10% of the total copies present and could reach 5% with reasonable conf idence, showing that this technique is more sensitive than direct sequencing (Rauscher & al., 2002). In summar y, ITS f ragments were a usef ul tool for deter mining parentage of many accessions, but obser ved
1197


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

ratios of alleles were not reliably infor mative about ploidy levels, nor could complete absence of an allele be conclusively demonst rated. Reinforcement of an allele due to backcrossing of the progeny with one parent could also be masked by effects of concer ted evolution and /or gene conversion. Significance of concerted evolution in ITS. -- In allotetraploids, a degree of homogenization occur red in the great majority of cases because parental ITS alleles were generally detected in non-Mendelian propor tions. In many cases, one parental allele was apparently "lost" (or, more accurately, was reduced to less than 5% ­10% of all copies present: Fig. 5). Comparison with mater nally inherited plastid microsatellites showed that in most cases the missing allele was the pater nal one, generally that derived from the D. incarnata group. For example, examination of a single accession of D. armeniaca, a recently described allotet raploid der ived f rom hybr idization bet ween D. euxina and D. incarnata (HedrČn, 2001b), showed that it lacked the euxina ITS allele but possessed a t y pical incarnata plastid haploty pe, indicating conversion to the ITS allele of the mater nal parent. Among allotetraploids there were two main exceptions to the r ule of mater nal conversion. Dact ylorhiza sphagnicola always had a majority of the D. incarnata (Xa) ITS allele in spite of having the B haploty pe characteristic of D. maculata, whereas the majorit y of samples of D. purpurella (including D. cambrensis) combined the A haploty pe of D. fuchsii with the incarnata-derived Xa ITS allele, which occur red in equal or more often greater frequency than the fuchsiiderived V allele. In these t wo unusual cases concer ted evolution appears to be favouring the pater nal ITS allele rather than the mater nal allele. Two contrasting mechanisms are most often proposed for concer ted evolution of rDNA: unequal crossing-over and gene conversion (Hillis & Dixon, 1991). The strong parental bias obser ved in Dact ylorhiza favours the hypothesis of gene conversion (Hillis & al., 1991) because it is an unlikely outcome of unequal crossing-over, nor is it clear how a mater nal effect could persist across the several generations needed to reduce one allele to an undetectable level. St udies of nuclear ribosomal ITS and the 18S rDNA intron in the marine macroalga Caulerpa (Durand & al., 2002) and of ITS and IGS in Drosophila (Polanco & al., 1998) similarly revealed differential evolution in both directions and rates of concer ted evolution that are incompatible with the u nequal crossing-over model. Moreover, homogenization of ITS demonstrably occurs in f lowering plants even when multiple ribosomal clusters occupy different ch romosomes (Wendel & al., 1995; Chase & al., 2003). Exploring origins and migration patterns of allotetraploids. -- One impor tant goal of this st udy was to evaluate the possibility of using ITS conversion rates
1198

to estimate the period of time elapsed since the initial hybridization event that preceded successf ul establishment of each allotetraploid lineage and thereby to explore likely implications of contrasting dates of origin for subsequent evolutionar y histories of the resulting lineages. It soon became apparent from our data that most of the allotetraploid samples f rom souther n Europe had lost one of their parental ITS alleles, whereas both parental alleles were t y pically still detectable in allotet raploids f rom the British Isles and Scandinavia (both examined in this st udy) and from nor ther n Russia (repor ted by Shipunov & al., 2004; Fig. 5). Although ack nowledging limits of sensitivity of our technique for detecting low-frequency alleles, ou r results were highly inter nally consistent; within most st udy populations, either all individuals had retained both parental ty pes or all had lost one. Because most nor ther n allotetraploids have retained at least some evidence of both parental alleles, indicating that concer ted evolution has not reached completion, we hy pothesize that they are younger than souther n allotet raploids (in making this assumption we recognize that several other factors, most notably cont rasts in effective population sizes, can also inf luence rates of change in the frequency of ITS alleles). During the Quater nar y, nor ther n Europe experienced several cycles of thick ice cover followed by recolonization. The whole of Scandinavia and much of the British Isles were covered with ice 18,000 years ago, and periglacial conditions persisted until 11,700 years ago. Thus, it is tempting to speculate that in Europe the nor ther n allotetraploids became established post-glacially, whereas the souther n allotet raploids that have largely eliminated one parental ITS allele may antedate the last glacial maximum. Combining the degree of gene conversion with infer red parentage suggests a quadripar tite classif ication of wester n European allopoly ploid dactylorchids according to their respective parentages and putative relative dates of origin. Older allotetraploids that lack one parental ITS allele can be divided into those derived from hybridization between D. incarnata s.l. and D. fuchsii (D. majalis) and those derived from hybridization between D. incarnata s.l. and D. maculata (D. elata). Similarly, younger allotetraploids that maintain evidence of both parental ITS alleles can be divided into those derived f rom hybridization bet ween D. incarnata s.l. and D. fuchsii (e.g., D. praetermissa, D. purpurella, D. traunsteineri s.l., D. baltica) and those derived f rom hybridization bet ween D. incarnata s.l. and D. maculata (e.g., D. occidentalis, D. sphagnicola). Application of a range of molecular techniques has f ur ther teased apar t the two categories of younger allotetraploids (e.g., HedrČn, 1996, 2001, 2002, 2003; HedrČn & al., 2001; Devos & al., 2003) and is par ticularly effective when combined with mor phometric data gathered from the same individuals collected for genetic


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

analyses (McLeod, 1995; Bateman, 2001; Shipunov & al., 2004; Shipunov & Bateman, 2005). Local origins have already been infer red for some of the younger allotetraploids, indicating that they evolved in sit u in nor ther n regions (HedrČn & al., 2001; HedrČn, 2002; Shipunov & al., 2004; Bateman, 2006, in prep.). However, we infer that the majority of allotetraploids have recolonized nor ther n areas by relatively recent migration from the south because souther n markers such as the C haploty pe are widespread in many allotetraploids but almost completely absent from their presumed progenitors among diploids and autopoly ploids. If so, the patter n of colonization infer red for the D. incarnata/maculata complex is unusual. In most other temperate clades that admix diploid and polyploid species, poly ploids have proven to be strong colonizers of Arctic regions (Abbot t & Broch man n, 2003), whereas their diploid progenitors have remained much f ur ther south. Although we differ in our respective opinions regarding the relative average colonization abilities of the diploids, autopoly ploids and allopoly ploids, it is evident that one categor y is not clearly superior to the others, whereas in many other groups poly ploids are competitively superior in the boreal zone. Possible reasons for this obser vation include the fact that, cont rar y to many other poly ploid groups such as the fer n Asplenium (Vogel & al., 1999), polyploid species in the Dact ylorhiza complex do not have colonization ability enhanced by apomixis. Also, Dijk & Grootjans (1998) argued that, at least in the Netherlands, D. majalis s.str. and D. praetermissa prefer more fertile soils than do D. maculata and D. incarnata, perhaps indicating that these allotetraploids are only able to successf ully colonize a nar rower range of habitats. Dact ylorhiza incarnata group. -- As previously obser ved with allozymes (HedrČn, 1996), AFLPs (HedrČn & al., 2001), PCR-R FLPs (Hed rČn, 2003) and ITS sequencing (Pridgeon & al., 1997; Bateman & al., 2003), remarkably little genetic variation was obser ved within the D. incarnata group (here def ined broadly to include D. euxina), even though we sampled several mor phologically circumscribed taxa that together span ned a large par t of its overall geographical range. Each individual analyzed yielded only the plastid haplotype E, and most individuals contained only the ITS allele Xa. However, allele VIII, which differs f rom allele Xa only in possessing a 2 bp deletion, occur red alongside allele Xa in a few samples f rom Wales and Ireland. The only sample f rom Turkey analyzed had two ITS alleles, both broadly resembling allele Xa but disting uishable by several substit utions. Extended analysis of D. incarnata s.l. f rom Turkey for allozymes and plastid markers indicated that the species is much more molecularly variable in southeaster n Europe than in nor ther n and wester n Europe (HedrČn, 2001b, in prep.). The endemic Turkish diploid D. euxina had two

unique haploty pes and a distinct ITS allele, although as expected both were most similar to those found in D. incarnata. Dact ylorhiza incarnata is also profoundly distinct from the D. maculata group as obser ved with allozymes (Hed rČn, 1996, 2001b), but a few cases of possible int rogression have been obser ved. The plastid haplot y pe and ITS allele X of D. incarnata were found in a few specimens of D. saccifera from Croatia. We also found on single occasions the fuchsii ITS allele V and fuchsii haploty pe A in two samples of D. incarnata subsp. pulchella. Fur ther more, Shipunov & al. (2004) demonstrated that the fuchsii V allele is widespread in some populations of D. incarnata in Russia. Using R FLPs, HedrČn (2003) and Devos & al. (2003) also revealed some evidence of hybridization and /or subsequent int rogression bet ween D. incarnata and members of the D. maculata group in Sweden. Thus, limited gene f low may still be possible between the two divergent parental groups, either directly or via allotetraploids as a bridge (a decidedly less readily detected process). Dact ylorhiza maculata group. -- Although mor phologically based st udies are divided on whether to recognize D. maculata, D. fuchsii and D. saccifera as separate species (cf. DufrÉne & al., 1991; Bateman & Denholm, 1989, 2003; StĹhlberg, 2007), there is growing molecular evidence that the for mer two represent lineages evolved in isolation for a considerable period of time, and most of us (not MH) argue that all three are best regarded as distinct species. They have distinct ITS sequences (Fig. 3: Pridgeon & al., 1997; Bateman & al., 2003) and, at least in Sweden, are readily disting uished using AFLP data (HedrČn & al., 2001). In this st udy we found markers in both the plastid and the nuclear genomes that clearly distinguish among the three taxa. Admittedly, we have also obser ved mixing of these markers in several accessions, but we believe that this is due to secondar y hybridization and /or introgression rather than incomplete lineage sor ting. In this context, an introgression zone including two distinct genoty pes within D. maculata s.str. has recently been documented in Sweden (StĹhlberg, 2007). As regards D. saccifera, increased sampling is required from the Balkans to decide whether the taxon is distinct from D. fuchsii ; the distinctive ITS alleles (Fig. 3) found in most accessions of D. saccifera indicates that it too existed in isolation from both D. fuchsii and D. maculata for a signif icant period of time. Another line of evidence for their distinctiveness comes from the allotetraploids, which allow us to infer genetic content of their parents. In allotetraploids, markers characteristic of D. fuchsii and D. maculata are rarely combined, indicating that at the time and place of the for mation of the allotetraploids D. fuchsii and D. maculata were clearly distinct and not hybridizing as extensively as they are today.
1199


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

A few samples identif ied as Dact ylorhiza maculata that had some D. fuchsii markers (ITS ty pe and haplot y pe) or vice versa provide circumstantial evidence of hybridization or int rogression. A r tif icial hybridization of these t wo species has revealed substantially greater fer tilit y in back-crosses than in the f irst generation (Bateman & Haggar, in press). Distinguishing between D. fuchsii and D. saccifera also proved challenging. The most problematic sit uation was encountered in the Alps region (easter n France, Switzerland and Austria: Fig. 6), where many samples exhibited both fuchsii and maculata markers, whereas such cases are rarer in the British Isles and souther n Scandinavia. A recent set of 20 accessions of Dact ylorhiza sampled from acidic sites that should have been assigned to D. maculata if they were found in wester n Europe (M. Chase, G. Fischer & D. Dock rel, unpubl.) showed that ever y one of them was a recent hybrid of D. fuchsii and D. maculata (recent because they all exhibited both ITS alleles). These genetic obser vations cor relate well with mor phology, as f ield botanists regard the two species as more diff icult to distinguish in central Europe (Heslop-Har rison, 1951; DufrÉne & al., 1991; Bour nČrias & Prat, 2005) than in marginal areas in the nor thwester n par t of the Continent (Bateman & Denholm, 1989, 2003; Pedersen, 1998, 2004). Fu r ther more, putative species mor phologically inter mediate between D. fuchsii and D. maculata, such as D. savogiensis and D. sudetica, have been described in the Alps and contiguous uplands (e.g., Delforge, 2001). Unfor t unately, the ploidy of these plants remains un k nown. The fact that several ITS ty pes were still detected in some accessions of the D. majalis group indicated that gene conversion had not reached completion, in contrast with souther n allotetraploids. If taken together with the absence of mixt ures of fuchsii and maculata markers in allotetraploids, this indicates that most hybridization /introgression events involving allotetraploids occur red recently, after the for mation of at least most allotetraploids. They most likely originated soon after the re-establishment of sympatr y between these species when they expanded postglacially out of separate glacial ref ugia. In this context, it is notewor thy that the alpine region, where the relationship between D. fuchsii and D. maculata appears especially complex, is considered as an impor tant zone of secondar y contact among post-glacial migrants (ter med a "sut ure zone" by Hewit t, 2000). However, all markers used in this st udy are susceptible to rapid f ixation, either because they are mater nally inherited (plastids) or because they are subject to concer ted evolution (ITS). Such markers can reveal ancient genetic exchange between species even though the species themselves apparently remain mor phologically distinct, as has recently been obser ved among the anthropomor phic species group within the genus Orchis s.st r. (M. Fay & al., unpublished data). Such complex
1200

sit uations are best explored f ur ther by examination of multi-locus markers such as AFLPs and /or a selection of biparentally in herited nuclear microsatellites/int rons in low-copy genes. The occur rence of some populations that combine markers ty pical of Dact ylorhiza fuchsii and D. maculata can be explained by recent local hybridization, despite their contrasting ecological preferences. For example, the karstic landscape of the Bur ren in wester n Ireland suppor ts mainly the calcicole D. fuchsii, but small pockets of peat-rich residual soils dotted across the limestone suppor t the calcif uge D. maculata, bringing the two species into intimate proximity. They also frequently meet in mixed habitats in Scandinavia (StĹhlberg, 2007). Admittedly, many other sites where the two species co-occur ref lect recent anthropogenic dist urbance. Nonetheless, combinations of markers were also found in several populations where only one of the two species was found, most notably in Iceland where D. fuchsii is not k nown to now occur. Some samples conf idently identif ied as either D. fuchsii or D. maculata were found to contain markers characteristic of the other species. For example, near Llangurig, Wales, plants with mor phology characteristic of D. maculata and growing in ty pically acid soils yielded markers of both species, even though D. fuchsii was not obser ved growing in the immediate vicinity. Introgression between the two taxa is a more likely explanation of such obser vations, probably occur ring in both directions. In ter ms of likely underlying processes, transfer of markers from the diploid Dact ylorhiza fuchsii to the tetraploid D. maculata is possible via un reduced gametes in D. fuchsii. Although D. maculata is generally accepted to be an autotet raploid (Hager up, 1944; Heslop-Har rison, 1951), few reliable chromosome counts are available (conf usion regarding which mor phological characters best distinguish between D. fuchsii and D. maculata casts doubts on some deter minations; e.g., Tanako & Kamenoto, 1984). Thus, D. maculata may still be diploid in some par ts of its range. Conversely, other obser vations indicate that in central Europe tetraploidy may occur in the ty pically diploid D. fuchsii (e.g., HedrČn, 2002; Bour nČrias & Prat, 2005; StĹhlberg, 2007). Also, Hager up (1944) noted the occasional development of haploid embr yos without fer tilization in both D. fuchsii and D. maculata s.str., results later conf ir med by Heslop-Har rison (1957). If such embr yos were viable they could per mit gene f low from (auto)tetraploids to diploids. Thus, the theoretical bar rier to gene f low between the putatively diploid D. fuchsii and tetraploid D. maculata is probably a less profound obstacle than is generally supposed; moreover, such a bar rier has been overcome in Taraxacum, for example (Men ken & al., 1995). Except for a unique ITS allele, Dact ylorhiza saccifera appears to be con nected to D. fuchsii by populations with


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

aty pical combinations of plastid markers. Dact ylorhiza fuchsii ITS alleles were found in D. saccifera in Croatia, Turkey and Greece, which was the only region where some samples of D. saccifera (1) did not have markers t y pically found in D. fuchsii and (2) exhibited distinct haploty pes. We also detected variable populations of D. saccifera from Greece. In addition, we found the D. saccifera ITS allele VI in a few samples of D. fuchsii, not only from Croatia where the two species co-occur, but also in Britain, 600 k m from the nearest extant populations of D. saccifera (as explained below, this apparent enigma could ref lect occasional hybridization between D. fuchsii and D. praetermissa). Several obser vations suppor t the origin of D. maculata in Nor th Africa or the Iberian Peninsula. The f irst is the occur rence only in the Macaronesian island of Madeira of D. foliosa, a diploid species that resembles D. maculata both mor phologically and especially genetically (Pridgeon & al., 1997; HedrČn & al., 2001). We also found a distinct ITS allele in the samples of D. elata from Morocco (one parent of which was probably maculata-like); this differs f rom the common D. maculata allele only in lacking a distinctive 8 bp indel and thus appears to be plesiomor phic relative to all ITS alleles recovered from the D. maculata group. We also identif ied several unique but maculata-like plastid haploty pes in dactylorchids from this region. Corresponding mor phological diversity is indicated by recognition by some authorities of three segregates from D. maculata s.str. that are endemic to Morocco, Algeria and the Iberian Peninsula--D. maurusia, D. battandieri and D. caramulensis, respectively (Delforge, 2005) --although they are at best only subtly mor phologically distinct. Dact ylorhiza saccifera may have had a ref ugium in Greece because this is the only area where genetically it is both diverse and relatively distinct from D. fuchsii ; also, Greece is the centre of the present range of D. saccifera. Although we did not include samples of D. saccifera from Italy, another ref ugial candidate, preliminar y results from another st udy have not revealed unusual genetic diversity in this region (M. HedrČn, unpublished data). We have obtained even less evidence regarding possible ref ugia for D. fuchsii because this species showed little variation in plastid microsatellites (A or, less frequently, N). The fact that an additional haploty pe, Q, is common in Russia (Shipunov & al., 2004) tentatively indicates an easter n ref ugium, although the Balkans and Italy also remain credible candidates. Allotetraploids. -- It is notewor thy that we detected little evidence of gene f low between allotetraploids, indicating presence of effective bar riers to gene exchange. This is perhaps not sur prising, given that extensive ar tif icial crosses conducted among Swedish dact ylorchids by Malmgren (1992) yielded fer tile F2 plants only when one of the parents was Dact ylorhiza incarnata s.l. or

D. sphagnicola, the latter with the D. incarnata ITS alleles predominating instead of the ty pical (for allotetraploids) alleles of the D. maculata group. Bateman & Haggar (in press) created ar tif icial hybrids between D. praetermissa and D. purpurella that showed high fer tility in both the f irst generation and backcrosses. Accessions of D. majalis s.str. or D. traunsteineri exhibiting maculata markers were rare, and no fuchsii markers were obser ved in D. elata or D. occidentalis accessions, even though allotetraploids of these two categories often grow suff iciently close to each other to expect occasional cross-pollinations. However, allotetraploid populations that mix fuchsii and maculata haploty pes have recently been repor ted f rom Sweden (HedrČn, 2003) and are also suspected to occur in Scotland (R.M. Bateman, unpub.). Species in the Dact ylorhiza maculata group, most commonly D. fuchsii, were mater nal parents of the great majority of allotetraploids. As obser ved in poly ploids of other families (Soltis & Soltis, 1999), allotetraploid dactylorchids of wester n Europe have several origins; the number of plastid haplotypes indicates at least ten independent allopoly ploid events. However, three haploty pes occur red in most allotetraploids, having successf ully spread across most of the range of the genus: the most common fuchsii haploty pe (A), the most common maculata haploty pe (B), and the C haplotype, the last concentrated in the south and of an uncer tain parental derivation. In addition, it is clear that, although it is always repor ted to be a tetraploid, D. maculata (or a genetically similar entity) was the mater nal parent of several allotetraploid taxa. The C haplotype was found in only one putatively diploid individual, a Greek Dact ylorhiza saccifera. However, this accession also contained the common saccifera ITS allele VI, whereas most allotetraploids that possess the C haploty pe have the fuchsii allele IIIb. Thus, our cur rent (albeit limited) sampling suggests that D. saccifera is not likely to be a parent of these allotetraploids. It seems more likely that the diploid species that originally donated the C haploty pe to D. majalis and similar allotetraploids is extinct or at least has become suff iciently rare to escape our Europe-wide sampling effor t. The C haploty pe has a central position in the minimum span ning tree between the A haplot y pe characteristic of D. fuchsii and the G and W haploty pes found in some D. saccifera from the Greek mainland (Fig. 1). The presumed diploid species that once exhibited the C haplot y pe was probably formerly widespread, considering that the C haploty pe has been found in allotetraploid samples stretching from the Pyrenees to the Tibetan plateau. However, this haploty pe declines in f requency nor thward, being rare in Scandinavia and absent from European Russia (Shipunov & al., 2004). This suggests either that the hy pothetical ancestral diploid became rare before the end of the last glaciation or that it failed to migrate nor thward following glaciation.
1201


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

In either case, its contribution to for mation of relatively young nor ther n allotet raploids was less impor tant, being replaced in this role by t y pical D. fuchsii. Vir t ual absence of the C haploty pe from sampled diploids and its preponderance in several allotetraploids suggests either that these allotetraploids were not for med in their present geographical locations or (less likely) raises the possibility that older allotetraploids may have contributed to the origin of younger allotetraploids (e.g., D. majalis s.str. to D. alpestris). Most allotetraploids with A or C haplotypes had fuchsii or saccifera ITS alleles, and most possessing the B haplotype had the maculata ITS allele; the only exceptions to this patter n were a few samples combining maculata haplotypes with fuchsii ITS alleles. Putative allotetraploid accessions containing both fuchsii and maculata ITS alleles were rarely detected, indicating that introgression between D. maculata and allotetraploids is an infrequent event, even though they share the same ploidy level. The only exceptions found in this st udy were from Iceland, where four out of f ive accessions contained both maculata and fuchsii markers, despite the supposed absence from Iceland of D. fuchsii. Also, Shipunov & al. (2004) repor ted the presence in nor ther n Russia of allotetraploid populations that do not have D. incarnata as one of their parents but rather appear to be derived from hybridization between D. fuchsii and D. maculata. Nonetheless, almost all allotetraploids lack evidence of prior hybridization or introgression between D. fuchsii and D. maculata, and allotetraploids do not cur rently appear to be operating as a genetic bridge lin king D. fuchsii and D. maculata. Moreover, the characteristic Dact ylorhiza saccifera ITS allele VI was rarely found in allotetraploids examined here, indicating a limited contribution of D. saccifera to their for mation. However, the distribution of this allele is unusual. It also occurs sporadically and ty pically at low frequencies across the range of D. fuchsii (e.g., Croatia, U.K.), frequently in D. praetermissa (U.K.; slightly more than 50% of the accessions sampled) and is present in single populations of D. majalis (France) and D. purpurella (Wales). Thus, although D. saccifera is presently limited to the easter n Mediter ranean and the Near East, it is possible that it once extended into wester n Europe. Alter natively, presence of the saccifera allele in other taxa, par ticularly D. fuchsii, could be the result of local hybridization with sympatric D. praetermissa. However, this hy pothesis similarly requires a subsequent major contraction in the range of D. praetermissa to its present nor thwester n European enclave, after presumably originating in, and migrating out of, the Mediter ranean region. This seems unlikely, given that D. praetermissa is here characterized as a young allotetraploid. HedrČn (2001b) infer red that Dact ylorhiza saccifera or a closely related taxon was one parent of the allotet1202

raploids characteristic of Turkey because it is the only member of the D. maculata group that cur rently occurs in the region. However, different ITS alleles were found in D. fuchsii and D. saccifera, suggesting that the act ual parent of these allotet raploids may instead be a hy pothesized diploid that is either extinct or as yet undiscovered. Ou r data suggest that several widely recog nized allotet raploid taxa have multiple origins, including the exceptionally widely dist ributed Dact ylorhiza majalis. In the case of D. purpurella, presence of both D. fuchsii and D. incarnata haploty pes indicates that hybridization events that accompanied poly ploidization occur red in both directions or that introgression with its parents has cont ributed to additional haplot y pes af ter the original allotetraploid was for med. The case of D. traunsteineri and the closely related D. lapponica is especially instr uctive. Samples of each taxon f rom the British Isles, Scandinavia (the t y pe region for D. lapponica) and the Alps (the ty pe region for D. traunsteineri) are readily distinguished using either haploty pes or ITS alleles, but within each region, there are no signif icant differences between the two supposed species, conclusions previously indicated by st udies of allozymes (HedrČn, 1996, 2002, 2003; Bateman, 2001) and AFLPs (HedrČn & al., 2001). It is clearly advisable to synonymize D. lapponica with D. traunsteineri across their respective (and vir t ually coincident) ranges. However, the systematist must then make the diff icult decision of whether to (1) recognize a single allopoly ploid species that has at least three independent evolutionar y origins (Table 3) or (2) recognize three separate species that are putatively distinct, one species located in each of the three geographical regions. Perhaps the most appropriate arbiter is whether putative segregated species can be recognized using mor phological characters with an acceptable level of conf idence. On this basis, Bateman (2006) assigned to D. traunsteinerioides those dactylorchid populations in the British Isles that had received considerable conser vation attention because they had previously been ascribed to D. traunsteineri and /or to D. lapponica. However, even given extensive population genetic data and focusing on a restricted geographical area, it can prove challenging to deter mine with suff icient conf idence the number of origins of a par ticular allotetraploid taxon. For example, Swedish populations of D. sphagnicola collectively have only one origin according to plastid markers (see also Hed rČn, 2003), but allozyme data indicate at least two origins (HedrČn, 1996), and f ine-scale analysis of additional plastid markers indicates multiple origins (HedrČn, NordstrĆm & StĹhlberg, unpub.). Inferring the current evolutionar y status of allotetraploids. -- One of the most impor tant questions raised by the Dact ylorhiza incarnata/maculata complex is why allotetraploids that we can demonstrate to have the


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

same pair of parental species can exhibit substantially different mor phological, ecological and distributional proper ties. Examples of such contrasts include D. sphagnicola versus D. occidentalis versus D. elata (all derived from hybridization bet ween D. maculata and D. incarnata) and D. traunsteineri s.l. versus D. purpurella versus D. majalis (all derived from hybridization between D. fuchsii and D. incarnata). There are two contrasting hy potheses that could, either separately or in combination, explain differentiation and specialization among allopoly ploids. (1) Post- origin differentiation of allotetraploids . This hy pothesis is predicated on (1) presumed ability of differential directional or disr uptive selection to f inet une to cont rasting ecologies the products of different poly ploidization events between the same two parental species, and (2) using cont rasting degrees of ITS gene conversion to provide relative dates of different polyploidization events occur ring between the same pair of parental species. For example, the Ir ish endemic D. occidentalis is a recently synthesized allotetraploid, whereas the more widespread Iberian / Nor th African D. elata is judged to be substantially older. Its greater age since for mation offers selection more time to operate on the D. elata phenoty pes and thereby to mould them to f it a distinct set of ecological parameters. This hy pothesis predicts that D. occidentalis (a taxonomically controversial species, once tentatively misidentif ied as an autopolyploid: cf. Bateman & al., 2003; Bateman, 2006) should still exhibit a blend of parental traits, whereas the longer existence of D. elata should have allowed it suff icient time to diverge from parental traits, thereby becoming more specialized and thus more readily recognizable as a bona f ide species, a process perhaps assisted by a greater degree of genomic re-organization and integration of the two parental genomes (Parakon ny & Kenton, 1995). Fur ther more, if older allotet raploids, such as D. elata and D. majalis s.str., did indeed originate before the last glacial maximum, then they would have responded to profound climate change by migrating f irst southward and then nor thward, presumably alongside their progenitors. If so, they would likely have passed through at least one genetic bottleneck, which would have f ur ther homogenized their genetic, mor phological and ecological characteristics (cf. Cozzolino & al., 2003b). In contrast, the more recently synthesized allotetraploids such as D. occidentalis and D. praetermissa, hypothesized to have originated during the Holocene, should appear more heterogeneous. A n analogous but probably older case is provided by allotet raploid species complexes in Nicotiana (Solanaceae). In section Polydiclieae (sensu K napp & al., 2004), evidence from plastid (Chase & al., 2003) and ITS (Clarkson & al., 2004) DNA sequences indicated that two allotet raploid species, N. clevelandii and N. quadrival-

vis, were generated f rom the same parental lineages at different times in their histor y. These two species now exhibit contrasting f loral mor phologies and ecologies and have only a slight range overlap in southwester n Nor th America. Given suff icient time, some such entities become distinct evolutionar y lineages that can undergo subsequent phyletic radiations; examples include Nicotiana section Repandae, which consists of four species with a common origin, and section Suavolentes, which consists of approximately 25 species with a common origin (Chase & al., 2003; Clarkson & al., 2004). (2) Pre- origin differentiation of parents of allotetraploids. A contrasting hy pothesis can also explain our abilit y to disting uish mor phologically and ecologically most of the independent lineages resulting f rom separate poly ploidization events bet ween the Dact ylorhiza incarnata and D. maculata groups, as indicated by genetic data. This focuses more on the considerable degrees of mor phological, ecological and, at least in the case of the D. maculata group, genetic differentiation that is evident among various named infraspecif ic taxa within the two parental groups (Bateman, 2001, 2006, in prep.). Within the British Isles alone, D. incarnata is represented by at least six named infraspecif ic entities: one a specialist of sphag num bogs and another favou r ing depressions in du ne systems, whereas the remaining fou r are character istic of alkaline fens and marshes, occasionally extending into alkaline/neut ral meadows (Heslop-Har r ison, 1953; Bateman & Den holm, 1985). Moreover, D. fuchsii exhibits the widest habitat tolerance of any British orchid species. In addition to named infraspecif ic specialists of upland and coastal past ures, populations in habiting chalk and limestone grasslands, alkaline/neut ral past ures, and alkaline/neut ral marshes and woodland can all be disting uished by subtle morphological differences (Bateman & Denholm, 1989). This degree of largely cor related variation in mor phology and habitat preference offers much potential for iteratively generating contrasting allotetraploid lineages from within the same pair of parental species. Consider, for example, the th ree moist u re-loving allotet raploids that are show n by genetic data to be the progeny of D. incarnata and D. maculata s.st r. As its name suggests, D. sphagnicola preferentially in habits acid sphag nu m bogs in Scandinavia and nor thwester n Conti nent al Eu rope, where the most li kely mater nal parent is the sphag nu m bog specialist D. maculata elodes, in her iting f rom it not only mor phologies subtly distinct f rom those of the parental nominate race but also its ext reme ecological preference (cf. Hed rČn, 2003). In cont rast, the I r ish endemic D. occidentalis tolerates soils var ying f rom slightly acidic to slightly al kaline, especially when subject to anth ropogenic dist u rbance. Both its mor pholog y and ecolog y suggest that it is more
1203


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

likely der ived f rom hybr idization bet ween D. incarnata incarnata and D. maculata ericetor um. Lastly, the relatively poorly researched, putatively older allotet raploid D. elata f rom southwester n I ber ia and nor thwester n Af r ica prefers alkaline past u res and seepages and is hy pothesized to represent hybr idization bet ween D. incarnata s.l. (this species also has been u nder-researched in the southwester n ext reme of its range) and one or more of the regional seg regates of D. maculata (D. caramulensis, D. battandieri or D. maurusia). The rare Moroccan endemic D. maurusia is of par ticular interest in this context, as it is mor phologically reminiscent of D. elata (e.g., Landweh r, 1977) and, u nusually for D. maculata s.st r., it in habits alkaline soils. Separate evolutionar y or igins are likely for D. elata populations in I ber ia and Nor th Af r ica (the sou rce of the holot y pe; Pedersen & al., 2003), given their distinct haplot y pes and cont rasting conver ted ITS alleles (Table 4), and the tendency to taxonomically separate French and Spanish populations f rom the nominate race as inf raspecif ic taxa on the g rou nds of their subtly distinct mor phologies (e.g., Nieschalk & Nieschalk, 1972; Landweh r, 1977; Delforge, 2005). This cont rasting hy pothesis thus relies on the assumption that these allotetraploids originated locally, in sympatr y with their parents, and ref lect both the detailed mor pholog y and habitat preferences of those parents. This scenario implies that our ability to distinguish subtly genetically distinct lineages derived repeatedly from the same t wo parental species relies more on selection honing the parents pr ior to poly ploid for mation than post-derivational selection honing allotetraploid lineages, thereby down-playing the evolutionar y (and taxonomic) impor tance of relative periods elapsed since the initiating hybridization event. This hy pothesis is best evaluated by st udying mor phologically and genetically diagnosable

allotetraploids that show unusually restricted distributions and so are assumed to be of recent origin (Bateman, 2006; Bateman & al., in prep.).

CONCLUSIONS
Systematic implications of the genetic patterns. -- Of the th ree species aggregates considered here, the least taxonomically controversial within wester n Europe, at least at the species level, has been the Dact ylorhiza incarnata group. With few exceptions (notably Delforge, 2005), authorities have been inclined to award species stat us only to D. cruenta among the named taxa within this group, and this elevation is not upheld by genetic data (HedrČn, 1996; HedrČn & al., 2001; Bateman & al., 2003). Treatments of the D. maculata group have historically ranged from recognition as a single species (most common if st udy focuses on regions suspected of sustaining relatively high levels of introgression) through frequent recognition of three core species (D. maculata, D. fuchsii and D. saccifera) to f ur ther division into local endemic species (Delforge, 2005). Not sur prisingly, classif ication of allotet raploids has been most cont roversial, var ying from most (Sunderman n, 1980) or many (SoŃ, 1980) infraspecif ic taxa allocated to a single aggregate species, D. majalis, through to highly divided treatments recognizing many species, most poorly mor phologically differentiated (Aver yanov, 1991; Delforge, 2005). Our own framework taxonomy (Table 4), which currently excludes local endemics, is a compromise between these extremes. It represents an attempt to synthesize previous, mor phology-based taxonomic circumscriptions (and associated knowledge of ecological preferences and geographical distributions) with more process-oriented

Table 4. Recommended framework classif ication of European members of the Dact ylorhiza incarnata and D. maculata groups and their derived polyploid complex. The plastid haplot ype and ITS allele(s) given here are considered t ypical of each ta xon. This summar y focuses on well- established species, incorporating regional endemics but excluding local endemics.

Taxon D. D. D. D. D. D. D. D. D. D. D. D. D. fuchsii (incl. cornubiensis, okellyi) maculata (incl. ericetorum, elodes) saccifera incarnata s.l. (all W European taxa) euxina elata (Nor th Africa) elata (Europe) occidentalis (incl. kerr yensis) sphagnicola majalis (incl. alpestris) praetermissa (incl. junialis) traunsteineri (incl. lapponica) purpurella (incl. cambrensis)

Ploidy and parentage 2X 4X (autotetraploid) 2X 2X 2X maculata â incarnata maculata â incarnata maculata â incarnata maculata â incarnata fuchsii â incarnata fuchsii/saccifera â incarnata fuchsii â incarnata fuchsii â incarnata

Plast id haplotype A B C, E Y, O B B B A, A, A, A

ITS al lele(s) V, IIIb I VI Xa Xb IIIa, completely conver ted I, most accessions completely conver ted I dominant, X in 1/3 or fewer copies Xa dominant, I in 1/3 or fewer copies V, IIIb, most accessions completely conver ted V, IIIb, VI V, IIIb, rarely with Xa dominant V, IIIb, rarely with Xa dominant

G, W K

C C C

1204


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

data on occur rence of gene f low (hybridization/introgression), both cur rently and, by inference, in the past. Dividing species more f inely risks generating taxa that cannot reliably be distinguished using mor phology or, in many cases, DNA data, thereby hampering communication and under mining conser vation initiatives (the conser vation implications of our data are explored elsewhere; Pillon & al., 2006). Alter natively, further amalgamating species into "super-species", in response to evidence of past or present gene f low among component species, obscures our hardwon knowledge of evolutionar y processes operating within this allopolyploid complex. In par ticular, mor phological, genetic and ecological differentiation evident among both diploids and tetraploids, and evolutionar y causes of that differentiation, would no longer be represented in an unnecessarily cr ude taxonomy (consider the extreme cases of lumping cer tain highly calcicolous lineages of D. fuchsii with the highly calcifugic D. maculata elodes or combining ecologically contrasting allopolyploids D. traunsteineri and D. sphagnicola). Our insights into the probable contrasting ages of different allotetraploid lineages, and their consequently contrasting genetic compositions and evolutionar y trajectories, would also be ignored. Cur rent co-existence of various diploids and tetraploids in at least some regions without genetic mixing indicates that bar riers to gene f low are operating. Moreover, differentiation of the taxa has strong ecological as well as geographical components, suggesting that these taxa are likely to operate as distinct evolutionar y units and hence suppor ting our use of the species categor y. Even where signif icant gene f low is evident between species in por tions of their present ranges (e.g., between D. fuchsii and D. maculata in the Alps), there is evidence that these species have in the past experienced periods of isolation that allowed them to develop substantially different plastid haploty pes and ITS alleles. Fur ther more, the signif icant cor relation bet ween ITS alleles and plastid haplot y pes in both diploids and tetraploids means it is unlikely that lineage sor ting rather than hybridization /int rogression is responsible for the heterogeneity of these markers obser ved in Dact ylorhiza. With regard to future f ield collecting, our primar y objectives are to intensify sampling in likely glacial ref ugia in Iberia/ Nor th Af rica, Italy, Greece and the Caucasus and to extend application of our markers eastward into Asia. Novel haploty pes were found in material from the Russian Caucasus (Shipunov & al., 2004), and a set of just four probable allotetraploids collected in Georgia revealed no less than three unique haplotypes (albeit clearly related to those previously found in Dact ylorhiza fuchsii). Only one of our 399 samples was located east of the Urals; this sample, f rom the Tibetan Plateau, reassuringly yielded a haplot y pe and ITS prof ile t y pical of the dominantly European D. majalis.

Limitations to the application of the genetic markers used in this study. -- The t wo sets of genetic markers used here, plastid DNA f ragment length variants and ITS n rDNA alleles, are both subject to being "capt ured": plastid DNA due to its uniparental (mater nal) patter n of inheritance and ITS because of concer ted evolution /gene conversion, which over time erodes evidence of its original biparental inheritance. Although parentage of young hybrids can be deter mined with conf idence using these markers, older hybrids will not appear to be hybrids because of conversion of one ITS allele. Moreover, because conversion usually favours the mater nal allele, mater nally in her ited plastid DNA markers are likely to generate sets of relationships concordant with those derived f rom mater nally biased conver ted ITS alleles, thereby f ur ther camouf laging evidence of past hybridization events. However, for the majorit y of accessions st udied here, we were able to use these markers to identify hybrids and deter mine which species was the mater nal parent. A n unexpected benef it of quantif ying ITS f requencies was that the degree of loss of the less favoured ITS allele indicates relative ages of allotet raploid taxa. This is especially advantageous when, as here, the same pair of parental taxa has generated multiple allotetraploid lineages at different times in the past. Admittedly, even these two complementar y sets of markers appear too conser vative to adequately inter pret some f ine-scale patter ns. For example, the considerable mor phological variation evident within the D. incarnata group has proven invisible to most markers used so far, with the exception of a single allozyme locus (cf. HedrČn, 1996; Bateman, 2001) and one promising plastid region (M. Hed rČn, unpublished data). Bet ter markers within the D. incarnata group are essential if we are to evaluate our hy pothesis that various allotetraploids are still being synthesized locally (e.g., HedrČn, 2003: Bateman, 2006; Bateman & al., in prep.). Similarly, if parental markers are too highly conser ved we can not detect cases of local hybridization and introgression. The results of this st udy provide a tantalizing glimpse into the complex evolution and ecology of these widespread European orchids, which nonetheless remain a serious challenge to the taxonomist.

ACKNOWLEDGEMENTS
We than k Roby n Cowan, Edith Kapinos, Lazlo Csiba and Olivier Maur in for assistance in the laborator y and Dirk Albach, Cr inan Alexander, Lau re Civey rel, Sid Clarke, Howard Clase, Helena Cot r im, Tom Cu r tis, Ian Den holm, Manf red Fischer, Trevor Hod kinson, Pete Hollingswor th, Fran k Horsman, Mike Lowe, Yi-bo Luo, R ichard Manuel, Ian Phillips, JÓao Pinto, Walter Rossi, Paula Rudall, Fred Ru msey, A ndy SchĆnswet ter, 1205


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

Peter Tr ibsch, and Steve Wald ren for collecting mater ial used in this st udy. Paula Rudall kindly cr itically read the manuscr ipt. We also than k the t wo anony mous reviewers, who made many valuable com ments. Fi nancial suppor t was provided by the Royal Botanic Gardens, Kew.

LITER ATURE CITED
Abbot t, R . J. & Brochmann, C. 2003. Histor y and evolution of the arctic f lora: in the footsteps of Er ic HultČn. Molec. Ecol. 12: 299 ­313. Alvarez, I. & Wendel, J.F. 2003. Ribosomal ITS sequences and plant phylogenetics. Molec. Phylog. Evol. 29: 417­ 434. Aver yanov, L .V. 1990. A review of the genus Dact ylorhiza. Pp. 159 ­206 in: A rdit ti, J. (ed.), Orchid Biolog y: Reviews and Perspectives, vol. V. Timber Press, Por tland. Baldw in, B.G., Sanderson, M. J., Por ter, J.M., Wojciechowsk i, M.F., Campbel l, C. S. & Donoghue, M. J. 1995. The ITS region of nuclear r ibosomal DNA: a valuable sou rce of evidence on A ngiosper m phylogeny. Ann. Missouri Bot. Gard. 82: 247­277. Bateman, R.M. 2001. Evolution and classif ication of Eu ropean orchids: insights f rom molecular and mor phological characters. J. Eur. Orchid. 33: 33 ­119. Bateman, R .M. 2006. How many orchid species are cu r rently native to the Br itish Isles? Pp. 89 ­110, pl. 1, in: Bailey, J.P. & Ellis, R.G. (eds.), Current Ta xonomic Research on the British and European Flora. Botanical Societ y of the Br itish Isles, London. Bateman, R.M. & Denholm, I. 1985. A reappraisal of the Br itish and Ir ish dact ylorchids, 2. The diploid marsh-orchids. Watsonia 15: 321­355. Bateman, R .M. & Denholm, I. 1989. A reappraisal of the Br itish and I r ish dact ylorchids, 3. T he spot ted- orchids. Watsonia 17: 319 ­349. Bateman, R.M. & Denholm, I. 2003. The heath spot ted-orchid (Dact ylorhiza maculata (L.) SoŃ) in the Br itish Isles: a cautionar y case-st udy in delimitating inf raspecif ic taxa and i nfer r i ng thei r evolut iona r y relat ionships. J. Eur. Orchid. 35: 3 ­36. Bateman, R.M. & Haggar, J. In press. Dact ylorhiza Necker ex Nevski. In: Stace, C.A. (ed.), Hybridization and the Flora of the British Isles. Botanical Societ y of the Br itish Isles, London. Bateman, R .M. & Hol l ingswor th, P.M. 2004. Mor phological and molecular investigation of the parentage and mater nit y of Anacamptis âalbuferensis (A. f rag rans â A. robusta), a new hybr id f rom Mallorca, Spain. Ta xon 53: 43 ­54. Bateman, R.M., Hol l ingsworth, P.M., Preston, J., Yi-Bo, L ., Pridgeon, A.M. & Chase, M.W. 2003. Molecular phylogenetics and evolution of Orchidinae and selected Habenar iinae (Orchidaceae). Bot. J. Linn. Soc. 142: 1­ 40. Bateman, R .M., Rudal l, P. J. & James, K.E. 2006. Phylogenetic context, gener ic aff inities and evolutionar y or igin of the enig matic Balkan orchid Gymnadenia f rivaldii Hampe ex Gr iseb. Ta xon 55: 107­118. Baumann, H. & KÝnkele, S. 1988. Die wild wachsenden Orchideen Europas. Kosmos, St ut tgar t. BournČrias, M. & Prat, D. 2005. Les OrchidČes de France, Belgique et Luxembourg, 2nd ed. Par thČnope, Par is. 1206

Campbel l, C. S., Wojciechowsk i, M.F., Baldw in, B.G., A l ice, L.A. & Donoghue, M. J. 1997. Persistent nuclear ribosomal DNA sequence poly mor phism in the Amelanchier agamic complex (Rosaceae). Molec. Biol. Evol. 14: 81­90. Chase, M.W. & Hi l ls, H.H. 1991. Silica gel: an ideal mater ial for f ield preser vation of leaf samples for DNA st udies. Ta xon 40: 215 ­220. Chase, M.W., K napp, S., Cox, A.V., Clarkson, J. J., Butsko, Y., Joseph, J., Savolainen, V. & Parokonny, A. S. 2003. Molecular systematics, GISH and the or igin of hybr id taxa in Nicotiana (Solanaceae). Ann. Bot. 92: 107­127. Clarkson, J. J., K napp, S., Garcia, V., Olmstead, R .G. & Chase, M.W. 2004. Phylogenetic relationships in Nicotiana (Solanaceae) infer red f rom multiple plastid DNA regions. Molec. Phylog. Evol. 33: 75 ­90. Clarkson, J. J., Lim, K.Y., Kovarik, A., Chase, M.W., K napp, S. & Leitch, A.R. 2005. Long-ter m genome diploidization in allopoly ploid Nicotiana section Repandae (Solanaceae). New Phytol. 168: 241­252. Corr iveau, J.L . & Coleman, A.W. 1988. Rapid screeni ng method to detect potential biparental in her itance of plastid DNA and results for over 200 angiosper m species. Amer. J. Bot. 75: 1443 ­1458. Cozzol ino, S., Cafasso, D., Pel legrino, G., Musacchio, A. & Widmer, A. 2003a. Fine-scale phylogeog raphical analysis of Mediter ranean Anacamptis palustris (Orchidaceae) populations based on chloroplast minisatellite and microsatellite var iation. Molec. Ecol. 12: 2783 ­2792. Cozzol ino, S., Noce, M.E., Musacchio, A. & Widmer, A. 2003b. Var iation at a chloroplast minisatellite locus reveals the sig nat u re of habitat f rag mentation and genet yic bot tlenecks in the are orchid Anacamptis palust ris (Orchidaceae). Amer. J. Bot. 90: 1681­1690. Delforge, P. 1993. Guide des OrchidČes d'Europe, d'Af rique du Nord et du Proche- Orient. Delachau x et NiestlČ, NeuchÁtel-Par is. Delforge, P. 2005. Guide des OrchidČes d'Europe, d'Af rique du Nord et du Proche- Orient, 3rd ed. Delachaux et NiestlČ, NeuchÁtel-Par is. Devos, N., Ty teca, D., RaspČ, O., Wesselingh, R.A. & Jacque mar t, A.-L . 2003. Pat ter ns of chloroplast diversit y among wester n Eu ropean Dact ylorhiza species (Orchidaceae). Pl. Syst. Evol. 243: 85 ­97. Dijk , E. & Grootjans, A.B. 1998. Perfor mance of fou r Dact ylorhiza species over a complex t rophic g radient. Acta Bot. Neerl. 47: 351­368. Doyle, J. J. & Doyle, J.L . 1987. A rapid DNA isolation procedu re for small quantities of f resh leaf tissue. Phytochem. Bull. 19: 11­15. DufrÉne, M., Gathoye, J.-L . & Ty teca, D. 1991. Biostatistical st udies on wester n Eu ropean Dact ylorhiza (Orchidaceae): the D. maculata group. Pl. Syst. Evol. 175: 55 ­72. Durand, C., Manuel, M., Boudouresque, C.-F., Meisnesz , A., Verlaque, M. & Le Parco, Y. 2002. Molecular data suggest a hybr id or igin for the invasive Caulerpa racemosa (Cauler pales, Chlorophy ta) in the Mediter ranean Sea. J. Evol. Biol. 15: 122­133. Fay, M.F. & Cowan, R . S. 2001. Plastid microsatellites in Cypripedium calceolus (Orchidaceae): genetic f inger pr ints f rom herbar iu m specimens. Lindleyana 16: 151­156. Forrest, A.D., Hol l ingswor th, M.L ., Hol l ingswor th, P.M., Sydes, C. & Bateman, R .M. 2004. Population genetic


TA XON 56 (4) · November 2007: 1185 ­1208

Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

st r uct u re in Eu ropean populations of Spiranthes romanzoffiana set in the context of other genetic st udies on orchids. Heredit y 92: 218 ­227. Franzke, A. & Mummenhoff, K. 1999. Recent hybr id speciation in Cardamine (Brassicaceae): conversion of nuclear r ibosomal ITS sequences ITS in st at u nascendi. Theor. Appl. Genet. 98: 831­ 834. Hagerup, O. 1944. On fer tilisation, poly ploidy and haploidy in Orchis maculata L. sens.l. Danish Bot. Ark . 11: 1­25. Hami lton, M.B. 1999. Fou r pr imer pairs for the amplif ication of chloroplast intergener ic regions with int raspecif ic variation. Molec. Ecol. 8: 521­525. HedrČn, M. 1996. Genetic differentiation, poly ploidization and hybr idizat ion i n nor ther n Eu ropean Dact ylorhiza (Orchidaceae): evidence f rom alloz y mes markers. Pl. Syst. Evol. 201: 31­55. HedrČn, M. 2001a. Conser vation pr ior ities in Dact ylorhiza, a taxonomically complex genus. Lindleyana 16: 17­25. HedrČn, M. 2001b. Systematics of the Dact ylorhiza euxina/ incarnata/maculata poly ploid complex (Orchidaceae) in Tu rkey: evidence f rom alloz y me data. Pl. Syst. Evol. 229: 23 ­ 44. HedrČn, M. 2002. Speciation pat ter ns i n the Dact ylorhiza incar nata/maculata poly ploid complex (Orchid aceae): evidence f rom molecular markers. J. Eur. Orchid. 34: 707­731. HedrČn, M. 2003. Plastid DNA var iation in the Dact ylorhiza incarnata/maculata poly poid complex and the or igin of allotet raploid D. sphagnicola (Orchidaceae). Molec. Ecol. 12: 2669 ­2680. HedrČn, M., Fay, M.F. & Chase, M.W. 2001. A mplif ied f rag ment leng th poly mor phisms (A FLP) reveal details of poly ploid evolution in Dact ylorhiza (Orchidaceae). Amer. J. Bot. 88: 1868 ­1880. Heslop -Harr ison, J. 1951. A compar ison of some Swedish and Br itish for ms of Orchis maculata sens.l. Svensk Bot. Tidsk r. 45: 608 ­ 635. Heslop -Harrison, J. 1953. Some obser vations on Dact ylorchis incarnata (L.) Ver mln. in the Br itish Isles. Proc. Linn. Soc. London 166: 51­ 82. Heslop -Harrison, J. 1954. A sy nopsis of the dact ylorchids of the Br itish Isles. Ber. Gebot. Forschungsinst. RÝbel ZÝrich 1953: 53 ­ 82. Heslop -Harrison, J. 1957. Apomictic potentialities in Dact ylorchis. Proc. Linn. Soc. Lond. 170: 174 ­178. Heslop -Harrison, J. 1968. Genetic system and ecological habit as factors in dact ylorchid var iation. Jahresber. Natur wiss. Vereins Wuppertal 21+22: 20 ­27. Hew it t, G.M. 2000. The genetic legacy of the Quater nar y ice ages. Nature 405: 907­913. Hi l l is, D.M. & Di xon, M.T. 1991. R ibosomal DNA: molecular evolution and phylogenetic inference. Quart. Rev. Biol. 66: 411­ 445. Hi l l is, D.M., Moritz , C., Por ter, C.A. & Baker, R . J. 1991. Evidence for biased gene conversion in concer ted evolution of r ibosomal DNA. Science 251: 308 ­310. Hodk inson, T.R ., Chase, M.W., Takahashi, C., Leicht, I. J., Bennet t, M.D. & Renvoi ze, S.A. 2002. The use of DNA sequencing (ITS and trnL-F ), A FLP, and f luorescent in sit u hybr idization to st udy allopoly ploid Miscanthus (Poaceae). Amer. J. Bot. 89: 279 ­286. Kelchner, S. 2000. The evolution of non-coding chloroplast

DNA and its application in plant systematics. Ann. Missouri Bot. Gard. 87: 482­ 498. K napp, S., Chase, M.W. & Clarkson, J. J. 2004. Nomenclat ural changes and a new sectional classif ication in Nicotiana (Solanaceae). Ta xon 53: 73 ­ 82. Landwehr, J. 1977. Wilde OrchideeĘn van Europa. Nat uu rmonu menten, A msterdam. Malmgren, S. 1992. Crossing and cultivation exper iments with Swedish orchids. Svensk . Bot. Tidsk r. 86: 337­346. McLeod, L . 1995. Combined Mor phomet ric and Isoz yme Analysis of Dact ylorhiza in Scotland. MSc thesis, Royal Bot anic Garden Edi nbu rgh and Edi nbu rgh Universit y, Edinbu rgh. Menken, S.B., Smit, E. & den Nijs, H.C.M. 1995. Genetic population st r uct u re in plants: gene f low bet ween diploid sexual and t r iploid asexual dandelions (Tara xacum section Ruderalia). Evolution 49: 1108 ­1118. Morjan, C.L . & R ieseberg, L .H. 2004. How species evolve collectively: implications of gene f low and selection for the spread of advantageous alleles. Molec. Ecol. 13: 1341­ 1356. Nieschalk , A. & Nieschalk , C. 1972. K r itische Bemerk u ngen z u r Taxonomie u nd Verbreit u ng von Dact ylorhiza elata (Poi r.) SoŃ (Hohes K naben k raut, OrchideengewÄchse). Philippia 1: 137­148. Parokonny, A. & Kenton, A.Y. 1995. Comparative physical mapping and evolution of the Nicotiana tabacum kar yot y pe. Pp. 301­320 in: Brand ham, P.E. & Ben net t, M.D. (eds.), Ke w Chromosome Conference IV. Royal Bot anic Gardens, Kew. Pedersen, H.ô. 1998. Species concept and g uidelines for inf raspecif ic taxonomic ran king in Dact ylorhiza (Orchidaceae). Nord. J. Bot. 18: 289 ­309. Pedersen, H. ô. 2004. Conf licting classif ications of Nordic poly mor phic orchids: a cr itical assessment. J. Eur. Orchid. 36: 869 ­916. Pedersen, H. ô., HedrČn, M. & Bateman, R .M. 2003. (1600) Proposal to conser ve the name Orchis majalis against O. elata, O. vestita, and O. sesquipedalis (Dact ylorhiza, Orchidinae, Orchidaceae). Ta xon 52: 633 ­ 634. Pi l lon, Y., Fay, M.F., Shipunov, A.B. & Chase, M.W. 2006. Species diversit y versus phylogenetic diversit y: a practical st udy in the taxonomically diff icult genus Dact ylorhiza (Orchidaceae). Biol. Conser vation 126: 4 ­13. Polanco, C., GonzŔlez , A.I., de la Fuente, A. & Dover, G.A. 1998. Multigene family of r ibosomal DNA in Drosophila melanogaster reveals cont rasting pat ter ns of homogenization for IGS and ITS spacer regions: a possible mechanism to resolve this paradox. Genetics 149: 243 ­256. Powel l, W., Morgante, M., McDev it t, R ., Vendramin, G.G. & Rafalsk i, J.A . 1995. Poly mor phic simple sequence repeat regions in chloroplast genomes: applications to the population genetics of pines. Proc. Natl. Acad. Sci. U.S.A. 92: 7759 ­7763. Pridgeon, A.M., Bateman, R .M., Cox, A.V., Hapeman, J.R . & Chase, M.W. 1997. Phylogenetics of subt r ibe Orchidinae (Orchidoideae, Orchidaceae) based on nuclear ITS sequences. 1. Intergener ic relationships and poly phyly of Orchis sensu lato. Lindleyana 12: 89 ­109. Provan, J., Powel l, W. & Hol l ingswor th, P.M. 2001. Chloroplast microsatellites: new tools for st udies in plant ecolog y and evolution. Trends Ecol. Evol. 16: 142­147. 1207


Pillon & al. · Evolut ion and diversif icat ion of Dact ylorhiza

TA XON 56 (4) · November 2007: 1185 ­1208

Ramsey, J. & Schemske, D.W. 1998. Pathways, mechanisms, and rates of poly ploidy for mation i n f lower i ng plants. Annu. Rev. Ecol. Syst. 29: 467­501. Rauscher, J.T., Doyle, J. J. & Brown, A.H.D. 2002. Inter nal t ranscr ibed spacer repeat-specif ic pr imers and the analysis of hybr idization in the Glycine tomentella (Leg u minosae) poly ploid complex. Molec. Ecol. 11: 2691­2702. Sang, T., Crawford, D. J. & Stuessy, T.F. 1995. Documentation of reticulate evolution in peonies (Paeonia) using inter nal t ranscr ibed spacer sequences of nuclear r ibosomal DNA: implications for biogeog raphy and concer ted evolution. Proc. Natl. Acad. Sci. U.S.A. 92: 6813 ­ 6817. Shipunov, A.B. & Bateman, R.M. 2005. Geomet r ic mor phomet r ics as a tool for u nderstanding Dact ylorhiza (Orchidaceae) diversit y in Eu ropean Russia. Biol. J. Linn. Soc. 85: 1­12. Shipunov, A.B., Fay, M.F. & Chase, M.W. 2005. The taxonomic stat us of Dact ylorhiza baltica (Orchidaceae) f rom Eu ropean Russia: evidence f rom molecular markers and mor pholog y. Bot. J. Linn. Soc. 147: 257­274. Shipunov, A.B., Fay, M.F., Pi l lon, Y., Bateman, R .M. & Chase, M.W. 2004. Dact ylorhiza (Orchidaceae) in European Russia: combined molecular and mor phological analysis. Amer. J. Bot. 91: 1419 ­1426. Soltis, D.E. & Soltis, P. S. 1999. Poly ploidy: recur rent for mation and genome evolution. Trends Ecol. Evol. 14: 348 ­352. SoŃ, R . de. 1980. Dact ylorhiza. Pp. 337­342 in: Tutin, T.G., Hey wood, V.H., Bu rges, N.A., Moore, D.M., Valentine, D.H., Walters, S.M. & Webb, D.A. (eds.), Flora Europaea, vol. 5. Cambr idge Universit y Press, Cambr idge.

StĹhlberg, D. 2007. Systematics, Phylogeog raphy and Polyploid Evolution in the Dact ylorhiza maculat a Complex (Orchidaceae). Doctoral Disser t ation, Lu nd Universit y, Lu nd. Sun, Y., Sk inner, D. Z ., Liang, G.H. & Hulber t, S.H. 1994. Phylogenetic analysis of Sorghum and related taxa using inter nal t ranscr ibed spacers of nuclear r ibosomal DNA. Theor. Appl. Genet. 89: 26 ­32. Sundermann, H. 1980. EuropÄische und mediterrane Orchideen. Eine Bestimmungsf lora mit Ber Ýck sichtig ung der ćkologie, 3rd ed. Sch mersow, Hildesheim. Swof ford, D.L . 2001. PAUP*: Phylogenetic Analysis Using Parsimony (*and Other Methods), Version 4. Sinauer, Su nderland, Massachuset ts. Taberlet, P., Giel ly, G. & Bouvet, J. 1991. Universal pr imers for amplif ication of th ree non-coding regions of chloroplast DNA. Pl. Molec. Biol. 17: 1105 ­1109. Tanako, R . & Kamenoto, H. 1984. Ch romosomes in orchids: cou nting and nu mbers. Pp. 323 ­ 410 in: A rdit ti, J. (ed.), Orchid Biolog y: Reviews and Perspectives, vol. 3. Cor nell Universit y Press, Ithaca. Vogel, J.C., Rumsey, F. J., Schnel ler, J. J., Barret t, J.A. & Gibby, M. 1999. W here are the glacial ref ugia in Eu rope? Evidence f rom pter idophy tes. Biol. J. Linn. Soc. 66: 23 ­ 37. Wendel, J.F., Schnabel, A. & Seelanan, T. 1995. Bidirectional interlocus concer ted evolution following allotepoly ploid speciation in cot ton (Gossypium). Proc. Natl. Acad. Sci. U.S.A. 92: 280 ­284.

1208


Appendix. Accessions sampled in this study and their ITS alleles and plastid haplotypes.
a

Country

Species 11861 SchĆnswetter & Tribsch 603A SchĆnswetter & Tribsch 603B SchĆnswetter & Tribsch 603C SchĆnswetter & Tribsch 603D SchĆnswetter & Tribsch 603E Bateman 139 SchĆnswetter & Tribsch 6301 WU SchĆnswetter & Tribsch 6301 WU SchĆnswetter & Tribsch 6301 WU SchĆnswetter & Tribsch 6301 WU Civeyrel & al. 574B Civeyrel & al. 574C Civeyrel & al. 583 Civeyrel & al. 584 Civeyrel & al. 590 Chase 13813a (spirit) Chase 13813b (spirit) Chase 13817a (spirit) Chase 13817b (spirit) Chase 13819 (spirit) A A A A N B A A A B B He He Ir Fay K1874.3 (spirit) Bateman 738 No voucher N A A A A A A 70"72" 70(2)72(1) 70(3)72(1) 70(3)75(1) 70 70"75" 70"75" "70"75 70(1)75(3) 70(1)75(1) "70"75 "70"75 75 75 75 70(2)75(1) ~70(1)>75(1) "70"75 A 72(1)75(1) A 70 80 80 80 80 80 80 80 80 80 80 72(3)80(1) 72(1)80(3) 72"80" 72(1)80(1) 72"80" 72"77" 72 80 72(1)80(2) 80 N 70(1)75(2) N ~70(2)75(3) 72(1)80(2) 72(1)80(1) A 70(1)75(1) 72(1)80(2) N "70"75 72(1)80(1) I-III V-I-III V-I V-I-III V VI-III VV-VI V-VI V-III V V V III I-V V-I-III I I-III I I I V-III V-I-III III 11862 11863 11864 11865 15166 14618 14619 14620 14621 12124 12125 12139 12140 12157 13813 13814 13817 13818 13819 15189 Fr Fr 15190 14607 14699 14606 Birr Castle NNW Braunwald Sorvilier Massif Central, Vallee de la Dourbie Massif Central, Vallee de la Dourbie Bourgogne, Flacey Fr Bourgogne, Flacey Fr Bourgogne, Flacey Fr Bourgogne, Sagy Fr Bourgogne, Sagy Fr Pyrenees, Espezel Fr Pyrenees, Pla des Ails Fr Pyrenees, Pla des Ails Fr Pyrenees, Nalzen Fr Pyrenees, Nalzen Fr Paklenica National Park Cr Paklenica National Park Cr Paklenica National Park Cr Paklenica National Park Cr NE Scharnitz Au Below Schneealpen Au Below Schneealpen Au Below Schneealpen Au Below Schneealpen Au Below Schneealpen Au A 70(1)75(1) "72"80 V-III

Number Origin

Voucher

Haplotype Gel reading of the ITS fragments (lengths in bp)c Dfuch Dmac Allelesc Major I V (III)

b

Minor Ratio 1:1 1:1

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

TA XON 56 (4) · November 2007: E1­E17

D. fuchsii

D. fuchsii

2:1 VI 2:1 3:1 3:1 III III V (III) V (III) V III II 1:1 3:1 2:1:1

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

2:1 V

D. fuchsii

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. fuchsii

E1


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major V III V-III V V-III V 80 80 80 80 80 70(1)75(1) A A A A A A A A A A 75 70(2)72(1)75(1) 70(2)75(1) 70(1)75(1) 70(1)75(1) ~70(1)75(1) 70 70(3)75(1) 70 70 80 80 80 80 80 80 80 80 80 80 80 V-III V V V V-III V-III III V-III-VI V-III V-III V-III V-III V V-III V V VI-III III III "70"75 75 70(1)75(1) 70"75" 70(1)75(1) 70 70(3)75(1) 70"72""75" 70 70 70(1)75(1) 80 80 80 80 80 80

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E2 Number Origin 3877 HedrČn 97108 HedrČn 97037 HedrČn 97096 HedrČn 97184 HedrČn 97186 HedrČn 97221 HedrČn 97282 A N A Chase 10240(1) Chase 10240(2) Chase 10240(3) Chase 13810 (spirit) Chase 14075 (spirit) Chase 10243(1) Chase 10243(2) Chase 10243(3) Chase 10243(4) Chase 10243(5) Fay 635a no voucher Fay 635b no voucher Fay 670a (spirit) UK UK UK Chase 13786a (spirit) Chase 13786b (spirit) Chase 13786c (spirit) A N A A A A N A 3971 3976 H356 3989 3996 5547 H174 H175 O-1375 10240 10241 10242 Nr. Cardiff Nr. Cardiff Uffington Uffington Uffington Uffington Uffington Anglesey, Rhos-y-gard Anglesey, Rhos-y-gard Anglesey, Malltraeth Avon, Cuckoo Lane Avon, Cuckoo Lane Avon, Cuckoo Lane UK UK UK UK UK UK UK UK UK UK 14075 10243 10244 10245 10246 10247 13526 13527 13746 13786 13787 13788 Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Gotland, Fide Parish Sw VÄstragotaland, Jattened Sw VÄstragotaland, Jattened Sw JÄmtland, Hammerdal Sw SĆdermanland, Svarta Sw Ostergotaland, Karna Sw Ostergotaland, Karna Sw Gotland, Gerum Sw Skane, Maryd Sw Gotland, Boge Sw A Voucher Minor Ratio 1:1 1:1 3:1 1:1 1:1 2:1:1 2:1 1:1 1:1 3:1 A A A A 70 ~"70"72(1)75(2) 70(2)75(1) 70(3)75(1) 80 80 80 80 V III-VI V-III V-III V 2:1 3:1

Appendix. Continued.

Species

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii (hyperchromic) 13810

D. fuchsii ? hybrid

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

TA XON 56 (4) · November 2007: E1­E17

D. fuchsii


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major V V-III V>III V V V-III 80 80 80 72(1)80(2) 70(1)<75(1) A A N N A "70"75 75 70(3)75(1) 70 70 ~72(1)80(2) 80 ~72(2)80(1) 80 80 80 III V-III V V-I V-I III I-III V-III V V (III) (III) V V III III 70"75" 70(2)75(1) 70>75 70 70"75" 70(1)75(1) "70"75 70(1)75(1) 70 ~70(1)>75(1) 80 80 80 80 80 80

Species 13789 Chase 13786d (spirit) Chase 13786f (spirit) Chase 13786g (spirit) Chase 13786h (spirit) Chase 14195a (spirit) A Chase 10284 Chase 11794 A Au SchĆnswetter & Tribsch 3065 herbario Gutermann SchĆnswetter & Tribsch 3065 herbario Gutermann SchĆnswetter & Tribsch 3065 herbario Gutermann SchĆnswetter & Tribsch 3065 herbario Gutermann SchĆnswetter & Tribsch 30533055 WU SchĆnswetter & Tribsch 30533055 WU SchĆnswetter & Tribsch 30533055 WU SchĆnswetter & Tribsch 3056 WU Au SchĆnswetter & Tribsch 6290 WU SchĆnswetter & Tribsch 6290 WU A Au Au Au Au Au Au Au A A A A A A A 13791 13792 13793 14195 O-784 10284 11794 O-1123 14593 14594 14595 14596 14597 14598 14599 14600 14747 14748 Steiermark, Hochschwabgebiet OberĆsterreich, Mollner Voralpen OberĆsterreich, Mollner Voralpen OberĆsterreich, Mollner Voralpen OberĆsterreich, Mollner Voralpen Steiermark, Rottenmanner Tauern Steiermark, Rottenmanner Tauern Steiermark, Rottenmanner Tauern Steiermark, Rottenmanner Tauern Bateman isozyme Kew 1998-2727 Kew 1997-6379, Surrey, North Downs UK Box Hill UK Sussex, Cradle Hill UK Avon, Cuckoo Lane UK Avon, Cuckoo Lane UK Avon, Cuckoo Lane UK Avon, Cuckoo Lane UK A

Number Origin

Voucher

Minor Ratio 2:1

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

TA XON 56 (4) · November 2007: E1­E17

D. fuchsii

1:1

D. fuchsii "Rachel"

D. fuchsii

1:1

D. fuchsii

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

3:1

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

N A A

70 70 70"75"

80 80 80

V V V III

D. fuchsii (collected as maculata s.l.)

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. fuchsii (collected as maculata s.l.)

Steiermark, Hochschwabgebiet Au

E3


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major V-III III-V I-III I "72"80 "72"80 "72"80 70(2)75(1) A A A A A A A N ~ 70(1)75(1) 70(1)"72"75(2) 70(3)75(1) 70"72" 70(2)72(1)75(3) "72"70 72(1)75(3) ~70(1)75(1) 72(1)80(2) ~72(1)80(1) 80 80 80 80 80 80 80 III III V-III V-I V-I III-V III-VI V V-III-VI V III-VI V-III VI VI I V-I I (III) (III) VI I 70(1)75(1) 70(1)75(3) 75 75 75 "70"75 ~70(1)75(2) 72 ~72(1)80(1) "72"80 80

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E4 Number Origin 14749 SchĆnswetter & Tribsch 3069 herbario W. Gutermann SchĆnswetter & Tribsch 3069 herbario W. Gutermann SchĆnswetter & Tribsch 3047 WU SchĆnswetter & Tribsch 6403 SchĆnswetter & Tribsch 6403 SchĆnswetter & Tribsch 6331 SchĆnswetter & Tribsch 6331 SchĆnswetter & Tribsch 6331 SchĆnswetter & Tribsch 6331 No voucher No voucher No voucher No voucher No voucher No voucher No voucher Bateman 738 UK UK UK Bateman & Rudall s.n. Bateman & Rudall s.n. Bateman & Rudall s.n. A A A N N A A 14750 14751 14692 14693 14622 14623 14624 14625 10285 10286 10287 10288 10289 10290 13750 14697 17040 17042 17043 Cornwall, St. Ives Cornwall, St. Ives Cornwall, St. Ives Steingletscher He Anglesey, Malltraeth UK Bix Bottom UK Bix Bottom UK Bix Bottom UK Bix Bottom UK Bix Bottom UK Bix Bottom UK Trento, Lagorai It Trento, Lagorai It Trento, Lagorai It Trento, Lagorai It Pyrenees, Pic du Canigou Fr Pyrenees, Pic du Canigou Fr Steiermark, Kalkalpen Au Steiermark, Schladminger Tauern Au Steiermark, Schladminger Tauern Au Q Voucher Minor Ratio 1:1 3:1 2:1 1:1 2:1 3:1 3:1 A A A A 70(2)75(1) 75 75 75 80 80 80 80 V-III III III III 2:1

Appendix. Continued.

Species

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii (collected as maculata s.l.)

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii

D. fuchsii ?

D. fuchsii (collected as maculata s.l.)

D. fuchsii v. cornubiensis

D. fuchsii v. cornubiensis

TA XON 56 (4) · November 2007: E1­E17

D. fuchsii v. cornubiensis


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major III III III V-III V-III V-III 80 72 80 80 72 72(1)75(1) G C E A A A B B B 72 72 ~70(1)73(1) "72"73 70(1)75(3) "70"75 75 80 80 80 80 80 ~70(1)72(2)75(1) 80 80 ~70(1)72(2)75(1) 80 72(3)80(1) 72"80" 72"80" V-III I V III-VI VI III-VI VI VI V-X V-III-VI X V-III-VI I-V I I (III) V (III) III VI III-VI 75 75 75 70(1)75(1) 70(1)75(1) 70(1)75(1) 70(1)75(1) 75 70"72""75" 72(1)75(3) 80 80 80 80 80 80

Species 17045 Bateman & Rudall s.n. Bateman & Rudall s.n. Bateman & Rudall s.n. No voucher Waldren 02-05 Waldren 02-19 TCD No voucher Bateman 583 HedrČn 98080 HedrČn 98066 HedrČn 010619 HedrČn 010619 HedrČn 010619 Bateman 74 Gutermann Gutermann Gutermann Gutermann Hollingsworth 268 Hollingsworth 267 Hollingsworth 269 Hollingsworth 273 Ic Ir Ir No Hollingsworth 274 Hodkinson 1 Hodkinson 3 HedrČn 97037 W G A A B A A A A A A 17046 17047 14604 15554 15555 14605 15162 11797 11800 H201 H202 H204 15986 14614 14615 14616 14617 11798 14608 14609 14701 14702 Donegal Donegal 5553 VÄrntresk Lambafell Lambafell Ic Lambafell Ic Lambafell Ic Lambafell Ic Paklenica National Park Cr Paklenica National Park Cr Paklenica National Park Cr Paklenica National Park Cr Kastoria Gr SW Pavliani Gr SW Pavliani Gr SW Pavliani Gr 41°0625 N, 33°4489 E Tu 40°3987 N, 31°2552 E Tu Barra, Eoligarry UK Burren, Co. Clare Ir Carron, Co. Clare Ir Mullach Mor, Co. Clare Ir Carron, Co. Clare, Co. Clare Ir Cornwall, Tintangel UK Cornwall, Tintangel UK Cornwall, Tintangel UK A

Number Origin

Voucher

Minor Ratio

D. fuchsii v. cornubiensis

D. fuchsii v. cornubiensis

D. fuchsii v. cornubiensis

D. okellyi

1:1 1:1 1:1 1:1

D. okellyi

TA XON 56 (4) · November 2007: E1­E17

D. okellyi

D. okellyi ?

D. hebridensis

D. bythinica

D. saccifera

3:1 1:1

D. saccifera

D. saccifera

D. saccifera

D. saccifera

D. cf. saccifera

D. cf. saccifera

D. cf. saccifera

D. cf. saccifera

D. maculata (islandica)

3:1

D. maculata (islandica)

D. maculata (islandica)

D. maculata (islandica)

B B M M A

75 70(1)75(3) 75 75 70(1)75(1)

72 72(3)80(1) 72 72 80

I I-V I I V-III 1:1 (III) 3:1

D. maculata (islandica)

D. maculata (ericetorum)

D. maculata (ericetorum)

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. maculata

E5


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major I I I I I I 72(1)77(1) 72(1)75(1) ~72(1)80(1) 72 75 75 B B B B B B B 75 75 75 75 75 75 75 72 72 72 72 72 72 72 72 72 I-II V-I V-I-III I I I I I I I I I I (III) V 75 "70"75 75 75 75 75 75 70(1)75(1) 70(1)75(3) 75 72 72 72 72 72"80" 72

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E6 Number Origin Vila Nova de Mil Fontes Pinto s.n. Cotrim & Pinto HedrČn 97214 B B HedrČn 97261 HedrČn 97264a Hedren 980711 HedrČn 97235 Fay 659 (spirit) Fay 655a (spirit) Fay 655a (spirit) Fay 655b (spirit) Fay 655c (spirit) Fay 628a no voucher Fay 628b no voucher Bateman s.n. (spirit) Bateman s.n. (spirit) Bateman s.n. (spirit) Bateman s.n. (spirit) UK UK Bateman s.n. (spirit) Fay s.n. (no voucher) B B B B A X A N B Between Bejar and Barca d`ţvila (Avila) SĆdermanland, Svarta Gotland, Kauparve Gotland, Kauparve SmÄland, Sjomaden SmÄland, Sjomaden VÄstragotaland, Karshult Uppland, Langbromossen Anglesey, Nant Isaf Anglesey, Rhos-y-gard Anglesey, Rhos-y-gard Anglesey, Rhos-y-gard Anglesey, Rhos-y-gard Great Orme, Llandudno Great Orme, Llandudno Hertfordshire, Bricketwood common Hertfordshire, Bricketwood common Hertfordshire, Bricketwood common Hertfordshire, Bricketwood common Hertfordshire, Bricketwood common Lake District, Duddon Valley UK UK UK UK UK UK UK UK UK UK UK Sw Sw Sw Sw Sw Sw Sw Sp Po P 10684 3995 H38 H192 H365 H367 H371 H375 13500 13528 13528 13529 13530 13780 13781 14070 14071 14072 14073 14074 6505 Voucher Minor Ratio 1:1 1:1 B B B 75 75 75 72 72 72 I I I

Appendix. Continued.

Species

D. maculata (caramulensis) 11795

D. maculata

D. maculata

D. maculata

D. maculata

D. maculata

D. maculata

D. maculata

D. maculata

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

TA XON 56 (4) · November 2007: E1­E17

D. maculata (ericetorum)


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major I I I-V I I I 72 72 78 80 70(1)73(2) Z B ? 75 75 75 75 75 75 75 75 75 75 80 80 80 80 80 78(1)80(2) 80 80 80 80 80 I I VIII III X-V III* III* III* III* III*-IV III* III* III* III* III* V V 75 75 75 75 75 75 75 75 73 75 72 72"80" 72"80" 72(2)80(1) 72 72

Species 6506 Fay s.n. (no voucher) Chase 15191 (spirit) Chase 15192 (spirit) Chase 15552 (spirit) Chase 15553 (spirit) Fay 660 (spirit) Chase 0-537 Bateman 608 Bateman 366 Gr Bateman 108 Bateman 55 Bateman 760 Bateman 760 Bateman 760 Bateman 760 Bateman 760 Bateman 760 Bateman 760 Bateman 760 Balls B2991 Ferguson & al. 6374195 Al Sp Sp Sp Fr Joad 1882 Bateman 322 Chase O-718 Sandwitch 6304 HedrČn 990610 E UK Mo Mo Mo Mo Mo Mo Mo Mo Mo Mo T F D D N B B N N 15191 15192 15552 15553 13501 537 14695 15160 15156 994 15121 15122 15123 15124 15125 15126 15127 15128 Herb5 Herb1 Herb6 15159 718 Herb3 H346 Jaen Massif Central, Val de Trebans Sierra de Nevada Barcelona, Vic Parador gue de Constantine Cherfchaouene Tashdiert Western Atlas Western Atlas Western Atlas Western Atlas Western Atlas Western Atlas Western Atlas Western Atlas W Newtonferry, North Uist Kastoria From cultivation Pico de Jorge, Madeira Po Madeira Po Anglesey, Nant Isaf UK Llangurig, Ponterwyd UK Llangurig, Ponterwyd UK Llangurig, Ponterwyd UK Llangurig, Ponterwyd UK Lake District, Duddon Valley UK B

Number Origin

Voucher

Minor Ratio

D. maculata (ericetorum)

D. maculata (ericetorum)

D. maculata (ericetorum)

2:1

D. maculata (ericetorum)

TA XON 56 (4) · November 2007: E1­E17

D. maculata (ericetorum)

D. maculata (ericetorum; coll. as a potential hybrid)

D. foliosa

D. foliosa

D. aristata

D. cordigera

D. ebudensis

2:1

D. elata

D. elata

D. elata

D. elata

D. elata

2:1

D. elata

D. elata

D. elata

D. elata

D. cf. elata

D. elata

? B B ? B

75 "73"75 75 75 "73"75

80 72"80" 72 72 72(3)80(1)

III* I I I I X (III*) X (III*)

D. elata

D. elata

D. elata

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. elata

E7


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major I I I-X I I I 72(3)80(1) 72(3)80(1) 72(1)80(1) 72 72(1)80(3) 73(3)75(1) B B B C C N N A A A 73(3)75(1) 73"75" 73(3)75(1) 75 75 70(2)75(1) 75 70(1)75(1) 70(1)75(1) 70(1)75(1) 72(1)80(3) 72(1)80(3) "72"80 ~72(1)80(3) 80 80 80 80 80 80 80 I-X I-X I-X I X-I X-I X-I X X-I III III V-III III V-III V-III V-III I X X X 75 75 73(1)75(3) "73"75 "73"75 "73"75 73(1)75(3) 73(1)75(3) 73(1)75(1) 75 73(3)75(1) 72"80" 72"80" 72"80" 72(3)80(1) 72 72

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E8 Number Origin H66 HedrČn 990610 Wood 559 Chase 10155 Chase 10156 Chase 10157 Chase 10158 Chase 10159 Chase 10160 Chase 10161 Bateman 118 HedrČn 97152 HedrČn 97260 HedrČn 97259 HedrČn 97204 HedrČn 980711 Civeyrel & al. 591A Civeyrel & al. 591B Civeyrel & al. 591C Civeyrel & al. 591D SchĆnswetter & Tribsch 3059 WU SchĆnswetter & Tribsch 3059 WU SchĆnswetter & Tribsch 3059 WU SchĆnswetter & Tribsch 6294 WU SchĆnswetter & Tribsch 6294 WU SchĆnswetter & Tribsch 6294 WU SchĆnswetter & Tribsch 6294 WU B B B B B B B B B B ? Herb2 10155 10156 10157 10158 10159 10160 10161 15157 3978 3999 H293 3992 H101 12158 12159 12160 12161 14567 14568 14569 14570 14571 14572 14573 Steiermark, Hochschwabgebiet Au Steiermark, Hochschwabgebiet Au Steiermark, Hochschwabgebiet Au Steiermark, Hochschwabgebiet Au OberĆsterreich, Kirchdorf Au OberĆsterreich, Kirchdorf Au OberĆsterreich, Kirchdorf Au Pyrenees, Espezel Fr Pyrenees, Espezel Fr Pyrenees, Espezel Fr Pyrenees, Espezel Fr VÄstragotaland, Karshult Sw SĆdermanland, Kila Sw SmÄland, Sjomaden Sw SmÄland, Madesjo Sw SmÄland, S. Ljuna Sw W Galway Ir Lough Bunny Ir Lough Bunny Ir Lough Bunny Ir Poulsallagh Ir Poulsallagh Ir Poulsallagh Ir Moher Ir Aveyron Fr Massif Central, Compregnac Fr B Voucher Minor Ratio 3:1 3:1 3:1 3:1 3:1 3:1 3:1 3:1 2:1 1:1 1:1 1:1 C C C C "70"75 "70"75 "70"75 "70"75 80 80 80 80 III III III III V V V V

Appendix. Continued.

Species

D. elata

D. elata

D. occidentalis

D. occidentalis

D. occidentalis

D. occidentalis

D. occidentalis

D. occidentalis

D. occidentalis

D. kerryensis

D. sphagnicola

D. sphagnicola

D. sphagnicola

D. sphagnicola

D. sphagnicola

D. majalis

D. majalis

D. majalis

D. majalis

D. majalis

D. majalis

D. majalis

D. majalis

D. majalis

D. majalis

TA XON 56 (4) · November 2007: E1­E17

D. majalis


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major III III III III III III 80 80 80 80 "70"75 A 75 C C C B C C G 75 75 75 70(2)73(1)75(1) 70(1)75(1) 70(3)75(1) 70(3)73(1)75(3) 80 80 80 80 80 80 80 80 78(1)80(2) III V-III V-III V-III 75 III III III III III III V-III-X V-III V-III V-III-X-? V X-VI V V "70"75 "70"75 75 75 75 75 75 70(1)75(1) 70(1)75(1) 80 80 80 80 80 80

Species 14574 SchĆnswetter & Tribsch 6294 WU SchĆnswetter & Tribsch 6294 WU Civeyrel & al. 586A Civeyrel & al. 586C Civeyrel & al. 586D Civeyrel & al. 596F Civeyrel & al. 596G Chase 16543 (spirit) Chase 16544 (spirit) HedrČn 990606 Bateman 459 HedrČn 97028 Luo 9 Bateman 48 Bateman 727 HedrČn 990606 HedrČn 990611 Bateman 69 Bateman 70 Bateman 144 Ge Ir Sw Sw Bateman 461 Bateman 117 HedrČn 97074 HedrČn 97189 C A B C N C C C C C C 14592 12144 12146 12147 12149 12150 16543 16544 H65 15170 3969 O-1382 8039 O-963 14698 H382 H40 15983 15984 15158 15171 15165 3972 3979 Gotland, Hall Ostergotaland, Karna Ballindooly, N Galway Murnauer Moos KitzbÝhel [type] Au Alpes, Col du Sarenne, Bottom Fr Alpes, Col du Sarenne, Top Fr Massif Central, Eygas Fr Massif, Central, La Batie II Fr Braunwald He Andorra An Nyingch, southeastern Xizhang (Tibet) Ch Skane, Torma-Hallestad Sw Skane, Saxtorp Sw Bayern Ge Massif Central, La Batie I Fr Bourgogne, St Martin Fr Bourgogne, St Martin Fr Pyrenees, Pla des Ails Fr Pyrenees, Pla des Ails Fr Pyrenees, Pla des Ails Fr Pyrenees, Pla des Ails Fr Pyrenees, Pla des Ails Fr Steiermark, Hochschwabgebiet Au Steiermark, Hochschwabgebiet Au C

Number Origin

Voucher

Minor Ratio

D. majalis

D. majalis

D. majalis

D. majalis

D. majalis

TA XON 56 (4) · November 2007: E1­E17

D. majalis

D. majalis

D. majalis

1:1 1:1

D. majalis

D. majalis

70(1)"72""73"75(1) 80

D. majalis

D. majalis

D. majalis

D. majalis

D. alpestris

D. alpestris

D. alpestris

D. alpestris

1:1 3:1

D. alpestris

D. alpestris

D. traunsteineri

B E C C A

75 73 70(1)73(2)75(2) 70(1)75(1)"73" 70(1)"73"75(1)

80 80 80 80 80

III X III-X-V V-III V-III X X 1:1 1:1

D. traunsteineri

D. traunsteineri

D. traunsteineri

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. traunsteineri

E9


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major V-III V-III-X V-III-X V-III V-III V-III-X 80 80 80 80 80 70(3)73(2)75(1) C C C C C C C C C 70"73""75" 70(3)73(2)75(1) 70"73""75" 70(2)73(1) 70(3)73(1) 70"73" ~70(1)"73"75(1) 70(3)73(1) 70(1)"73"75(2) 80 80 80 80 80 80 80 80 80 80 V-III-X X V X-III X-V V-X-III V V-X-III V V-X V-X V V-III V-X III-V X X X X-III III-X X V III 70(1)75(1) 70(3)73(1)75(2) 70(2)73(1)75(1) 70(2)75(1) ~70(3)75(1) 70(2)73(1)75(2) 70>73>75 73 70"73" "70"73(3)75(1) 70(1)73(2)"75" 80 80 80 80 80 80

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E10 Number Origin 3980 HedrČn 98165 Hedren 97210 HedrČn 97250 HedrČn 97313 HedrČn 97310 HedrČn 010716 Fay 658a (spirit) Fay 658b (spirit) Fay 658c (spirit) Fay 658d (spirit) Fay 658e (spirit) Fay 658f (spirit) Fay 658g (spirit) Fay 633a (spirit) Fay 633b (spirit) Fay 633c (spirit) Fay 633d (spirit) Fay 633e (spirit) Fay 633f (spirit) Fay 633g (spirit) Bateman 468 Bateman 300 UK UK UK Bateman 325 Bateman Bateman C C C A E E B A A A A 3993 3998 5554 H327 H413 13493 13494 13495 13496 13497 13498 13499 13517 13518 13519 13520 13521 13522 13523 15161 15167 15169 15988 15989 Oxfordshire, Cothill fen Oxfordshire, Cothill fen Hants, Mapledurwell fen NE Yorks UK Hants, Exbury UK Anglesey, Rhos-y-gard UK Anglesey, Rhos-y-gard UK Anglesey, Rhos-y-gard UK Anglesey, Rhos-y-gard UK Anglesey, Rhos-y-gard UK Anglesey, Rhos-y-gard UK Anglesey, Rhos-y-gard UK Anglesey, Nant Isaf UK Anglesey, Nant Isaf UK Anglesey, Nant Isaf UK Anglesey, Nant Isaf UK Anglesey, Nant Isaf UK Anglesey, Nant Isaf UK Anglesey, Nant Isaf UK Gotland, Lojsthajd Sw Medelpad, Granboda Sw Medelpad, Ange Sw Uppland, Ed Sw SĆdermanland, Svarta Sw VÄstragotaland, Radane Sw A Voucher Minor Ratio 1:1 2:1:1 2:1 3:1 2:1 2:1 3:1 3:1 2:1 C C C C 70(1)"73"75(1) 70"73""75" 70(1)73(1)75(2) "70""73"75 80 80 80 80 V-III V III-V-X III V-X X X-III 2:1:1 1:1

Appendix. Continued.

Species

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri

D. traunsteineri bowmanii

D. traunsteineri

D. traunsteineri

D. traunsteineri

TA XON 56 (4) · November 2007: E1­E17

D. traunsteineri


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major V V-III V V V-III III-V 80 80 80 80 70(1)73(1) 70(1)73(3) A A A A E A E E A 70(1)73(3) ~70(1)73(2) 70(1)73(1) 72(1)73(1) 73 72(1)75(1) "70"73 73 70(1)73(3) 80 80 80 80 80 80 80 80 80 80 80 X X-VI X X X-V V V V-X V-X X V-X X-V X-V X-V V-X V III X X X 70"73""75" 70(3)75(1) 70"75" 70"73" 70(3)"73"75(1) 70(1)"73"75(2) 70 70(1)73(1) ~70(1)73(1) "70"73 80 80 80 80 80 80 III-X

Species O-990 Bateman 53 Bateman 52 Bateman 54 HedrČn 97269 HedrČn 97297 HedrČn 97305 Chase 9485 Bateman 51 Fay 671 (spirit) Fay 677a (spirit) Fay 677b (spirit) Fay 677c (spirit) Fay 677d (spirit) Fay 677e (spirit) Fay 677f (spirit) Chase 15193 (spirit) Chase 15194 (spirit) Chase 16547 (spirit) Chase 16549 (spirit) Chase 16550 (spirit) Bateman 46 Fay 668a (spirit) UK UK UK UK Fay 668b (spirit) Fay 668c (spirit) Fay 668d (spirit) Fay 669 (spirit) A A A A A A A A A C C O-996 989 4000 5550 5552 9485 987 13748 13763 13764 13765 13766 13767 13768 15193 15194 16547 16549 16550 964 13754 13755 13756 13757 13758 Anglesey, Newborough Warren Anglesey, Newborough Warren Anglesey, Newborough Warren Anglesey, Newborough Warren Anglesey, Newborough Warren UK Aberlady Bay, East Lothian UK Ynyslas, Dyfed UK Ynyslas, Dyfed UK Ynyslas, Dyfed UK Ynyslas, Dyfed UK Ynyslas, Dyfed UK Nr Llanfrothen, Pont Croesor UK Nr Llanfrothen, Pont Croesor UK Nr Llanfrothen, Pont Croesor UK Nr Llanfrothen, Pont Croesor UK Nr Llanfrothen, Pont Croesor UK Nr Llanfrothen, Pont Croesor UK Anglesey, Malltraeth UK Caithness, Thurso East UK Tartu Es Lycksele Sw Jamtland, Hammerdal Sw HÝrjdalen, Linsell, Hamra Sw E Skye, Raasay UK Wester Ross, Loch a Nhuilinn UK Wester Ross, Loch Kersary UK C

Number Origin

Voucher

Minor Ratio 3:1

D. traunsteineri

D. traunsteineri

D. lapponica

D. lapponica

D. lapponica

3:1 2:1 1:1

TA XON 56 (4) · November 2007: E1­E17

D. lapponica

D. baltica (longifolia)

D. purpurella

D. purpurella

D. purpurella

D. purpurella

1:1 3:1 3:1 1:1

D. purpurella

D. purpurella

D. purpurella

D. purpurella

D. purpurella

D. purpurella

D. purpurella

1:1

D. purpurella

D. purpurella

D. purpurella

3:1 A E A A A "70"73 73 "70"73 ~70(1)73(1)"75" ~70(1)73(1)"75" 80 80 80 80 80 X X X V-X V-X V III III V

D. purpurella

D. purp urella

D. purpurella

D. purpurella

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. purpurella

E11


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major V III III III X-VI 80 80 80 80 80 70(1)"72"73(1) A A A A C A C A C A 72"73" 75 72"73" 70(1)"73"75(3) "73"75 70(1)"73"75(3) 73(1)75(3) 72(3)73(1) "72"73 80 80 80 80 ~70(1)72(1)75(1) 80 80 80 80 80 80 80 X-III III III III V-X V-X VI III VI V-III-VI III-V III V-III III-X VI-X X VI X X X X VI VI VI III VI X X-III 70"73""75" 75 75 75 7273 73(2)75(1) "72"75 "72"75 "72"75 70(1)"75"73(1) 80 80 80 80 80

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E12 Number Origin Donegal Hodkinson 2 HedrČn 98081 HedrČn 98057 Denholm s.n. Chase 10235 (1) Chase 10235 (2) Chase 10235 (3) Chase 10235 (4) Chase 10235 (5) Chase 14052 (spirit) Chase 14053 (spirit) Chase 14054 (spirit) Chase 14056 (spirit) Chase 14057 (spirit) Chase 14059 (spirit) Bateman 469 Bateman 471 Bateman 470 Bateman 472 Bateman 76 Chase 16548 (spirit) Fr Tu Tu Civeyrel & al. 567 HedrČn 98020 A A A A A A A C C C 11801 11793 1124 10235 10236 10237 10238 10239 14052 14053 14054 14056 14057 14059 15173 15175 15174 15176 6098 16548 12122 11799 H31 Trabzon, Uzungol 40°4449 N, 39°3351 E Pyrenees, St Marcel-Aulon Ynyslas, Dyfed UK Novia Scotia, St Johns Ca Dorset, Hinton St Mary UK Hants, Brambridge UK Dorset, Hinton St Mary UK Hants, Brambridge UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Nr. Cardiff UK Hertfordshire, Blagrove common UK 41°0281 N, 34°0411 E Tu 40°3637 N, 31°1617 E Tu Ir A Voucher Minor Ratio 2:1 1:1 1:1 3:1 1:3 3:1 3:1 A C C 70(3)72(1)75(2) 75 75 80 80 80 V-III-VI III III

Appendix. Continued.

Species

D. purpurella

D. nieschalkiorum

D. nieschalkiorum (ilgazica)

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa

D. praetermissa junialis

D. praetermissa junialis

D. praetermissa junialis

D. praetermissa ?

D. praetermissa ? (collected as maculata s.l).

D. urvilleana

TA XON 56 (4) · November 2007: E1­E17

D. urvilleana


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major V V-III III-X I I V-X 80 80 80 73 E E E E E E 73 73 73 73 73 73 E E E E 73 73 73 73 80 80 80 80 80 "78"80 78(1)80(3) 80 80 80 80 X X X X X X X X X X-VIII X X X X VIII X 70 70(1)"73"75(1) ~73(1)75(1) 75 75 70(1)73(1) 73 73 73 80 72 72 80 80 80

Species 3985 HedrČn 97098 HedrČn 97226 A HedrČn 010710 HedrČn 010710 E Fischer & al. no voucher Fischer & al. no voucher Civeyrel & al. 585A Civeyrel & al. 585B Civeyrel & al. 585C Chase 16545 (spirit) Chase 16546 (spirit) E E E E X N A 3997 H134 H436 H440 H135 11857 11858 Au Fr Fr Fr Fr Fr Ir Ir Ir Sw Sw Sw Sw Sw Sw Sw Tu UK Fay 661a (spirit) HedrČn 97034 HedrČn 97082 HedrČn 97132 HedrČn 97164 HedrČn 97194 HedrČn 97213 12141 12142 12143 16545 16546 15179 15180 15181 3970 3975 3986 3987 3991 3994 0-1378 H4 13502 Erzurum, E Erzurum Anglesey, Nant Isaf Gotland, Harudden SĆdermanland, Svarta Ostergotaland, Slaka VÄstragotaland, Radane Uppland, Bladaker Gotland, Hall Skane, Troll-Ljungby Roscommon, Lough Rea Roscommon, Lough Rea Roscommon, Lough Rea Jura, Sarvagnat Jura, Sarvagnat Pyrenees, Pla des Ails Pyrenees, Pla des Ails Pyrenees, Pla des Ails Burgenland, Zitzmannsdorfer Wiese Burgenland, Zitzmannsdorfer Wiese Au Gotland, Gylvik Sw Vastragotaland, Dimbo Sw Vastragotaland, Dimbo Sw Gotland, Gylvik Sw Uppland, Frotuna Sw Gotland, Viklau Sw A

Number Origin

Voucher

Minor Ratio 1:1

D. sp. (allotet)

D. sp. (allotet)

D. sp. (allotet)

D. sp. (allotet)

D. sp. (allotet)

TA XON 56 (4) · November 2007: E1­E17

D. sp. (allotet)

1:1

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

3:1

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

E E E E E

73 73 73 73(1)77(1) 73

80 80 80 80 78(1)80(3)

X X X X X-VIII X-XII 1:1 3:1

D. incarnata

D. incarnata

D. incarnata

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. incarnata

E13


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major X X X X X X 80 80 80 80 80 73 E E E E E E E E E E 73 73 73 73 73 73 73 73 73 73 80 80 80 80 80 80 80 80 80 80 80 X X X X X X X X X X X X X X X X 73 73 73 73 73 73 73 73 73 73 73 80 80 80 80 80 80

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E14 Number Origin 13503 Fay 661b (spirit) Fay 661c (spirit) Fay 632 (spirit) Fay 666 (spirit) Fay 675 (spirit) Bateman 582 Chase 15185 (spirit) Chase 15186 (spirit) Chase 15187 (spirit) E E HedrČn 97078 HedrČn 97178 HedrČn 97272 HedrČn 97287 Bateman 57 Chase 10163 Chase 10165 Bateman 115 Bateman 116 HedrČn 97075 HedrČn 97192 Bateman 612 Ge UK Bateman 462 Bateman 56 E E E E E E E E E 13504 13516 13751 13759 15178 15185 15186 15187 O-1385 O-1388 3974 3988 5546 5549 6097 10163 10165 15163 15164 3973 3990 14696 15172 O-988 Wester Ross, Kernsary Murnauer Moos Cambs, Chippenham fen Ostergotaland, Kaga Sw UK Gotland, Hall Sw Lough Bunny Ir Lough Carra, Mayo Ir Lough Bunny Ir Mulloch Mor Ir Wester Ross, Lochdroma UK JÄmtland, Hammerdal Sw HÝrjedalen, Linsell Sw Ostergotaland, Karna Sw Gotland, Hall Sw Gotland Sw Gotland Sw Dyfed, Ynyslas UK Dyfed, Ynyslas UK Dyfed, Ynyslas UK Barra, Eoligarry UK Nr. Llanfrothen UK Anglesey, Newborough Warren UK Anglesey, Rhos-y-gard UK Anglesey, Nant Isaf UK Anglesey, Nant Isaf UK E Voucher Minor Ratio E E E 73 73 73 80 80 80 X X X

Appendix. Continued.

Species

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata

D. incarnata s.l.

D. incarnata s.l.

D. cruenta

D. cruenta

D. cruenta

D. cruenta

D. cruenta

D. cruenta

D. cruenta

D. cruenta

D. cruenta

D. ochroleuca

D. ochroleuca

D. ochroleuca

D. ochroleuca

TA XON 56 (4) · November 2007: E1­E17

D. pulchella


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major X X X X X X 80 80 80 80 73 73 E K Y J E S1 S2 S1 73 73 73 73 73 75 75 75 S1 75 80 80 80 80 80 73 80 80 80 80 80 X X X X X X X X X XI X III III III III V 73 73 73 73 "70"73 73 73 73 73 73 80 80 80 80 80 80

Species 13505 Fay 662 (spirit) Chase 13783a (spirit) Chase 13783b (spirit) Chase 13783c (spirit) Chase 15182 (spirit) Bateman 45 Fay 667a (spirit) Fay 667b (spirit) Chase 15183 (spirit) Chase 15184 (spirit) HedrČn 97286 HedrČn 97303 HedrČn 000609 HedrČn 98025 Chase O-960 HedrČn 98078 E E E E E E E E E E E 13783 13784 13785 15182 O-965 13752 13753 15183 15184 5548 5551 H39 12821 H37 O-960 12822 O-1373 O-1374 O-1377 14610 14611 14612 14613 15168 Huesca, Valle de echo Paklenica National Park Paklenica National Park Paklenica National Park Cr Cr Cr Sp Paklenica National Park Cr Gotland, Hejnum Sw No voucher No voucher No voucher No voucher Bateman 315 Gotland, Valar Sw Gotland, Valar Sw 40°3637 N, 31°1617 E Tu Kew 1982-1515, Troodos Monts Cy Trabzon, Zigana Pass Tu 40°4152 N, 40°4164 E Tu Artvin, Savsat Tu Lycksele Sw JÄmtland, Hammerdal Sw Dyfed, Ynyslas UK Dyfed, Ynyslas UK Anglesey, Newborough UK Anglesey, Newborough UK East Lothian, Aberlady Bay UK Dyfed, Ynyslas UK Surrey, Thursley Heath UK Surrey, Thursley Heath UK Surrey, Thursley Heath UK Anglesey, Nant Isaf UK A

Number Origin

Voucher

Minor Ratio

D. pulchella

D. pulchella

D. pulchella

D. pulchella

D. pulchella

TA XON 56 (4) · November 2007: E1­E17

D. coccinea

D. coccinea

D. coccinea

D. coccinea

D. coccinea

D. borealis

D. borealis

D. armeniaca

D. euxina

D. euxina

D. iberica

D. iberica

D. sambucina

D. sambucina

D. sambucina

D. sambucina

D. sambucina

S1 S1 S1 S1

75 75 75 75

80 80 80 80

III III III III

D. sambucina

D. sambucina

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

D. sambucina

E15


Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major III IX IX IX VIII VIII 78 78 _ V1 73 V3 73 V6 73 V1 73 V4 73 H H U I C 73 73 75 73(1)75(1) "72""73"75 _ _ 78 _ _ 80 80 80 80 80 VIII VIII IX IX IX VIII IX IX X X III X-III 75 76 76 76 78 78 78 78 78 80

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E16 Number Origin 15188 O-760 Rossi Bateman 522 Bateman 34 Luo & Luo 657 Luo & Luo 667 Luo & Luo 668 Luo & Luo 677 Bateman 71 Bateman 66 Hartwell s.n. HedrČn 98033 HedrČn 980706 Albach 430 Albach 430 Albach 431 Albach 431 Chase 10283 UK UK UK Fay 664 (spirit) Fay 665 (spirit) Fay 679 (spirit) V5 73 V1 73 V2 73 V5 73 V3 73 R3 R2 15177 15987 13070 13074 13075 13079 15985 15982 O-576 15979 15980 O-1396 14602 14603 15155 14691 10283 13507 Anglesey, Nant Isaf Nr Llanfrothen 13513 Anglesey, Nant Isaf Kew 1980-2844, W Suffolk, Elmswell UK Georgia Go Georgia Go Georgia Go Georgia Go Sw 40°4037 N, 40°4259 E Tu UK UK Fifeness, Fife UK Alpes, Col du Sarenne Fr Jinchuan County, northwestern Sichuan Ch Hongyuan County, northwestern Sichuan Ch Hongyuan County, northwestern Sichuan Ch Li Xian County, northwestern Sichuan Ch Crete, NW Spili Gr Sicily, NE Etna It Siena It R1 Massif Central, Vallee de la Dourbie Fr S1 Voucher Minor Ratio 1:1 III C C A 70(3)73(1) 75>72 73(1)75(2) 80 72>80 72(2)80(1) V-X I-III I-X (III) 2:1 VI-X 3:1

Appendix. Continued.

Species

D. sambucina

D. romana

D. romana

D. romana

D. viridis

D. viridis

D. viridis

D. viridis

D. viridis

D. viridis

D. viridis

D. viridis

D. viridis

D. viridis

D. sp.

D. sp.

D. sp.

D. sp.

D. transiens

D. hybrid

D. maculata â traunsteineri 13508

TA XON 56 (4) · November 2007: E1­E17

D. fuchsii â maculata ?


Appendix. Continued.

Country

Haplotype Gel reading of the ITS fragments (lengths in bp) Dfuch Dmac Alleles Major X-III III-X 73(2)75(1) 73(1)75(2) 80 80

Species 12145 Civeyrel & al. 586B Civeyrel & al. 596E E 12148 Pyrennees, Pla des Ails Fr Pyrennees, Pla des Ails Fr E

Number Origin

Voucher

Minor Ratio 2:1 2:1

D. majalis â incarnata

D. majalis â incarnata

Note: Empt y cells denote u navailable infor mation.

a

Abbrevations: Al, Alger ia; A n, A ndor ra; Au, Aust r ia; Ch, China; Cr, Croatia; Cy, Cy pr us; Es, Estonia; Fr, France; Ge, Ger many; Go, Georgia; Gr, Grecce; He, Switzerland; Ic, Iceland; Ir, Ireland; It, Italy; Mo, Morocco; No, Nor way; Po, Por t ugal; Sp, Spain; Sw, Sweden; Tu, Tu rkey; U K, United K ingdom

TA XON 56 (4) · November 2007: E1­E17

b

All in K u nless other wise noted.

c

A band was considered as minor (indicated bet ween quotation marks " " ), if it was more than th ree times weaker than the st rongest band. Similarly an allele is minor if its quantit y is at least th ree times smaller than the one of the main allele, other wiese it was considered as major. Number in brackets for the gel reading indicate the relative propor tion of a given band. Brackets are used for alleles that were not obser ved but for which we could not reject the presence due to band overlap. The ratio given only takes into accou nt the major alleles.

*Same leng th as III, but this allele has substit utional differences when sequenced.

Pillon & al. · Evolut ion and diversif icat ion in Dact ylorhiza

E17