The issue of determining the most appropriate rank for each accepted taxon fuels ongoing controversy throughout systematics. The particularly marked escalation of such issues in modern Australian orchid systematics merits examination, not only because of wider implications in taxonomy but also because of direct effects on studies of comparative biology and conservation management.
This paper briefly reviews the causes of recent taxonomic turmoil for Australian orchids and outlines new research opportunities and conservation implications arising from an improved understanding of their molecular phylogenetics.
DNA sequencing and intensified field work have contributed towards a much improved understanding of Australian orchid systematics. Great progress has been made in discerning monophyletic groups or clades. Fresh interpretations of morphological evolution have been made possible by comparisons with the results of DNA analyses. Significant conceptual shifts from polymorphic species concepts to biological and phylogenetic concepts have also elevated the discovery and description of new species. Consequently, over the past decade, the number of Australian orchid species recognized by taxonomists has risen from approx. 900 to 1200. Similarly, the number of genera recognized by some taxonomists has increased from 110 to 192, resulting in 45% of Australian species/subspecies being assigned a new generic epithet since 2000. At higher taxonomic levels, much of the recent controversy in Australian orchid systematics reflects a divergence in views about where to split and assign formal names within unequivocally monophyletic groups. Differences regarding typification in the case of Caladenia have added additional confusion and complexity. However, new insights into and research opportunities concerning speciation processes in orchids have arisen from the wealth of new data and discrimination of species. Robustly supported molecular analyses of most clades enable comparative biological studies of Australian orchids to be conducted as never before. Outstanding subjects exist for exploring pollination by sexual deception and understanding the intricacies of mycorrhizal relationships and orchid conservation biology.
As the most species-rich plant family, it is perhaps not surprising that Orchidaceae continue to furnish ongoing taxonomic novelties and provide challenging material for evolutionary studies. Yet some of the statistics to emerge recently from Australia appear surprising even in the context of growing interest in orchid research. For example, over the past decade, the number of widely recognized Australian orchid species has risen from approx. 900 to 1200 (Jones, 2006). The number of genera has increased even more substantially, from 110 to 192. Taken at face value, this results in 45 % of Australian species/subspecies being assigned to a new genus since 2000. To add to even greater taxonomic turmoil, many species have been included under up to four different generic names published in <2 years between 2000 and 2002 (e.g. Caladenia filifera and its recent synonyms: Jones and Clements, 2002a, 2003; Szlachetko, 2003; Hopper and Brown, 2004a).
This paper briefly reviews what underlies such taxonomic turmoil for orchids in Australia and touches upon new research opportunities arising from an improved molecular phylogenetic understanding of Australian orchids. For illustrative purposes, I will focus on tribe Diurideae (Kores et al., 2001), and subtribes Caladeniinae and Thelymitrinae in particular (Figs 1 and 2). These are well-known and well-researched taxa, serving to highlight divergent approaches to defining species, genera and higher taxa, and presenting considerable evolutionary research opportunities.
A similar situation to that for Diurideae applies more broadly in recent Australian orchid systematics; examples include Pterostylis being treated as 16 genera or one (Szlachetko, 2001b; Jones and Clements, 2002a; Hopper and Brown, 2004b; Janes, 2008) and Dendrobium as six genera in Australia or one (Clements and Jones, 2002a; Clements, 2006; Burke et al., 2008). Indeed, the basic issue of determining appropriate ranks for accepted taxa underlies ongoing controversy throughout systematics (e.g. Chase, 1999; APG II, 2003). However, the particularly marked escalation of such issues in modern Australian orchid systematics, and their complex underpinnings, merit the following examination.
MOLECULAR PHYLOGENETICS OF DIURIDEAE
Chase et al. (2003) provided a revised classification of higher level taxa within Orchidaceae, based on molecular phylogenetics, which is followed here. There is now strong support for recognizing Diurideae as one of four tribes in subfamily Orchidoideae (Fig. 1). Morphological classifications have consistently misplaced and misclassified members of this mostly terrestrial orchid subfamily, even when attempting a phylogenetic interpretation (e.g. Dressler, 1993). Specifically, molecular analyses have confirmed that several taxa were traditionally misplaced within Diurideae (Kores et al., 2001; Clements et al., 2002). These include (Fig. 1) the southern Andean Codonorchis (now forming the monotypic tribe Codonorchideae), the temperate South American Chloraeinae and the monogeneric Australasian Pterostylidinae (now transferred to tribe Cranichideae).
Thus, the now predominantly Australasian Diurideae constitute seven subtribes (Chase et al., 2003) in two major clades (Kores et al., 2001; Clements et al., 2002): the first clade comprising subtribes Cryptostylidinae, Diuridineae and Thelymitrinae and the second clade containing Acianthinae, Caladeniinae, Prasophyllinae and Rhizanthellinae.
Chase et al. (2003) noted historical precedents for recognizing more subtribes in Diurideae than in other Orchideae or in Cranichideae. The circumscription of Thelymitrinae in particular was enlarged considerably from concepts such as that of Dressler (1993) by Chase et al. (2003), who included as synonyms within Thelymitrinae such traditionally recognized subtribes as Drakaeinae, Lyperanthinae and Megastylidinae (in part). This is also an advance on the thinking of subtribal delimitation embraced by Kores et al. (2001; Figs 1 and 2).
The logic applied in these analyses to naming taxa hinges principally upon monophyly (naming only well-supported clades, not polyphyletic aggregations). A second consideration is minimizing change in names while consistently applying the criterion of monophyly. This nomenclatural stability criterion is illustrated in Fig. 2 and below in the discussion of the genus Caladenia. Retaining as much as possible of the classificatory structure and content of existing names, consistent with the scientific evidence for monophyly, has benefits in terms of taxonomic literature searches, comparative analyses and predictive models in a diversity of biological disciplines.
A consideration of branch length in cladograms has been used to justify separation of sister genera such as Caleana and Paracaleana (Jones et al., 2002). However, such logic has been used sparingly in Australian orchid phylogenetics. It might be argued that focusing on morphologically well-circumscribed groups (rather than those that are particularly well supported statistically in molecular phylogenies but lack supporting morphological characters) would often lead to a more operational classification. This proposition has not been embraced in recent Australian orchid taxonomy, principally because it confounds shared morphological characters due to descent vs. those shared due to convergent evolution. In short, paraphyletic taxa are recognized and given the same rank and recognition as those that are monophyletic. The predictive power of classifications constructed in this way is diminished because unrelated taxa are grouped and sister taxa may be separated. An outcome is that evolutionary pathways for morphological characters are obscured rather than rigorously elucidated.
Indeed, fresh interpretations of morphological evolution have been made possible through the rigour of DNA analyses and their interpretation. For example, the evolution of insectiform flowers that sexually deceive male wasps as pollinators has occurred at least three times in Diurideae, in Caladeniinae (Fig. 3), Thelymitrinae (Drakaea and allies, Fig. 4) and Cryptostylidinae (Kores et al., 2001).
Circumscription of this subtribe has been particularly problematic (reviewed by Hopper and Brown, 2004a). Indeed, various authors have included such diverse genera within the genus Caladenia itself as Adenochilus, Chiloglottis, Cyrtostylis, Glossodia, Lyperanthus and Rimacola (Fig. 2). Molecular phylogenetic studies have shown that such broad concepts of Caladenia render the genus polyphyletic due to the inclusion of members of up to three of the seven subtribes of Diurideae (Kores et al., 2001).
Even with clarification of monophyletic groups within Diurideae, authors differ in their views on how best to circumscribe Caladeniinae. Kores et al. (2001) included all genera of their core Caladeniinae clade in the subtribe, a view supported by Chase et al. (2003). In contrast, Jones et al. (2002; also Clements and Jones, 2002b; Clements et al., 2002) placed Eriochilus and Adenochilus in their own subtribes. Adenochilus and Eriochilus form a grade relative to the clade of Caladenia, Cyanicula, Elythranthera, Ericksonella, Glossodia, Leptoceras, Pheladenia and Praecoxanthus sensu lato (s.l.). For both Adenochilus and Eriochilus, then, this is a classic quandary regarding appropriate formal treatment of a monogeneric sister to a larger clade (Backlund and Bremer, 1998). It can be argued that recognition of such groups as higher taxa (e.g. subtribes or tribes) has merit for information retrieval and nomenclatural stability if they have been embedded in the literature for a long time, if such recognition has been consistently applied elsewhere among their sister taxa (Hopper and Brown, 2004a) or if one takes molecular and/or morphological branch length into account. The alternative view is that such an approach increases the redundancy and decreases the cladistic information content of a classification. In any event, splitting Adenochilidinae and Eriochilidinae from Caladeniinae is a recent phenomenon (Jones et al., 2002), so arguments for nomenclatural stability, better information retrieval or historical precedent do not apply. Thus, the broader modern circumscription of Caladeniinae, including Adenochilus and Eriochilus, should prevail (Kores et al., 2001; Chase et al., 2003; Hopper and Brown, 2004a).
CIRCUMSCRIPTION OF CYANICULA, PHELADENIA AND ERICKSONELLA
Cyanicula was named by Hopper and Brown (2000) on the basis of morphological and preliminary molecular data (subsequently published by Kores et al., 2001). Cyanicula has tubers encased in a multilayered tunic (Fig. 3), its hairs lack an enlarged basal cell and its flowers are usually blue, whereas Caladenia has its tubers partially encased in a few-layered tunic and its hairs have an enlarged basal cell. The flowers of Caladenia are of many colours other than blue (Fig. 3).
Although a name was not coined, Cyanicula was first proposed as a genus distinct from Caladenia by Drummond (1838). His view was not adopted by Lindley (1840) and subsequent workers, who instead chose to name a section (Pentisia) in Caladenia for two species now in Cyanicula. Bentham (1873) considered recognizing Caladenia section Pentisia as a genus distinct from typical Caladenia, but elected not to do so, maintaining the taxon as a section within Caladenia. In the 1980s, following an analysis of morphological evidence and specimens across taxa included by Dressler (1993) in Diurideae (S. D. Hopper, unpubl. res.), I became convinced of the generic distinctness of those blue-flowered species of Caladenia with tubers encased in multilayered fibrous tunics, as Drummond (1838) had first recorded, and with trichomes lacking an enlarged basal cell. Together with Andrew Brown, I subsequently foreshadowed Cyanicula informally in Hoffman and Brown (1992). Pridgeon (1993, 1994) pursued anatomical studies to test this hypothesis, recommending that ‘Caladenia be circumscribed more narrowly and that C. gemmata and C. sericea be segregated and redefined.’ In light of molecular phylogenetic data published subsequently by Kores et al. (2001), the latter two species were placed in Cyanicula (Hopper and Brown, 2000).
Molecular phylogenetic data on Caladenia deformis were not available to Hopper and Brown (2000) when the species was included within Cyanicula. Subsequently, based on nuclear internal transcribed spacer (ITS) DNA sequences, this species was named as the monotypic genus Pheladenia, weakly supported as sister to a clade comprising Glossodia and Elythranthera (Jones et al., 2001; Clements and Jones, 2002b; P. Kores and M. Molvray, University of Oklahoma, pers. comm., 2002). However, plastid matK sequences indicated an alternative sister relationship between Pheladenia and a clade of Caladenia and Cyanicula (P. Kores and M. Molvray, pers. comm., 2002). Further research, sampling more taxa and DNA regions, is needed to refine understanding of sister relationships here, including exploration of the possibility of a hybrid origin for Pheladenia.
A final species misplaced within Caladenia proved to be what is now the monotypic Ericksonella (Hopper and Brown, 2004a). Only minor floral morphological differences separate these genera, but molecular phylogenetic data provide moderate support for a sister relationship of Ericksonella to Cyanicula rather than to Caladenia (Jones et al., 2001; Clements and Jones, 2002b). This is intriguing because ‘like Caladenia, and unlike Cyanicula, Ericksonella has tubers enclosed only in their upper half by a persistent shaggy fibrous tunic, it produces dropper tubers, and its trichomes have a swollen basal cell’ (Hopper and Brown, 2004a, p. 209). The pace of molecular evolution in this case has considerably outstripped that of morphological evolution. Indeed, molecular phylogenetics of Diurideae and Caladeniinae in particular highlight just how complex the pattern of morphological evolution has been (Kores et al., 2001). Part of the taxonomic turmoil under review is therefore due to the recent breakthroughs provided by molecular phylogenetics in rigorously establishing monophyletic groups. However, even in light of this new evidence, the perennial decisions as to what to recognize as formal taxa within clades remain hotly debated and divergent, nowhere more so than in circumscribing Caladenia itself.
CALADENIA: ONE GENUS OR SIX?
As treated by Hopper and Brown (2004a), Caladenia is the largest genus of Australasian orchids, containing 250–300 species, the majority of which are split across southern Australia between the Southwest Australian Floristic Region (SWAFR sensuHopper and Gioia, 2004) and southeastern Australia. A few species extend north into Queensland and across the Tasman Sea to New Zealand, and the high mountains of the Indonesian Archipelago are home to the most northerly representatives of the genus (Jones, 2001, 2006; Phillips et al., 2009a).
Molecular phylogenetics has established six major clades within Caladenia (Figs 2 and 3). Some authors regard these clades as genera themselves, whereas they are treated as subgenera here and elsewhere (reviewed by Hopper and Brown, 2001, 2004a) to maximize nomenclatural stability [as advocated in the Preamble to the International Code of Botanical Nomenclature (ICBN): McNeil et al., 2006] and to facilitate information retrieval through web-based searches on generic names, for example. This view is supported by others, notably Chase et al. (2003) and Phillips et al. (2009a), but not by Jones et al. (2001; also Jones, 2006), who argue for circumscribing Caladenia much more narrowly as a genus of just six species (System 2 in Fig. 3).
More work is needed to document adequately relationships within each of the Caladenia s.l. clades (e.g. Farrington et al., 2008). Disputed matters of typification greatly influence the nomenclature applied to many species if all six clades are recognized as genera (Fig. 3). For example, Jones et al. (2001, p. 391) argued that Clements' (1989) lectotypification of Caladenia on C. carnea was pre-dated by Pfitzer (1889) who is claimed to have ‘clearly selected C. flava R. Br. as the type of section Eucaladenia.’ A contrary view (Hopper and Brown, 2004a, p. 180) argued that the word ‘type’ or equivalent was not used by Pfitzer (1889). He ‘merely cited a representative and well-illustrated species for each of the five sections of Caladenia he recognised. Hence his citation of C. flava in the treatment of section Eucaladenia [= section Caladenia] was not a lectotypification. The first to do so in conformity with the Code was indeed Clements (1989). Caladenia carnea is the type species of Caladenia, not C. flava.’ Hopper and Brown (2004a) elaborated on other aspects of this complex situation regarding typification of taxa within Caladenia, some of which do not appear to have been fully understood by subsequent authors (e.g. Alrich and Higgins, 2008).
The situation in Caladenia is replicated elsewhere among Australian orchids. Many traditionally recognized genera now known to be monophyletic have been split into segregate genera wherever robustly supported clades can be recognized based on nuclear ribosomal DNA (nrDNA) analyses (for an overview advocating such fine splitting, see Jones, 2006). This applies to genera of other Australasian terrestrial orchids recognized by myself and others (e.g. Chase et al., 2003), including Arthrochilus, Chiloglottis, Corybas, Cyrtostylis, Microtis, Prasophyllum and Pterostylis. The recent turmoil is thus due, in part, to a widely accepted view of phylogenetic relationships but competing typifications and concepts of how finely to make splits within well-supported clades.
In the case of Caladenia, individual and mediated attempts to reach consensus among originators of the competing taxonomies in Fig. 3 have failed thus far. Some consensus favouring the broader approach (Hopper and Brown, 2000, 2001, 2004a, b; Chase et al., 2003; Phillips et al., 2009a) is emerging in how Australian herbaria are managing their collections (e.g. http://florabase.calm.wa.gov.au; http://www.flora.sa.gov.au; http://plantnet.rbgsyd.nsw.gov.au; http://www.rbg.vic.gov.au/dbpages/viclist/cd.) and in non-taxonomic botanical publications (e.g. Dickson and Petit, 2006). However, limited support for narrower generic concepts is seen in other non-taxonomic publications, so it is still too early to tell which view will ultimately gain the most widespread support.
Entwisle and Weston (2005) provided guidelines for deciding among competing taxonomies, including monophyly of named taxa, minimizing change (especially in groups with wide public interest), scientific defensibility and using names in majority-use among herbaria. These guidelines were applied in constructing the Australian Virtual Herbarium (http://www.anbg.gov.au/chah/avh/index.html) and suggest that the broader view of Caladenia and other genera of Australian Orchidaceae (Pterostylis, etc.) may well prevail.
POLYMORPHIC VS. OTHER CONCEPTS OF SPECIES
An extraordinary phase of the discovery and description of new species of Australian orchids has occurred over the past few decades. Is this a case of ‘species mongering’, the term applied by Joseph Hooker (1860) to excessive splitting of taxa – a practice that he vehemently opposed, both publicy and in correspondence to Charles Darwin? Almost certainly, the answer is Yes. Early in their careers, Hooker and Darwin upheld a polymorphic concept of species, derived ultimately from essentialism. The art of the taxonomist was to discern real species characters, and to discriminate these from minor morphological ‘variations’. Darwin (1859) and Hooker (1860) eventually saw that varieties might be steps along the way to becoming species, and that, therefore, there may not be fundamental differences between the characters of species and varieties (although Hooker did not fully support this view in subsequent taxonomic publications). This made the matter of species recognition dependent on the whim of educated opinion, which was bound to differ among taxonomists.
Twentieth century thinkers attempted to introduce greater scientific rigour to species concepts (Mayden 1997; Mallet, 2001; Sites and Marshall, 2003). The Biological Species Concept of Mayr in particular gave emphasis to reproductive relationships among populations. Many other nuances have appeared subsequently, but all struggle with the fact that phylogenetic transitions in breeding relationships, ecology and morphology are quantitative and complex, rarely sufficiently abrupt for rigorous evidence to enable agreement among independent workers as to the attainment of species status (Sites and Marshall, 2003). Nevertheless, there are elements of improved predictivity and falsifiability of species concepts that arise from the inclusion of reproductive, genetic and ecogeographical parameters in addition to patterns of morphological divergence (Hopper and Brown, 2001). Essentially, this philosophical shift underlies the recent upsurge in naming of orchid species in Australia. Some consensus has been reached by those practitioners advocating modern species concepts in independent laboratories across Australia, although argument prevails over the most appropriate rank in difficult cases (Hopper and Brown, 2004a). Old polymorphic species such as Bentham's (1873) extraordinarily broad concepts of Caladenia patersonii or C. filamentosa have indeed been split (compare, in Western Australian taxa for example, George, 1971 with Hopper and Brown, 2001). However, with greater field effort and an increasing cadre of knowledgeable orchid enthusiasts, many previously uncollected species have also been discovered.
SPECIATION STUDIES IN SEXUALLY DECEPTIVE ORCHIDS
Recent taxonomic revisions of Australian orchids that rely for pollination on sexual deception of male wasps (e.g. Hopper and Brown, 2006, 2007; Jones, 2006) have enabled penetrating new studies on speciation mechanisms. It has become increasingly evident that pollination by deception is a recurrent theme in orchid evolution (Schiestl, 2005; Jersáková et al., 2006; Peakall, 2007). Moreover, food-deceptive orchids appear to have speciated at a slower rate than the explosive pattern exhibited by sexually deceptive orchids (Cozzolino and Widmer, 2005; Cozzolino and Scopece, 2008).
In revising the genus Drakaea, Andrew Brown and I conducted the usual morphological studies of fresh flowers and herbarium material. However, we also used experimental ‘bait flowers’ at several locations across the geographical range of taxa to determine pollinator specificity as an additional means to identify biologically distinct species in the genus. These studies, following those of Stoutamire (1974, 1975, 1981) and Peakall (1988, 1990), established that Drakaea pollinators display remarkably high levels of species specificity for a genus of Australian plants (Hopper and Brown, 2007) (Fig. 4). When given a choice of flowers of several Drakaea species, each species of wasp may be highly selective. The existence of six new species was documented in this genus of ten endemics to the SWAFR. A similar approach was used by Jones (2006), Bower and Brown (2009) and colleagues to discriminate among species of Chiloglottis and by Phillips et al. (2009b) for species of Caladenia.
Exploring how such speciation has occurred is an active and challenging focus of contemporary Australian orchid research. It will involve a complex interplay of ecogeographic divergence, genetic system responses to the effects of small disjunct population structures, adaptation to pollinators and mycorrhizal relationships.
IMPLICATIONS FOR CONSERVATION BIOLOGY
There are many practical advantages arising from a molecular phylogenetic and biological approach to species delimitation, not the least being improved management for conservation and cultivation, and more precise communication and conduct of popular and scientific studies of orchids. Hopper and Brown (2001) elaborated upon this theme, from which the following is abstracted and updated. An example of a broad polymorphic concept of C. filamentosa was advocated by authors from Bentham (1873) to George (1971). These authors included in C. filamentosa populations of diverse ecological tolerances and phenologies, from winter-flowering plants on remote granite outcrops well inland in the arid pastoral region of southwest and South Australia, to early summer-flowering plants in high rainfall country on consolidated dunes of the south coast of Western Australia. Knowing that C. filamentosa s.l. is on a conservation reserve provides little practical help to a manager concerned about fire regimes, for instance, because the literature would indicate that the flowering season could be any time over a 6-month period, and the habitat occupied could be anything from well-drained soils on the highest eminences, to seasonally waterlogged flats low in the landscape.
Similarly, the appropriate watering regime for cultivated plants is difficult to predict if all the grower knows is that specimens are of C. filamentosa in the broad sense. Both the conservation manager and the grower are forced to find out for themselves exactly what are the biological attributes of the particular plants of C. filamentosa that they are looking after, often to the detriment of the orchids if initial guesses about fire or watering regimes are wrong.
The poor predicability of such a broad polymorphic species concept contrasts with that in more recent treatments of the C. filamentosa complex. Land managers or orchid growers will know if they have plants of C. abbreviata, that they are late flowering denizens of the south-coastal dune country of Western Australia requiring consistently damp soil moisture conditions. Caladenia remota is a winter/spring flowering inhabitant of inland arid zone granite rocks subjected to prolonged periods of dry soil, C. dimidia occurs in the Western Australian wheatbelt on good soils, etc.
To highlight another significant implication of the species concept adopted, a broad view of C. filamentosa gives no suggestion of conservation problems, because members of the complex range all across southern Australia, and some are abundant in conservation reserves. In contrast, recent biologically informed treatments segregate out from the C. filamentosa complex some taxa that are rare, C. dorrienii, C. elegans and C. melanema (Hopper et al., 1990; Brown et al., 1998).
Of course, it is not advocated that all infraspecific taxa should be elevated to specific rank for this reason. Such an approach leads to unacceptably narrow species and an often unworkable taxonomy. However, where taxa meet biological criteria regarding strong reproductive isolation in sympatry and are morphologically distinct, species rank may be considered most appropriate.
Perhaps the legacy of greatest concern regarding past concepts of polymorphic Australian orchid species is the diminished value of much of the information in the popular and scientific literature for applications in conservation and horticulture. Many authors, despite making meticulous observations over considerable periods, have used names such as C. filamentosa or C. patersonii without citing voucher specimens or even precise locations from where their material came. Without voucher specimens, it is often difficult or impossible to relate much valuable and expensively acquired data to the species segregated out from species complexes. Conflicting results obtained by researchers working on the same topic may be due to the fact that they are working on different species within a complex.
The situations detailed above for species concepts in the C. filamentosa complex could be repeated for several other complexes in Australia, such as Drakaea (Fig. 4). Taxonomic concepts for these orchids based on biology and morphological judgements have significant practical implications that have considerably advanced the cultivation, conservation and study of the plants involved (Brown et al., 1998; Phillips et al., 2009a; Swarts and Dixon, 2009).
DNA sequencing and intensified fieldwork have contributed towards a much improved understanding of Australian orchid systematics. Great progress has been made in delimiting monophyletic groups. Fresh interpretations of morphological evolution have been enabled through the rigour of DNA analyses. Significant conceptual shifts from polymorphic species concepts of the kind advocated by Hooker (1860) and Bentham (1873) to biological and phylogenetic concepts have also elevated the discovery and description of new species.
At higher taxonomic levels, much of the recent turmoil in Australian orchid systematics reflects a divergence in views about the need for nomenclatural stability as prescribed in the Preamble to the ICBN (NcNeill et al., 2006). For example, in light of molecular phylogenetics, it is more prudent and offers greater nomenclatural stability to retain a broad concept of Caladenia consistent with monophyly (System 1 in Fig. 3) than to apply a narrow concept of Caladenia by splitting it into six genera (System 2 in Fig. 3). Philosophical differences on this point have influenced where authors choose to split and assign formal names within clades. Hence there has been an extraordinary recent proliferation of generic names. Differences regarding typification in the case of Caladenia and several mistakes made in publications have also added to unnecessary confusion and complexity (see Hopper and Brown, 2004a, for details). There persist major disagreements over rank within such groups including the generic limits of Caladenia.
However, new insights and research opportunities on speciation processes in orchids have arisen from the wealth of new data and discrimination of species now underway. Robustly supported molecular analyses of most clades now enable rigorous comparative biological studies of Australian orchids. Outstanding subjects exist for exploring pollination by sexual deception and for understanding the intricacies of mycorrhizal relationships. Such insights are contributing significantly to understanding the conservation biology of the many threatened orchids of Australia (Dixon and Hopper, 1996; Swarts and Dixon, 2009).
I am indebted to Andrew Brown for collaborative research, to Mike Fay and Mark Chase for the invitation to present at the Kew symposium and comments on the manuscript, and to Rhian Smith for assistance with aspects of manuscript preparation. Richard Bateman and Henrik Æ. Pedersen provided comments as referees that materially improved the paper. Mike Fay as Special Issue Editor expedited final publication.