Molecular Phylogentics Systematics in Dendrobieae (Orchidaceae)

Molecular systematic research, using DNA sequences of the internal transcribed spacer (ITS) region of the 18-26S nuclear ribosomal repeat unit, was conducted on a broadly representative sample of the tribe Dendrobieae. The results provide independent support, in addition to evidence from plastid DNA analysis and morphology, for the phylogenetic reassessment of the taxon. At a broad level, Dendrobieae are polyphyletic with Dendrobium sect. Oxystophyllum being deeply embedded within one of the outgroup taxa, subtribe Eriinae: Podochileae. The remaining taxa form a weakly supported monophyletic group consisting of three major clades, Epigeneium, sister to predominantly Asian and Australasian clades. This author has formally recognized these as Epigeneiinae, Dendrobiinae s.s., and Grastidiinae, respectively. Detailed studies using species representative of all major historical taxonomic groups within those subtribes provide strong support for the continued recognition of the genera Cadetia, Diplocaulobium, and Flickingeria, all of which are deeply embedded within Grastidiinae and far removed from Dendrobium s.s. in Dendrobiinae. These studies have also identified numerous other strongly supported clades that group species predominantly on the basis of synapomorphic vegetative rather than floral characters. The recognition of these morphologically distinct monophyletic groups as genera is considered to be phylogenetically more informative, predictive, and realistic than any of the previously offered alternatives.

Members of Dendrobiinae species vary considerably in their floral and vegetative morphology, especially in the southern part of their range. The subtribe has already been the focus of a broad taxonomic study (Brieger 1981) and phylogenetic studies based on analysis of chloroplast DNA restriction sites (Yukawa et al. 1993(Yukawa et al. , 1996 and chloroplast DNA sequences (Yukawa et al. 2000).
However, the systematic position and presumed monophyly of Dendrobiinae sensu Schlechter ( 1926) was challenged when Brieger (1981) published an account of the group as part of a revised classification of the family. He noted that historically the circumscription of the genera had taken a peculiar course. In particular, Lindley had united a great number of taxa into two huge collective genera, Dendrobium and Eria. A consequence of this approach was that, with the description of many additional species, it became necessary to create a sometimes complex, infrageneric classification to accommodate the various forms. Brieger ( 1981) took the classification of subtribes Dendrobiinae and Eriinae to new levels of confusion by combining them into one large subtribe containing six informal categories that he termed "Genera-Series," based on overall perceived similarities of vegetative and floral morphology. During this process, he reinstated, recognized, and described many new genera. The systematic position of Dendrobiinae and Eriinae also came under scrutiny. Using patterns of embryological development from reproductive biology studies of the family, a plethora of embryo types were discovered, especially within Dendrobiinae (Clements 2002). Considering the highly conserved nature of embryological developmental types in other tribes observed within the family, the differences that exist in the representative subsample of Dendrobiinae and Eriinae seemed significant. These results suggested that Dendrobiinae, and more particularly Dendrobium, was polyphyletic with respect to Eriinae (Podochileae). At about the same time, results from phylogenetic analyses of rbcL and matK plastid sequences in Dendrobiinae (Yukawa et a!. 1993(Yukawa et a!. , 1996(Yukawa et a!. , 2000Yukawa and Uehara 1996) also arrived at a similar set of conclusions, where Dendrobiinae were shown to comprise four major clades: (i) Pseuderia, which was shown to belong to Podochilinae (Yukawa et al. 1996); (ii) Epigeneium; (iii) Dendrobium clade 1 containing species predominantly inhabitant of mainland Asia, including the type of the genus; and (iv) Dendrobium clade 2 containing taxa including Cadetia, Diplocaulobium, and Flickingeria widespread mostly in Australasia and the Pacific Islands. It was on the basis of these unpublished developmental biology and published molecular results, as well as observed morphological and biological differences, that genera such as Dockrillia (Clements and Jones 1996), Grastidium (Clements and Jones 1997), Inobulbon (Clements and Jones 1998a), Tetrodon (Clements and Jones 1998a), Winika (Clements et al. 1997), Cannaeorchis (Clements andJones 1998b), Bouletia, Cepobaculum, Ceraia, Ceratobium, Davejonesia, Dendrobates, Durabaculum, Eleutheroglossum, Thelychiton andTropilis (Clements and, from the Malesian and Australasian regions, were recognized, reinstated, or described. Since that time, considerable research using both molecular and morphological data has been undertaken on the group, resulting in proposals for considerable change in our interpretation of these taxa (Clements 2003). This paper aims to reiterate, update, and expand on results already published, as well as focus on the status of tribe Dendrobieae and major phylogenetic relationships within these taxa.

Plant Material
Material of 143 species, representative of the majority of the principal taxonomic groups within Dendrobieae, especially those within Dendrobium, was used in this study (Table 1 ). Outgroups where chosen on the basis of previous studies by Cameron et al. ( 1999), including those roughly equivalent to the ones used by Yukawa et a!. (1996Yukawa et a!. ( , 2000 and comprised species of Cypripedium, Paphiopedilum (Cypripedioideae ), Nervilia (Nervilieae ), Liparis (Malaxideae ), Bryobium, Eria (Eriinae: Podochileae), Adelopetalum, Blepharochilum, Bulbophyllum, Oxysepala, Serpenticaulis (Bulbophyllinae: Dendrobieae), Drymoanthus (Aeridinae: Vandeae) (see Table 1 from Clements 2003). Species used in this study were carefully chosen from among representatives of the ca. 350 species in the study group for which sequences where available. Fresh leaf samples were collected either in the field or from cultivated plants, mostly from the Australian National Botanic Gardens (formerly Canberra Botanic Gardens) collection, and vouchers were deposited at the Australian National Herbarium (CANE).

Morphological Data
For each species, data was collected on plant habitat, plant habit, and inflorescence development, as well as detailed floral morphology. Protocorm and seedling developmental morphology were also recorded for many species.

DNA Extraction, PCR, and Sequencing Procedures
Genomic DNA extraction and sequencing procedures used are the same as those described in the preparation of material for the analyses of Diurideae (Clements et al. 2002). The complete sequence of the ITS-1-5.8S-ITS-2 region for each sample was determined using Sequencher vers. 3.0 software (Gene Codes Corporation, Ann Arbor, Michigan, USA) to edit and assemble the sequencing chromatograms. Sequences used have been submitted to GenBank and given accession numbers (see Clements 2003).

Method of Alignment and Sequence Analysis
Sequences were first aligned using the ECLUSTALW program supplied by the Australian National Genomic Information Service (ANGIS). Multiple alignment parameters were set at the default values; a gap opening penalty of 10 (range 1-1 00), a gap extension penalty of 5 (range 0.10-100), and gap separation penalty of 8 (range 1-50). Sequences were secondarily manually aligned, using BioEdit vers. 4.7.8, and the aligned files converted to a PAUP/NEXUS file.      (Maddison and Maddison 1992) for analysis using PAUP* vers. 4.0b4a (Swofford 1998). The most-parsimonious trees were determined using a heuristic search algorithm with 100 replicates of random taxon addition, tree bisection reconnection (TBR) branch swapping and the MULTREES option, generating a consensus tree. Successive weighting (Farris 1969) was applied through recalculation of the rescaled consistency indices until a stable topology was attained. Bootstrap analyses were done for both unweighted and successive weighted trees, to determine the relative support for the clades (Felsenstein 1985) and all minimal-length trees were saved. At the tribal level, an assessment of Dendrobieae, centered on Dendrobiinae, was first undertaken through a sequence analysis of 23 species, representative of most major taxonomic groups used in previous studies of the group (Yukawa et al. 1993) including seven outgroup species, viz. Bryobium, Eria (Eriinae: Podochileae), Adelopetalum, Bulbophyllum and Oxysepala (Bulbophyllinae: Dendrobieae), Liparis (Malaxideae) and Drymoanthus (Aeridinae: Vandeae). A second and more comprehensive analysis was then undertaken based on 142 species, including representatives of all major historically recognized taxa within Dendrobiinae plus a number of outgroup taxa, viz. Cypripedium, Paphiopedilum (Cypripedioideae), Nervilia (Nervilieae), Liparis (Malaxideae), Bryobium, Eria (Eriinae: Podochileae), and Adelopetalum, Blepharochilum, Bulbophyllum, Oxysepala, Serpenticaulis (Bulbophyllinae: Dendrobieae ).

RESULTS
An assessment of Dendrobieae using 23 representative species produced an alignment that contains 786 nucleotide sites of which 324 were potentially parsimony informative. Analysis of this computer-generated alignment produced three equally, near identical parsimonious trees; tree length (L) = 1440, consistency index (CI) = 0.5743, retention index (RI) = 0.4853, and rescaled consistency index (RC) = 0.2787. One of these trees is shown and compared with that of Yukawa et al. (1993; Fig. 1). These results show, for the ITS tree, a polyphyletic Dendrobiinae with: (i) D. govidjoae and D. sinuatum, two species representative of Dendrobium sect. Oxystophyllum, deeply embedded within a strongly supported (100% bootstrap support), monophyletic Eriinae (Podochileae); (ii) a strongly supported (96% bootstrap support), monophyletic Epigeneium, sister to the reminder of Dendrobiinae; and (iii) the remaining representatives of Dendrobiinae separated into two major groups, the Australasian and Asian clades. Excluding Dendrobium sect. Oxystophyllum, these results reveal only weak support ( < 50% bootstrap support) for the monophyly of the remainder of the ingroup, including Epigeneium. Conversely, there is moderate support for the monophyly of Epigeneium, the Australasian (85% bootstrap support) and Asian (91% bootstrap support) clades. These results also show that Dendrobieae are polyphyletic, when Bulbophyllinae are included within that concept, where Eriinae (Podochileae) are embedded between the two main elements of tribe Dendrobieae.
Results from a second more comprehensive study were based on the analyses of 142 species from an alignment of 846 nucleotide sites, of which 529 were potentially parsi-mony informative, produced 66 equally parsimonious trees; L = 4797, CI = 0.2760, RI = 0.6318, and RC = 0.1744. One randomly selected tree is shown as a phylogram in Fig.  2. These results show that the outgroup taxa, Cypripedium and Paphiopedilum, are isolated from the remainder of the study group. Epidendroideae are monophyletic, and within them are contained representatives of strongly supported tribes in Nervilieae, Malaxideae, Podochileae, as well as a weakly supported Dendrobieae comprised of four major clades, Bulbophyllinae, Epigeneium, an Asian and an Australasian clade. Epigeneium and Dendrobium sect. Oxystophyllum together with Bulbophyllinae remain isolated from the main body of Dendrobiinae.

DISCUSSION
These phylogenetic analyses, based primarily on ITS nrDNA sequence data, in conjunction with morphological data, provide a further insight into the overall relationships of most of Dendrobieae, including elements of the Asian clade, Dendrobiinae, not already accounted for in a recent publication by Clements (2003). Importantly, these results also outline the basic phylogeny of major taxa within the Australian clade, Grastidiinae, further emphasizing the importance to recognize this taxon separate from Dendrobiinae.
At the broader level, using a similar range of species to that used in previous studies of chloroplast rbcL and matK sequences, results strongly correlate with the phylogeny of Dendrobieae presented by Yukawa et al. (1996Yukawa et al. ( , 2000. Both sets of results show remarkable similarity, where representatives of Epigeneium grouped together in a strongly supported clade (96-100% bootstrap support) isolated from other ingroup, Bulbophyllinae, Eriinae, Malaxidinae, and outgroup taxa, Cypripedioideae. The strong correlation between   (1993). The right-hand side tree is one of three most-parsimonious trees generated, derived from a data matrix based on nuclear ITS sequences, showing the composition of Dendrobiinae with Oxystophyllum, Epigeneium, and Australasian and Asian clades identified relative to seven outgroup species: L = 1440, CI = 0.5743, RI = 0.4853. Numbers above branches are branch length estimates (ACCTRAN optimalization); bootstrap percentages greater than 50% are given below in bold.
these results provides further evidence that the historical treatment of Dendrobiinae based primarily on floral morphology is inconsistent with estimates of phylogenetic relationships. One possible outcome from the assessment of these results is the recognition of three clades as subtribes, viz. Epigeneiinae, Dendrobiinae, and Grastidiinae (Clements 2003), with Bulbophyllinae as sister-group to these three subtribes, as proposed by Dressler (1993) and others. Likewise, the presumed close relationship among these four taxa and Podochileae, of which Eriinae are a part, also seems sound. Significantly, these and previous ITS sequence analyses results (Clements 2003) also show Dendrobium govidjoae and D. sinuatum related to representatives of Eriinae (Podochileae). Both species have historically been treated either as members of Dendrobium sect. Oxystophyllum (Mi-quel 1859;Smith 1905a, b;Kraenzlin 1910;Schlechter 1912) or as a section of the genus Aporum (Reichenbach 1861;Hooker 1890;Brieger 1981). When treated within Dendrobium, sect. Oxystophyllum has been interpreted as allied to sects. Aporum and Rhopalanthe (Schlechter 1912) on account of similarities of vegetative features, most notably the equitant leaves and the abbreviated, lateral inflorescences, formed in the axils of the leaf sheaths, along the stem. The position of sect. Oxystophyllum outside Dendrobieae renders the tribe paraphyletic unless Podochileae and Bulbophyllinae were included. Amalgamation of these taxa would lead to the creation of a very unwieldy taxon that would be very difficult to define using morphology. The more logical alternative is to reject sect. Oxystophyllum as part of Dendrobieae. It belongs to Eriinae, within Podochi- ., leae. The possession of several morphological characters that are putative synapomorphies and phylogenetic inferences based on ITS (Fig. 1, 2) leads to the proposal that Oxystophyllum be recognized at generic rank within the subtribe Eriinae (Podochileae), rather than in Dendrobieae (Clements 2003). The description of Eriinae accordingly has been modified to include: pollinia, eight (rarely four) in two sets of four (two), with caudicles.

STATUS OF SCHLECHTER'S CLASSIFICATION OF DENDROBIINAE
With the recognition of Oxystophyllum, treated by Schlechter as a section of Dendrobium (subgen. Xerobium), further consideration of the infrageneric status of the tribe is required. Table 2 shows Schlechter's system of classification of Dendrobiinae, in which genera are spilt into two groups based on possession of either four or eight pollinia. This study concentrates only on those taxa with four pollinia, the three sections in Cadetia, and the 41 sections in Dendrobium. The remaining genus, Pseuderia, for which no material was available for analysis, has recently been shown to belong to Podochilinae (Yukawa et al. 1996) and will not be considered further here. A comparison between results generated by these ITS analyses and Schlechter's classification demonstrates inconsistency with the ITS phylogenetic estimates (Table 2).
A more detailed examination of these results demonstrates the nature of this conflict. For example, Schlechter's concept of Dendrobium subgen. Athecebium included 12 sections. Results based on analysis of ITS sequence data support sects. Desmotrichum (= Flickingeria), Microphynathe, and Diplocaulobium (with sect. Goniobulbon embedded) as a strongly supported clade within Grastidiinae; sect. Bolbidium is embedded within subgen. Rhopalobium (Dendrobiinae); sect. Euphlebium is paraphyletic within Grastidiinae; sect. Rhizobium is part of Grastidiinae; sect. Sarcopodium ( = Epigeneium) is sister to all other Dendrobieae; sects. Dendrocoryne and Latourea are both paraphyletic within Grastidiinae; sect. Inobulbum is monophyletic within a branch of the basal dichotomy within Grastidiinae; and Callista represents a well-supported subgroup within Dendrobiinae. On this basis Schlechter's system of classification is demonstrably inconsistent with the phylogeny. A second example, Sarcopodium was isolated from Dendrobium based on the possession of a single-jointed pseudobulb on a creeping rhizome from which arises a multiflowered inflorescence. Despite having habit and floral characteristics more akin to Coelogyne and Bulbophyllum, respectively, Schlechter was strongly opposed to the recognition of Sarcopodium at generic rank as proposed by Lindley (1850), Rolfe (1910), and Kraenzlin (1910). ITS data demonstrate significant support for Sarcopodium (as Epigeneium and with 96% bootstrap support), isolated from all other taxa in Dendrobium subgen. Athecebium. These results correlate strongly with those based on rbcL and matK sequences (Yukawa et al. 1993(Yukawa et al. , 1996(Yukawa et al. , 2000, where in one case (Yukawa et aL 1996) there was strong support (96% bootstrap support) for the isolation of these taxa. Because of the unavailability of the name Sarcopodium LindL, a later homonym of Sarcopodium Ehrenberg ex A. T. Brongniart (Hyphomycetes), Hawkes (1956) proposed the alternate name Katherinea for this orchid genus, overlooking an earlier published legitimate name, Epigeneium. The name Epigeneium has since been taken up and applied, using a broader circumscription (Summerhayes 1957;Seidenfaden 1980;Garay and Romero-Gonzalez 1999); although, some authors have preferred to maintain both Epigeneium and Katherinea (Balakrishnan and Chowdhury 1966). To account for its position outside of the two major groups within Dendrobieae, as well as the possession of a combination of morphological synapomorphies, this distinctive taxon has been recognized as a separate subtribe, Epigeneiinae, within Dendrobieae (Clements 2003).
Focusing on the Australasian clade (Grastidiinae ), the presence of the long accepted and established genera, Cadetia, Diplocaulobium, and Flickingeria, deep within the well-supported clade (85% bootstrap support) further supports the traditional circumscription of Dendrobium as nonmonophyletic (Fig. 2). Again, there is a very strong correlation with the results generated, based on rbcL and matK (Yukawa et aL 1993;Yukawa and Uehara 1996) or the combined results using matK and ITS sequence data ) based on the same or a similar group of species. Cadetia was first described by Gaudichaud (1829), but it was Schlechter (1912) who firmly established it as distinct from Dendrobium, within Dendrobiinae. This has been the typical interpretation since. Schlechter characterized Cadetia by the possession of the Pleurothallis-habit, formation of a spur through cohesion anteriorly in the lower half of the lateral sepals, and papillae in front of the column and labellum. Possession of these characters, coupled with very strong molecular support (100% bootstrap support), add weight to the continued recognition of Cadetia. The genus Diplocaulobium (Kraenzlin 191 0), represented in Schlechter's system of classification by sects. Goniobulbum and Diplocaulobium, has gradually gained acceptance as being separate from Dendrobium. This is based on possession of closely spaced single internode pseudobulbs, mostly solitary inflorescences, pedicels arising from a conspicuous bract at the base of a single terminal leaf, flowers ephemeral (lasting one day, rarely longer), with distinct mentum and articulate labellum (Lavarack et al. 2000). With D. glabrum, which was placed in sect. Goniobulbum by Schlechter, deeply embedded within a monophyletic Diplocaulobium, recognition of sect. Goniobulbum is superfluous.
Flickingeria, originally described under the illegitimate name Desmotrichum (Blume 1825), has similarly regained acceptance as a genus, having in the interim been treated as a section of Dendrobium (Schlechter 1912). The genus Flickingeria is characterized by irregularly branched, creeping rhizomes, terminating in an erect pseudobulbous internode with a terminal nonsheathing leaf; single or multiflowered inflorescences, terminal or subterminal to the pseudobulb, adaxial or abaxial to leaf, and covered by persistent sheathing bracts; flowers ephemeral, the labellum midlobe fimbriate plicate, or bilobate (Seidenfaden 1980). Since the type species for Dendrobium is within the Asian clade, one option is to accept all species in the Australasian clade as representatives of genera other than Dendrobium. Alternatively, Cadetia, Diplocaulobium, and Flickingeria could be reincorporated into a broadly defined Dendrobium that includes both the Australian and Asian clades. However, there are several problems with this later option. Since there is only very weak support from ITS sequence analysis for monophyly of the Australasian and Asian clades, the inclusion of these well-established and morphologically well-defined genera into an extremely broad circumscription of Dendrobium is not advisable. Such a move would also be retrogressive and cause much nomenclatural confusion. An alternative is to treat the Australasian and Asian clades as two separate genera, but like the expansion of Dendrobium, this seems problematic. Such a proposal requires, in the case VOLUME 22 Molecular Systematics of Dendrobieae 475  ceme with terminal, inward-facing flowers; and, ephemeral flowers (Clements and Jones 1997). These characters, or combinations thereof, are absent from the remaining species in the Australasian clade so the aggregation of all species into a single genus would create far more confusion than presently exists. Moreover, it would be extremely difficult to find morphological features to characterize such a genus. Yukawa et al. (1996Yukawa et al. ( , 2000, Yukawa and Uehara (1996), and  arrived at similar conclusions with respect to the integrity and maintenance of representatives within the Australasian clade, including the genera Cadetia, Diplocaulobium, and Flickingeria; although, they refrained from formally recognizing any other genera in the clade. On the basis of molecular phylogenetic inference and morphological considerations, the taxonomic and nomenclatural changes have been proposed for the recognition of three subtribes within Dendrobiinae, viz. Dendrobiinae, Epigeneiinae, and Grastidiinae (Clements 2003).
If we accept these four-Cadetia, Diplocaulobium, Flickingeria, and include Grastidium-as genera, for which the arguments are very compelling, then consideration must also be given to the recognition at similar rank of all other major clades within the tribe supported by the ITS data (Fig. 2). Historically, Cadetia, Diplocaulobium, Flickingeria, and Grastidium are not the only generic names available for representatives within Dendrobieae. In fact, the name Dendrobium was conserved in favor of two earlier legitimate names, Callista and Ceraia (Loureiero 1790). Thereafter, the genera Aporum, Pedilonum (Blume 1825), Cadetia (Gaudicaud 1829), Thelychiton (Endlicher 1833), Froscula, Tropilis (Rafinesque 1837-1838), Aclinia (Griffith 1851), Coelandria (FitzGerald 1882), and Sayeria and Inobulbon (Kraenzlin 1910) were all described on the basis of perceived morphological differences from Dendrobium. These genera are scattered throughout the two major subtribes within Dendrobieae. Treatment of the monophyletic Cadetia, Diplocaulobium, Flickingeria, and Grastidium within Grastidiinae at the generic rank, necessitates the recognition of all the other sister groups at a similar rank (Fig. 2). Accordingly, the taxonomic and systematic treatments of most taxa within the Australasian clade are already well advanced (see Appendix 1). A full account of the phylogeny of these taxa based on the results of molecular analyses is in preparation.
The Asian clade (Dendrobiinae in the strict sense, Clements 2003), is monophyletic (94% bootstrap support in Fig.  1, but only 52% in Fig. 2), which correlates with the results of Yukawa et al. (1996Yukawa et al. ( , 2000Yukawa et al. ( , 2001 and Wongsawad et al. (2001). The presence of Dendrobium moniliforme, the type species for Dendrobium (Holttum et al. 1979) automatically ties that clade to the name Dendrobiinae. Apart from the branching order of the clade, these results reveal the strongest supported groups within the tree. Conversely, there is much weaker support along the main spine of the tree, within and between the three main taxa, making up Dendrobieae, excluding Bulbophyllinae and Oxystophyllum. In addition, those taxa not adequately covered include species in the numbered clades 1, 2, 3, 4, and 6 ( Fig. 3). The situation is mirrored in Grastidiinae where the strongest clades are again internal with little divergence and supported along the spine of the clade.
These results also reveal that many traditionally recog-nized sections are not monophyletic. This was clearly demonstrated in the recently published paper (Clements 2003), focusing on Dendrobium sect. Pedilonum. In that study, sects. Calcarifera, Dendrobium, Oxyglossum, Pedilonum, and Rhopalanthe were all shown to be either para-or polyphyletic (Fig. 3). Results presented here reaffirm those observations as well as highlight others. In addition to those listed above, sect. Formosae is paraphyletic and sect. Dendrobium is polyphyletic in their present circumscription. A similar situation exists within Grastidiinae. For example, sect. Latourea is paraphyletic with D. coxii (formerly of sect. Euphlebium) within it, which also renders sect. Euphlebium paraphyletic. The relationship of D. spectabile, traditionally treated within sect. Latourea (Cribb 1983), is also problematic. In the present studies, D. spectabile is placed in the clade containing Diplocaulobium. Conversely, these results provide evidence as to why many distinctive groups found particularly in Australia, New Zealand, and New Caledonia have been recognized as separate genera. It is recognized that additional research is still needed to fully resolve the relationships and status between several of these taxa, the monophyly of still other taxa requires further resolution. For example, Dockrillia, Grastidium, Kinetochilus, and Thelychiton are all paraphyletic based on ITS sequences. Inclusion of sequence data from one or both plastid genes (matK and rbcL) as well as those from the ITS region is already underway and should help resolve these areas of ambiguity. Similar studies have already been undertaken by  and Wongsawad et al. (2001). Unfortunately, their results cannot be reanalyzed because the sequence data generated for use in those analyses has not been submitted to GenBank. Similar combined analyses using at least one plastid gene as well as the ITS sequence data, using increased species sampling, are underway and should help resolve the identified problem areas and lead to an improved understanding of tribe Dendrobieae. In the meantime, the classification of Dendrobieae, as it currently stands based on the results of this and previously published research is outlined in the Appendix 1.