SEARCH

SEARCH BY CITATION

Keywords:

  • Africa;
  • Asia;
  • biogeography;
  • cladistics;
  • Europe;
  • homoplasy;
  • ITS;
  • North America;
  • orchids;
  • rDNA

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

Internal transcribed spacer (ITS nuclear rDNA) data have been obtained from 190 terrestrial orchid species, encompassing all genera and the great majority of the widely recognized species of Orchidinae, a heterogeneous selection of species of Habenariinae, and single species of Satyriinae and Disinae (the latter serving as outgroup). The resulting parsimony-based phylogeny reveals 12 well-resolved clades within the Orchidinae, based on Anacamptis s.l., Serapias, Ophrys, SteveniellaHimantoglossum s.l. (including ‘Comperia’ and ‘Barlia’, most species being 2n = 36), Neotinea s.l., TraunsteineraChamorchis, Orchis s.s., PseudorchisAmerorchisGalearisNeolindleyaPlatanthera s.l. (most 2n = 42), Dactylorhiza s.l., Gymnadenia s.l. (most 2n = 40, 80), Ponerorchis s.l.Hemipilia s.l.AmitostigmaNeottianthe, and Brachycorythis (most 2n = 42). Relationships are less clearly resolved among these 12 clades, as are those within Habenariinae; the subtribe appears either weakly supported as monophyletic or as paraphyletic under maximum parsimony, and the species-rich genus Habenaria is clearly highly polyphyletic. The triphyly of Orchis as previously delimited is confirmed, and the improved sampling allows further generic transfers to Anacamptis s.l. and Neotinea s.l. In addition, justifications are given for: (1) establishing Steveniella as the basally divergent member of an appreciably expanded Himantoglossum that incorporates the former genera ‘Barlia’ and ‘Comperia’, (2) reuniting ‘Piperia’ with a broadly defined Platanthera as section Piperia, necessitating ten new combinations, (3) broadening Ponerorchis to include Chusua, and Hemipilia to include single ‘orphan’ species of Ponerorchis and Habenaria, and (4) recognizing ‘Gymnadeniacamtschatica as the monotypic Neolindleya camtschatica within the PseudorchisPlatanthera clade. Few further generic transfers are likely in Orchidinae s.s., but they are anticipated among habenariid genera, on acquisition of additional morphological and molecular evidence; one probable outcome is expansion of Herminium. Species-level relationships are also satisfactorily resolved within most of the major clades of Orchidinae, with the notable exceptions of Serapias, the derived sections of Ophrys, Himantoglossum s.s., some sections within Dactylorhiza, the former genus ‘Nigritella’ (now tentatively placed within Gymnadenia s.l.), Hemipilia s.l., and possibly Ponerorchis s.s. Relationships among the 12 major clades broadly accord with bona fide records of intergeneric hybridization. Current evidence supports the recently recognized 2n = 36 clade; it also indicates a 2n = 40 clade that is further diagnosed by digitate root-tubers, and is derived relative to the recently recognized clade of exclusively Asian genera (Ponerorchis s.l.Hemipilia s.l.AmitostigmaNeottianthe). This in turn appears derived relative to the Afro-Asiatic Brachycorythis group; together, these two clades identify the plesiomorphic chromosome number as 2n = 42. If the African genus Stenogolottis is correctly placed as basally divergent within a monophyletic Habenariinae, the tribe Orchideae and subtribes Orchidinae and Habenariinae could all have originated in Africa, though in contrast the Asiatic focus of the basally divergent members of most major clades of Orchidinae suggests an Asiatic radiation of the subtribe. Morphological characters informally ‘mapped’ across the molecular phylogeny and showing appreciable levels of homoplasy include floral and vegetative pigmentation, flower shape, leaf posture, gynostemium features, and various pollinator attractants. Qualitative comparison of, and reciprocal illumination between, degrees of sequence and morphological divergence suggests a nested set of radiations of progressively decreasing phenotypic magnitude. Brief scenarios, both adaptive and non-adaptive, are outlined for specific evolutionary transitions. Recommendations are made for further species sampling, concentrating on Asian Orchidinae (together with the Afro-Asiatic Brachycorythis group) and both Asian and Southern Hemisphere Habenariinae, and adding plastid sequence data. Taxonomic changes listed are: Anacamptis robusta (T.Stephenson) R.M.Bateman, comb. nov., A. fragrans (Pollini) R.M.Bateman, comb. nov., A. picta (Loiseleur) R.M.Bateman, comb. nov., Neotinea commutata (Todari) R.M.Bateman, comb. nov., N. conica (Willdenow) R.M.Bateman, comb. nov., Platanthera elegans Lindley ssp. maritima (Rydberg) R.M.Bateman, comb. nov., P. elegans Lindley ssp. decurtata (R.Morgan & Glicenstein) R.M.Bateman, comb. nov., P. elongata (Rydberg) R.M.Bateman, comb. nov., P. michaelii (Greene) R.M.Bateman, comb. nov., P. leptopetala (Rydberg) R.M.Bateman, comb. nov., P. transversa (Suksdorf) R.M.Bateman, comb. nov., P. cooperi (S.Watson) R.M.Bateman, comb. nov., P. colemanii (R.Morgan & Glicenstein) R.M.Bateman, comb. nov., P. candida (R.Morgan & Ackerman) R.M.Bateman, comb. nov. and P. yadonii (R.Morgan & Ackerman) R.M.Bateman, comb. nov. © 2003 The Linnean Society of London, Botanical Journal of the Linnean Society, 2003, 142, 1–40.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

During the last decade, molecular phylogenetics has undergone an exponential increase in both the range of techniques available (e.g. Soltis, Soltis & Doyle, 1998; Hollingsworth, Bateman & Gornall, 1999; Chase, Fay & Savolainen, 2000a) and the spectrum of topics to which they are applied. Early studies understandably addressed broader phylogenetic questions, notably the origins of major clades of land-plants (summarized in Kenrick & Crane, 1997) or the relationships among, and re-delimitation of, angiosperm families (Soltis et al., 2000; see also Angiosperm Phylogeny Group,1998). Predictably, early molecular studies of the Orchidaceae sought to discover the closest sister-group to the family (e.g. Chase et al., 1994; Dressler & Chase, 1995) or the relationships of major groups within the family (e.g. Kores et al., 1997). In our opinion, despite much progress, both issues have yet to be unequivocally resolved (cf. Rasmussen, 1999; Rudall & Bateman, 2002).

More recently, it became clear that evolutionary interpretations of plant groups would benefit from more focused phylogenetic surveys that maximized species-level coverage within specific clades at lower taxonomic levels. The Orchidaceae is a prime candidate for such studies (e.g. Chase, 1999; Pridgeon et al., 1999, 2001; Bateman, 2001), given that it is arguably (a) the second-largest extant plant family and (b) one of the most recent species-rich plant families to undergo a major evolutionary radiation. However, Chase (2001) recently postulated a rather earlier origin (c. 100 Ma), an assertion based largely on molecular clock evidence (e.g. Sanderson, 1997).

Here, we greatly extend the sampling of recent pioneering nrDNA ITS sequence analyses of the subtribe Orchidinae (Bateman, Pridgeon & Chase, 1997; Pridgeon et al., 1997; Bateman, 1999a, 2001; Bateman et al., 2001). Although this is the most species-rich and ecologically dominant subtribe of Orchidaceae in the much-studied European orchid flora, those clades more basally divergent within the subtribe also have representative species in Asia and North America. Our objective is to encompass all of the widely recognized extant species assigned to the subtribe, with the intention of gaining a better understanding of patterns of molecular and morphological character change and, ultimately, of the underlying evolutionary processes.

REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

MORPHOLOGY

Thus far, morphological cladists have focused on broad-brush (and typically non-algorithmic) analyses of the entire orchid family. Sampling was therefore rather sparse, and the coded taxa were required to represent high taxonomic levels in Burns-Balogh & Funk (1986) and Dressler (1993). Dressler's (1993: fig. 7.6) benchmark, but rather character-deficient, phylogeny of the Orchideae and Diseae at subtribal level yielded a polytomy of the Orchidinae, Habenariinae and Diseae. The genus-based phylogenetic analysis of Freudenstein & Rasmussen (1999: fig. 1) generated a fairly well-supported, derived Orchideae–Diseae clade, but no meaningful resolution among the four representative genera that were included (Disa, Satyrium, Platanthera, Dactylor-hiza). Sadly, despite extensive morphological study over two centuries, there has been no work on the Northern Hemisphere Orchidinae to match the morphological phylogenetic monographs of South African orchids by H. P. Linder and H. Kurzweil (e.g. Linder & Kurzweil, 1994, 1999; Kurzweil & Linder, 1999).

ALLOZYME AND RFLP ANALYSES

Schlegel et al. (1989) obtained allozyme data from a range of members of the genus Orchis as then conceived. In retrospect, the incongruity of Anacamptis morio and Neotinea ustulata, both traditionally ascribed to Orchis s.l., is clearly evident in their much longer terminal branches relative to those subtending the six species of Orchis s.s. that were analysed. In a more detailed analysis, Rossi et al. (1994) then correctly inferred considerable genetic distances among four species groups: the derived portion of our Anacamptis s.l. clade (e.g. A. morio), the basal portion of our Anacamptis clade (e.g. A. laxiflora), the anthropomorphic members of Orchis s.s. (including the former monotypic genus ‘Aceras’) and the more derived portion of that clade (represented by O. quadripunctata and O. brancifortii). However, Rossi et al. did not examine members of the third clade that was formerly included in Orchis, namely Neotinea s.l.

Cozzolino et al. (1998) then performed an RFLP (restriction fragment length polymorphism) study of plastid DNA that examined nine members of the former Orchis s.l. plus the related ‘Acerasanthropophorum and Anacamptis pyramidalis (but again lacking Neotinea s.l.). Also included in the ingroup was Dactylorhiza saccifera, and the outgroup consisted of Serapias vomeracea and Cephalanthera rubra. However, various sequencing studies show that Cephalanthera is far too distant from Orchis s.l. to form a reliable outgroup (a point noted by Cozzolino et al., 1998; Aceto et al., 1999), and that the second ‘outgroup’, Serapias, is in fact more closely related to Orchis s.l. than is one of the supposed ingroups, Dactylorhiza (cf. Pridgeon et al., 1997). Not surprisingly, the resulting preferred most-parsimonious tree (Cozzolino et al., 1998: fig. 1b) deviates considerably from our own recent research (Bateman, 2001; Bateman et al., 2001), especially near the base, yielding congruent topologies only with the pairing of Orchis mascula with O. pauciflora and of species within our Anacamptis s.l. clade. Moreover, Anacamptis pyramidalis was subsequently suspected by Aceto et al. (1999) of having undergone plastid capture, though recent unpublished trnL analyses suggest that this assum-ption was probably incorrect (R. Bateman and P. Hollingsworth, unpubl. obs.).

PLASTID DNA SEQUENCING

Early molecular phylogenetic studies of the Orchidaceae that included representative species of Orchideae used the plastid genes rbcL and ndhF. The first rbcL study (Chase et al., 1994: fig. 2) analysed Disa tripetaloides and Platanthera ciliaris, which proved to be well-supported sister-genera functioning as an aggregate placeholder for the Orchideae. Three approximately contemporaneous studies of ndhF used Habenaria repens as the placeholder for the Orchideae (Neyland & Urbatsch, 1995, 1996a,b). This species has since proved a less than ideal representative (see ‘Habenariids’ below), and was replaced by Orchis quadripunctata as the placeholder for Orchideae in an nad1 analysis of epidendroids (Freudenstein, Senyo & Chase, 2000a).

Extending earlier studies of rbcL, with its limited number of variable sites, Kores et al. (1997: fig. 1) added to D. tripetaloides and H. repens the newly sequenced Satyrium nepalense and Orchis quadripunctata. The resulting tree gave strong jack-knife support to monophyly of the Orchideae but no support to the preferred pectinate topology within the clade of (((Satyrium, Disa) Habenaria) Orchis). The next, more densely sampled rbcL study of the Orchidaceae (Cameron et al., 1999: fig. 4) restored to the analysis P. ciliaris and added Ophrys apifera, overturning all of  the relationships inferred by Kores et al. to yield the topology (((Satyrium, Platanthera) Habenaria) (Ophrys, Orchis) Disa), albeit again without strong support for the internal nodes. Kores et al. (2000: fig. 1) analysed ten species (identities not specified) of nine genera of Orchideae in a matK study focusing on Diurideae, again yielding a unique and highly contrasting topology of (((((Gymnadenia, Platanthera) Orchis) (Stenoglottis, Holothrix)) ((Habenaria, Cynor-kis) Gennaria)) Satyrium), nested within five Diseae. Kores et al. (2001) further explored these relationships by combining the plastid regions matK and trnL, again including ten species of Orchideae and generating the topology (((((Ophrys, Holothrix) (Habenaria, Peristylus)) Satyrium) ((Disa, Monadenia) (Corycium, Pterigodium))) Disperis).

Thus, plastid sequences have contributed only modestly and equivocally to understanding of the phylogeny of the Orchideae, the primary advance being confirming the morphologically inferred monophyly of the Orchideae as a whole rather than gaining reliable insights on relationships within the group. This limitation largely reflected inadequate sampling of Orchideae in studies designed primarily to address relationships of orchids outside the subtribe.

NUCLEAR DNA SEQUENCING

Nuclear ribosomal ITS DNA of Orchidinae was first sequenced by Cozzolino et al. (1996); five species of Orchis s.l. were resolved using ITS1 into two of the three groups subsequently recognized by Pridgeon et al. (1997): ‘O.coriophora, ‘O.morio and ‘O.laxiflora (our Anacamptis s.l.) vs. O. simia and O. purpurea (our Orchis s.s.).

A greatly increased scale of sampling for molecular analysis in the Orchidinae was heralded by the nuclear ribosomal DNA matrix of Pridgeon et al. (1997) and Bateman et al. (1997). The 88 sequences of the ITS1–5.8S–ITS2 assembly (cf. Baldwin, 1992; Baldwin et al., 1995; Hershkovitz, Zimmer & Hahn, 1999) provided the forerunner of the present study. Data were analysed in detail (Fig. 1) and subsequent discussion topics included sequence evolution, inter- and infrageneric relationships and hybridization frequencies, comparison with the results of past studies (both non-phylogenetic and phylogenetic), key synapomorphies (including tubers and karyotypes), homoplastic characters (including floral morphology and pigmentation), and misleading information detected in the extensive literature on the group. Their primary taxonomic conclusion was that Orchis as then widely delimited was triphyletic. After discussing levels of sequence disparity necessary for particular taxonomic ranks and reviewing the taxonomic history of the genus, the authors controversially but logically reassigned the majority of species in the former Orchis to expanded concepts of Neotinea and especially of Anacamptis. Other taxonomic rationalizations were the inclusions of (a) Nigritella within an otherwise apparently paraphyletic Gymnadenia, (b) the monotypic Aceras within Orchis s.s., and (c) the monotypic Coeloglossum within Dactylorhiza (which required subsequent nomenclatural conservation: Cribb & Chase, 2001) (Fig. 1).

image

Figure 1. Summary of the ITS phylogenies of the Orchidinae presented by Pridgeon et al. (1997: figs 1–4), as estimated (a) without and (b) with indels. Thick branches indicate> 80% bootstrap support; cross-hatched groups are paraphyletic (see also Bateman et al., 1997: fig. 1). The number of named taxa analysed for each major clade is indicated in parentheses.

Download figure to PowerPoint

Bateman (1999a: fig. 19.3; see also Bateman et al., 1998) presented a preliminary ITS tree derived from an expanded matrix of 129 species, including five genera absent from the previous ITS studies. The short discussion focused on patterns of morphological homology within the clade, the wisdom (or otherwise) of ‘mapping’ such characters across molecular trees, and the potential of comparing morphological and molecular trees to detect saltational evolutionary events. The matrix was upgraded to 172 species of Orchideae for inclusion in Genera Orchidacearum 2 (Pridgeon et al., 2001), and further upgraded to the current matrix (together with Dactylorhiza cf. hatagirea) by Bateman (2001) in order to generate a Neighbour Joining tree.

An independent study of ITS1 and ITS2 by Aceto et al. (1999) presented trees for 30 species of Orchidinae (all but Orchis patens s.s. were analysed independently for the present study), confirming the triphyly of Orchis s.l. and the existence of a 2n = 36 clade already elucidated by Pridgeon et al. (1997). However, the placements of the Neotinea and Himantoglossum clades, and of ‘Aceras’, differed, perhaps because Gennaria diphylla, a highly divergent long-branch taxon (Pridgeon et al., 1997; Bateman et al., 2001), was chosen by Aceto et al. (1999) as the sole outgroup. This choice also gave ambiguous polarization for crucial characters such as karyotype.

The six genera of Orchidinae analysed by Cameron et al. (1999) for rbcL (albeit most represented by different species) were also incorporated in a later and taxonomically broader molecular study focusing on Diseae and using nuclear ITS sequences (Douzery et al., 1999: fig. 1). Nine sequences of Orchidinae and five of Habenariinae carried forward from the earlier study (Bateman et al., 1997; Pridgeon et al., 1997) were combined with a further six habenariids and 30 Diseae, revealing a paraphyletic Diseae s.l. (Diseae s.s. sandwiched within a polyphyletic Coryciinae, with Satyriinae relatively derived) subtending a monophyletic Orchidinae as sister to a monophyletic Habenariinae. Support for monophyly of Orchidinae is strong, but that for Habenariinae is very weak (bootstrap values < 50%), causing Pridgeon et al. (2001) to sink Habenariinae into Orchidinae. Interestingly, the topology of Cameron et al. (1999) resembles the morphological subtribal phylogeny of Dressler (1993), but with a highly contrasting rooting that reverses the polarity of most key characters. However, except for the near-basal divergence of Disa, the results of this study differed strongly from those of the preceding, less well-sampled rbcL analyses. Within the Orchidinae/Habenariinae, with the exception of the pairings of Pseudorchis albida plus Platanthera chlorantha and Habenaria sagittifera plus Herminium lanceum, the tree showed disconcertingly little topological similarity to the preceding, more intensively sampled ITS analysis (Bateman et al., 1997, 1998; Pridgeon et al., 1997; Bateman, 1999a).

Luo Y.-B. and A. Pridgeon (pers. comm., 2000) similarly used the ITS framework of Pridgeon et al. as the basis of an unpublished study to examine in greater detail single additional sequences from south-east Asian representatives of the Habenariinae genera Habenaria and Peristylus, and the Orchidinae genera Platanthera, Galearis, Gymnadenia, Neottianthe, Amitostigma and Ponerorchis (including ‘Chusua’), together with eight putative species of Hemipilia. Relationships among the south-east Asian clade, consisting of AmitostigmaHemipilia, resemble those observed in the present study, but generic relationships within the ostensible sister group of PlatantheraGymnadeniaDactylorhizaPseudorchisOrchis s.s.Galearis again are strongly contradicted by the earlier ITS study of Pridgeon et al. (1997). Cozzolino et al. (2001) further developed their ITS matrix to preferentially encompass Middle Eastern segregate species previously encompassed by Orchis s.l., focusing their interpretations on pollination biology. Lastly, ITS sequences have proved to be sufficiently good taxonomic markers to identify to species level individual pollinia found attached to pollinating insects (Widmer et al., 2000).

Other recent studies have investigated additional regions of nuclear rDNA. Focusing on higher taxonomic levels, an analysis of representatives of c. 60 genera spanning the Orchidaceae (Freudenstein, Senyo & Chase, 2000b) compared sequences for ITS2 with those for 26S rDNA, noting that sequence alignment is more challenging for ITS. The results strongly supported the monophyly of five subfamilies (including Orchidoideae s.l., incorporating the former Spiranthoideae: Salazar, 2003) and many tribes, but also emphasized the difficulty of resolving relationships among the tribes. An 18S study of the Orchidaceae focusing on identifying multiple origins of mycoheterotrophy (Molvray, Kores & Chase, 2000: fig. 1) included four genera of Orchideae in an apparently weakly supported topology of (((Habenaria, Platanthera) Orchis) Satyrium), the sister-group relationship of Habenaria and Plat-anthera differing from the various topologies generated using ITS.

MATERIAL AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

A total of 190 sequences was analysed, representing 187 named taxa. Most of the 88 ITS sequences analysed by Pridgeon et al. (1997; sample details given in their table 2) were carried forward, via Bateman (1999a), into this more detailed analysis. The few exclusions included two of three apparently contrasting ITS sequences extracted from a single individual of Dactylorhiza praetermissa and one of the pair of molecularly identical individuals of Anacamptis laxiflora. Also, the putative Southern Hemisphere outgroup Holothrix scopularia was deemed too difficult to align (as well as being viewed as a doubtful bona fide member of the Orchidinae–Habenariinae clade), and Habenaria quinqueseta was judged to possess an unacceptably large proportion of missing values. In addition, a new sequence was substituted for the original Orchis pauciflora due to its surprising (though admittedly ultimately upheld) phylogenetic position, separate from the other two yellow-flowered Orchis species.

A further 109 sequences were added during the present study. Accession numbers of these samples additional to Pridgeon et al. (1997: table 2) are given here in Table 1. Some sequences were derived from plant material supplied by several collaborators, and the ITS sequence for Orchis patens s.s. was downloaded from GenBank deposits made by Aceto et al. (1999: ITS1 = Z94097, ITS2 = Z94098; data are not available for the intervening 5.8S region). Together, these sequences encompass all well-founded genera of the re-delimited Orchidinae, and for most genera they incorporate the majority of the species widely recognized by terrestrial orchid specialists (cf. Delforge, 1995, 2001). Coverage of the newly expanded concept of the Habenariinae is broad at the generic level but remains undesirably sparsely sampled at the species level (cf. Dressler, 1993; Linder & Kurzweil, 1999). Taxonomy for Orchidinae largely follows the classification of Delforge, unashamedly a ‘splitters’ classification at the species level.

Table 1.  Plant materials of Orchidaceae tribe Orchideae analysed in this study in addition to those listed by Pridgeon et al. (1997: table 2). 1= new combination made in this paper; 2= name currently illegitimate, new combination yet to be made; 3= supposed species here demonstrated to be a hybrid
TaxonReference/voucher
Tribe Orchideae
Subtribe Orchidinae
 Amerorchis rotundifolia (Banks ex Pursh) HulténChase O-0850
 Amitostigma gracilie SchlechterBateman 488 (Lee)
 Anacamptis (Orchis) fragrans (Pollini) R.M.Bateman, Pridgeon and M.W.Chase1Bateman 282
 Anacamptis (Orchis) israelitica (H.Baumann and Dafni) R.M.Bateman, Pridgeon and M.W.ChaseBateman 107 (Tattersall)
 Anacamptis (Orchis) longicornu (Poiret) R.M.Bateman, Pridgeon and M.W.ChaseBateman 105 (Tattersall)
 Anacamptis (Orchis) palustris (Jacquin) R.M.Bateman, Pridgeon and M.W.ChaseHedrén 5557
 Anacamptis (Orchis) picta (Loiseleur) R.M.Bateman, Pridgeon and M.W.Chase1Bateman 430 (Lowe)
 Anacamptis (Orchis) robusta (Stephenson) R.M.Bateman, Pridgeon and M.W.Chase1Bateman 281
 Anacamptis (Orchis) syriaca (Boissier ex H.Baumann & Künkele) R.M.Bateman, Pridgeon and M.W.ChaseBateman 106 (Tattersall)
 Chamorchis alpina (L.) L.C.M.RichardBateman 182 (Gössmann)
 Chusua (Ponerorchis) cf. chidori (Makino) P.F.HuntBateman 437
 Chusua donii NevskiLuo 053
 Chusua (Ponerorchis) jooiokiana (Makino) P.F.HuntBateman 587 (Lee)
 Dactylorhizabowmanii3 (= D.  traunsteineri ssp. bowmanii M.Jenkinson)Bateman 468
 Dactylorhiza cordigera (Fries) SoóBateman 108 (Manuel)
 Dactylorhiza euxina (Nevski) CzerepanovHedrén s.n.
 Dactylorhiza fuchsii (Druce) Soó var. hebridensis (Wilmott) Heslop-Harrison f.Bateman 341
 Dactylorhiza cf. hatagirea (D.Don) Soó[excluded from analysis]Bateman 191 (Long)
 Dactylorhiza incarnata (L.) Soó ssp. cruenta (O.F.Möller) P.D.SellBateman 057
 Dactylorhiza insularis (Sommier) LandwehrBateman 431 (Lowe)
 Dactylorhiza markusii (Tineo) H.Baumann and KünkeleBateman 109 (Manuel)
 Dactylorhiza occidentalis (Pugsley) Delforge var. kerryensis (Wilmott) R.M.Bateman and DenholmBateman 118
 Dactylorhiza praetermissa s.s. (Druce) SoóBateman 469
 Dactylorhiza praetermissa (Druce) Soó var. junialis (Vermeulen) SenghasBateman 470
 Dactylorhiza saccifera (Brongniart) SoóBateman 172 (Manuel)
 Dactylorhiza sambucina (L.) SoóBateman 112 (Manuel)
 Dactylorhiza traunsteineri (Sauter ex Reichenbach p.) Soó ssp. ebudensis (Wiefelspütz)2Bateman 055
 Galearis diantha (Schlechter) P.F.HuntLuo 074
 Gymnadenia (Nigritella) nigra s.s. (L.) Reichenbach f.Hedrén 97322
 Gymnadenia densiflora (Wahlenberg) DietrichBateman 165
 Gymnadenia (Nigritella) miniata (Crantz) HayekBateman 154
 Gymnadenia odoratissima (L.) L.C.M.RichardBateman 138
 Gymnadenia orchidis LindleyLuo 078
 ‘Habenariapurpureopunctata LangLuo 014
 Hemipilia crassicalcarata ChienLuo 245
 Hemipilia cruciata FinetLuo 055
 Hemipilia flabellata Bureau and FranchetLuo 228
 Hemipilia henryi RolfeLuo 186
 Hemipilia kwangsiensis Tang and Wang ex LangLuo s.n.
 Hemipilia limprichtii SchlechterLuo 123
 Hemipilia quinquangularis Tang and WangLuo 221
 Himantoglossum (Comperia) comperianum (Steven) P.DelforgeBateman 114 (Tattersall)
 Himantoglossum calcaratum (G.Beck) SchlechterBateman 170 (Manuel)
 Neolindleya camtschatica (Chamisso) NevskiBateman 578 (Lee)
 Neotinea (Orchis) commutata (Todaro) R.M.Bateman, Pridgeon and M.W.Chase1Bateman 547
 Neotinea (Orchis) conica (Willdenow) R.M.Bateman, Pridgeon and M.W.Chase1Bateman 195 (Phillips)
 Neotinea (Orchis) tridentata (Scopoli) R.M.Bateman, Pridgeon and M.W.Chase (replaces Chase O-0914 of Pridgeon et al., 1997)Bateman 548
 Neottianthe calcicola SchlechterLuo 178
 Neottianthe cucullata (L.) SchlechterBateman 177 (Manuel)
 Ophrys atlantica MumbyBateman 208 (Phillips)
 Ophrys benacensis (Reisigl) O. and E.Danesch and EhrendorferBateman 571
 Ophrys biancae (Todaro) MacchiatiBateman 527
 Ophrys candica W.Greuter, Matthäs and RisseBateman 037
 Ophrys dyris MaireBateman 093 (Courtis)
 Ophrys elegans (Renz) H.Baumann and KünkeleBateman 201 (Rowland)
 Ophrys exaltata TenoreBateman 365 (Manual)
 Ophrys fuciflora (F.W.Schmidt) MoenchBateman 167 (Manuel)
 Ophrys iricolor DesfontainesBateman 033
 Ophrys levantina Gölz and ReinhardBateman 230 (Rowland)
 Ophrys lunulata ParlatoreBateman 528
 Ophrys oxyrrhynchos TodaroBateman 540
 Ophrys pallida RafinesqueBateman 553
 Ophrys sphegifera WilldenowBateman 211 (Phillips)
 Ophrys spruneri NymanBateman 001
 Ophrys umbilicata DesfontainesBateman 203 (Rowland)
 Ophrys vasconica (O. and E.Danesch) P.DelforgeBateman 316 (Lowe)
 Orchis brancifortii Bivona-BernardiBateman 103 (Tattersall)
 Orchis canariensis LindleyBateman 131 (Welsh)
 Orchis cazorlensis LacaitaBateman 479 (Lowe)
 Orchis galilaea (Bornmüller and Schulze) SchlechterBateman 098 (Tattersall)
 Orchis ichnusae (Corrias) J. and P.Devillers-TerschurenBateman 102 (Tattersall)
 Orchis langei K.RichterChase O-0921
 Orchis olbiensis Reuter ex GrenierBateman 271
 Orchis ovalis F.W.Schmidt ex MayerBateman 156
 Orchis pallens L.Bateman 314 (Lowe)
 Orchis patens DesfontainesCozzolino Z94097-8
 Orchis pauciflora Tenore (replaces Chase O-0710 of Pridgeon et al., 1997)Bateman 005
 Orchis prisca HautzingerBateman 286 (Clarke)
 Orchis provincialis Balbis ex Lamarck and DC.Bateman 100 (Tattersall)
 Orchis punctulata Steven ex LindleyBateman 099 (Tattersall)
 Orchis scopulorum SummerhayesBateman 104 (Tattersall)
 Orchis spitzelii Sauter ex W.D.J.KochBateman 168 (Manuel)
 Orchis troodii (Renz) DelforgeBateman 200 (Rowland)
 Platanthera bakeriana KraenzlinLuo 035
 Platanthera (Piperia) colemanii (R.Morgan & Glicenstein) R.M.Bateman1Temple s.n.
 Platanthera (Piperia) elongata (Rydberg) R.M.Bateman1Temple s.n.
 Platanthera (Piperia) foetida Geyer ex HookerChase O-0930 (Hapeman)
 Platanthera holmboei H.Lindberg f.Bateman 443 (Lowe)
 Ponerorchis brevicalcarata (Finet) SoóLuo 161
 Pseudorchis straminea (Fernald) SojakLiden s.n.
 Serapias bergonii E.G.CamusBateman 294 (Ettlinger)
 Steveniella satyrioides (Steven) SchlechterGüner 10238 (K)
Subtribe Habenariinae Bentham
 Bonatea speciosa WilldenowRoux s.n.
 Brachycorythis macrantha (Lindley) SummerhayesChase O-0593
 Cynorkislowiana’ Reichenbach f.2Chase O-0583
 Cynorkis sp.Bateman 372
 ‘Habenariarepens NuttallChase O-0381
 Habenaria arenaria LindleyChase O-1135
 Habenaria delavayi FinetLuo 161
 Habenaria socotrana Balfour f.Bateman 199
 Habenaria tibetica SchlechterBateman 186 (Long)
 Habenaria tridactylites LindleyBateman 194 (Ettlinger)
 Herminium alaschianum MaximowiczBateman 188 (Long)
 Herminium (Peristylus) coeloceras SchlechterLuo 052
 Herminium (Platanthera) latilabris (Lindley)2Bateman 392 (Noltie)
 Herminium monorchis (L.) R.BrownBateman 121
 Stenoglottis longifolia Hooker f.Bateman 373
 Stenoglottis woodii SchlechterBateman 374
Tribe Diseae Dressler
Subtribe Satyriinae Schlechter
 Satyrium nepalense D.DonChase O-0539

The complete ITS1–5.8S–ITS2 assembly (e.g. Hershkovitz et al., 1999) was determined for each species, following extraction and sequencing protocols described by Pridgeon et al. (1997). Sequence editing employed Sequence Navigator (Applied Biosystems, Inc.), and alignment was performed by several pairs of eyes. After considerable experimentation, Disa uniflora was eventually designated as the outgroup. A single outgroup is acceptable in this case, given that the monophyly of the re-delimited Orchidinae had already been thoroughly tested (Pridgeon et al., 1997; Cameron et al., 1999; Bateman et al., 2001). Although Satyrium nepalense was added to the matrix as a potential intermediate outgroup, following the recognition on ITS evidence of a monophyletic Satyriinae immediately above the Diseae by Douzery et al. (1999) and Kores et al. (2000), its placement in some resulting trees above the long-branch Habenariinae Gennaria diphylla (cf. Bateman, 2001) rendered it unsatisfactory as an outgroup, since it demonstrated the topological instability characteristic of a ‘wildcard’ taxon sensuNixon & Wheeler (1992).

In total, 708 base positions were considered in the analysis. Short regions close to the priming sites in 18S and especially 26S were excluded due to the relatively poor quality of the sequence reads. Ambiguous gaps were scored as missing data, though 36 small but unambiguous indels were included in the analysis as bistate characters, following the ‘simple coding’ procedure of Simmons & Ochoterena (2000; cf. Giribet & Wheeler, 1999).

After some ultimately unproductive experimentation, tree-building followed a relatively simple and rapid protocol. A heuristic search was performed using PAUP* 4 (Swofford, 2000). All characters were treated as unordered and of equal weight. The initial heuristic search was 1000 random addition replicates with TBR branch swapping, saving a maximum of ten trees per replicate. The resulting 730 most-parsimonious trees were then used as starting trees for a further search, again using TBR swapping. Memory constraints prevented the second search from swapping to completion, limiting the number of trees saved to 10 000. Standard tree descriptors and the strict consensus tree were then obtained. To assess the support for individual nodes, a heuristic fast bootstrap search was subsequently performed for 10 000 replications using PAUP*.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

Of the 744 molecular characters included, 214 were constant. A further 87 bases were variable but uninformative, leaving 443 parsimony-informative characters. A heuristic search of the 190 coded taxa resulted in> 10 000 trees of a Fitch (1971) length of 2886 (including autapomorphies), Consistency Index of 0.35 (0.32 excluding uninformative characters) and Retention Index of 0.81. A corresponding Neighbour Joining tree based on the same alignment was published by Bateman (2001: fig. 15; see also Temple, 2001).

Base composition, compositional bias and transition/transversion ratios of the original matrix were discussed by Pridgeon et al. (1997) and will be further reviewed by us elsewhere (R. Bateman et al., unpubl. data); here, we will instead focus on perceived relationships inferred from the phylogeny, discussed in the context of mapped morphological and other non-sequence characters. To this end, a randomly selected most-parsimonious tree, optimized under Acctran, is presented as Figure 2, and the strict consensus tree bearing fast bootstrap percentages constitutes Figure 3.

imageimage

Figure 2. Selected most-parsimonious ITS tree, presented as a phylogram under Acctran optimization, for 186 species of Orchidinae and selected Habenariinae plus a Diseae outgroup.

imageimage

Figure 3. Strict consensus tree of the 10 000 most-parsimonious trees found when the number of trees was limited during the present analysis. Figures are all fast bootstrap support values exceeding 50%, based on 10 000 replicates.

RELATIONSHIPS WITHIN THE MAJOR CLADES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

OVERVIEW

The ITS phylogenies, plus as yet unpublished trnL phylogenies for a subset of the coded taxa (R. Bateman and P. Hollingsworth, unpubl. obs.), delimit 12 well-supported clades of unequivocal Orchidinae (Fig. 2), each encompassing from one to five genera as recircumscribed by Bateman et al. (1997) and further modified in the present study. There is also a relatively poorly resolved basal habenariid group, the two subtribes being separated by the ambiguously placed genus Brachycorythis, which is unusual in occurring in both the Northern (Chen, Tsi & Luo, 1999) and Southern (Linder & Kurzweil, 1999) Hemispheres. In contrast, most of the perceived relationships among those major clades receive weaker bootstrap support.

Here, we consider briefly each of those 12 clades in turn, beginning with Anacamptis s.l. (arbitrarily depicted here as the ‘apex’ of the tree), and concluding with the more ambiguous Brachycorythis and Habenariinae plus Satyrium and Disa. We then summarize higher level relationships and broader patterns of character distribution. A few of the clades are discussed in greater detail, where further generic transfers are required, notably (1) upgrading to species level some subspecies of Anacamptis s.l. and Neotinea s.l. recognized by Bateman et al. (1997), (2) incorporating the former monotypic genera ‘Comperia’ and ‘Barlia’ into an expanded Himantoglossum, (3) incorporating ‘Piperia’ into a yet further expanded Platanthera, and (4) returning ‘Gymnadeniacamtschatica to the monotypic genus Neolindleya. Also, Ponerorchis s.l. appears triphyletic; P. cf. chidori is retained in Ponerorchis pending further species sampling, but ‘P.’ brevicalcarata is transferred to the morphologically similar Hemipilia, along with ‘Habenariapurpureopunctata (for alternative taxonomic views see Bateman et al., 2001; Luo & Chen, in press). In contrast, we maintain Chamorchis and Traunsteinera as separate (arguably monotypic) genera, despite their surprising relationship as sister taxa. Suggested taxonomic rearrangements of the Habenariinae are not formalized here, as current taxonomic sampling is viewed as insufficiently dense. More detailed accounts of the remaining clades of Orchidinae (albeit discussing far fewer coded taxa) were given by Bateman et al. (1997).

Within the text, the symbol ‘∼’ is used to indicate an inclusive clade of three or more species shown on the figures that is bracketed by the two explicitly stated end-members (e.g. the clade of 21 closely related species of Ophrys bracketed by O. apiferaO. biancae on Fig. 2). Levels of bootstrap support for particular clades on Figure 3 are summarized as poor (<50%), moderate (51–80%) or strong (>80%). Comments on the correlation between degrees of ITS sequence divergence and possible species-level status are highly tentative, and should be considered carefully in conjunction with discussions placed under ‘Species delimitation’ and ‘Future research’.

ANACAMPTIS (16 CODED TAXA)

This moderately supported clade encompasses all species of the former Orchis that possess 36 (or, in the case of A. papilionacea, 32) chromosomes (for a more detailed account see Tichy & Del Prete, 2001). Since the analysis of Pridgeon et al. (1997), it has attracted further attention from Cozzolino et al. (2001) and R. Bateman and P. Hollingsworth (unpubl. obs.). Molecular, morphological and cytological (D’Emerico, Pignone & Bianco, 1996; D’Emerico, Pignone & Scrugli, 2000) divergences are on average high within the clade (Fig. 2), which shows clear karyological differences from both Serapias and Ophrys (D’Emerico et al., 2000a).

Groups that are monophyletic on molecular evidence and also possess morphological cohesion include champagneuxiiboryi (including the pairing of A. boryi with A. israelitica, which share the characteristic of basipetal flowering: Cozzolino et al., 2001), fragranssancta and laxiflorarobusta; the latter group, and the group immediately above A. papilionacea, are each supported by single indels. The more morphologically individualistic A. pyramidalis, A. papilionacea and A. collina are, unsurprisingly, more ambiguously placed (A. collina also yielded strongly divergent allozyme patterns according to Rossi et al., 1994). The new deeper placements of A. pyramidalis and A. collina undermine the ‘papilionacea subgroup’ tentatively recognized by Bateman et al. (1997). Similar positional ambiguities, and a contrasting topology, are evident in the ITS phylogenies of Aceto et al. (1999: fig. 1) and Cozzolino et al. (2001: fig. 1). The morphologically predicted placement of A. syriaca as unequivocal sister to the morio group demonstrates that the specimen analysed by Cozzolino et al. (2001) as ‘A. syriaca’ was, as they subsequently suspected, more akin to A. papilionacea.

The only species of Orchidinae analysed by Cozzolino et al. but absent from our tree is A. dinsmorei; as expected, this proved to be the Asia Minor segregate of A. palustris. The tentative suggestion of Bateman et al. (1997) that the vegetatively distinct laxiflorapalustrisrobusta clade (represented in their tree only by A. laxiflora) could merit generic segregation receives some support from the subsequent tree (Fig. 2), since their inclusion reduces bootstrap support for Anacamptis as currently delimited to 59%, compared with 89% for the remainder of the clade. However, trnL data suggest that these species are bona fide members of Anacamptis (R. Bateman and P. Hollingsworth, unpubl. obs.).

The eight-step ITS disparity between A. coriophora s.s. and A. fragrans, and the five-step disparity between A. palustris s.s. and A. robusta, suggest that Bateman et al. (1997) may have been premature in treating fragrans and robusta as subspecies rather than as bona fide species; they are therefore elevated to species level in Table 2. However, a much smaller ITS disparity was inferred in a more detailed survey of the palustris group conducted by Cafasso et al. which has in addition revealed biogeographically informative variation in plastid minisatellites (Cafasso et al., 2001). Moreover, the decision of Bateman et al. to include A. morio ssp. picta as a synonym of A. morio s.s., taken in the absence of molecular data for picta, is not upheld; rather, picta appears from ITS data to be more closely related to A. champagneuxii and to the relatively large-spurred species A. longicornu, newly incorporated in our matrix (see also Cafasso et al., 2000; Cozzolino et al., 2001).

Table 2.  Modifications to the species-level revisions of the expanded genera Anacamptis s.l. and Neotinea s.l. presented in Bateman et al. (1997)
Anacamptis robusta (T.Stephenson) R.M.Bateman, comb. nov.
Basionym: Orchis palustris Jacquin var. robusta T.Stephenson, J. Bot. (Lond.) 69: 179 (1931).
Synonyms: Orchis laxiflora Lamarck ssp. robusta (T.Stephenson) H.Sundermann, Europ. Medit. Orch. 40 (1980); Orchis robusta (T.Stephenson) Gölz and Reinhard, Ber. Schweiz. Bot. Ges., 85: 288 (1975); Anacamptis palustris (Jacquin) R.M.Bateman, Pridgeon and M.W.Chase ssp. robusta (T.Stephenson) R.M.Bateman, Pridgeon and M.W.Chase, Lindleyana 12: 120 (1997).
Anacamptis fragrans (Pollini) R.M.Bateman, comb. nov.
Basionym: Orchis fragrans Pollini, Elem. Prov. Ver. 2: 155 (1811)
Synonyms: Orchis coriophora L. ssp. fragrans (Pollini) Sudre, Fl. Toulous. 187 (1907); Anacamptis coriophora (L.) R.M.Bateman, Pridgeon and M.W.Chase ssp. fragrans (Pollini) R.M.Bateman, Pridgeon and M.W.Chase, Lindleyana 12: 120 (1997).
Anacamptis picta (Loiseleur) R.M.Bateman, comb. nov.
Basionym: Orchis picta Loiseleur, Mem. Soc. Linn. Paris 6: 431 (1827).
Synonym: Orchis morio L. ssp. picta (Loiseleur) Arcangeli, Comp. Fl. Ital. ed. 2: 167 (1894) [placed under Anacamptis morio (L) R.M.Bateman, Pridgeon and M.W.Chase by Bateman et al., 1997: 122].
Neotinea commutata (Todari) R.M.Bateman, comb. nov.
Basionym: Orchis commutata Todari, Orch. Sic. 24 (1842).
Synonyms: Orchis tridentata Scopoli ssp. commutata (Todari) Nyman, Consp. 691 (1882); Neotinea tridentata (Scopoli) R.M.Bateman, Pridgeon and M.W.Chase ssp. commutata (Todari) R.M.Bateman, Pridgeon and M.W.Chase, Lindleyana 12: 122 (1997).
Neotinea conica (Willdenow) R.M.Bateman, comb. nov.
Basionym: Orchis conica Willdenow, Spec. 4: 14 (1805).
Synonyms: Orchis tridentata Scopoli ssp. conica (Willdenow); Neotinea tridentata (Scopoli) R.M.Bateman, Pridgeon and M.W.Chase ssp. conica (Willdenow) R.M.Bateman, Pridgeon and M.W.Chase, Lindleyana 12: 122 (1997).

SERAPIAS (7 CODED TAXA)

As in Ophrys (below), monophyly of Serapias is strongly supported by ITS evidence; the large base-pair divergence, plus three synapomorphic indels, reinforce its unequivocal morphological cohesion. Although well covered taxonomically, even less molecular resolution is evident within Serapias than within Ophrys, and adding S. bergonii to the analysis failed to modify the original topology of Pridgeon et al. (1997). The present results cannot unequivocally distinguish the morphologically distinct tetraploid S. lingua (possibly an autotetraploid derivative of S. parviflora) from the remaining, diploid species. A recent study of heterochromatin patterns (D’Emerico et al., 2000) revealed similarly low levels of divergence, though their conclusion that this indicates a recent evolutionary origin of the genus (p. 491) is inconsistent with the long ITS branch subtending the genus (Fig. 2); rather, the low divergence in karyotypic features indicates recent origin of the species within the genus, even for the more morphologically distinct species aggregates (cf. Lorenz, 2001). Hopefully, current research being pursued by S. Cozzolino and co-workers using genetic fingerprinting techniques will yield clearer phylogenetic insights into the 29 putative ‘species’ recognized by Delforge (2001).

OPHRYS (33 CODED TAXA)

Although totalling less than one sixth of the 215 putative species listed by Delforge (2001), the 33 species analysed here encompass all of his informal groups and subgroups; they more than double the species-level coverage achieved by Pridgeon et al. (1997). The monophyly of the genus is strongly supported by both base divergence and no less than five indels (Figs 2a, 3a), but most relationships within the genus are poorly resolved and interpretations are therefore tentative.

Some ITS trees (e.g. Bateman, 2001: fig. 15) suggest that a heterogeneous and apparently paraphyletic group of species with relatively simple labella gave rise to a more derived and species-rich group with complex three-dimensional labella and beaked connectives, although in Figure 2a the two groups tentatively appear monophyletic, each supported by an indel. Two distinctive and enigmatic species, O. insectifera and O. tenthredinifera, vary considerably in placement, effectively acting as ‘wild-card’ taxa within the genus (Fig. 3a: Nixon & Wheeler, 1992). Notable features of Figure 2a are the monophyly of the morphologically distinct fuscalutea group (including the controversial Ophrys pallida, which superficially resembles O. bombyliflora as well as the O. fusca aggregate) and the basal divergence of Ophrys apifera within the group rich in morphologically similar ‘species’ that is defined by possessing a strongly three-dimensional labellum bearing a complex speculum.

Within this group, the tree tentatively suggests non-monophyly of several species groups recognized by Delforge (1995), including the tenthredinifera, scolopax, bornmuelleri and fuciflora groups. It also fails to separate the groups progressively segregated from Ophrys sphegodes sensu latissimo, not only from each other but also from the Ophrys bertolonii group (including O. bertoloniiformis and O. benacensis). Some groups suggest a strong geographical influence, such as the pairing of the Sicilian species O. oxyrrhynchos (fuciflora group) and the very similar O. biancae (supposedly bornmuelleri group).

Soliva, Kocyan & Widmer (2001) combined ITS and trnL–F data for 32 species of Ophrys (15 shared with our set of 33 species) and six outgroups (surprisingly excluding the sister genus of Ophrys, Anacamptis). Analysis of trnL–F yielded even fewer phylogenetically informative characters than ITS, so not surprisingly Soliva et al. similarly detected limited phylogenetic structure. Their maximum parsimony topology differed from Figure 2 in two features: (1) the O. fuscalutea group was subtended first by O. tenthredinifera, then by O. bombyliflora, then by O. speculum, a morphologically improbable topology (cf. Devillers & Devillers-Terschuren, 1994); interestingly, the maximum likelihood tree of Soliva et al. (2001) more closely resembled our parsimony tree in this respect. (2) Ophrys insectifera was weakly supported as being associated with the O. apifera clade rather than the O. fusca clade. Moreover, Soliva et al. suggested that the 14 species analysed by them that yielded sequences identical to those of other species reflected hybridization (i.e. sequence divergence followed by homogenization), in contrast to Bateman (2001) who invoked conspecificity (i.e. a historical absence of divergence).

STEVENIELLAHIMANTOGLOSSUM S.L., INCLUDINGCOMPERIAANDBARLIA’ (7 CCODED TAXA)

In our previous study (Pridgeon et al., 1997) we analysed four putative species of Himantoglossum s.s., together with the widespread Mediterranean species Barlia robertiana. In this study we have added a fifth species of Himantoglossum s.s., H. calcaratum, which as expected yielded similar ITS sequences to the other eastern Mediterranean segregates of H. hircinum. The  substantial separation of these segregates from H. hircinum s.s. envisaged by Teschner (1980) is not supported. All four of these taxa, plus H. affine, have been given subspecific status in some treatments (e.g. Sundermann, 1973, 1980; Moore, 1980). Interestingly, H. caprinum and H. calcaratum were the only taxa segregated from H. hircinum by Moore (1980), yet in Delforge's (1995, 1999, 2001) more finely divided reclassification of the genus, calcaratum was treated as a synonym of caprinum. Here, H. calcaratum is depicted as moderately supported sister to H. hircinum plus H. adriaticum, differing from its Adriatic companion H. adriaticum by only two ITS substitutions but from the more geographically widespread H. caprinum by four substitutions.

The rare Canary Island endemic ‘Barliametlesicsiana was sampled and sequenced too late for inclusion in the present parsimony analysis. It showed a surprisingly high divergence from its ‘mainland’ European sister-species ‘B.’ robertiana of 13 substitutions plus three indels, suggesting longer isolation than for Macaronesian endemics of other clades of Orchideae such as Orchis canariensis (see ‘Orchis’ below).

Of even greater interest are the results obtained from adding to the matrix two widely recognized monotypic genera, Comperia comperiana and Steveniella satyrioides, which have attracted considerable speculation regarding both their systematic / phylogenetic position and their nomenclature (e.g. Delforge, 1999, 2000, 2001; Baumann, Künkele & Lorenz, 2000; Bateman, 2001). Both putative genera were originally described (albeit illegitimately) by von Steven (1809) as members of an exceptionally broad concept of Orchis but were later erected as the largely undisputed monotypic genera Comperia (illegitimately by Koch, 1849; then legitimately by Ascherson & Graebner, 1907) and Steveniella (by Schlechter, 1918; see also Sundermann, 1972; Baumann, Künkele & Lorenz, 2000; Delforge, 2000). In the case of Steveniella satyrioides, this taxonomic odyssey occurred via an impressive series of ‘stepping stone’ genera; the species was first transferred (with some prescience) to Himantoglossum by Sprengel (1826) before passing through, in chronological order, Peristylus, Platanthera, Coeloglossum, Habenaria (Baumann et al., 2000) and the Orchis (Anacamptis) coriophora group sensuPridgeon et al. (1997) (Delforge, 2000, 2001).

The majority of recent treatments have placed Steveniella and Comperia together, and Barlia and Himantoglossum s.s. together, but have not recognized a close relationship between the two pairs of genera (Moore, 1980; Sundermann, 1980; Buttler, 1991). This approach is epitomised by Flora Europaea (ed. Moore, 1980), wherein the two pairs of genera were described by different contributing authors, further decreasing the probability of recognizing any morphological similarities between them. Himantoglossum s.s. and Barlia were linked to ‘higher’ Orchidinae, typically Orchis s.l. (especially ‘Aceras’) and Serapias and/or Anacamptis pyramidalis. In contrast, Steveniella and Comperia were compared with the ‘lower’ Orchidinae, typically Neotinea maculata (often with Traunsteinera globosa) and Dactylorhiza s.s. (often with ‘Coeloglossumviride).

An alternative hypothesis grouped Comperia with Barlia and Himantoglossum s.s., once again placed close to Orchis/‘ Aceras’ and Serapias/Anacamptis s.s. (Landwehr, 1977; Delforge, 1995, 1999). This treatment left Steveniella isolated and again viewed as relatively primitive. Landwehr (1977) placed Steveniella between Neottianthe cucullata/Neotinea maculata and Traunsteinera globosa. Delforge (1995) similarly placed Steveniella between Neotinea s.s./Traunsteinera and Anacamptis pyramidalis, but in his text he noted especially a similarity to the Anacamptis (formerly Orchis) coriophora group (a comparison echoed by Pridgeon et al., 1997: 101). He subsequently transferred Steveniella to that group as ‘Orchis prosteveniella’ (Delforge, 2000, 2001).

Our intriguing results reveal that these four genera constitute a probable clade, with the three putative monotypic genera diverging sequentially, in the order SteveniellaComperiaBarlia, subtending a far less resolved group of less morphologically differentiated species of Himantoglossum s.s. Moreover, ComperiaBarliaHimantoglossum appears to be a well-resolved clade (87% bootstrap) supported by two indels (one shared with Serapias plus Anacamptis s.l.) and morphologically defined by vegetative vigour, toothed lateral petals and large, elongate, trilobed labella. In contrast, the smaller, 1–2-leaved, small-flowered Steveniella superficially resembles members of the Neotinea s.l. and NeottiantheHemipilia clades (with which it shares an indel), though its gynostemium structure, purplish-brown galea, strongly three-lobed labellum and short, robust spur do, as Sprengel perceived, suggest some similarities to the more derived himantoglossids. Although the position of Steveniella in the ITS tree is tenuous, collapsing to a polytomy in the strict consensus tree (Fig. 3a), we can confidently reject the many previous hypotheses of its relationship with other genera that were advanced in the absence of phylogenetic data.

Chromosomally, Himantoglossum s.s. and Barlia have recently been carefully examined and shown to have the 36 chromosomes expected of members of the  large 2n = 36 clade identified by Pridgeon et al. (1997), though H. hircinum possesses an additional B-chromosome (Carpineri & Rossi, 1987; D’Emerico, Bianco & Medagli, 1992). Steveniella has been reported to have the improbable karyotype of 2n = 38  by Sundermann & von der Bank (1977), but this assertion may simply reflect the possible presence of  another B-chromosome analogous to that of H. hircinum. More difficult to explain is the count of 2n = 30 reported for Comperia by Ströhlein & Sundermann (1972), which may represent a counting error (Delforge, 1999) or a true autapomorphic reduction in chromosome number by fusion, analogous to the decrease from 2n = 36 to 2n = 32 unequivocally observed in Anacamptis (formerly Orchis) papilionacea in the morphologically heterogeneous Anacamptis clade (D’Emerico et al., 1990; Pridgeon et al., 1997).

Steveniella and ‘Comperia’ are confined to Asia Minor where they have almost mutually exclusive distributions: Steveniella to the north and east, and ‘Comperia’ to the south and west (e.g. Baumann & Künkele, 1982; see also Baumann et al., 2000). Barlia and Himantoglossum s.s. similarly occur in Asia Minor, but also include species that extend west throughout the Mediterranean and north-west into Continental Europe.

Each of the four genera as previously conceived possesses morphological autapomorphies within the clade: note for example the single fully expanded leaf, flat labellum and bilobed spur of Steveniella, the four filiform labellar extensions and elongate spur of ‘Comperia’, the well-developed lateral labellum lobes and early flowering period of ‘Barlia’, and the extremely elongate, helically twisted central labellum lobe of Himantoglossum s.s. (admittedly, H. formosum is intermediate between the H. hircinum aggregate and ‘Barlia’ in these characters, e.g. Delforge, 2001).

Taxonomically, the topology shown in Figure 2a presents a similar dilemma to that posed to Pridgeon et al. (1997) by that of Neotinea maculata vs. the more derived species of Neotinea formerly attributed to Orchis s.l. (i.e. the tridentatalactea group: see below). Monophyly permits continued recognition of the three more-or-less monotypic genera; also, each possesses morphological (and supposed chromosomal) autapomorphies and is subtended by a substantial ITS branch length of c. 25 steps. Branches of similar or lesser lengths subtend the genera Pseudorchis, Amerorchis, Galearis and Neolindleya at the base of the Platanthera clade, Ponerorchis, Amitostigma and Neottianthe at the base of the Hemipilia clade, and separate Traunsteinera from the highly morphologically dissimilar Chamorchis. On the other hand, monotypic genera by definition contain no phylogenetic structure, and the phylogenetic structure among the three more derived genera is reflected in the morphological synapomorphies outlined above. More pragmatically, amalgamating the genera nullifies some of the nomenclatural controversies that survived even the formal conservation of Himantoglossum against Loroglossum in 1972 (cf. Garay, 1997; Delforge, 1999; Baumann et al., 2000; Bateman, 2001).

Prompted in part by ITS results already published by ourselves (Bateman et al., 1997; Pridgeon et al., 1997), Delforge (1999, 2001) recently stimulated further controversy by incorporating Barlia and Comperia into Himantoglossum and subsequently Steveniella into Orchis s.l. (= our Anacamptis) as Orchis prosteveniella (Steven) Delforge, noting the apparent illegitimacy of ‘Orchis satyrioidesvon Steven (1809) (Delforge, 2000, 2001). On the current balance of evidence, and on phylogenetic logic, we concur with his decisions regarding ‘Barlia’ and ‘Comperia’, but our data clearly reject his treatment of Steveniella. Similarly, the molecular isolation and morphological distinctiveness of Steveniella suggest that it is better left as a monotypic genus than sunk into a further expanded Himantoglossum as H. satyrioides (Steven) Sprengel (1826: see the revised classification in Table 3).

Table 3.  Revised classification of the SteveniellaHimantoglossum s.l. group (see also Nelson, 1968; Delforge, 1999; Baumann et al., 2000). Asterisked taxa have not yet generated ITS sequences
 (1) Himantoglossum hircinum (L.) Sprengel, Syst. Veg. 3: 694 (1826).
 Basionym: Satyrium hircinum L., Sp. Plant. 944 (1753).
 Synonym: Orchis hircina (L.) Crantz, Stirp. Austr.2: 484 (1769).
 (2) Himantoglossum adriaticum H.Baumann, Orchidee29: 171 (1978).
 Synonym: Himantoglossum hircinum (L.) Sprengel ssp. adriaticum (H.Baumann) Sundermann, Europ. medit. Orchideen: 40 (1980).
 (3) Himantoglossum caprinum (M.-Bieb.) Sprengel, Syst. Veg. 3: 694 (1826).
 Basionym: Orchis caprina M.-Bieb., Fl. Taurico-Caucasica 3: 602 (1819).
 Synonym: Himantoglossum hircinum (L.) Sprengel ssp. caprinum (M.Bieb.) K.Richter, Pl. Europ.1: 276 (1890).
 (4) Himantoglossum calcaratum (G.Beck) Schlechter, Fedd. Repert.15: 287 (1918).
 Basionym: Aceras calcaratum G.Beck, Ann. Hofmus. Wien2: 55 (1887).
 Synonym: Himantoglossum hircinum (L.) Sprengel ssp. calcaratum (G.Beck) Soó, Bot. Arch. (Berlin)23: 90 (1929, published as ‘1928’).
 (5) Himantoglossum affine* (Boissier) Schlechter, Feddes Repert.15: 287 (1918).
 Basionym: Aceras affinis Boissier, Fl. Orientalis (Orchidacea) 5: 56 (1884).
 Synonym: Himantoglossum hircinum (L.) Sprengel ssp. affine (Boissier) Sundermann, Europ. medit. Orchideen: 40 (1980).
 (6) Himantoglossum formosum* (Steven) K.Koch, Linnaea22: 287 (1849).
 Basionym: Orchis formosa Steven, Mém. Soc. Nat. Mosc.4: 66 (1813).
 (7) Himantoglossum robertianum (Loiseleur) P.Delforge, Nat. Belg.80: 401 (1999).
 Basionym: Orchis robertiana Loiseleur, Fl. Gallica 1: 606 (1807, as ‘1806’).
 Synonym: Barlia robertiana (Loiseleur) W.Greuter, Boissiera13: 192 (1967).
 (8) Himantoglossum metlesicsianum (Teschner) P.Delforge, Nat. Belg.80: 401 (1999).
 Basionym: Barlia metlesicsiana Teschner, Orchidee33: 117 (1982).
 (9) Himantoglossum comperianum (Steven) P.Delforge, Nat. Belg.80: 401 (1999).
 Basionym: Orchis comperiana Steven, Mém. Soc. Nat. Mosc.8: 259 (1829).
 Synonym: Comperia comperiana (Steven) Ascherson and Graebner, Syn. Mitteleur. Fl. 3: 620 (1907).
(10) Steveniella satyrioides (Steven) Schlechter, Feddes Repert.15: 295 (1918).
 Basionym: Orchis satyrioides Steven, Mém. Soc. Nat. Mosc.2: 176 (1809), nom. illegit.
 Synonym: Himantoglossum satyrioides (Steven) Sprengel, Syst. Veg. 3: 694 (1826), nom. illegit. (for full synonymy see Baumann et al., 2000).

After presenting much well-informed discussion, Delforge (1999; fig. 10) followed Teschner (1980) in producing a wholly speculative ‘phylogeny’ and a derived ‘natural’ key that were not based on any data matrix but rather were extrapolated from our original tree (Pridgeon et al., 1997). In the absence of relevant ITS sequences, we share Delforge's speculation that H. affine is placed below the H. hircinumcaprinum group; we also agree that the rare and elusive H. formosum could be placed below H. affine, although on morphological evidence it is equally likely to be placed on the Barlia lineage rather than that of Himantoglossum s.s. However, we defer formal judgement on these species pending acquisition of molecular data. Here, ITS sequences conclusively refute Delforge's (1999, 2001) speculative BarliaComperia clade, which appears to have been based on the shared characteristic of a stigmatic surface that is higher than wide. Our data suggest that this is a shared primitive character (symplesiomorphy) rather than a shared derived character (synapomorphy) and thus is unsuitable for grouping these taxa; it renders Delforge's speculative phylogeny and his related key unequivocally ‘unnatural’.

Delforge (1995, 1999, 2001) further speculated that his BarliaComperia alliance was ‘probably derived from a taxon of the Orchis spitzelii species group’, but our ITS tree reveals a vast disparity of 109 substitutions (=c. 16% total ITS sequence divergence) between Himantoglossum (formerly Comperia) comperiana and Orchis spitzelii, clearly demonstrating that several floral similarities between these species (e.g. elongate robust labella, broad conical spurs, pale purple flowers bearing darker purple markings) are evolutionary convergences and hence are unsuitable for classification at this level (cf. Bateman et al., 1997; Chase, 1999; Bateman, 2001).

NEOTINEA S.L. (7 SEQUENCES REPRESENTING 6 CODED TAXA)

The formerly monotypic Neotinea was expanded by Pridgeon et al. (1997) and Bateman et al. (1997) to encompass the small number of small-flowered, essentially trilobed-lipped species of the ustulata group that were formerly included in Orchis s.l. A similar topology separating the Neotinea group was subsequently obtained by Aceto et al. (1999: fig. 1). These former members of Orchis could in theory have been treated as a genus separate from the more narrowly delimited original concept of Neotinea, given the relatively long molecular branch, distinct vegetative markings and reputed autogamy of N. maculata, but there are clear similarities in the size and morphology of the flowers. The clade has 100% bootstrap support and two synapo-morphic indels.

New insights have also been generated among closely related species. Firstly, as expected, N. conica proved similar to N. lactea, though the four-step disparity tentatively suggests that Bateman et al. (1997) may have erred in awarding subspecific status to the former. Secondly, Neotinea commutata and the characteristically smaller-flowered but otherwise morphologically similar N. tridentata were both sampled from a mixed population on Sicily. Surprisingly, they proved to be highly divergent on ITS evidence (see also Cafasso et al., 2000), apparently refuting the suggested origin of N. commutata as an autotetraploid of N. tridentata (Delforge, 1995). Even more surprisingly, N. commutata yielded an identical sequence to the specimen of supposed O. tridentata from west-central Italy analysed by Pridgeon et al. (1997), rather than with the coexisting Sicilian tridentata which resolved as sister-species to the morphologically similar N. conica and N. lactea (they also share a characteristic indel). This result indicates that the specimen of N. ‘tridentata’ originally published by Pridgeon et al. (1997) was attributable to N. commutata. Neotinea commutata and N. conica are upgraded to species status in Table 2. The Table excludes the taxonomically contentious late-flowering form of N. ustulata, ‘ssp.’aestivalis (Kümpel, 1988; Kümpel & Mrkvicka, 1990; Tali & Kull, 2001), which is currently the subject of detailed molecular investigations (K. Tali, R. Bateman, M. Chase and M. Fay, unpubl. obs.). Preliminary results, including identical ITS sequences, suggest that its original status as a variety is more appropriate.

ORCHIS S.S. (26 CODED TAXA)

Sampling within true Orchis has been much improved since the publication of Pridgeon et al. (1997). The main remaining omission is O. stevenii, which on recent thin-layer chromatographic evidence may best be treated as a subspecies of O. militaris (Breiner, 1999). The basic structure originally observed within the clade by Pridgeon et al. is upheld. The clade benefits from 97% bootstrap support, aided by an indel. Orchis (formerly Aceras) anthropophora and then O. italica are basally divergent (in contrast, O. anthropophora diverged above O. italica in the tree of Aceto et al., 1999; the two were shown as sisters in the tree of Cozzolino et al., 2001); in our tree they render paraphyletic the more primitive ‘anthropomorphic’ group. Similar relationships were discerned by Qamaruz-Zaman (2000) using AFLP data. The remaining five anthropomorphic species, sampling now including the often dominantly yellow-flowered eastern Mediterranean species O. punctulata and O. galilaea, form a well-supported clade (punctulatamilitaris), which is sister to the species-rich group of broad-lipped Orchis that are morphologically convergent (and were formerly classified) with the Anacamptis morio group (A. longicornuboryi).

This clade in turn reveals a near-trichotomy of three well-supported clades: (1) the yellow-flowered O. pallens group, (2) O. mascula and its segregates, and (3) the O. spitzeliianatolicaquadripunctata complex (100% bootstrap support, including an indel). This topology confirms the placement predicted by Bateman et al. (1997) for O. spitzelii and its relatives, but shows less molecular divergence than might be expected from the considerable morphological diversity evident within the group. Orchis patens s.s. also belongs to this clade, as sister to the morphologically similar O. canariensis (see also Aceto et al., 1999). The substantial allozymic divergence between O. quad-ripunctata and the morphologically similar O. brancifortii reported by Rossi et al. (1994) is reflected in our ITS tree, though only a single base substitution was detected by Cafasso et al. (2000).

The most surprising result, first noted by Bateman (1999a), was the inclusion of the yellow-flowered O. pauciflora in the otherwise purple-flowered O. mascula group, reliably separated from the other yellow-flowered species O. provincialis and O. pallens. We further investigated this unexpected placement by sequencing a second specimen of O. pauciflora from a different country. Nonetheless, Aceto et al. (1999) and Cozzolino et al. (2001) obtained the converse placement, provincialis rather than pauciflora grouping with O. mascula, suggesting that further accessions of both species should be sequenced for ITS (see also hybridization evidence in Pellegrino et al., 2000, 2001). Nonetheless, either topology refutes Bateman et al.∏s (1997) treatment of pauciflora as a subspecies of the morphologically similar O. provincialis. This may help explain the substantial anatomical differences noted between O. provincialis and O. pauciflora by Del Prete & Miceli (1999), although our tree does not support the sectional/subsectional re-arrangements of either these authors or Hautzinger (1978). Further complicating the issue of yellow-flowered Orchis, Vermeulen (1977) earlier had erroneously argued that it is O. pallens that should be classified with O. mascula, separately from the remaining yellow-flowered species, which were placed by Vermeulen in their own subsection, Provincialae.

Also within the mascula group, the ITS divergence of some recent species-level segregates of O. mascula, such as O. langei (syn. O. mascula ssp. hispanica, a name that may in fact encompass two distinct genetic entities), O. scopulorum and O. olbiensis, was supported by relatively substantial AFLP divergence documented by Qamaruz-Zaman (2000), and at least one Portuguese population of O. olbiensis has since proven to be tetraploid (Bernardos et al., 2002). This evidence suggests that Bateman et al. (1997) may have been in error to award these taxa only subspecific status. However, the relatively long ITS branch subtending O. ovalis is inconsistent with the more modest divergence evident in the AFLP data, which show it as negligibly distinct from O. mascula s.s., as is O. ichnusae (see also Cafasso et al., 2000) and probably also the as-yet unsequenced O. tenera. The morphology of the mascula group was recently reviewed by Essink et al. (1999).

TRAUNSTEINERACHAMORCHIS (2 CODED TAXA)

Perhaps the most surprising outcome of this expanded matrix is the pairing of two arguably monotypic alpine/boreal specialists, Traunsteinera and Cham-orchis. Traunsteinera is tall with large leaves evenly spaced along the stem, which terminates in a compressed conical spike of pale pink flowers with trilobed labella and short spurs that superficially resemble those of Neotinea s.l. In contrast, Chamorchis has a short stem that is subtended by basally concentrated filiform leaves and bears a more elongate spike of pale green flowers and entire, spur-less labella; superficially they are reminiscent of the flowers of Dactylorhiza (formerly Coeloglossum) viridis or Pseudorchis albida.

Among recent authors, only Landwehr (1977), benefiting from the shrewd taxonomic advice of P. Vermeulen, implied a close relationship between Chamorchis and Traunsteinera, by depicting them on adjacent plates in his iconograph. Most authors understandably placed Chamorchis adjacent to Plat-anthera and/or Pseudorchis (e.g. Moore, 1980; Sundermann, 1980; Buttler, 1991; Delforge, 1995), though others implied affinities with ‘Coeloglossum’ (Buttler, 1991) and, less intuitively, with Neottianthe (Moore, 1980; Delforge, 1995). In contrast, Traunsteinera was most frequently linked to Orchis s.s. (often including ‘Aceras’: Moore, 1980; Sundermann, 1980; Buttler, 1991) and Neotinea maculata (Moore, 1980; Sundermann, 1980; Buttler, 1991; Delforge, 1995); also less plausibly to Steveniella by Landwehr (1977).

In fact, the sister-group relationship of these two morphologically distinct but co-occurring genera is strongly upheld by the ITS data (bootstrap support 100%); moreover, their ITS sequences are sufficiently similar to be more typical of distantly related species within a single genus than of separate but closely related genera (cf. Bateman et al., 1997), and they have similar karyotypes distinct from those of other Orchidinae (D’Emerico & Grünanger, 2001). The main argument for maintaining Chamorchis and Traunsteinera as separate genera lies in morphological characters, notably the paucity of obvious synapomorphies relative to autapomorphies of the existing genera.

PSEUDORCHISAMERORCHISGALEARISNEOLINDLEYAPLATANTHERA S.L. (INCLUDINGPIPERIA’) (18 CODED TAXA)

This clade receives less bootstrap support (<50%) than the preceding clades, with the geographically widespread, short-branch genus Pseudorchis being the most tenuous, basally divergent genus in the group (Fig. 2). The negligible ITS divergence between Ps. albida and the recently segregated Ps. straminea lends little support to recent morphometric (Reinhammar, 1995) and allozyme (Reinhammar & Hedrén, 1998) studies indicating that the latter should be treated as a full species, segregated from Ps. albida, though there is evidence of ecological segregation (Reinhammar, Olsson & Sørmeland, 2002).

The next group to diverge consists of the monotypic Amerorchis, which as anticipated by Pridgeon et al. (1997) is placed (albeit weakly; cf. Bateman, 2001) as sister to Galearis; these taxa are barely sufficiently divergent, both morphologically and molecularly (it has one autapomorphic indel), to justify Hultén's (1968) generic segregation of Amerorchis. The two Asian species of Galearis form a derived pairing relative to the North American G. spectabilis and to Amerorchis, surprisingly indicating that the North American lineage gave rise to the East Asian species; further sampling of Galearis is desirable to test this hypothesis.

The relationship among GalearisAmerorchis, Neolindleya and Platanthera is essentially unresolved (Fig. 3). Neolindleya camtschatica (Kränzlin, 1899 in 1897–1904) is more often attributed to Gymnadenia (e.g. Chen et al., 1999), but clearly it has no close relationship to that genus. It is sufficiently distinct from Platanthera on both morphological and molecular grounds to merit its status as a monotypic genus, analogous to Amerorchis. This result was a surprise, as ‘G.’ camtschatica has a pink-purple flower colour and narrow spur in accordance with Gymnadenia but has vegetative characters more consistent with one of the more robust species of Dactylorhiza (albeit also possessing the autapomorphy of crenulate leaf margins), and a distinct elongate, notched labellum resembling that of D. (formerly Coeloglossum) viridis. A confirmatory chromosome count for Neolindleya is desirable for comparison with Galearis, Amerorchis and Platanthera (all 2n = 42: Pridgeon et al., 1997).

Although sampling of Platanthera s.l. is sparse compared with the detailed ITS analysis of 36 species presented by Hapeman & Inoue (1997), the few new additions reveal some interesting relationships. For example, the clade immediately above P. sonoharai is supported by 82% bootstrap and two indels. At the species level, the Chinese Platanthera bakeriana is shown to be sister to the Japanese P. florentia, and the previously noted negligible divergence between P. bifolia and P. chlorantha is extended to include the questionably distinct green-flowered species P. holmboei, which characterizes montane regions of the eastern Mediterranean (its western Mediterranean equivalent, P. algeriensis, has yet to be sequenced successfully but also appears questionably distinct from P. chlorantha).

The supposed North American genus ‘Piperia’ is, as we anticipated, nested well within Platanthera s.l. (Figs 2, 3). The three species analysed show appreciable ITS divergence but nonetheless constitute a tight monophyletic group (100% bootstrap plus an indel) that is more appropriately recognized as a section Piperia within the genus Platanthera. It can therefore be viewed as analogous to other former genera now encompassed by Platanthera s.l. such as Limnorchis, Lysiella and Tulotis (cf. Rydberg, 1901a,b; Landwehr, 1977; Hapeman & Inoue, 1997; Lee, 1998; Chen et al., 1999). Admittedly, of these former genera, Piperia is perhaps the most morphologically distinct (to quote Luer, 1975: 162, ‘the species of Piperia differ distinctly from those of Platanthera or Habenaria. As much justification exists to maintain the genus Piperia as to maintain the genera Amerorchis, Dactylorhiza and Galearis distinct from Orchis’). Section Piperia is separated from the remaining five sections of Platanthera (as outlined by Hapeman & Inoue, 1997) by its very short, inconspicuous caudicles, leaves typically senescing during, rather than subsequent to, anthesis, and most notably by tubers that are globose (albeit somewhat elongate) rather than fusiform (Ackerman, 1977: table 1). Section Piperia appears to represent the most derived species group of Platanthera found in North America; the other, more derived species of the genus shown in Figure 2 either occur in Asia or Europe.

Piperia’ was both established and monographed by Rydberg (1901a,b), who recognized ten species; one further species was soon added by Suksdorf (1906). More conservative treatments by Luer (1975) and Ackerman (1977) listed totals of three and four species, respectively, the latter figure being echoed by Dressler (1993). Ackerman (1977) thoroughly revised morphological, taxonomic and biogeographical knowledge of the genus. He also noted chromosome counts of 2n = 42 for all four species, a value that similarly characterizes Platanthera s.s. (Pridgeon et al., 1997), and he reported thin-layer chromatography data that suggested low degrees of divergence among the species. He recognized only one species that possessed short (≤ 5 mm) spurs, ‘P.’ unalascensis, but soon acknowledged that this represented a highly morphologically heterogeneous species aggregate (cf. Coleman, 1995). Hence, two short-spurred species of ‘Piperia’ were segregated by Morgan & Ackerman (1990), and a further two species by Morgan & Glicenstein (1993).

Largely following these authors, we provisionally recognize ten species in Platanthera section Piperia, together with two additional subspecies of P. elegans (Table 4). Species 1–5 have spurs> 5 mm long and species 6–10 have spurs ≤ 5 mm long, although these are not natural groups; the long-spurred P. elongata is sandwiched between the short-spurred P. colemanii and the more widely distributed short-spurred P. foetida (formerly Piperia unalascensis). Spur length is a highly evolutionarily malleable character in the Orchidaceae. The section Piperia is largely confined to the states of the western seaboard of North America (P. elegans ssp. decurtii and P. yadonii being especially local and rare), though the most widely distributed species, P. foetida, has also established outliers in the extreme north-west and north-east of North America. Ackerman (1977) suggested that the close morphological similarity among ‘Piperia’ species reflects both their relatively recent origins and lack of extinction of intermediates, plausible hypotheses that are best tested by further molecular investigations of the section.

Table 4.  Revised classification of Platanthera section Piperia (formerly the genus Piperia). Asterisked taxa have not yet generated ITS sequences
(1a) Platanthera elegans Lindley ssp. elegans*, Gen. Spec. Orchid. 285 (1835).
 Synonym: Piperia elegans (Lindley) Rydberg, Bull. Torrey Bot. Club28: 270 (1901a).
(1b) Platanthera elegans Lindley ssp. maritima* (Rydberg) R.M.Bateman, comb. nov.
 Basionym: Piperia maritima Rydberg, Bull. Torrey Bot. Club28: 641 (1901b).
(1c) Platanthera elegans Lindley ssp. decurtata* (R.Morgan and Glicenstein) R.M.Bateman, comb. nov.
 Basionym: Piperia elegans Rydberg ssp. decurtata R.Morgan and Glicenstein, Lindleyana8: 93 (1993).
 (2) Platanthera elongata (Rydberg) R.M.Bateman, comb. nov.
 Basionym: Piperia elongata Rydberg, Bull. Torrey Bot. Club28: 270 (1901a).
 Synonym: Piperia elegans (Lindley) Rydberg var. elata (Jepson) Luer, Native Orchids of North America and Canada: 167 (1975).
 (3) Platanthera michaelii* (Greene) R.M.Bateman, comb. nov.
 Basionym: Habenaria michaeli Greene, Bull. Calif. Acad. Sci.1: 282 (1885).
 Synonyms: Piperia michaeli (Greene) Rydberg, Bull. Torrey Bot. Club28: 640 (1901b); Piperia elongata Rydberg ssp. michaelii (Greene) Ackerman, Bot. J. Linn. Soc.75: 266 (1977).
 (4) Platanthera leptopetala* (Rydberg) R.M.Bateman, comb. nov.
 Basionym: Piperia leptopetala Rydberg, Bull. Torrey Bot. Club28: 637 (1901b).
 (5) Platanthera transversa* (Suksdorf) R.M.Bateman, comb. nov.
 Basionym: Piperia transversa Suksdorf, Allg. Bot. Zeit. Syst.12: 43 (1906).
 (6) Platanthera foetida Geyer ex Hooker f., J. Bot. Kew Misc.7: 376 (1855).
 Basionym: Spiranthes unalascensis Sprengel, Systema Vegetabilium3: 708 (1826).
 Synonyms: Herminium unalasc(hk)ense (Sprengel) Reichenbach.f., Icon. Fl. Germ. Helv.13–14: 107 (1838), Habenaria unalasc(h)ensis (Sprengel) S.Watson, Proc. Amer. Acad. Arts12: 277 (1877); Platanthera unalascensis (Sprengel) Kurtz, Bot. Jahrb.19: 408 (1894); Piperia unalasc(h)ensis (Sprengel) Rydberg, Bull. Torrey Bot. Club28: 270 (1901a).
 (7) Platanthera cooperi* (S.Watson) R.M.Bateman, comb. nov.
 Basionym: Habenaria cooperi S.Watson, Proc. Amer. Acad. Arts12: 276 (1877).
 Synonym: Piperia cooperi (S.Watson) Rydberg, Bull. Torrey Bot. Club28: 636 (1901b).
 (8) Platanthera colemanii (R.Morgan and Glicenstein) R.M.Bateman, comb. nov.
 Basionym: Piperia colemanii R.Morgan and Glicenstein, Lindleyana8: 89 (1993).
 (9) Platanthera candida* (R.Morgan and Ackerman) R.M.Bateman, comb. nov.
 Basionym: Piperia candida R.Morgan and Ackerman, Lindleyana5: 207 (1990).
(10) Platanthera yadonii* (R.Morgan and Ackerman) R.M.Bateman, comb. nov.
 Basionym: Piperia yadonii R.Morgan and Ackerman, Lindleyana5: 209 (1990).

Our revised classification of Platanthera section Piperia (Table 4) requires ten new combinations, although fortunately the generitype, Platanthera elegans, is one of the two exceptions.

DACTYLORHIZA S.L. (28 SEQUENCES REPRESENTING 26 CODED TAXA)

Sampling of the Dactylorhiza clade (which includes the former monotypic genus Coeloglossum: Bateman et al., 1997; Pridgeon et al., 1997) has been doubled relative to our previous study. The previous pattern of relationships is retained here, though the positions of the additional taxa provide further fascinating insights that will be discussed in greater detail elsewhere. Interpretation of the genus is complicated by relatively low levels of ITS divergence, and by extensive and iterative polyploidy (e.g. Averyanov, 1990; Hedrén, 1996; Hedrén, Fay & Chase, 2001; Tyteca, 2001; Bateman & Denholm, 2003; R. Bateman and P. Hollingsworth, unpubl. obs.; all taxa above D. saccifera in Fig. 2 except D. foliosa and D. fuchsii are polyploid).

Firstly, the analysis of Bateman (2001) suggested that the south-east Asian D. hatagirea is the basally divergent member of the genus, and moreover is sufficiently distinct to appreciably weaken (from 89% to < 50%) the formerly strong bootstrap support for the monophyly of the genus (cf. Pridgeon et al., 1997); this and other similar species (e.g. ‘Orchiswardii of Chen et al., 1999) warrant more detailed study. However, incongruities in the sequence data obtained by us for D. cf. hatagirea encouraged us to omit the taxon from the present analysis, leaving a more cohesive Dactylorhiza with a synapomorphic indel and bootstrap support of 85%. The addition of D. euxina below the relatively primitive diploids of the D. incarnata aggregate (albeit a relationship poorly supported by bootstrap) is no surprise, unlike the inclusion in this clade of the Scottish allotetraploids D. purpurella and D. ebudensis, which cluster with the traunsteineri complex on morphometric (McLeod, 1995) and AFLP (Hedrén, Fay & Chase, 2001) evidence.

Next to branch off is D. iberica, which was incorrectly represented by an additional sequence for D. maculata in the analyses of Pridgeon et al. (1997) and Bateman (2001). Re-analysis of the species, conducted too late for inclusion in the present tree, places D. iberica in the weakly resolved basal portion of Dactylorhiza alongside other diploids such as D. (formerly Coeloglossum) viridis, D. aristata and the questionably distinct D. romanaD. sambucina aggregate, which includes the triploid D. insularis. The phylogenetic position of D. romana gives little support to arguments that this species has a primitive karyotype (cf. D’Emerico, Pignone & Scrugli, 2002). Within the remaining, well-supported clade, D. saccifera does not group with the morphologically similar D. fuchsii. Dactylorhiza occidentalis, now viewed by us as a probable autopolyploid, falls into the morphologically heterogeneous but indel-supported clade of the diploid D. foliosa plus the presumed autopolyploid D. maculata (cf. Hedrén, Fay & Chase, 2001; Bateman & Denholm, 2003). Out of a plexus of the D. majalis complex of allopolyploids (D. cordigera∼‘bowmanii’) emerges a well-supported clade containing the diploid D. fuchsii aggregate, plus allopolyploids of the D. traunsteineri complex and D. elata from the western Mediterranean.

Credible interpretation of these intriguing patterns will require careful integration of ITS sequences with trnL sequences (e.g. Bateman & Denholm, 2003; R. Bateman and P. Hollingsworth, unpubl. obs.), AFLPs (Hedrén et al., 2001), allozymes (e.g. Hedrén, 1996; R. Bateman and P. Hollingsworth, unpubl. obs.), plastid microsatellites (Y. Pillon, M. Chase and M. Fay, unpubl. obs.), and morphometric ordinations (e.g. Bateman & Denholm, 1989et pro, also unpubl. obs.; Tyteca, 2001). For the present (and despite considerable research effort) Dactylorhiza remains perhaps the most tantalizing of the dominantly European clades of Orchidinae, its phylogenetic history obscured partly by a combination of iterative hybridization and chromosomal instability and partly by suboptimal species delimitation and misidentifications of chosen study organisms.

GYMNADENIA S.L. (8 CODED TAXA)

As in Dactylorhiza, sampling in Gymnadenia s.l. has doubled since the publication of Pridgeon et al. (1997), who took the controversial step of sinking the two most widespread species of the morphologically distinct ‘Nigritella’ back into synonymy with Gymnadenia s.s., which would otherwise have been unequivocally paraphyletic. Further species of ‘Nigritella’ were transferred to Gymnadenia by Delforge (1998) and Teppner & Klein (1998), although this was done without the benefit of additional sequence data. Indeed, the decision to synonymise these genera has since been challenged by contradictory data from allozymes (Hedrén, Teppner & Klein, 2000) and trnL sequences (Bateman & Denholm, 2003; R. Bateman, P. Hollingsworth and M. Hollingsworth, unpubl. obs.).

The strongly supported Gymnadenia s.l. clade is presently the subject of several research projects, including morphometrics (R. Bateman and I. Denholm, unpubl. obs.), population-level ITS and trnL sequences (R. Bateman, P. Hollingsworth and M. Hollingsworth, unpubl. obs.), allozymes (K. Marhold, pers. comm., 1999; Hedrén, Klein & Teppner, 2000), karyotypes (E. Klein, pers. comm., 1997; K. Marhold, pers. comm., 1999; D’Emerico & Grünanger, 2001) and nuclear microsatellites (Campbell, 2000; V. Campbell et al., unpubl. obs.). All available evidence reinforces the then tentative decision of Bateman et al. (1997) to separate G. densiflora and G. borealis from G. conopsea as full species on the basis of significant and consistent ITS divergence (albeit less divergence than was originally inferred by Bateman et al.). Although tetraploid karyotypes have been reported in both G. densiflora (Mrkvicka, 1993) and G. conopsea s.s. (K. Marhold, pers. comm., 1999), Gymnadenia s.s. appears dominantly diploid in the U.K. (e.g. J. Bailey, pers. comm., 1998) and elsewhere in western Europe.

In contrast, the more widely distributed species of ‘Nigritella’ analysed by us are triploids and tetraploids (e.g. Teppner & Klein, 1985, 1998); only recently have we sequenced more localized diploids such as ‘N.’ lithopolitanica and ‘N.’ rhellicani (Teppner & Klein, 1985; D’Emerico, 2001). Although not included in the present analysis, they yielded ITS sequences identical to those of the tetraploids (R. Bateman et al., unpubl. obs.). Moreover, Hedrén (1999) recently demonstrated that the Scandinavian endemic apomict בGymnigritellaruneiTeppner & Klein (1989) arose by stabilized hybridization between the triploid ‘Nigritellanigra ssp. nigra and the diploid Gymnadenia conopsea s.l., suggesting genetic compatibility and thereby further justifying the incorporation of ‘Nigritella’ into Gymnadenia. Putative species of the ‘Nigritella’ complex remain undifferentiable on ITS (Bateman, 2001; R. Bateman et al., unpubl. obs.) and trnL (R. Bateman et al., unpubl. obs.) evidence, despite recent recognition of clear thin-layer chromatography differences separating the nigra and miniata (=rubra) complexes (Breiner, 1999).

On present evidence, most of the Gymnadenia s.s. species form a near-polytomy. The often sympatric G. conopsea s.s. and G. odoratissima have identical or near-identical ITS sequences, despite their clear and reliable morphological differences, and the most robust species, G. densiflora, appears somewhat unintutively to be the sister-species of the relatively diminutive members of the well-supported ‘Nigritella’ complex (cf. Bateman & DiMichele, 2002), a position contradicted by evidence from allozymes (Hedrén et al., 2000) and trnL sequences (Bateman & Denholm, 2003; R. Bateman et al., unpubl. obs.).

PONERORCHIS S.L.HEMIPILIA S.L.AMITOSTIGMANEOTTIANTHE (15 CODED TAXA)

No members of this archetypal Asian clade were sequenced by Pridgeon et al. (1997), constituting the most important sampling gap in the original study. The group has since been actively targeted for study by us, most notably by Luo Yi-Bo; our results can be compared with the strict consensus ITS trees of Luo (1999; also taken into account are unpublished trees prepared by Y.-B. Luo and A. Pridgeon).

Both these studies, together with the present investigation and Bateman (2001), agree on the existence of a  major dichotomy separating Neottianthe plus Amitostigma from Hemipilia s.s. plus ‘Habenariapurpureopunctata. However, considerable topological instability is conferred upon the clade by Ponerorchis s.l. (including Chusua). Luo (1999) tentatively found Ponerorchis to be paraphyletic relative to the AmitostigmaNeottianthe clade. Reworking of the data matrix with additional species (Y.-B. Luo and A. Pridgeon, unpubl. obs.) showed Ponerorchis to be diphyletic; a tight cluster of species surrounded the type species, P. graminifolia (not included in the present study but similar to P. jooiokiana), with ‘P.’ brevicalcarata more closely associated with Hemipilia. A similar topology resulted from Bateman's (2001) Neighbour Joining tree, with P. cf. chidori (distinguished by two indels) linked to the P. jooiokiana group. However, the maximum parsimony tree of the same matrix presented here (Fig. 2b) shows P. cf. chidori as basal within the entire group, though other most-parsimonious trees from the same analytical set placed this species immediately above P. jooiokiana; it appears to be a ‘wildcard’ taxon.

Sampling for other genes and careful re-examination of morphology are desirable to clarify relationships in this group. It seems most likely that this will lead to a narrower concept of Ponerorchis constructed around the type species, possibly with new genera being erected around P. cf. chidori and P. brevicalcarata. A further complication is provided by the Tibetan endemic ‘Habenariapurpureopunctata sensuLang & Tsi (1978), which molecularly resembles P. brevicalcarata, being depicted either as sisters (Fig. 2b) or as potentially paraphyletic to Hemipilia s.l. (Bateman, 2001). Bateman provisionally assigned both species to Hemipilia, although a long branch that includes three indels separates them from Hemipilia s.s.

In contrast, Luo & Chen (in press) controversially argued that ‘Habenariapurpureopunctata should become a new monotypic genus, Hemipiliopsis Y.B. Luo and S.C. Chen, distinguished primarily by its unusually poorly developed rostellar lobes with laterals that extend directly forward and by its naked viscidia positioned directly above the entrance of the large, saccate spur. However, these are all evolutionarly malleable characters within Orchidinae (see below); moreover, this taxonomic option does not take adequate account of the close relationship with P. brevicalcarata evident in Figure 2, which contradicts the suggestions of Luo & Chen (in press) that ‘Habenariapurpureopunctata is most closely related to Brachycorythis.

Hemipilia s.s. has 100% bootstrap support and three non-homoplastic indels. With the exception of the clear basal divergence of H. kwangsiensis, relationships within the well-sampled Hemipilia s.s. are poorly resolved and terminal branches are short, suggesting recent species-level divergence.

Further sampling and morphological study are prerequisites for eventual taxonomic revision of this interesting and under-researched clade. Overall, the majority of the genera currently recognized in this wholly Asiatic group survive the test of monophyly, despite the long-term taxonomic controversies that they continue to engender (cf. Vermeulen, 1972; von Soó, 1974; Chen, 1982; Banjeri & Pradham, 1984; Chen & Luo, 1999; Chen et al., 1999; P. J. Cribb et al., pers. comm., 1999; Luo, 1999; Luo & Chen, 1999, 2000; Bateman et al., 2001; H. A. Pederson, pers. comm., 2001; Luo, Shen & Zhu, in press; Y.-B. Luo and A. Pridgeon, unpubl. obs.).

BRACHYCORYTHIS (1 CODED TAXON)

The single ITS sequence for Brachycorythis, obtained from the relatively unspecialized west-central African species B. macrantha, was inherited from the study of Douzery et al. (1999). This hydrophilic genus of c. 35 species has an unusually wide distribution that is focused on central Africa (Summerhayes, 1955; Pridgeon et al., 2001) and diminishes in diversity southward toward Madagascar and the Cape (Linder & Kurzweil, 1999), and eastward as far as China and Taiwan (Chen et al., 1999). The genus has previously been compared with Platanthera s.l., Gymnadenia and Habenaria (Summerhayes, 1955).

Although it has an elongate-globose underground tuber, the affinities of Brachycorythis with Orchideae are not immediately apparent from its above-ground organs. In most species, the helically arranged ovate leaves are evenly distributed along the robust monopodial stem and grade imperceptibly into exceptionally large bracts in the often lax inflorescence, although a few species, segregated by some authors as Schwartzkopffia (Kränzlin, 1900), are mycoheterotrophic (Summerhayes, 1955). The flowers have spreading tepals; the petal bases show varying degrees of fusion to the gynostemium, which bears two pollinia, each with a naked viscidium. The labellum is generally shallowly three-lobed and constricted at the base, where it generates a spur that varies from elongate to saccate. Asiatic species such as B. henryi tend to have larger, more cylindrical flowers with entire labella (Chen et al., 1999). Wider geographical sampling, encompassing its Asiatic range, is highly desirable.

HABENARIINAE PLUS OUTGROUPS (22 CODED TAXA)

Our sample size of habenariids has tripled to 20 over the past five years, though it remains a minor fraction of the c. 1040 species estimated to comprise the putative subtribe (including c. 110 species previously placed within the then more broadly delimited Or-chidinae by Dressler, 1993). Previous authors of morphological studies, which generally emphasized gynostemial features, have noted that relationships within the subtribe are highly equivocal (e.g. Kurzweil, 1987; Kurzweil & Weber, 1992; Luo et al., in press), and they remain so given present DNA evidence. Our previous analyses (Pridgeon et al., 1997; Bateman, 1999a, 2001; Bateman et al., 2001) indicated varying degrees of paraphyly of the group relative to the Orchidinae, yet it was shown as weakly monophyletic (with Stenoglottis basally divergent) in the taxonomically broader ITS study of Douzery et al. (1999), implying that outgroup choice may profoundly affect the results.

Using the present data matrix and sequence alignment, Bateman's (2001) Neighbour Joining tree tentatively suggested monophyly of the habenariids, with the exception of the long-branch monotypic genus Gennaria, which was placed between the single representatives of Satyriinae and Diseae. However, short internal branches and long terminal branches indicated the statistical weakness of this association, given the present inadequate sampling of species. Maximum parsimony yielded some trees indicating complete monophyly of the habenariids (Fig. 2b) but others indicating substantial paraphyly relative to Orchidinae, resulting in a bootstrap consensus tree with no strongly supported relationships other than those between closely related species pairs (Fig. 3b). Moreover, the scattered indels all have multiple origins, being duplicated in other more derived groups within Orchidinae s.s.

Review of these recent studies indicates that the highly species-rich genus Habenaria will require extensive taxonomic fragmentation and reconstruction around monophyletic generic segregates (a parallel process to that experienced by the Orchidinae, which was progressively dismantled during the past century: Bateman et al., 1997). Herminium appears likely to survive as a monophyletic group (with H. lanceum basally divergent) if the genus is expanded to incorporate some Asian habenariid species currently assigned to Peristylus (and, even more retrogressively, to Platanthera in the case of ‘P.’ latilabris). As currently delimited, Peristylus contains some temperate species such as ‘P.’ coeloceras that, on evidence of gynostemium morphology, should be transferred to Herminium, and other tropical species such as P. goodyerioides that show greater affinities with some species of Habenaria (Luo, 1999; Luo et al., in press; see also Lang, 1999).

Other putative generic segregates, notably Pecteilis and Bonatea, occur as derived members of clades otherwise consisting of species of Habenaria s.l., and the current placement of Cynorkis is questionable, the two species sequenced (neither confidently identified to species level) only forming a monophyletic group in a minority of most-parsimonious trees. Cynorkis shows strong floral convergence with the bona fide Orchidinae genus Amitostigma; both genera possess short, broad tepals, slender spurs and planar labella divided into four lobes that expand towards their tips (cf. Vermeulen, 1972: fig. 18; Linder & Kurzweil, 1999: figs 25, 26). Once again, Stenoglottis is placed as basal to the habenariids, the two species sequenced showing relatively low divergence. Substantial variation in gynostemium structure recently observed within Habenaria should provide useful phylogenetic data (Luo et al., in press).

Focusing on the base of the tree, several anomalies are evident, both between the two ITS analyses and between both molecular studies and conventional taxonomy. The most positionally unstable genera (Gennaria, Satyrium, the now excluded Holothrix and ‘Habenariarepens, which is either misnamed or misidentified) tended to be relatively difficult to align and to be subtended by comparatively long branches, raising the spectre of possible long-branch effects. However, experimentation with our matrix suggests that the range of outgroups used may be the primary factor leading to the contrasting topologies; much denser sampling of species within these genera is desirable. The ostensible basal genera of Orchidinae and Habenariinae, Brachycorythis and Stenoglottis, respectively, especially merit further study, as they appear to define the most appropriate boundary between the Habenariinae and Orchidinae (Stenoglottis was depicted as belonging to the Habenariinae by virtue of only a single step in the ITS tree of Douzery et al., 1999).

RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

Whereas most of the clades described in the previous section are viewed as well-supported, comparison of Figures 1–3 in general, and in particular of the selected most-parsimonious tree in Figure 2 with the strict consensus tree of Figure 3, demonstrates that the relationships among most of those clades are far more ambiguous (see also Bateman et al., 1997, 2001; Pridgeon et al., 1997; Bateman, 2001). Such ambiguity is also evident in the ITS phylogeny first published by Aceto et al. (1999) and further developed by Cozzolino et al. (2001; see below).

In our original analysis (Bateman et al., 1997; Pridgeon et al., 1997), exclusion (Fig. 1a) and inclusion (Fig. 1b) of insertion–deletion events (indels) generated profoundly contrasting topologies that agreed only in the well-supported monophyly of the recircumscribed Orchidinae and in the only moderately supported relationships within the AnacamptisHimantoglossum clade, which is delimited by a chromosomal reduction from n= 21 to n= 18 and a tendency toward viscidial fusion. In Figure 1a, the sister-groups of this clade are Neotinea s.l. and Orchis s.s.Traunsteinera, respectively, together generating a clade that is weakly supported by ITS but is well marked by morphological characters such as globose tubers, sheathed inflorescences and membranous bracts (Vermeulen, 1947; Pridgeon et al., 1997).

Pridgeon et al. viewed this topology as more credible in overall patterns of relationship than that generated from nucleotides plus indels (Fig. 1b), wherein OrchisTraunsteinera is shown as sister to a fusiform/digitate-tubered clade and Neotinea appears to be the most basally divergent clade of Orchidinae. In contrast, the representation of the Platanthera, Dactylorhiza and Gymnadenia groups was viewed as more credible in Figure 1b than in Figure 1a, with the fusiform-tubered Platanthera appealingly depicted as an intermediate form leading to the relatively derived digitate tubers (see also Blinova, 2000), chromosomal reduction (from n = 21 to n = 20), and tendencies for both natural hybridization and polyploidy that unite Dactylorhiza s.l. with Gymnadenia s.l. The polarity of these key characters is, of course, the converse in Figure 1a, where these genera appear primitive rather than derived (a topology admittedly somewhat more consistent with biogeographical patterns: see below).

This triumvirate of major clades (Orchis s.s., Gymnadenia s.l.Dactylorhiza s.l., Platanthera group) is also evident in our new expanded tree (Fig. 2), with the highly ambiguously placed TraunsteineraChamorchis shown as sister to the three (i.e. no longer linked directly to Orchis s.s., as was previously believed). This relationship is supported by only a single step but makes better morphological sense than their previous placement (Fig. 1a,b), as Orchis s.s. possesses inflorescence sheaths and membranous bracts whereas Traunsteinera and especially Chamorchis are more plesiomorphic and thus akin to Pseudorchis and Platanthera in these features. The new tree also recovers, in its original topology, the derived clade that possesses the cytological synapomorphy of n= 18 (albeit with only weak bootstrap support: Fig. 3a). Interestingly, D’Emerico et al. (2001) noted that Himantoglossum s.l. and Anacamptis s.l. have similarly symmetrical karyotypes, whereas those of Serapias and Ophrys are highly asymmetrical and highly divergent, both from each other and from Himantoglossum s.l. and Anacamptis s.l., which apparently represent the relatively primitive condition. Moreover, mycorrhizal inoculation patterns under cultivation also reveal the relative phylogenetic isolation of Ophrys and Serapias in particular (R. Manuel, pers. comm., 2001).

Neotinea s.l. is on ITS evidence once again depicted as sister to the 2n = 36 clade, as in the studies of Pridgeon et al. (1997) and Bateman (2001), but contrary to Bateman (1999a), wherein the genus was tentatively placed as sister to Orchis s.s. A sister-group relationship with Orchis s.s. is more consistent with allozyme (W. Rossi, pers. comm., 2001) and karyotype (D’Emerico, 2001) evidence, though the more suspect placement is that of Orchis s.s, which is shown, somewhat unintuitively, as sister to TraunsteineraPlatantheraDactylorhizaGymnadenia, thereby once again refuting Vermeulen's (1947) concept of a clade united by globose root tubers (sensuStern, 1998), sheathed inflorescences and membranous bracts; rather, these characters are optimized as plesiomorphic across this particular topology.

At this point, it is useful to briefly consider the less densely sampled ITS trees of Orchidinae published by Aceto et al. (1999) and Cozzolino et al. (2001), which revealed similarly composed major clades but differed in the representation of the relative positions of some of those clades (admittedly, none of the incongruencies between their results and ours are supported by significant bootstrap values in either set of  trees). Most notably, Cozzolino et al. (2001: fig. 1) depicted Himantoglossum s.l. as sister to Ophrys rather than to Ophrys plus Serapias plus Anacamptis s.l., and Neotinea s.l. as sister to Platanthera, thus implying that this oval-tubered genus is nested within the fusiform-tubered clade. Neither of these placements is likely on morphological evidence, and we suspect that they reflect the use of very few ‘placeholder’ species for Himantoglossum s.l., Ophrys, and the Platanthera group, compounded by the absence of representatives of the major clades that separate these relatively derived Orchidinae from the disiid and satyrinid outgroups (i.e. the AmerorchisNeottianthe group, Brachycorythis, and the habenariids). In our analysis, the basal groups usefully reduced topological instability among major clades higher in the tree.

Perhaps the greatest novelty of the updated ITS tree presented here (Fig. 2b) is the inclusion of the Asian NeottiantheHemipilia clade and Asian–South African Brachycorythis, revealing their clear position as basally divergent to the remaining clades within the Orchidinae s.s. Given the great morphological and karyotypic diversity evident among the habenariids, and the equivocal relationships obtained in this and other recent studies, our decision to include in the analysis these additional clades of Orchidinae significantly improved the chances of obtaining realistic polarities for morphological and cytological characters (Luo, 1999; Luo & Chen, in press). The few relevant karyotypes available, of 2n = 42 for both Amitostigma gracile (Tanaka, 1965; D’Emerico, 2000), Neottianthe cucullata (Sundermann, 1980) and Brachycorythis helfei/B. obcordata (Larsen, 1966; Mehra & Kashyap, 1978), argue against primitiveness for the 2n = 36 clade. Inflorescences are unsheathed, bracts non-membranous and tubers relatively small and globose – interestingly, these are plesiomorphic features that also characterize Northern Hemisphere temperate Habenariinae such as Herminium. Relationships within the Habenariinae remain decidedly equivocal.

Brief scrutiny of the bootstrap tree (Fig. 3) reveals that the ‘spine’ of the cladogram lacks any moderately supported or well-supported branches, other than the unusually close positioning of the Dactylorhiza s.l. and Gymnadenia s.l. clades. Indels are few and mostly homoplastic; of the exceptions, one links Dactylorhiza with Gymnadenia and the other groups all clades above Brachycorythis.

EVOLUTIONARY IMPLICATIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

NON-MOLECULAR HOMOPLASY AND EVOLUTIONARY TRENDS

The above brief discussion focuses on a few morphological and cytological characters – in other words, potential synapomorphies – as ‘mapped’ across the ITS trees. Ideally, such characters should themselves be subjected to parsimony analysis (e.g. Bateman, 1999a, 2001), but this has not yet been achieved for the Orchidinae, which is species-rich and thus especially time-consuming to code. Thus far, our discussion has focused on highly conserved non-molecular characters of low homoplasy and therefore of high phylogenetic information content, notably karyotype and tuber morphology. In this section, in contrast, we examine highly homoplastic characters that are more likely to play central roles in lower-level (and potentially coadaptive) speciation events.

PIGMENTATION

Our knowledge of the floral pigmentation of the Orchidinae owes much to the pioneering quantitative work of Uphoff (1979, 1980) and the subsequent, more detailed, survey of Strack, Busch & Klein (1989). Both studies were weakened by the presence of substantial proportions of pigments that could not be precisely identified, by having sampled from concentrated regions within the phylogeny rather than more evenly, and by attempting to draw taxonomic conclusions from such highly homoplastic characters under the mistaken assumption that simplicity equates with plesiomorphy. Nonetheless, some intriguing trends are evident. Firstly, Ophrys and Serapias, distinctive long-branch genera demonstrating particularly close evolutionary relationships with pollinators, have developed their own specific pigments. In contrast, most species of Orchis s.s., Neotinea s.l., Himantoglossum p.p. and Anacamptis s.l. appear evolutionarily labile; many rely on various combinations of cyanins and especially orchicyanins, and sister-species within these genera repeatedly demonstrate the ability to switch readily from one dominant pigment to another. Most species express most pigments in trace amounts, suggesting that subtle switches allow the differential enhancement or suppression of biosynthetic pathways, in turn permitting rapid coevolutionary responses that replace suites of one to three dominant pigments.

Discrete anthocyanin markings on the vegetative parts of the orchids, restricted to the upper epidermis of the leaves in most species, originated at least three times in the Orchidinae. They characterize a wide but phylogenetically disparate range of Dactylorhiza species, both diploid and tetraploid, usually showing varying degrees of polymorphism within individual species and often forming transversely elongated patches. Leaf markings are more often longitudinally elongate in Orchis s.s., wherein mode of expression differs subtly in three phylogenetically disparate groups (O. italica, the O. mascula aggregate, and the O. anatolica aggregate), possibly reflecting independent origins of the character. In the Neotinea s.l. clade, leaf markings are confined to the basally divergent Neotinea maculata, where yet again they are polymorphic within the species, and occur as distinctive, strongly elongate dashes that decorate the stem as well as the leaves. In contrast, Steveniella satyrioides typically bears diffuse anthocyanins throughout its above-ground organs. The function of vegetative anthocyanins remains speculative; two contrasting but credible inferences are relatively long-distance attraction of pollinators or avoidance of herbivory.

OTHER POLLINATOR-RELATED CHARACTERS

The genus Serapias and the often sympatric species Anacamptis papilionacea seem to have converged on another supplementary method of probable pollinator attraction, namely a greatly expanded bract that is prominently veined and pigmented. Also, characters summarizing contrasting degrees of lobing of the labellum incur predictably high levels of homoplasy (Aceto et al., 1999: fig. 2), making them unsuitable for high-level classification despite a recent determined attempt to resurrect them (Buttler, 2001).

Within the labellum, various forms of pilose and papillose ornamentations have evolved in several distinct lineages. Large multicellular trichomes characterize the labellum of Ophrys, wherein they tend to be concentrated toward the margin, and the labellum of Serapias, wherein they tend to be concentrated toward the near-cylindrical throat. Most members of the former Orchis and of the ‘Comperia’∼Himantoglossum clade bear discrete dark spots of anthocyanin in the throat (again, presumably acting as pollinator attractants), but in the ‘core’ clade of anthropomorphic Orchis species (punctulatamilitaris) and in the more derived species of Himantoglossum s.s. (hircinum s.s.caprinum) these features are physically elevated as clusters of substantial epidermal papillae that presumably improve the orientation and/or traction of the pollinating insect upon landing. Less prominent but still brightly pigmented papillae characterize some species in other clades, including Traunsteinera globosa, Orchis spitzelii, Neotinea tridentata s.l., Anacamptis papilionacea and Anacamptis coriophoraA. fragrans.

Another character that strongly influences pollination syndromes in Orchideae is the labellar spur; specifically, its dimensions, posture and nectar-producing ability. A moderately large, downcurved spur is the plesiomorphic condition in the tribe, but this has been profoundly modified on many occasions (see also Cozzolino et al., 2001). Unusually short spurs characterize Gennaria, Neotinea maculata, Traunsteinera, Pseudorchis, ‘Coeloglossum’, ‘Nigritella’ and the HimantoglossumSteveniella clade. Spurs are even less developed in other groups, being rudimentary or non-existent in Herminium, ‘Aceras’, Chamorchis, Serapias and Ophrys. In contrast, long, downcurved spurs evolved independently in Platanthera (found in most species, nectiferous), Gymnadenia s.s. (found in  most species, nectiferous) and Anacamptis pyramidalis (not nectiferous), presumably to facilitate lepidopteran pollination. Wider, upcurved and elongate spurs lacking nectar evolved independently in the Dactylorhiza sambucina group, the Orchis masculaanatolica clade, the Anacamptis laxiflorarobusta clade and the A. boryilongicornu clade (Fig. 2), whereas ‘Habenariapurpureopunctata has innovated an apical sac (Luo & Chen, 2000, in press).

Fusion (typically congenital fusion) of floral structures is a frequent evolutionary trend in orchids (Rudall & Bateman, 2002, 2003). The presence of a galea (hood formed of the sepals and lateral petals) is also highly homoplastic, although it may prove more phylogenetically informative if dissected into more precise characters describing (a) relative degrees of connivence and fusion, and (b) whether these phenomena affect only three perianth segments (median sepal plus lateral petals) or all five (including the two lateral sepals). Certainly, preliminary study suggests that the basal fusion of the three sepals is a useful (and novel) character for distinguishing in the field Anacamptis s.l. (most species of which were formerly assigned to Orchis s.l) from Orchis s.s. (Bateman, 1999b).

Among the gymnostemium-based characters that have long been favoured by most orchid taxonomists, the presence of two viscidia (adhesive discs = retinacles) and absence of an enclosing bursicle (pollinarium sac) are primitive characters within tribe Orchideae (cf. Dressler, 1993). Fusion of the pair of pollinarium bases into a single viscidium occurred three times within subtribe Orchidinae: in Anacamptis pyramidalis, in the genus Serapias, and in the Himantoglossum s.s.–‘Barlia’ clade. Fused viscidia were foreshadowed by juxtaposed but rarely fused paired viscidia in the sister group, the former ‘Comperia’, and a similar condition has been observed in the former ‘Aceras’ (e.g. see the well-informed downgrading of gynostemium characters by Delforge, 1999).

There is considerable disagreement in the literature regarding whether some genera (e.g. Herminium, Chamorchis, Traunsteinera, Galearis, Amerorchis, ‘Nigritella’, ‘Coeloglossum’) lack bursicles or possess rudimentary bursicles; consequently, several of these genera, plus Dactylorhiza, were the focal points of arguments over the boundaries of two putative subtribes within the tribe Orchideae, the Gymnadeniinae and the Serapiadinae (cf. Delforge, 1995). Several other, relatively derived genera unequivocally possess well-developed bursicles, apparently reflecting at least three separate origins of the feature: the 2n = 36 clade, Neotinea s.l.Orchis s.s., and Dactylorhiza s.s. In addition, Dactylorhiza apparently experienced (a) partial loss of the bursicle in D. viridis (the former ‘Coeloglossum’), and (b) reputed division into two bursicles in a few other species (Nelson, 1976). Among these genera there is a reliable positive correlation between juxtaposed or fused viscidia and simple rather than bilobed bursicles. A bursicle wholly divided into two segments is a genus-level apomorphy of Ophrys, and a strongly elongate gynostemium is a genus-level apomorphy of Serapias. Deeply divided pollinia characterize Platanthera section Piperia (fig. 1 of Ackerman, 1977) and have also been observed in Neotinea maculata (plate 13 of Ross-Craig, 1971). Lastly, the grouping of Platanthera and Gymnadenia with Habenaria that was inferred using rostellar structure by Rasmussen (1985) is clearly refuted by the DNA data.

Many orchid specialists continue to give pre-eminence to gynostemium characters in typological classifications of the Orchideae and/or in phylogenetic hypotheses that are not rooted in numerical data-matrices (e.g. Vöth, 1999; Wucherpfennig, 1999; Buttler, 2001). For example, Vöth's (1999) insightful ecological interpretations of similarities in the floral morphologies of Neottia, Listera (both Epidendroideae: Neottieae), Chamorchis, ‘Coeloglossum’ and ‘Aceras’ were converted by him into a phylogenetic hypothesis of these taxa, mistakenly indicating a close relationship between the last three taxa (cf. figs 2 and 3 of Vöth). Although less homoplastic than spur dimensions and labellum shape, it appears that gynostemium characters are also modifiable (and reversable) under the influence of evolutionary interactions between flowers and pollinators. Cozzolino et al. (2001) presented a more detailed exploration of the presumed phylogenetic implications of pollination biology, notably the apparently recent separation of sister-species of the former Orchis s.l., which were attributed to switching among specialized pollinators.

Such observations offer a salutary warning regarding the dangers of a priori weighting of taxonomic characters, and further emphasize the importance of applying to systematic data the congruence text of homology through explicit, quantified cladistic analysis. However, they are not an acceptable justification for omitting potentially adaptive morphological characters from phylogenetic analyses (Bateman, 1999a, 2001).

OTHER EVOLUTIONARY TRENDS WITHIN THE MAJOR CLADES

Although more detailed quantitative exploration is desirable, some potential evolutionary trends are evident within specific clades as one traces the divergence of species distally from the base. Firstly, there is a tendency for increased concentration of leaves at the  base of the flowering stem; leaves are relatively widely spaced up the stems of the Anacamptis laxiflorarobusta, Ophrys insectifera, Pseudorchis and Traunsteinera. Interestingly, the first three examples are basally divergent members of clades that more typically possess leaves that are few in number and concentrated as basal rosettes (there is also a continued trend of reduction in leaf number in the Platanthera s.s. and GalearisAmerorchis clades). Also, flowers tend to become larger and more striking in appearance within major clades (exceptions being Anacamptis s.l. and Gymnadenia s.l.).

Moving on to more cryptic characters, increasing chromosomal asymmetries were reported within all three clades formerly attributed to Orchis (Yokata, 1987; D’Emerico et al., 1996; D’Emerico, 2000), whereas microanatomical characters such as seed morphology and developmental embryogeny appear to have remained resolutely constant; all reputedly possess ‘Orchis’ group 4b seeds (sensuWildhaber, 1972), and the ‘Orchideae type’ embryological model extends across the Orchidinae and Habenariinae (Clements, 1999). Pollen grains vary subtly in surface sculpture among species (Schill & Pfeiffer, 1977), but no clear phylogenetic patterns are evident. The intriguing suggestion by Aceto et al. (1999: 74) that basally divergent genera are more species-rich than more derived genera is not upheld by Figure 2, and is of course dependent on partly subjective choices of generic circumscription by the observer.

Not all of the evolutionary transitions implied by the cladogram are likely to reflect subtle adaptative ‘tweaking’; punctuationist interpretations of the initial subfamilial radiation of Orchidaceae were offered by both Bateman (1999a; see also Bateman, 2001; Bateman & DiMichele, 2002; Rudall & Bateman, 2002) and Aceto et al. (1999). As noted by Bateman (1999a), the strong contrast in degrees of molecular and morphological divergence implies that ‘Nigritella’ probably represents a relatively recent but successful establishment, in Europe's uplands, of a mutant of Gymnadenia, most likely originating by a single-gene mutation that radically simplified the perianth segments and rendered the flower both non-resupinate and pseudopeloric (Bateman, 1985; Rudall & Bateman, 2002). If so, this evolutionary event qualifies as non-adaptive dichotomous saltation sensuBateman &  DiMichele (1994, 2002). A similar mode of origin, probably from within Galearis, is likely for the pseudopeloric ‘Aceratorchistschiliensis (cf. Chen et al., 1999: 4). Such evolutionary–developmental events could also have initiated the highly distinct floral morphologies of, for example, Ophrys and Serapias.

HYBRIDIZATION

The frequent records of hybrids in the Orchidinae require considerable critical appraisal when, as in the vast majority of cases, they are based entirely on qualitative morphological comparison (e.g. Averyanov, 1990). Infrageneric hybrids are often under-recorded due to preclusion of identification by morphological overlap between the putative parents (Bateman et al., 1997; Bateman, 2001; Bateman & Hollingsworth, 2003). In contrast, past records of intergeneric hybrids are probably vastly inflated, due to (a) recognition of phylogenetically spurious genera, and (b) erroneous identification of morphological extremes of single species, including teratological mutants, as hybrids (e.g. Rudall & Bateman, 2002).

Those few genuine bigeneric hybrids – in other words, those that survived our past and present taxonomic rearrangements of the Orchidinae – are generally sister-genera. The most frequent hybrids between the 12 major clades of Orchidinae discussed above also occur between sisters, such as Gymnadenia×Dactylorhiza, Pseudorchis×Dactylorhiza, Anacamptis s.l.×Serapias (e.g. Sundermann, 1980; Ettlinger, 1999), and Traunsteinera with members of the fusiform/digitate-tubered clade (Peitz, 1972), supporting their revised ITS placements as sister-clades. Convincing records of hybrids are very rare between members of any pair of the six most distinct groups on Figure 2 (Habenariinae, Brachycorythis, Neottianthe group, Neotinea plus Orchis s.s. (paraphyletic), PlatantheraGymnadenia, HimantoglossumAnacamptis). One recent example involved two natural specimens of Orchis mascula×Anacamptis morio found in Cumbria, UK (Halliday, 1997: 579). The plants were short-lived and the flowers imperfectly formed, implying a predictably high degree of infragenomic incompatibility upon combining these two highly divergent lineages. Similar genetic instability is evident in a photograph of a hybrid generated artificially between Orchis mascula and Dactylorhiza fuchsii (Ettlinger, 1999: 202), suggesting that the long-term evolutionary prognosis is generally poor for orchid lineages blending two such disparate genomes. Other recent experiments in artificial hybridization have shown that the maternal parent has a far greater influence over the resulting hybrid morphology than does the paternal parent (e.g. R. Manuel, pers. comm., 2002; see also Bateman & Hollingsworth, 2003).

BIOGEOGRAPHY

Renz (1980: fig. 6; see also Chen, 1982) presented a simple evolutionary–biogeographical scenario for the Orchideae that involved an origin in south-east Asia followed by two migrations: of the Disperideae (since shown using ITS sequences to be sister to the remainder of Diseae by Douzery et al., 1999) plus Diseae plus Satyriineae (presumably also plus Habenariinae, though this was not explicitly stated) across south-western Asia and into a centre of radiation in southern Africa, and of the Orchidinae across western Asia into a centre of radiation in Europe.

Comparing the topology of Figure 2 with the current distributions of the analysed species yields an equivocal result, due to the relative instability of the sparsely sampled Brachycorythis and the various habenariid genera. In other trees (e.g. Bateman, 2001) the north-western African Gennaria lies phylogenetically outside the other habenariids, but here it groups tentatively (and morphologically improbably) with the co-occurring H. tridactylites. Bonatea, Cynorkis and Pecteilis each cluster with contrasting species of Habenaria s.l. (Fig. 2b), though many of these supposed relationships collapse in the consensus tree (Fig. 3b). The Asian ‘Peristylus’ is firmly nested within the similarly Asian Herminium s.l. Crucially from a biogeographical viewpoint, the southern African Stenoglottis and Afro-Asiatic Brachycorythis are interpolated between the aforementioned unequivocal Habenariinae genera and the basal-most bona fide clade of Orchidinae, NeottiantheHemipilia; indeed, these two genera occupy similar positions in the maximum-parsimony and maximum-likelihood trees generated from ITS data by Douzery et al. (1999). Having the Afro-Asiatic Brachycorythis tentatively basal within the Orchidinae and the southern African Stenoglottis tentatively basal within the Habenariinae raises the possibility of an ‘out of Africa’ (and into central Asia) hypothesis for the origin of both subtribes.

Overall, the difficulty of resolving the relationships among the major clades within Orchidinae suggests a  relatively early and rapid evolutionary radiation (sensuBateman, 1999a). Above Brachycorythis, the next divergent clade, NeottiantheHemipilia, is exclusively Asian. The PseudorchisPlatanthera clade exhibits the most complex biogeography – one that implies a migration following phylogenetic divergence into Pseudorchis, GalearisAmerorchis, Neolindleya and Platanthera– that tracked across north-east Asia and into North America before eventually (and somewhat improbably) colonizing Europe (Bateman et al., 1997). The present centre of diversity of Dactylorhiza is Europe, but the more basally divergent species extend into central/eastern Asia and the basal-most species analysed (D. cf. hatagirea, excluded from this analysis pending confirmation of its ITS sequence, but see Bateman, 2001) is exclusively central/eastern Asian, suggesting a western migration of the clade.

In contrast, the remaining major clades (Neotinea, Orchis s.s., TraunsteineraChamorchis, and the four clades collectively delimited by 2n = 36) are exclusively European, the intriguing exceptions being Steveniella and ‘Comperia’. These two taxa, localized within Asia Minor, are the two basal-most members of the SteveniellaHimantoglossum s.l. clade, which is in turn the basal-most group in the 2n = 36 clade. In terms of overall biogeographical patterns, the phylogenetic interpolations in Figure 2 of (a) Neotinea s.l. beneath the 2n = 36 clade, and (b) Orchis s.s. and TraunsteineraChamorchis below the more geographically widespread PlatantheraDactylorhizaGymnadenia clade, are not optimally parsimonious.

TAXONOMIC IMPLICATIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

GENERIC RE-DELIMITATION

The original ITS analysis of Pridgeon et al. (1997) and the subsequent, monophyly driven taxonomic rearrangements formalized by Bateman et al. (1997) caused predictable controversy (cf. Delforge, 1998, 1999, 2001; Gerbaud, 1998, 1999; Hughes, 1998; Teppner & Klein, 1998; Bateman, 1999b, 2001; Ettlinger, 1999; Marren, 1999: 86–88; Quentin, 1999; Wucherpfennig, 1999; Breiner & Breiner, 2000; Hardwick, 2000; Hedrén et al., 2000; Vöth, 2000; Bateman et al., 2001; Buttler, 2001; Grünanger, 2001; Kocyan & Widmer, 2001; Kretzschmar, et al., 2001, 2002; Temple, 2001; Dusak & Pernot, 2002). Nonetheless, given subsequent confirmation by independent ITS analysis of a similar range of species (Aceto et al., 1999), these taxonomic revisions have begun to permeate the secondary literature; for example, they have been used in recent detailed plant atlases for counties within the UK (e.g. French, Murphy & Atkinson, 1999), France (Dusak & Pernot, 2002) and various Mediterranean islands (Kretzschmar, Kretzschmar & Eccarius, 2001, 2002). They have also been wholly adopted in Genera Orchidacearum (Pridgeon et al., 2001), which in turn provides the nomenclature recommended by two key British societies, the Royal Horticultural Society and the Hardy Orchid Society. Generic transfers implemented by Bateman et al. (1997) involved the incorporation of Nigritella into Gymnadenia, Coeloglossum into the conserved Dactylorhiza, Aceras into Orchis s.s., and the expansion of the formerly monotypic Neotinea and Anacamptis to encompass tranches of the formerly triphyletic Orchis s.l.

Publication of the previous phylogenies (Bateman et al., 1997; Pridgeon et al., 1997; later Bateman, 1999a, 2001; Bateman et al., 2001) prompted further generic transfers of ‘Nigritella’ species by other workers: firstly the widespread Scandinavian species ‘N.’ nigra s.s. by Delforge (1998; a species since sequenced for the present study), and then a series of more finely divided and geographically localized Alpine species (examples of which were only recently sequenced by us) by Teppner & Klein (1998) and Foelsche et al. (1999). Gerbaud (1999) since resurrected Nigritella, this time as a monophyletic subgenus of Gymnadenia s.l., but he did not attempt the recognition of several corresponding subgenera within Gymnadenia s.s. necessary to erect a bona fide phylogenetic classification of the group. In contrast, Hedrén et al. (2000) used a substantial body of allozyme data to argue that Nigritella and Gymnadenia s.s. are sister-taxa (albeit close sisters) and should therefore be maintained as separate genera.

More controversially, Delforge (1999) used Pridgeon et al.∏s (1997) evidence of sister-group relationships to sink ‘Barlia’ into Himantoglossum s.s., also sinking ‘Comperia’ on suspicion of a sister-group relationship with ‘Barlia’ that has been conclusively disproved by the present study (Fig. 2). Nonetheless, as discussed in detail above, the overall balance of evidence supports Delforge's decision to incorporate both ‘Barlia’ and ‘Comperia’ into a revised concept of Himantoglossum s.l. that encompasses a substantially increased range of morphological variation. The less certain phylogenetic placement of Steveniella (Figs 2, 3) currently precludes its incorporation into an even further expanded Himantoglossum, even if it was considered sufficiently morphologically similar to the other species in the clade.

Synonymization of ‘Piperia’ into Platanthera is clearly required by the tenets of monophyly, as continued recognition of ‘Piperia’ as a genus would render Platanthera paraphyletic. Floral and vegetative similarities linking the two former genera are reflected in  the fact that the first species of ‘Piperia’ to be described was assigned to Platanthera by its author, Lindley (1835). However, this placement of ‘Piperia’ increases perceived levels of homoplasy in two suites of morphological characters routinely prioritized for high-level orchid classification: pollinarium structure (many authors) and tuber morphology (Dressler, 1993; Pridgeon et al., 1997), given that the very short caudicles and globose tubers ostensibly are characters shared with other species outside, but not within, Platanthera s.l.

On balance, Chamorchis and Traunsteinera merit continued generic recognition, unless a taxon is subsequently found that bridges the morphological and molecular gaps that currently readily distinguish them.

Decisions regarding other apparent taxonomic requirements of the present tree (Fig. 2) within the Orchidinae require additional data. The more likely changes concern the ChusuaNeottianthe clade: they include (a) confirming the distinction made here between Ponerorchis s.s. and Chusua s.s., and carefully examining the more narrowly delimited Chusua s.s. to determine whether it is monophyletic or paraphyletic, and (b) either incorporating ‘Habenariapurpureopunctata into an expanded Hemipilia s.l. or establishing this species as a new monotypic genus that is sister to Hemipilia (see also Luo & Chen, in press; Luo et al., in press). ‘Gymnadeniacamtschatica should be assigned to the pre-existing but rarely used genus name, Neolindleya.

Other options considered for taxonomic changes but ultimately rejected included erecting Dactylorhiza hatagirea as a new sister-genus to the other more derived dactylorchids, and redelimiting the Anacamptis laxiflorarobusta clade as a new genus. Even if all these suggested transfers, past, present and future, were implemented and accepted they would affect only an estimated 12% of the total number of species widely recognized within the Orchidinae s.s.

INFRAGENERIC CLASSIFICATION

There are also clear implications for potential hierarchical classifications within the more species-rich genera. Bateman et al. (1997) and Pridgeon et al. (1997) discussed at length the severe taxonomic problems that had developed when over two centuries of taxonomic research attempted to shoe-horn the triphyletic Orchis s.l. into various infrageneric classifications. Indeed, new results presented here require further dismantling of Vermeulen's (1972, 1977) classification of Orchis s.l., since his subsections Punctulatae, Provincialae and Masculacae have all proved to be non-monophyletic (see Bateman et al., 1997; note that all of the ‘placeholder’ species listed in their fig. 10 have now been sequenced). We are currently preparing a revised infrageneric classification of Orchis s.s. and have completed that of Anacamptis s.l. (R. Bateman and P. Hollingsworth, unpubl. obs.). Infrageneric classification is also now feasible for the relatively well-studied Platanthera–‘Piperia’ clade, as has already been achieved informally for Platanthera s.s. by Hapeman & Inoue (1997).

SPECIES DELIMITATION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

The issue explicitly raised by Bateman et al. (1997: see also Cafasso et al., 2000; Bateman, 2001) of the relevance of degrees of ITS sequence divergence to species delimitation can be further discussed in the light of the present results. Several disparate species within the Orchideae have now yielded multiple sequences, often acquired from conspecific individuals sampled in different countries: these include Habenaria socotrana, Neottianthe cucullata, several members of the GymnadeniaDactylorhiza clade (R. Bateman et al., unpubl. obs.; M. Chase et al., unpubl. obs.), Neotinea tridentata, Orchis paucifora and Anacamptis laxiflora. All but the supposed N. tridentata (see above) yielded sequences that were identical or apparently differed by only a single substitution, indicating both the high repeatability of the laboratory analyses and the high conservation of ITS sequences within species. Admittedly, considerably higher levels of ITS divergence were evident between some of the 25 pairs or triplets of conspecific individuals included in the recent densely sampled study of tribe Epidendroideae subtribe Laeliinae by van den Berg et al. (2000). Also, a single individual of the allotetraploid D. praetermissa was inferred by Pridgeon et al. (1997) to contain three ITS variants diverging by up to 13 bases, but re-examination of the data suggests that these putative levels of divergence were greatly over-estimated. Overall, significant ITS divergence appears a strong indicator (though not proof) of species-level differences.

However, the converse argument is not upheld; some pairs of orchid species that are clearly separable using morphological criteria reliably yield identical ITS sequences. Examples include Platanthera bifolia and P. chlorantha, Gymnadenia conopsea s.s. and G. odoratissima, and among the neottioid orchids several Epipactis species, both allogamous and autogamous (e.g. E. helleborine, E. purpurata, E. phyllanthes: P. Hollingsworth et al., unpubl. obs.). Thus, genera that have recently been highly (and controversially) split on morphological grounds, notably Ophrys, Serapias, ‘Nigritella’ and the more derived, iteratively polyploid species of Dactylorhiza, will require other more effective molecular (and morphometric) tools for delimiting bona fide species and determining their relationships. For example, Hedrén et al. (2001) used AFLP data to argue controversially that all of the tetraploid Dactylorhiza species should be amalgamated into a single species, D. elata (but see Appendix), and that the geographically widespread diploid D. fuchsii should be made conspecific with the tetraploid D. maculata (contraBateman & Denholm, 2003). More such studies are needed in order to distinguish genuine but recent infrageneric radiations from pseudo-speciation events perpetrated by orchid taxonomists adopting an extreme ‘splitters’ approach (cf. Bateman, 2001; Delforge, 2001).

SUPRAGENERIC CLASSIFICATION

Moving to higher phylogenetic levels, it is also tempting to generate classifications that group combinations of the 12 well-founded major clades discussed individually above. However, the low bootstrap values associated with most of the internal nodes along the ‘spine’ of the cladogram suggest that further formal classification would be premature and that a simultaneous analysis of multiple data sources (e.g. adding morphological and plastid sequence data to the ITS matrix: Bateman, 1999a; R. Bateman et al., unpubl. obs.) is required. Also, although the taxonomic history of the Orchidinae was well reviewed by Klinge (1898), Vermeulen (1947) and Bateman et al. (1997), there nonetheless remains an opportunity to compare in greater detail the suprageneric classifications of such orchidological luminaries as (in chronological order) Linnaeus, Haller, Swartz, Brown, Richard, Lindley, Reichenbachs p. and f., Klinge, Schlechter, Camus, Vermeulen, Dressler and Delforge. All unknowingly (or, in the case of Delforge, 2001, knowingly) deviated considerably from the dictats of monophyly due to typological reliance on a relatively small number of often highly homoplastic morphological characters.

For example, Dressler (1993) classified the Orchidinae primarily according to tuber morphology into four tentative ‘alliances’ (tabulated with distributional data as table 1 of Pridgeon et al., 1997). Of these, Alliance 4, consisting of the African genera Holothrix and Bartholina, certainly does not belong in the Orchidinae, nor is its ITS placement in the Habenariinae by Douzery et al. (1999) convincing, as the single Holothrix species is subtended by an improbably long-terminal branch; we believe that the group may even lie outside the tribe Orchideae. Figure 2 demonstrates that Dressler's Alliance 1, incorporating genera with minimal morphological expression of tubers such as Galearis and Amerorchis, is a subgroup of the more typically fusiform-tubered Platanthera clade, which occurs in Dressler's Alliance 2 alongside digitate-tubered genera of the Dactylorhiza and Gymnadenia clades. Also erroneously placed in Alliance 2 are two far more primitive groups of Orchidinae, the Asian Ponerorchis s.l. and the dominantly African Brachycorythis (see also Douzery et al., 1999). Dressler's globose-tubered Alliance 3 is the most genus-rich and morphologically heterogeneous assembly, mixing primitive Asian genera such as Amitostigma and Hemipilia with the derived 2n = 36 clade of HimantoglossumAnacamptis and the globose-tubered sister-groups (Neotinea s.l.Orchis s.s. and TraunsteineraChamorchis) of the fusiform–digitate tubered clade. Moreover, this study suggests that the globose tubers of Platanthera subgenus Piperia, included as ‘Piperia’ by Dressler in his Alliance 3, represent a reversal from the fusiform tubers that characterize the remaining subgenera of Platanthera (see above). Also included in Alliance 3 is the as-yet un-sequenced South African genus Schizochilus, which in our view is more likely to be a habenariid.

Thus, Dressler's (1993) morphologically based class-ification contains valuable insights into the phylogeny of the Orchidinae but has required considerable modification following DNA sequencing. The fairly radical generic re-delimitation accompanying our earlier ITS phylogeny (Pridgeon et al., 1997; see also Bateman, 2001) has been followed by finer tuning to reflect the present ITS phylogeny (Figs 2, 3). Current evidence suggests that the ITS phylogeny is now stabilizing for the Orchidinae if not the Habenariinae, presumably due to the much-improved taxonomic sampling of the Orchidinae. Phylogenetic knowledge of the tribe would now benefit most from confirmation by sequencing other regions of the plant genome, particularly to increase resolution along the spine of the tree. Such work is now underway (R. Bateman et al., unpubl. obs.).

FUTURE RESEARCH

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

Much of the additional sampling advocated by Bateman et al. (1997) and Pridgeon et al. (1997) has since been achieved. Consequently, the Orchidinae is now among the most thoroughly sampled tribes/subtribes of the Orchidaceae, along with the Cypripedioideae (Cox et al., 1997, 1998; Pridgeon et al., 1999) and the Epidendroideae subtribe Laeliinae (van den Berg et al., 2000). Moreover, the Orchidaceae is in turn one of the most thoroughly sampled and actively researched large families of plants.

Nonetheless, we will continue to gather additional ITS sequences with the aim of covering all species widely recognized within the group. The new data can be integrated with sequences for other taxa, either gathered by us recently but too late for inclusion in the present analysis (Himantoglossum metlesicsianum, Neotinea ustulata var. aestivalis, Dactylorhiza iberica, the diploid species of ‘Nigritella’), or requiring re-analysis due to possible misidentification of original samples (Orchis provincialis, Dactylorhiza hatagirea, Habenaria repens). Additional taxa are also required for a complete ITS-based phylogenetic analysis. Remaining priorities for further ITS sequencing within the Orchidinae include five supposed monotypic genera from Asia: Aorchis, Aceratorchis, Chondradenia, Symphyosepalum and Pseudodiphryllum (e.g. Vermeulen, 1972; Dressler, 1993; Chen et al., 1999). The need for further species-level sampling within genera already sequenced by us also focuses on  (Eur)asia. ‘Pseudorchisfrivaldii and, arguably, Himantoglossum formosum remain generically ambivalent, the former with Gymnadenia (Delforge, 1995) and the latter with the former genus Barlia (Bateman et al., 1997). Also, resolution is required of ambiguities in the NeottiantheHemipilia clade, notably the apparent paraphyly of Ponerorchis as provisionally re-delimited here.

Further molecular phylogenetic research is also desirable on putatively related southern African orchids, as their phylogenetic positions are crucial to interpretations of both the putative monophyly and respective relationships of the Orchidinae s.s. and Habenariinae (cf. Douzery et al., 1999; Linder & Kurzweil, 1999), which were controversially synonymised in Genera Orchidacearum (Pridgeon et al., 2001). Brachycorythis may encompass at least parts of Neobolusia, the monotypic Dracomonticola and Schizochilus. Moreover, Holothrix and Bartholina may even lie outside the Habenariinae. The erroneous identification of ‘H.’ repens is noteworthy, as this species was used as representative of the habenariids in early molecular phylogenetic studies of the Orchideae (e.g. Neyland & Urbatsch, 1995, 1996a,b). Similarly, the use of Peristylus coeloceras in both this and earlier studies (Douzery et al., 1999) to exemplify the genus was misleading, since morphological (Luo et al., in press) and sequence (this study) evidence shows that P. coeloceras and other high-altitude temperate species are distinct from Peristylus s.s. and should be transferred to Herminium, thereby rendering the latter genus monophyletic. Certainly, the habenariids require far more intensive (and preferably internationally collaborative) sampling for both sequencing and karyotypic studies (cf. D’Emerico, 2000). Current lists of orchids native to Africa, the centre of distribution of Habenaria, are fraught with taxonomic hazards (cf. Minasiewicz & Olszewski, 1999).

Tree rooting also remains ambiguous. We previously believed that rooting of the Habenariinae in this study would be improved by adding to our ITS matrix at least one Satyrium sequence, since the genus was phylogenetically interpolated between the Orchidinae–Habenariinae and Diseae in the phylogenetic studies of Linder & Kurzweil (1994, 1999; see also Kurzweil & Linder, 1999) and Douzery et al. (1999). However, the more detailed sampling of Bellstedt, Linder & Harley (2001) led to a near-polytomy among Disa, Satyrium and Habenaria, and in some most-parsimonious trees obtained during this study the single Satyrium sequence analysed by us nested within the habenariids. We also believe that ITS sequence alignment ambiguities, which are most problematic among the Habenariinae in our study (ultimately precluding satisfactory incorporation of Holothrix) and proved even more troubling among the Diseae in the analysis of Bellstedt et al. (2001), could be improved by reciprocal illumination with computer-generated secondary structures (cf. Hershkovitz et al., 1999; admittedly, this procedure requires sequence similarities in excess of 60%).

Indeed, reciprocal illumination has already greatly aided our understanding of the phylogeny of the Orchidinae in other ways. For example, our original ITS sequences revealed the strongly conserved phylogenetic signal of cytological characters (cf. Brandham, 1999; D’Emerico, 2001); these observations in turn allowed accurate predictions of the broad-brush phylogenetic positions of additional taxa characterized by 2n = 36, 40 or, to a lesser degree, 42. Reappraisal of morphological characters then reveals additional synapomorphies, such as the basally fused sepals of Anacamptis s.l. that apparently separate it from Orchis s.s. with unfused sepals (Bateman, 1999b; R. Bateman and P. Hollingsworth, unpubl. obs.). Interesting micromorphological distinctions have been observed between the pollinaria of Serapias plus Orchis s.s., Platanthera, and Dactylorhiza plus Gymnadenia (Barone Lumaga et al., 2000). Ongoing construction of matrices for the plastid regions trnL and matK, and for morphology (R. Bateman and P. Hollingsworth, unpubl. obs.) plus anatomy and ontogeny (Kurzweil, 1987, 1990, 1999, 2000; Kurzweil & Weber, 1991, 1992; Stern, 1998; Luo & Chen, in press; Luo et al., in press), will aid the more distant goal of inter-matrix comparison followed by simultaneous analyses of all three matrices (cf. Bateman, 1999a, 2001).

At lower taxonomic levels, the weaknesses elucidated by Bateman (2001) of the current trend towards ‘literative speciation’– upgrading infraspecific taxa to full species without gathering additional scientific data to test direct or indirect evidence of gene flow among populations – are clear; contrast for example the species-rich classifications of Dactylorhiza by Averyanov (1990) and of all European orchids by Delforge (1995, 2001) with earlier, more conservative comparable publications. As noted above and by Bateman et al. (1997), there are considerable dangers in delimiting species using degrees of ITS divergence as the only molecular evidence. Rather, combining morphometrics with the rapidly improving range of molecular fingerprinting techniques (cf. Qamaruz-Zaman et al., 1998; Soltis et al., 1998; Hollingsworth et al., 1999; Qamaruz-Zaman, 2000; Soliva et al., 2000; Tyteca, 2001) offers a more powerful approach to delimiting species, identifying putative hybrids, and placing phylogenetically troublesome polyploids, especially those within the DactylorhizaGymnadenia clade (Hedrén et al., 2001; Bateman & Denholm, 2003; R. Bateman et al., unpubl. obs., M. Chase et al., unpubl. obs.).

Lastly, we urgently need more rigorous ecological observations to generate less speculative coevolutionary studies, not only with pollinators of Orchideae (cf. Dafni, 1987; Paulus & Gack, 1990; Nilsson, 1992; van der Cingel, 1995; Neiland & Wilcock, 1998; Johnson & Steiner, 2000; Gumbert & Kunze, 2001) but also with their perennially under-researched mycorrhizal symbionts (Rasmussen, 2000; J. Leake and S. McKendrick, pers. comm., 2001).

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

We thank Stephen Bungard, Sid Clarke, Jason Courtis, Ian Denholm, Derek Turner Ettlinger, Orpah Farrington, Alfred Gössmann, Adil Güner, Mikael Hédren, Tony Hughes, Yong No Lee, David Long, Mike Lowe, Richard Manuel, Henry Noltie, Ian Phillips, Mark Rowland, Barry Tattersall, Bill Temple and Josie Welsh for supplying specimens sequenced since the publication of our 1997 analysis, Michelle Hollingsworth for generating additional sequences, Sue Meades for advice on the nomenclature of ‘Piperia’, Richard Manuel for helpful discussions on terrestrial orchids in cultivation, and Paula Rudall, Dave Roberts and three anonymous referees for their constructive comments on the manuscript.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE
  • Aceto S, Caputo P, Cozzolino S, Gaudio L, Moretti A. 1999. Phylogeny and evolution of Orchis and allied genera based on ITS DNA variation: morphological gaps and molecular continuity. Molecular Phylogenetics and Evolution 13: 6776.
  • Ackerman JD. 1977. Biosystematics of the genus Piperia Rydb. (Orchidaceae). Botanical Journal of the Linnean Society 75: 245270.
  • Angiosperm Phylogeny Group (Chase MW, Bremer K, Stevens P & 26 co-authors). 1998. An ordinal classification of the flowering plants. Annals of the Missouri Botanical Garden 85: 531553.
  • Ascherson P, Graebner P. 1907. Synopsis der Mitteleuropäischen Flora 3 (Monocotyledones). Leipzig: Engelmann.
  • Averyanov LV. 1990. A review of the genus Dactylorhiza. In: ArdittiJ, ed. Orchid biology: reviews and perspectives, V. Portland, Oregon: Timber Press, 159206.
  • Baldwin BG. 1992. Phylogenetic utility of the internal transcribed spacers of nuclear ribosomal DNA in plants: an example from the Compositae. Molecular Phylogenetics and Evolution 1: 316.
  • Baldwin BG, Sanderson MJ, Porter JM, Wojciechowski MF, Campbell CS, Donoghue MJ. 1995. The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82: 247277.
  • Banerji ML, Pradhan P. 1984. Orchids of Nepal Himalaya Vaduz: Cramer.
  • Barone Lumaga MR, Caputo P, Cozzolino S, Marallo R. 2000. Pollinaria and massulae surface morphology in the subtribe Orchidinae. In: Abstracts of the Orchis 2000 Conference, San Zeno di Montagna, Italy, 56.
  • Bateman RM. 1985. Peloria and pseudopeloria in British orchids. Watsonia 15: 357359.
  • Bateman RM. 1999a. Integrating molecular and morphological evidence for evolutionary radiations. In: HollingsworthPM, BatemanRM, GornallRJ, eds. Molecular systematics and plant evolution. London: Taylor & Francis, 432471.
  • Bateman RM. 1999b. DNA-based phylogenies are no threat to field botany. BSBI News 82: 3133.
  • Bateman RM. 2001. Evolution and classification of European orchids: insights from molecular and morphological characters. Journal Europäischer Orchideen 33: 33119.
  • Bateman RM, Denholm I. 1989. A reappraisal of the British and Irish dactylorchids, 3. The spotted-orchids. Watsonia 17: 319349.
  • Bateman RM, Denholm I. 2003. The Heath Spotted-orchid (Dactylorhiza maculata (L.) (Sóo) in the British Isles: a cautionary case-study in delimiting infraspecific taxa and inferring their evolutionary relationships. Journal Europäischer Orchideen 35: 501526.
  • Bateman RM, DiMichele WA. 1994. Saltational evolution of form in vascular plants: a neoGoldschmidtian synthesis. In: IngramDS, HudsonA, eds. Shape and form in plants and fungi. Linnean Society of Symposium Series 16. London: Academic Press, 63102.
  • Bateman RM, DiMichele WA. 2002. Generating and filtering major phenotypic novelties: neoGoldschmidtian saltation revisited. In: CronkQCB, BatemanRM, HawkinsJA, eds. Developmental genetics and plant evolution. London: Taylor & Francis, 109159.
  • Bateman RM, Hollingsworth PM. 2003. Morphological and  molecular delimitation of Anacamptis (Orchidaceae: Orchideae), with particular reference to A. x albuferensis R.M. Bateman, nothosp. nov. (A. robusta x A. fragrans), a new hybrid from Mallorca, Spain. Taxon in press.
  • Bateman RM, Hollingsworth PM, Preston J, Luo Yi-Bo Pridgeon AM, Chase MW. 2001. Orchidinae: phylogenetics. In: PridgeonAM, CribbPL, ChaseMW, RasmussenFN, eds. Genera Orchidacearum, 2 Orchidoideae. Oxford: Oxford University Press, 225233.
  • Bateman RM, Hollingsworth PM, Preston J, Pridgeon AM, Chase MW. 1998. Phylogenetics of the Orchidinae (Orchidaceae): chromosomal changes, morphological convergences and polyphyletic plesiomorphs. Second International Monocot Congress (Sydney) Abstracts, 78.
  • Bateman RM, Pridgeon AM, Chase MW. 1997. Phylogenetics of subtribe Orchidinae (Orchidoideae, Orchidaceae) based on nuclear ITS sequences. 2. Infrageneric relationships and taxonomic revision to achieve monophyly of Orchis sensu stricto. Lindleyana 12: 113141.
  • Baumann H, Künkele S. 1982. Die wildwachsenden Orchideen Europas. Stuttgart: Kosmos.
  • Baumann H, Künkele S, Lorenz R. 2000. Zur Nomenklatur von Steveniella satyrioides (Spreng.) Schlechter. Journal Europäischer Orchideen 32: 101106.
  • Baumann H, Künkele S, Lorenz R. 2002. Taxonomische Liste der Orchideen Deutschlands. Journal Europäischer Orchideen 34: 129206.
  • Bellstedt DU, Linder HP, Harley EH. 2001. Phylogenetic relationships in Disa based on non-coding trnL–trnF chloroplast sequences: evidence of numerous repeat regions. American Journal of Botany 88: 20882100.
  • Van Den Berg C, Higgins WE, Dressler RL, Whitten WM, Soto Arenas MA, Culham A, Chase MW. 2000. A phylogenetic analysis of Laeliinae (Orchidaceae) based on sequence data from Internal Transcribed Spacers (ITS) of nuclear ribosomal DNA. Lindleyana 15: 96114.
  • Bernardos S, Amich F, Gallego F, Crespi AL. 2002. Contributions to the knowledge on the orchid flora of northern Portugal. Journal Europäischer Orchideen 34: 3550.
  • Blinova I. 2000. Sprossbildung einiger Orchidaceae der Kola-Halbinsel (Nordwest-Russland) [Shoot structure of some Orchidaceae of the Kola-PI (NW Russia)]. Flora 195: 3544.
  • Brandham P. 1999. Cytogenetics. In: PridgeonAM, CribbPJ, ChaseMC, RasmussenFN, eds. Genera Orchidacearum. 1. General introduction, Apostasioideae, Cypripedioideae. Oxford: Oxford University Press, 6780.
  • Breiner E, Breiner R. 2000. Some new results of taxconomic research on Nigritella nigra s.l. Colloqium Societé Orchidée Français 6: 7780.
  • Breiner R. 1999. Farbstoff und Farbe – ihre taxonomische Relevanz bei den Orchidaceae. Journal Europäischer Orchideen 31: 423440.
  • Burns-Balogh P, Funk VA. 1986. A phylogenetic analysis of the Orchidaceae. Smithsonian Contributions to Botany 61: 179.
  • Buttler KP. 1991. Field guide to orchids of Britain and Europe Swindon, UK: Crowood Press.
  • Buttler KP. 2001. Taxonomy of Orchidaceae tribus Orchideae, a traditional approach. Journal Europäischer Orchideen 33: 732.
  • Cafasso D, Caputo P, Galesi R, Scrugli A, Cozzolino S. 2000. Incipient speciation processes in Mediterranean orchids. Abstracts of the Orchis 2000 Conference, San Zeno di Montagna, Italy, 7–8
  • Cafasso D, Pellegrino G, Musacchio A, Widmer A, Cozzolino S. 2001. Characterization of a minisatellite repeat locus in the chloroplast genome of Orchis palustris (Orchidaceae). Current Genetics 39: 394398.
  • Cameron KM, Chase MW, Whitten WM, Kores PJ, Jarrell DC, Albert VA, Yukawa T, Hills HG, Goldman DH. 1999. a phylogenetic analysis of the Orchidaceae: evidence from rbcL nucleotide sequences. American Journal of Botany 86: 208244.
  • Campbell V. 2000. Aspects of molecular ecology and population biology of Gymnadenia conopsea (L.) R.Br. in Britain. Unpublished PhD Thesis, Brighton, UK: University of Sussex.
  • Carpineri R, Rossi W. 1987. Numeri chromosomici per la flora italiana. Informazione Botanica Italica 19: 314318.
  • Chase MW. 1999. Molecular systematics, parsimony, and orchid classification. In: PridgeonAM, CribbPJ, ChaseMC, RasmussenFN, eds. Genera Orchidacearum. 1. General introduction, Apostasioideae, Cypripedioideae. Oxford: Oxford University Press, 8188.
  • Chase MW. 2001. The origin and biogeography of Orchidaceae. In: PridgeonAM, CribbPJ, ChaseMC, RasmussenFN, eds. Genera Orchidacearum. 2. Orchidoideae. Oxford: Oxford University Press, 15.
  • Chase MW, Cameron KM, Hills HG, Jarrell DC. 1994. DNA sequences and phylogenetics of the Orchidaceae and other lilioid monocots. In: PridgeonAM, ed. Proceedings of the 14th World Orchid Conference. Edinburgh: HMSO, 6173.
  • Chase MW, Fay MF, Savolainen V. 2000. Higher-level classification in the angiosperms: new insights from the perspective of DNA sequence data. Taxon 49: 685704.
  • Chen S.-C., Tsi Z.-H. Luo Y.-B. 1999. Native orchids of China in colour. Beijing: Science Press.
  • Chen Sing-chi. 1982. The origin and early differentiation of the Orchidaceae. Phytotaxonomica Sinica 20: 122.
  • Van Der Cingel NA. 1995. An atlas of orchid pollination: European orchids. Rotterdam: Balkema.
  • Clements MA. 1999. Embryology. in: PridgeonAM, CribbPJ, ChaseMC, RasmussenFN, eds. Genera Orchidacearum. 1. General introduction, Apostasioideae, Cypripedioideae. Oxford: Oxford University Press, 3858.
  • Coleman RA. 1995. The wild orchids of California. Ithaca, NY: Cornell University Press.
  • Cox AV, Abdelnour GJ, Bennett MD, Leitch IJ. 1998. Genome size and karyotype evolution in the slipper orchids (Cypripedioideae: Orchidaceae). American Journal of Botany 85: 681687.
  • Cox AV, Pridgeon AM, Albert VA, Chase MW. 1997. Phylogenetics of the slipper orchids (Cypripedioideae: Orchidaceae): nuclear rDNA sequences. Plant Systematics and Evolution 208: 197223.
  • Cozzolino S, Aceto S, Caputo P, Gaudio L, Nazzaro R. 1998. Phylogenetic relationships in Orchis and some related genera: an approach using chloroplast DNA. Nordic Journal of Botany 18: 7987.
  • Cozzolino S, Aceto S, Caputo P, Widmer A, Dafni A. 2001. Speciation processes in Eastern Mediterranean Orchis s.l. species: molecular evidence and the role of pollination biology. Israel Journal of Plant Sciences 49: 91103.
  • Cozzolino S, Caputo P, Aceto S, Rossi W, De Luca P. 1996. Testing the usefulness of ITS1 sequence as a tool to infer relationships in Orchis L. Delpinoa 33/34: 7785.
  • Cribb PJ. 1999. Morphology. In: PridgeonAM, CribbPJ, ChaseMC, RasmussenFN, eds. Genera Orchidacearum. 1. General introduction, Apostasioideae, Cypripedioideae. Oxford: Oxford University Press, 1323.
  • Cribb PJ, Chase MW. 2001. Conservation of Dactylorhiza (1481). Proposal to conserve the name Dactylorhiza Necker ex Nevski over Coeloglossum Hartm. (Orchidaceae). Taxon 50: 581582.
  • D’Emerico S. 2001. Orchideae: Cytogenetics. In: PridgeonAM, CribbPJ, ChaseMC, RasmussenFN, eds. Genera Orchidacearum. 2. Orchidoideae. Oxford: Oxford University Press, 216224.
  • D’Emerico S, Bianco P, Medagli P. 1992. Karyological studies on Orchidaceae. Tribe Ophrydeae, subtribe Serapiadinae. Caryologia 45: 301311.
  • D’Emerico S, Bianco P, Medagli P, Ruggiero L. 1990. Karyological studies of some taxa of the genera Himantoglossum, Orchis, Serapias and Spiranthes (Orchidaceae) from Apulia (Italy). Caryologia 43: 267276.
  • D’Emerico S, Cozzolino S, Pellegrino G, Pignone D, Scrugli A. 2002. Karyotype structure, supernumery chromosomes and heterochromatin distribution suggest a pathway of karyotype evolution in Dactylorhiza (Orchidaceae). Botanical Journal of the Linnean Society 138: 8591.
  • D’Emerico S, Grünanger P. 2001. Giemsa C-banding in some Gymnadenia species and in Chamorchis alpina from Dolomites (Italy). Journal Europäischer Orchideen 33: 405414.
  • D’Emerico S, Pignone D, Bianco P. 1996. Karyomorphological analyses and heterochromatin characters disclose phyletic relationships among 2n= 32 and 2n= 36 species of Orchis (Orchidaceae). Plant Systematics and Evolution 200: 111124.
  • D’Emerico S, Pignone D, Scrugli A. 2000. Giemsa C-banded karyotypes in Serapias L. (Orchidaceae). Botanical Journal of the Linnean Society 133: 485492.
  • D’Emerico S, Pignone D, Scrugli A. 2001. Karyomorphology and relationships between three genera within Orchidinae (Orchidaceae). Journal Europäischer Orchideen 33: 391394.
  • Dafni A. 1987. Pollination in Orchis and related genera: evolution from reward to deception. In: ArdittiJ, ed. Orchid biology: reviews and perspectives, IV. Ithaca, NY: Cornell University Press, 79104.
  • Del Prete C, Miceli P. 1999. Histoanatomical and taxonomical observations on some Central Mediterranean entities of Orchis section Labellotrilobatae P. Vermeul. subsections Masculae Nevski and Provinciales Nevski (Orchidaceae). Caesiana 12: 2144.
  • Delforge P. 1995. Orchids of Britain and Europe. London: HarperCollins.
  • Delforge P. 1998. Contribution taxonomique et nomenclaturale au genre Gymnadenia (Orchidaceae). Naturalistes Belges 79: 251256.
  • Delforge P. 1999. Contribution taxonomique et nomenclaturale au genre Himantoglossum (Orchidaceae). Naturalistes Belges 80: 387408.
  • Delforge P. 2000. Contribution taxonomique et nomenclaturale au genre Orchis L 1753: rémarques sur la validité du genre Steveniella Schlechter 1918. Naturalistes Belges 81: 176190.
  • Delforge P. 2001. Guide des Orchidées d’Europe (edn 2). Lausanne, Switzerland: Delachaux et Niestlé.
  • Devillers P, Devillers-Terschuren J. 1994. Essai d’analyse systématique du genre Ophrys. Naturaliste Belges 75: 273400.
  • Douzery EJP, Pridgeon AM, Kores P, Linder HP, Kurzweil H, Chase MW. 1999. Molecular phylogenetics of Diseae (Orchidaceae): a contribution from nuclear ribosomal ITS sequences. American Journal of Botany 86: 887899.
  • Dressler RL. 1993. Phylogeny and classification of the orchid family. Portland, Oregon: Timber Press.
  • Dressler RL, Chase MW. 1995. Whence the orchids? In: RudallPJ, CribbPJ, CutlerDF, HumphriesCJ, eds. Monocotyledons: systematics and evolution. Kew: Royal Botanic Gardens, 217226.
  • Dusak F, Pernot P. 2002. Les Orchidées sauvages d’île-de-France. Mèze, France: Pathénope (Biotope).
  • Essink J, Essink L, Kerremans N, Kerremans F, Schot MH. 1999. Bespreking Orchis mascula complex. Euorchis 11: 3654.
  • Ettlinger DMT. 1999. Illustrations of British and Irish orchids. Royden Cottage, Dorking, Surrey, UK: published by author.
  • Fitch WM. 1971. Toward defining the course of evolution: minimum change for a specific tree topology. Systematic Zoology 20: 406416.
  • Foelsche G, Foelsche W, Gerbaud M, Gerbaud O. 1999. Gymnadenia cenisia (Foelsche et Gerbaud) Foelsche et Gerbaud: une Nigritelle de France et d’Italie. Orchidophile 30: 235240.
  • French CN, Murphy RJ, Atkinson MGC. 1999. Flora of Cornwall. Cambourne, UK: Wheal Seton Press.
  • Freudenstein JV, Rasmussen FN. 1999. What does morphology tell us about orchid relationships? A cladistic analysis. American Journal of Botany 86: 225248.
  • Freudenstein JV, Senyo DM, Chase MW. 2000a. Mitochondrial DNA and relationships in the Orchidaceae. In: WilsonKL, MorrisonDA, eds. Monocots: systematics and evolution. Melbourne: CSIRO, 421429.
  • Freudenstein JV, Senyo DM, Chase MW. 2000b. Phylogenetic implications and comparative utility of 26S and ITS2 sequences in Orchidaceae. Americal Journal of Botany (Abstracts) 87: 127128.
  • Garay LA. 1997. De nominibus Orchidacearum Incunabulorum. Harvard Papers in Botany 11: 4950.
  • Gerbaud O. 1998. Gymnadenia R. Br. et Nigritella Rich. un seul et même genre? Cahors Societé Français Orchidées 4: 8093.
  • Gerbaud O. 1999. Considerations sur les Nigritelles et les hybrides qu’elles forment entre elles. Naturalistes Belges 80: 280, 372376.
  • Giribet G, Wheeler WC. 1999. On gaps. Molecular Phylogenetics and Evolution 13: 132143.
  • Grünanger P. 2001. Orchidacee d’Italia. Quaderni di Botanica Ambietale e Applicata 11 (2000): 380.
  • Gumbert A, Kunze J. 2001. Colour similarity to rewarding model plant affects pollination in a food deceptive orchid, Orchis boryi. Biological Journal of the Linnean Society 72: 419433.
  • Halliday G. 1997. A flora of Cumbria. Lancaster, UK: University of Lancaster.
  • Hapeman JR, Inoue K. 1997. Plant–pollinator interactions and floral radiation in Platanthera (Orchidaceae). In: GivnishTJ, SytsmaKJ, eds. Molecular evolution and adaptive radiation. Cambridge: Cambridge University Press, 433454.
  • Hardwick C. 2000. The cultivation of Coeloglossum viride. Hardy Orchid Society Newsletter 15: 17.
  • Hautzinger L. 1978. Genus Orchis L. (Orchidaceae) Sectio Robustocalcare Hautzinger. Annales Naturhistorica Museum Wien 81: 3173.
  • Hedrén M. 1996. Genetic differentiation, polyploidization and hybridization in northern European Dactylorhiza (Orchidaceae): evidence from allozyme markers. Plant Systematics and Evolution 201: 3155.
  • Hedrén M. 1999. Kommentarer om brudkullan och dess urpsrung [Comment on Gymnigritella runei (Orchidaceae) and its origin.]. Svensk Botanisk Tidsskrift 93: 145151.
  • Hedrén M, Fay MF, Chase MW. 2001. Amplified fragment length polymorphisms (AFLP) reveal details of polyploid evolution in Dactylorhiza (Orchidaceae). American Journal of Botany 88: 18681880.
  • Hedrén M, Klein E, Teppner H. 2000. Evolution of polyploids in the European orchid genus Nigritella; evidence from allozyme data. Phyton 40: 239275.
  • Hershkovitz MA, Zimmer EA, Hahn WJ. 1999. Ribosomal DNA sequences and angiosperm systematics. In: HollingsworthPM, BatemanRM, GornallRJ, eds. Molecular systematics and plant evolution. London: Taylor & Francis.
  • Hollingsworth PM, Bateman RM, Gornall RJ, eds. 1999. Molecular systematics and plant evolution. London: Taylor & Francis.
  • Hughes A. 1998. A layman's interpretation of gene sequencing [in Orchidinae.]. Hardy Orchid Society Newsletter 10: 1219.
  • Hultén E. 1968. Flora of Alaska and neighboring territories. Stanford, California University Press.
  • Johnson SD, Steiner KE. 2000. Generalization versus specialization in plant pollination systems. Trends in Ecology and Evolution 15: 140143.
  • Kenrick P, Crane PR. 1997. The origin and early diversification of land plants: a cladistic study. Smithsonian series in comparative evolutionary biology. Washington DC: Smithsonian Institution Press.
  • Klinge J. 1898. Dactylorchidis, Orchidis subgeneris, monographiae prodromus. Trudy imperatorskago S.-Peterburgsko botanischeskago sada 17: 145202.
  • Koch K. 1849. Beiträge zu einer Flora des Orientes. Linnaea 22: 177336.
  • Kocyan A, Widmer A. 2001. The old genus Orchis viewed from a molecular systematic perspective. In: NelsonE, ed. Orchis. Luzern: Felder & Vogel, 161172.
  • Kores PJ, Cameron K, Molvray M, Chase MW. 1997. The phylogenetic relationships of Orchidoideae and Spiranthoideae (Orchidaceae) as inferred from rbcL plastid sequences. Lindleyana 12: 111.
  • Kores PJ, Molvray M, Weston PH, Hopper SD, Brown AP, Cameron KM, Chase MW. 2001. A phylogenetic analysis of Diurideae (Orchidaceae) based on plastid DNA sequence data. American Journal of Botany 88: 19031914.
  • Kores PJ, Weston PH, Molvray M, Chase MW. 2000. Phylogenetic relationships within the Diurideae (Orchidaceae): Inferences from plastid matK DNA sequences. In: WilsonKL, MorrisonDA, eds. Monocots: systematics and evolution. Melbourne: CSIRO. 449456.
  • Kränzlin F. 18971904. Orchidacearum Genera et Species. Berlin: Meyer & Müller.
  • Kretzschmar H, Kretzschmar G, Eccarius W. 2001. Orchideen auf Rhodos. Bad Hersfeld: Hohl-Drunk.
  • Kretzschmar H, Kretzschmar G, Eccarius W. 2002. Orchideen a uf Kreta, Kasos, Karpathos. Bad Hersfeld: Hohl-Drunk.
  • Kümpel H. 1988. Über eine spätblühende Orchis ustulata-Sippe. Haussknechtia 4: 2324.
  • Kümpel H, Mrkvicka AC. 1990. Untersuchungen zur Abtrennung der Orchis ustulata subsp. aestivalis (Kümpel) Kümpel & Mrkvicka. Jahresbericht naturwisschaften Vereins Wuppertal 22: 306324.
  • Kurzweil H. 1987. Developmental studies in orchid flowers. II. Orchidioid species. Nordic Journal of Botany 7: 443451.
  • Kurzweil H. 1990. Floral morphology and ontogeny in the orchid subtribe Disinae. Botanical Journal of the Linnean Society 102: 6183.
  • Kurzweil H. 1999. Floral ontogeny of orchids: a review. Beiträge Biologie Pflanzen 71: 45100.
  • Kurzweil H. 2000. The value of early floral ontogeny in the systematics of Orchidaceae. In: WilsonKL, MorrisonDA, eds. Monocots: systematics and evolution. Melbourne: CSIRO, 436440.
  • Kurzweil H, Linder HP. 1999. A phylogenetic analysis of the genus Satyrium (Orchidaceae). Beitrage Biologie Pflanzen 71: 101181.
  • Kurzweil H, Weber A. 1991. Floral morphology of southern African Orchideae. 1. Orchidinae. Nordic Journal of Botany 11: 155178.
  • Kurzweil H, Weber A. 1992. Floral morphology of southern African Orchideae. 2. Habenariinae. Nordic Journal of Botany 12: 3961.
  • Landwehr J. 1977. Wilde orchideeën van Europa. Amsterdam.
  • Lang K-Y. 1999. Smithorchis , Diphylax , Herminium , Tsaiorchis , Pecteilis , Peristylus , Habenaria , Diplomeris. In: LangK-Y, ed. Flora Republicae Popularis Sinicae 17. Beijing: Science Press: 334–355, 394487.
  • Lang K-Y, Tsi Z-H. 1978. Some new taxa of Orchidaceae from Tibet, China. Acta Phytotaxonomica Sinica 16: 126129.
  • Larsen K. 1966. Studies in the flora of Thailand. 40. Cytology of vascular plants. II. Dansk Botanisk Arkiv 23: 375399.
  • Lee Yong No. 1998. Flora of Korea, 3rd edn. Seoul: Kyohak.
  • Linder HP, Kurzweil H. 1994. The phylogeny and classification of the Diseae (Orchidoideae: Orchidaceae). Annals of the Missouri Botanical Garden 81: 687713.
  • Linder HP, Kurzweil H. 1999. Orchids of southern Africa. Rotterdam: Balkema.
  • Lindley J. 1835. The genera and species of orchidaceous plants. Part IV. Ophrydeae. London: Ridgways.
  • Lorenz R. 2001. Die Gattung Serapias. Italien: Arten und Verbreitung. Journal Europäischer Orchideen 33: 235368.
  • Luer C. 1975. The native orchids of the United States and Canada, excluding Florida. New York: New York Botanical Garden.
  • Luo Y-B. 1999. Studies on the orchid genus Hemipilia. Unpublished PhD Thesis, Beijing: Institute of Botany, Chinese Academy of Science.
  • Luo Y-B, Chen S.-C. 1999. Observations on putative pollinators of Hemipilia flabellata Bur. et Franch. (Orchidaceae) in north-west Yunnan Province, China. Botanical Journal of the Linnean Society 131: 4564.
  • Luo Y-B, Chen S.-C. 2000. The floral morphology and ontogeny of some Chinese representatives of orchid subtribe Orchidinae. Botanical Journal of the Linnean Society 134: 529548.
  • Luo Y-B, Chen S.-C. in press. Hemipiliopsis: a new genus of Orchidaceae. Novon in press.
  • Luo Y-B, Shen S-C, Zhu G-H. in press. Floral morphology and ontogeny of some selected Chinese representatives of Orchidaceae subtribe Habenariinae. Annals of the Missouri Botanical Garden in press.
  • Marren P. 1999. Britain's rarest flowers. London: Poyser Press.
  • McLeod L. 1995. Combined morphometric and isozyme analysis of Dactylorhiza in Scotland. Unpublished MSc Thesis, Edinburgh: Royal Botanic Garden Edinburgh and Edinburgh University.
  • Mehra PN, Kashyap KK. 1978. In IOPB chromosome number reports, 60. Taxon 27: 223231.
  • Minasiewicz J, Olszewski TS. 1999. A provisional checklist of the continental African Orchidinae: 2. Orchidoideae 1. Orchideae: Orchidinae–Platantherinae. Fragmenta Flora Geobtanica 44: 315353.
  • Molvray M, Kores PJ, Chase MW. 2000. Polyphyly of mycoheterotrophic orchids and functional influences on floral and molecular characters. In: WilsonKL, MorrisonDA, eds. Monocots: systematics and evolution. Melbourne: CSIRO, 441448.
  • Moore DM, ed. 1980. Microspermae. CCIII. Orchidaceae. In: TutinTG, HeywoodVH, BurgesNA, MooreDM, ValentineDH, WaltersSM, WebbDA, eds. Flora Europaea V, Monocotyledones. Cambridge: Cambridge University Press, 325350.
  • Morgan R, Ackerman JD. 1990. Two new piperias (Orchidaceae) from western North America. Lindleyana 5: 205211.
  • Morgan R, Glicenstein L. 1993. Additional California taxa in Piperia (Orchidaceae). Lindleyana 8: 8995.
  • Mrkvicka AC. 1993. Statistische Untersuchungen an Gymnadenia conopsea (L.) R. Br. s.l. Mitteilungsblatt Arbeitskreis Heimische Orchideen Baden-Württemberg 25: 361367.
  • Neiland MRM, Wilcock CC. 1998. Fruit set, nectar reward, and rarity in the Orchidaceae. American Journal of Botany 85: 16571671.
  • Nelson E. 1968. Monographie und Ikonographie der Orchidaceen-Gattung Serapias, Aceras, Loroglossum, Barlia. Zürich: G & A Claraz-Schenkung.
  • Nelson E. 1976. Monographie und Ikonographie der Orchidaceen-Gattung Dactylorhiza. Zürich: G & A Claraz-Schenkung.
  • Neyland R, Urbatsch LE. 1995. A terrestrial origin for the Orchidaceae suggested by a phylogeny inferred from ndhF chloroplast gene sequences. Lindleyana 10: 244251.
  • Neyland R, Urbatsch LE. 1996a. Evolution in the number and position of fertile anthers in Orchidaceae inferred from ndhF chloroplast gene sequences. Lindleyana 11: 4753.
  • Neyland R, Urbatsch LE. 1996b. Phylogeny of subfamily Epidendroideae (Orchidaceae) inferred from ndhF chloroplast gene sequences. American Journal of Botany 83: 11951206.
  • Nilsson LA. 1992. Orchid pollination biology. Trends in Ecology and Evolution 7: 255259.
  • Nixon KC, Wheeler QD. 1992. Extinction and the origin of species. In: NovacekM, WheelerQD, eds. Extinction and phylogeny. New York: Columbia University Press, 119143.
  • Paulus HF, Gack C. 1990. Pollinators as prepollinating isolation factors: evolution and speciation in Ophrys (Orchidaceae). Israeli Journal of Botany 39: 4379.
  • Pedersen HR. 2000. On the status and synonymy of the names Dactylorhiza majalis and D. incarnata (Orchidaceae) and their typification. Taxon 49: 539543.
  • Peitz E. 1972. Zusammenstellung aller bisher bekannten Bastarde der in Deutschland verbreiteten Orchideen. Jahresberichte des Naturwisschaftlichen Vereins in Wuppertal 25: 167200.
  • Pellegrino G, Cozzolino S, Grünanger P, Musacchio A. 2001. Ribosomal DNA (ITS) as a molecular tool in the study of orchid hybridisation. Journal Europäischer Orchideen 33: 369376.
  • Pellegrino G, Nazarro R, Musacchio A, Menale B, Cozzolino S. 2000. Hybrid between Orchis mascula and O. provincialis: a not common finding. Journal Europäischer Orchideen 32: 539546.
  • Pridgeon AM, Bateman RM, Cox AV, Hapeman JR, Chase MW. 1997. Phylogenetics of the subtribe Orchidinae (Orchidoideae, Orchidaceae) based on nuclear ITS sequences. 1. Intergeneric relationships and polyphyly of Orchis sensu lato. Lindleyana 12: 89109.
  • Pridgeon AM, Cribb PJ, Chase MC, Rasmussen FN, eds. 1999. Genera Orchidacearum. 1. General introduction, Apostasioideae, Cypripedioideae. Oxford: Oxford University Press.
  • Pridgeon AM, Cribb PJ, Chase MC, Rasmussen FN, eds. 2001. Genera Orchidacearum. 2. Orchidoideae, 1. Oxford: Oxford University Press.
  • Qamaruz-Zaman F. 2000. Conservation genetics of selected British orchids. Unpublished PhD Thesis, University of Cambridge/Royal Botanic Gardens, Kew.
  • Qamaruz-Zaman F, Fay MF, Parker JS, Chase MW. 1998. Molecular techniques employed in the assessment of genetic diversity: a review focusing on orchid conservation. Lindleyana 13: 259283.
  • Quentin P. 2000. Proposition de révision des genres Orchis, Aceras, Anacamptis et Neotinea. L’Orchidophile 31: 179182.
  • Rasmussen FN. 1985. Orchids. In: DahlgrenRMT, CliffordHT, YeoPE, eds. The families of the monocotyledons: structure, evolution and taxonomy. Berlin: Springer Verlag, 249274.
  • Rasmussen FN. 1999. The development of orchid classification. In: PridgeonAM, CribbPJ, ChaseMC, RasmussenFN, eds. Genera Orchidacearum. 1. General Introduction, Apostasioideae, Cypripedioideae. Oxford: Oxford University Press, 312.
  • Rasmussen FN. 2000. Ins and outs of orchid phylogeny. In: WilsonKL, MorrisonDA, eds. Monocots: systematics and evolution. Melbourne: CSIRO. 430435.
  • Reinhammar L-G. 1995. Evidence for two distinctive species of Pseudorchis (Orchidaceae) in Scandinavia. Nordic Journal of Botany 15: 469481.
  • Reinhammar L-G, Hedrén M. 1998. Allozyme differentiation between lowland and alpine populations of Pseudorchis albida s.lat. (Orchidaceae) in Sweden. Nordic Journal of Botany 18: 714.
  • Reinhammar L-G, Olsson EGA, Sørmeland E. 2002. Conservation biology of an endangered grassland plant species, Pseudorchis albida, with some references to the closely related alpine P. straminea (Orchidaceae). Botanical Journal of the Linnean Society 139: 4766.
  • Renz J. 1980. Entwicklungstendenzen bei den Orchidoideae: Einige Betrachtungen. Jahresberichte des naturwisschaftlichen Vereins in Wuppertal 33: 3543.
  • Ross-Craig S. 1971. Drawings of British plants XXVIII Hydrocharitaceae, Orchidaceae. London: Bell.
  • Rossi W, Corrias B, Arduino P, Cianchi R, Bullini L. 1994. Multilocus electrophoresis and European orchid systematics: the genus Orchis and related genera. In: PridgeonAM, ed. Proceedings of the 14th World Orchid Conference. Edinburgh: HMSO, 7883.
  • Rudall PJ, Bateman RM. 2002. Roles of synorganisation, zygomorphy and heterotopy in the evolution of flowers: the gynostemium and labellum of orchids and other lilioid monocots. Biological Reviews 77: 403441.
  • Rudall PJ, Bateman RM. 2003. Evolutionary change in flowers and inflorescences: evidence from naturally occurring angiosperm terata. Trends in Plant Science, 8: 7682.
  • Rydberg PA. 1901a. Studies on the Rocky Mountain flora. Part 5. Bulletin of the Torrey Botanical Club 28: 266284.
  • Rydberg PA. 1901b. The North American species of Limnorchis and Piperia, north of Mexico. Bulletin of the Torrey Botanical Club 28: 605643.
  • Salazar GA. 2003. Phylogeny and classification of Subtribe Spiranthinae (Orchidaceae, Orchidoideae). Unpublished PhD Thesis, University of London/Royal Botanic Gardens, Kew.
  • Sanderson MJ. 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Molecular Biology and Evolution 14: 12181231.
  • Schill R, Pfeiffer W. 1977. Untersuchungen an Orchideenpollinien unter besonderer Berücksichtigung ihrer Feinskulpturen. Pollen et Spores 19: 5118.
  • Schlechter R. 1918. Mitteilungen über einige Europäische und Mediterrane Orchideen 1. Feddes Repertorium 15: 273302.
  • Schlegel M, Steinbrück G, Hahn K, Röttger B. 1989. Interspecific relationship of ten European orchid species as revealed by enzyme electrophoresis. Plant Systematics and Evolution 163: 107119.
  • Sell PD, Murrell G. 1996. Flora of Great Britain and Ireland. Cambridge: Cambridge University Press.
  • Simmons MP, Ochoterena H. 2000. Gaps as characters in sequence-based phylogenetic analyses. Systematic Biology 49: 369381.
  • Soliva M, Gautschi B, Salzmann C, Tenzer I, Widmer A. 2000. Isolation and characterization of microsatellite loci in the orchid Ophrys araneola (Orchidaceae) and a test of cross-species amplification. Molecular Ecology 9: 21782179.
  • Soliva M, Kocyan A, Widmer A. 2001. Molecular phylogenetics of the sexually deceptive orchid genus Ophrys (Orchidaceae) based on nuclear and chloroplast DNA sequences. Molecular Phylogenetics and Evolution 20: 7888.
  • Soltis DE, Soltis PS, Chase MW, & 13 co-authors. 2000. Angiosperm phylogeny inferred from 18S rDNA, rbcL and atpB sequences. Botanical Journal of the Linnean Society 133: 381461.
  • Soltis DE, Soltis PS, Doyle JJ. 1998. Molecular systematics of plants, 2nd edn. Amsterdam: Kluwer.
  • Von Soó R. 1974. The currently valid names and recent systematic position of species previously related to the genus Orchis in east and south-east Asia and in North America. Acta Botanica Academy of Sciences of Hungary Tomus 20: 340353.
  • Sprengel K. 1826. Caroli Linnaei Systema Vegetabilium 3. Gottingen: Dietrichian Library.
  • Stern WL. 1998. Vegetative anatomy of the subtribe Orchidinae (Orchidaceae). Botanical Journal of the Linnean Society 124: 121136.
  • Von Steven C. 1809. Decas plantarum nondum descriptarum Iberiae et Rossiae Meridionalis. Mémoires Societe Impériale Naturale Moscou 2: 175183.
  • Strack D, Busch E, Klein E. 1989. Anthocyanin patterns in European orchids and their taxonomic and phylogenetic relevance. Phytochemistry 28: 21272139.
  • Ströhlein H, Sundermann H. 1972. Chromosomenzählungen bei Erdorchideen. Jahresbericht naturwisschaften Vereins Wuppertal 25: 7580.
  • Suksdorf WN. 1906. Washingtonische pflanzen, 2. Allgemaine Botanische Zeitschrift für Systematik 12: 4243.
  • Summerhayes VS. 1955. A revision of the genus Brachycorythis. Kew Bulletin 10: 221264.
  • Sundermann H. 1972. Steveniella satyrioides (Steven) Schlechter, eine der seltensten Pflanzen der Welt. Orchidee 23: 247250.
  • Sundermann H. 1973. Himantoglossum (Loroglossum) hircinum–caprinum–calcaratum–affine. Acta Botanica Academia Scientifica Hungarica 19: 367374.
  • Sundermann H. 1980. Europäische und Mediterrane Orchideen – ein Bestimmungsflora, 3rd edn. Hildesheim: Schmersow.
  • Sundermann H, Von Der Bank C. 1977. Neue Beitrage zur Zytotaxonomie der Erdorchideen. II. Ophrys, Orchis, Neotinea, Steveniella und Gennaria. Orchideen 28: 167175.
  • Swofford DL. 2000. PAUP*: Phylogenetic Analysis Using Parsimony, Version 4.0 Washington, DC: Smithsonian Institution.
  • Tali K, Kull T. 2001. Highly variable flowering time in Orchis ustulata: consequences for population dynamics. Nordic Journal of Botany 21: 457465.
  • Tanaka R. 1965. Chromosome numbers of some species of Orchidaceae from Japan and its neighbouring areas. Journal of Japanese Botany 40: 6577.
  • Temple W. 2001. Does DNA reveal all about the evolution of terrestrial orchids? Hardy Orchid Society Newsletter 19: 1215; 20: 7–12; 21: 10–14.
  • Teppner H, Klein E. 1985. Karyologie und Fortpflanzungsmodus von Nigritella (Orchidaceae–Orchideae), inkl. N. archiducis-joannis spec. nov. und zweier Neukombinationen. Phyton (Horn) 25: 147176.
  • Teppner H, Klein E. 1989. Gymnigritella runei spec. nova (Orchidaceae–Orchideae) aus Schweden. Phyton (Horn) 29: 161173.
  • Teppner H, Klein E. 1998. Etiam atque etiam –Nigritella versus Gymnadenia: Neukombinationen und Gymnadenia dolomitensis spec. nova (Orchidaceae: Orchideae). Phyton (Horn) 38: 220224.
  • Teschner W. 1980. Sippendifferenzierung und Bestäubung bei Himantoglossum Koch. Jahresbericht naturwisschaften Vereins Wuppertal 33: 104115.
  • Tichy H, Del Prete C. 2001. Chromosome numbers of European orchids: interpreting data. Journal Europäischer Orchideen 33: 217234.
  • Tyteca D. 2001. Systematics and biostatistics of Dactylorhiza in western Europe: some recent contributions. Journal Europäischer Orchideen 33: 179199.
  • Uphoff W. 1979. Anthocyanins in the flowers of European orchids. Experientia 35: 10131014.
  • Uphoff W. 1980. Anthocyane und Evolution der Europäischen orchideen. Jahresberichte naturwisschaftlichen Vereins in Wuppertal 33: 123135.
  • Vermeulen P. 1947. Studies on dactylorchids. Utrecht, Netherlands: Schotanus & Jens.
  • Vermeulen P. 1972. Übersicht zur Systematik und Taxonomie der Gattung Orchis s.str. Jahresberichte naturwisschaftlichen Vereins in Wuppertal 25: 2236.
  • Vermeulen P. 1977. Orchideeën, systematisch ingedeeld. In: LandwehrJ, ed. Wilde orchideeën van Europa. Amsterdam, 551557.
  • Vöth W. 1999. Die Verwandtschaft zwishen Aceras und Orchis. Journal Europäischer Orchideen 31: 263276.
  • Vöth W. 2000. Gymnadenia, Nigritella und ihre Bestäuber. Journal Europäischer Orchideen 32: 547573.
  • Widmer A, Cozzolino S, Pellegrino G, Soliva M, Dafni A. 2000. Molecular analysis of orchid pollinaria and pollinaria-remains found on insects. Molecular Ecology 9: 19111914.
  • Wildhaber OJ. 1972. Zur Karpologie von Orchis. Jahresberichte naturwisschaftlichen Vereins in Wuppertal 25: 6166.
  • Wucherpfennig W. 1999. Gedanken zu einer Neuordnung der  Gattung Orchis L.: Nachtrag. Journal Europäischer Orchideen 31: 329346.
  • Yokota M. 1987. Karyotypes and phylogeny in Orchidinae and allied subtribes. In: SaitoK, TanakaR, eds. Proceedings of the 12th World Orchid Conference 1987. Tokyo: 12th World Orchid Conference Organizing Committee, 7079.

APPENDIX: NOMENCLATURAL NOTE

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. REVIEW OF PREVIOUS PHYLOGENETIC STUDIES OF THE ORCHIDEAE
  5. MATERIAL AND METHODS
  6. RESULTS
  7. RELATIONSHIPS WITHIN THE MAJOR CLADES
  8. RELATIONSHIPS AMONG THE MAJOR CLADES: CORRELATING MOLECULAR AND MORPHOLOGICAL EVIDENCE
  9. EVOLUTIONARY IMPLICATIONS
  10. TAXONOMIC IMPLICATIONS
  11. SPECIES DELIMITATION
  12. FUTURE RESEARCH
  13. ACKNOWLEDGEMENTS
  14. REFERENCES
  15. APPENDIX: NOMENCLATURAL NOTE

Terrestrial orchids continue to ‘benefit’ more than most plant groups from concerted attempts to find and establish older specific and infraspecific epithets by judicious application of the laws of nomenclatural priority. Of the many taxa analysed here, the most recent controversy has surrounded the replacement of the well-established D. majalis (Rchb. f. 1828) P.F. Hunt and Summerh. with D. comosa (Scop. 1772) P.D. Sell that was advocated by Sell in Sell & Murrell (1996; 364). Pedersen (2000) attributed the lectotype of D. comosa to the diploid D. incarnata, but the arguments of Baumann et al. (2002) that the lectotype is attributable to D. praetermissa are more convincing. However, Baumann et al. (2002) then proceeded to re-open the long-running nomenclatural debate concerning whether Platanthera montana (F.W. Schmidt 1793) Rchb. f. has priority over the widely recognized P. chlorantha (Custer 1827) Rchb. p. Given that in both cases holotypes are absent and the identity of the lectotype is ambiguous, the consequent nomenclatural complexity of the suggested changes is difficult to justify. In contrast, the generic transfers made in this paper are rooted in unambiguous phylogenetic relationships.