Molecular phylogenetics and character evolution of morphologically diverse groups, Dendrobium section Dendrobium and allies

The genus Dendrobium, one of the largest genera in Orchidaceae, exhibits enormous vegetative diversification. Such a situation has hindered the establishment of consistent classification systems. To clarify phylogenetic relationships in Dendrobium section Dendrobium and allied groups, we performed molecular phylogenetic analyses of 210 taxa. The results showed that many sections are not monophyletic. Most of the morphological characters that have been believed to reflect phylogenetic relationships are, in fact, the results of convergence. Consequently, recircumscription of the infrageneric classification is required.

Abstract

It is always difficult to construct coherent classification systems for plant lineages having diverse morphological characters. The genus Dendrobium, one of the largest genera in the Orchidaceae, includes ∼1100 species, and enormous morphological diversification has hindered the establishment of consistent classification systems covering all major groups of this genus. Given the particular importance of species in Dendrobium section Dendrobium and allied groups as floriculture and crude drug genetic resources, there is an urgent need to establish a stable classification system. To clarify phylogenetic relationships in Dendrobium section Dendrobium and allied groups, we analysed the macromolecular characters of the group. Phylogenetic analyses of 210 taxa of Dendrobium were conducted on DNA sequences of internal transcribed spacer (ITS) regions of 18S–26S nuclear ribosomal DNA and the maturase-coding gene (matK) located in an intron of the plastid gene trnK using maximum parsimony and Bayesian methods. The parsimony and Bayesian analyses revealed 13 distinct clades in the group comprising section Dendrobium and its allied groups. Results also showed paraphyly or polyphyly of sections Amblyanthus, Aporum, Breviflores, Calcarifera, Crumenata, Dendrobium, Densiflora, Distichophyllae, Dolichocentrum, Holochrysa, Oxyglossum and Pedilonum. On the other hand, the monophyly of section Stachyobium was well supported. It was found that many of the morphological characters that have been believed to reflect phylogenetic relationships are, in fact, the result of convergence. As such, many of the sections that have been recognized up to this point were found to not be monophyletic, so recircumscription of sections is required.

Introduction

It is always difficult to construct coherent classification systems for plant lineages having diverse morphological characters. Dendrobium Sw. (Orchidaceae) represents such difficult groups and so far has been established in many alternative systems (e.g. Lindley 1830; Bentham and Hooker 1883; Kränzlin 1910; Schlechter 1912; Brieger 1981; Clements 2003; Wood 2006; Schuiteman 2011). This genus is one of the largest orchid genera, with around 1100 species (Wood 2006). The distribution range extends from Sri Lanka and India throughout tropical Asia and Oceania, north to Japan, east to Tahiti and south to New Zealand. Enormous diversification of the vegetative organs in accordance with habitat shifts and lack of accessory structures of the pollinarium, a cardinal character in orchid classification, has hindered establishment of consistent classification systems covering all major groups of this genus. Previous studies of systematics based on the morphological characteristics of the group were reviewed by Wood (2006). Given the limits to what can be understood of affinities using morphological characters, Yukawa et al. (1993) analysed the molecular phylogenetics of the subtribe Dendrobiinae (Lindley 1830), which includes the genus Dendrobium and putatively related genera Cadetia, Diplocaulobium, Flickingeria, Epigeneium and Pseuderia based on chloroplast DNA restriction site variation. This analysis resulted in presentation of the first probable phylogenetic relationship between members of this genus. Yukawa et al. (1993) demonstrated that Dendrobium is not monophyletic and comprises two major clades (Asian and Australasian clades: sensu Clements 2003). The Asian clade is predominantly diversified west of Weber's line, and the Australasian clade, containing genera Cadetia, Diplocaulobium and Flickingeria, is distributed mostly in Australasia and the Pacific Islands. Subsequent studies on representative members of subtribe Dendrobiinae (e.g. Yukawa et al. 1996, 1999, 2000; Yukawa 2001; Clements 2003, 2006; Schuiteman 2011) incorporated additional macromolecular markers and taxa in their analyses. In addition to providing further support for the above-mentioned phylogeny, these studies identified other infrageneric monophyletic groups.

Dendrobium section Dendrobium is one of the largest sections in the genus Dendrobium, comprising ∼60 species (Wood 2006) distributed across almost the entire geographical range of the genus, with the exception of Micronesia and Melanesia. A number of species are considered important as crude drug sources and are highly sought-after genetic resources with potential value in medicine (Takamiya et al. 2011, 2013). Yukawa et al. (1993) demonstrated that section Dendrobium is nested within the Asian clade. Wongsawad et al. (2001, 2005) analysed sequences for the maturase-coding gene (matK) located in the plastid genome and the internal transcribed spacer regions (ITS) of the nuclear ribosomal DNA of 78 Asian clade species including 35 of section Dendrobium (Wongsawad et al. 2001) and 93 Asian clade species including 42 members of section Dendrobium (Wongsawad et al. 2005). Based on these analyses, Wongsawad et al. demonstrated that section Dendrobium is not monophyletic; that its core clade includes species of sections Breviflores, Densiflora, Holochrysa and Stuposa; and that sections Amblyanthus, Breviflores, Densiflora, Formosae, Holochrysa, Oxyglossum and Pedilonum are not monophyletic. These relationships were confirmed by Xiang et al. (2013). However, given that these studies did not include several species of section Dendrobium and only included a small number of species in sections Aporum, Calcarifera, Calyptrochilus, Crumenata, Distichophyllae, Oxyglossum, Pedilonum, Platycaulon, Stachyobium and Stuposa, which are likely to be closely related to section Dendrobium, our understanding of the relationships between section Dendrobium and other groups within the Asian clade remains incomplete. In this study, we conducted comprehensive phylogenetic analyses of representative species in the Asian clade using the ITS and matK regions to clarify the relationships and the taxonomic position of section Dendrobium.

Methods

Plant materials

The samples for analysis consisted of 210 Asian clade species (214 samples), including 56 species belonging to section Dendrobium. As an outgroup, we chose 10 species of the Australasian clade, based on the results of Yukawa (2001). Plant materials were collected from the living collection of Tsukuba Botanical Garden, National Museum of Nature and Science. All voucher specimens were deposited at the Herbarium, National Museum of Nature and Science (TNS). Voucher information and GenBank accession numbers for all ITS and matK sequences used in this study are listed in Appendix 1. We adopted an infrageneric classification based on that proposed by Schlechter (1912) and modified by Wood (2006).

DNA extraction and sequencing

Genomic DNA was isolated from fresh leaves or flowers by a DNeasy Plant Mini Kit (Qiagen, Hamburg, Germany) following the manufacturer's instructions, and was used as a template for PCR amplification. ITS1-5.8S-ITS2 regions were amplified and sequenced as described by Sun et al. (1994) and Hidayat et al. (2005). The amplification reactions included GC buffer I or II and LA Taq DNA polymerase (Takara Bio, Shiga, Japan). The PCR profile consisted of an initial 2-min premelt at 94 °C and 30 cycles of 50 s at 94 °C (denaturation), 1 min at 60 °C (annealing) and 30 s at 72 °C (extension), followed by a final 7-min extension at 72 °C. matK regions were amplified and sequenced as described by Hidayat et al. (2005). The amplification reactions included Ex Taq buffer and Ex Taq DNA polymerase (Takara Bio). The PCR profile consisted of an initial 5-min premelt at 94 °C and 30 cycles of 30 s at 94 °C (denaturation), 30 s at 53 °C (annealing) and 3 min at 72 °C (extension), followed by a final 7-min extension at 72 °C.

The PCR products were cleaned using a Montage PCR centrifugal filter device (Millipore, Billerica, MA, USA) and then sequenced in both forward and reverse directions using primers described by Hidayat et al. (2005). Sequences were obtained with an ABI PRISM 377 sequencer (Applied Biosystems, Foster City, CA, USA) following the manufacturer's instructions for auto-cycle sequencing reactions.

Molecular phylogenetic analysis

Except for ITS and matK sequences of Dendrobium ovatum obtained from GenBank (accession ID for ITS: HM054721; matK: HM055325), all sequence data were obtained by our own analyses. Two hundred and twenty-four DNA sequences were aligned with ClustalW software. A phylogenetic analysis based on the maximum parsimony (MP) was performed using PAUP* version 4.0b10 (Swofford 2002) for three data sets: ITS, matK and a combination of the two. Gaps were treated as missing data. All characters were equally weighted and unordered (Fitch 1971). Each data set was analysed by a heuristic search method with tree bisection–reconnection branch swapping and the MULTREES option on 100 replications of random addition sequence with the stepwise addition option, and each replicate after 1 × 10⁶ rearrangements was assessed. Bootstrap support values were obtained from 1000 replicates using ‘fast’ stepwise addition. Although the fast stepwise addition analyses are expected to provide a lower support value than obtained when comprehensive branch-swapping is performed (DeBry and Olmstead 2000; Mort et al. 2000; Barkman et al. 2004), we used this option because bootstrap analyses with a full heuristic search method were not computationally feasible with a large data set. Bootstrap percentages (BPs) of 50–74 were defined as weak, 75–84 as moderate and 85–100 as strong, as in Chase et al. (2000) and Kim et al. (2010). The number of steps, consistency index (CI) and retention index (RI) were calculated with one of the most parsimonious trees in each analysis using the TREE SCORES command in PAUP*. To test congruence among data partitions, the incongruence length difference test (Farris et al. 1994, 1995), also designated the partition homogeneity test in PAUP*, was employed to measure character conflicts under a parsimony framework among data sets using 100 heuristic search replications.

The same data sets were analysed by Bayesian analysis using MrModeltest ver. 2.3 (Nylander 2004) to determine the sequence evolution model that best described the data. The GTR + I + G model was chosen for ITS, matK and the combined data by hierarchical likelihood ratio tests. The chosen model was used to perform a Bayesian analysis using the program MrBayes ver. 3.1.2 (Huelsenbeck and Ronquist 2001; Ronquist and Huelsenbeck 2003). For analysis, two simultaneous runs of four chains each were carried out with the Markov chain Monte Carlo algorithm for 8 000 000 generations, sampling every 100 generations. Data from the first 20 000 generations were discarded as the ‘burn-in’ period. The 50 % majority rule consensus phylogeny and posterior probability (PP) of nodes were calculated from the remaining samples. Clades with PP ≥95 were regarded as strongly supported (Martínez-Azorín et al. 2011).

Ancestral state reconstruction of morphological characters

Among various morphological characters, stem shape and leaf number were examined with ancestral state reconstruction analysis due to their complex pattern of character states even among closely related taxa. Character states for each taxon were obtained from the literature (Seidenfaden 1985; Wood 2006) and observation by the authors. A parsimony reconstruction was performed with the ‘Unordered’ model in Mesquite 2.75 (Maddison and Maddison 2011). The Bayesian tree based on the combined sequence data (ITS and matK) was used as the standard tree topology (see Fig. 1).

Figure 1.

Consensus phylogram of 210 taxa of Dendrobium section Dendrobium and allied groups obtained from 96 596 Bayesian trees from the combined ITS and matK sequence data set. Values below and above branches indicate percentage bootstrap values from maximum parsimony analysis and Bayesian posterior probabilities, respectively.

Results

Parsimony analyses of ITS, matK and combined sequences of the two

Table 1 shows statistics from ITS, matK and combined sequences. We compared topologies between strict consensus trees of ITS and matK data sets [see Supporting Information]. Visual inspection of the topologies between the two regions did not show significant incongruence, although an incongruence length difference test did not support congruence (P value = 0.01). Inconsistent topologies where bootstrap support is weak can be ignored as a soft incongruence (Johnson and Soltis 1998). Moreover, strongly supported clades by matK were generally strongly supported by ITS except for six inconsistent regions [see Supporting Information]. Since analyses of combined data sets provide more resolution and internal support for relationships than individual data sets (Soltis et al. 1998), we combined ITS and matK data for further analysis.

Table 1.

Characteristics of DNA datasets used in this study.

Characteristics	ITS	matK	Combined data
Number of taxa	224	224	224
Total number of characters	906	1630	2536
Number of constant characters	303	1092	1395
Number of parsimony uninformative characters	102	279	381
Number of parsimony informative characters	501	259	760
Tree length	5522	1092	6692
Consistency index (CI)	0.226	0.6016	0.2847
Retention index (RI)	0.715	0.7915	0.7198
Number of trees	140	4638	103
Bayesian model of evolution	GTR + I + G	GTR + I + G	GTR + I + G

In the majority of cases, BPs for trees based on the combined data set were higher than those obtained from analysis based on ITS or matK alone. For example, six clades supported by BP at <85 % in both genetic regions when based on either individual data set had BP >85 % in the combined case.

Bayesian analyses of ITS, matK and combined sequences of the two

Comparing the Bayesian trees and strict consensus tree of the most parsimonious trees constructed from ITS and matK sequences, we found that in terms of the differences in results based on the sequence used, the clades of strict consensus trees that were strongly supported by bootstrap value in both ITS and matK trees were also strongly supported by the PP value [see Supporting Information]. However, when we compared the strict consensus tree of the most parsimonious trees and the Bayesian trees with the ITS sequence, we found phylogenetic relationships in four parts where they were not strongly supported by PP value, even if they were by BP value. Specifically, these were the clade comprising Dendrobium amoenum, D. gratiosissimum, D. wardianum, D. lituiflorum, D. transparens, D. luteolum, D. ruckeri and D. tetrachromum (BP86, PP90); the clade comprising D. aff. densiflorum, D. densiflorum, D. cf. griffithianum and D. griffithianum (BP87, PP84); the clade comprising D. farmeri and D. palpebrae (BP87, PP75); and the clade comprising D. speckmaieri and D. ypsilon (BP89, PP92). That said, the PP values for these regions were still fairly high. Furthermore, as in the case for MP analysis, Bayesian analysis revealed that there are no inconsistencies between ITS and matK data sets, so the two regions were combined for subsequent analysis.

The combined tree obtained from a majority-rule consensus of 96 596 trees produced by two runs of the Markov chain Monte Carlo algorithm is presented in Fig. 1. With the exception of one clade consisting of D. aff. densiflorum, D. densiflorum, D. cf. griffithianum and D. griffithianum (BP91, PP92), all of the clades that were strongly supported in bootstrap tests based on MP analysis were also strongly supported by PP values in Bayesian analysis. In this study, major clades whose BP values from MP analysis were >50 % and those whose PP values from Bayesian analysis were >95 %—i.e. clades at least weakly supported by bootstrap values and strongly supported by PP values—were assigned letters A through M. Clade A comprises the majority of sections Dendrobium and Holochrysa, as well as sections Breviflores and Stuposa, and Dendrobium chrysotoxum in section Densiflora. Clade B comprises Dendrobium albosanguineum in section Dendrobium and Dendrobium braianense and D. capillipes in section Holochrysa. Clade C comprises the majority of section Densiflora and section Amblyanthus. Clade D consists of some members of sections Pedilonum and Calcarifera along with sections Dolichocentrum, Oxyglossum and Calyptrochilus. Clade E comprises sections Aporum, Crumenata and Bolbodium. Clade F comprises section Platycaulon and some members of sections Pedilonum and Calcarifera. Clade G comprises D. cf. jenkinsii, D. jenkinsii and D. lindleyi in section Densiflora. Clade H comprises section Conostalix and the majority of section Distichophyllae. Clade I comprises the majority of section Formosae. Clade J comprises D. dearei, D. parthenium, D. sanderae and D. schuetzei in section Formosae. Clade K comprises section Stachyobium and Dendrobium oligophyllum in section Distichophyllae. Clade L comprises section Herbaceae. Clade M comprises section Fytchianthe. No clear relationship to other clades was indicated for Dendrobium gibsonii in section Holochrysa, D. senile in section Dendrobium, or D. jerdonianum and D. trigonopus in section Formosae.

Ancestral state reconstruction of morphological characters

In the analysis of stem shape morphology, the terete, non-succulent stem was supported as the plesiomorphic state in the Asian clade (Fig. 2A). The other stem character states, namely, pseudobulbous, heteroblastic, succulent stem; clavate, non-heteroblastic, succulent stem; basally bulged stem; entirely flattened stem; and thin, wiry stem, were inferred to evolve twice, more than 14 times, twice, three times, and more than seven times, respectively. As for the evolution of leaf number, a character state bearing more than five leaves was suggested to be plesiomorphic in the Asian clade (Fig. 2B). The other character states of leaf number, namely, between three and five; two; and one, were inferred to evolve more than 10 times, once, and once, respectively.

Figure 2.

Ancestral state reconstruction of morphological characters ((A) stem shape; (B) leaf number) of 210 taxa of Dendrobium section Dendrobium and allied groups. All analyses were based on the parsimony reconstruction.

Discussion

Phylogenetic relationships at the section level

Section Dendrobium is divided into three subclades. The first is the core clade, which includes the type species Dendrobium moniliforme (Clade A). The second is D. albosanguineum, included in Clade B, and the final subclade is D. senile, with a phylogenetic position that is unresolved. The possibility is suggested (PP72) that D. senile and D. gibsonii in section Holochrysa may form a clade. Furthermore, while there is a support of PP100 suggesting that these two species are sister groups to Clade A, due to the weak bootstrap support in MP analysis (BP < 50), they were not included in Clade A for the purposes of this study. In addition, Clade A also contains all species in section Breviflores, D. chrysotoxum in section Densiflora, all species in section Stuposa, and the majority of species in section Holochrysa, thus rendering section Dendrobium polyphyletic and paraphyletic. These results support the findings of Wongsawad et al. (2001, 2005) and Xiang et al. (2013).

Five species in section Breviflores (Dendrobium aduncum, D. bifurcatum, D. hercoglossum, D. linguella and D. trantuanii) are divided into three subclades within Clade A, making them a polyphyletic group. These results are consistent with the findings of Wongsawad et al. (2005) and Xiang et al. (2013). Two species in section Stuposa (Dendrobium stuposum and D. praecinctum) represent a monophyletic group within Clade A (BP100, PP100), supporting the findings of Wongsawad et al. (2005). Section Holochrysa was found to not be monophyletic, and its eight species (Dendrobium atavus, D. chryseum, D. dixanthum, D. fimbriatum, D. henryi, D. lohohense, D. moschatum and D. pulchellum) in Clade A are divided into six lineages. Further, a subclade comprising D. braianense and D. capillipes is placed within Clade B and the phylogenetic position of D. gibsonii is unresolved. These results are consistent with the analyses by Yukawa (2001), Wongsawad et al. (2001, 2005) and Xiang et al. (2013).

Section Densiflora was also found to not be monophyletic but rather to consist of three groups comprising the majority of Clade C, including the type species D. densiflorum, D. chrysotoxum belonging to Clade A, and Clade G, which is made up of D. jenkinsii, D. cf. jenkinsii and D. lindleyi. The analyses by Yukawa (2001), Wongsawad et al. (2001, 2005) and Xiang et al. (2013) also demonstrated the polyphyly of section Densiflora. Section Amblyanthus was found to nest within Clade C. Four species of section Amblyanthus were found to be divided into two subclades within Clade C, and were polyphyletic, supporting the results of Yukawa (2001) and Wongsawad et al. (2005).

Section Pedilonum was found to not be monophyletic, its members forming eight polyphyletic groups within Clade D and three polyphyletic groups within Clade F. The type species of section Pedilonum, D. secundum, was included in Clade F. Our results were consistent with analyses by Yukawa (2001), Wongsawad et al. (2001, 2005) and Clements (2003, 2006), which also revealed section Pedilonum to not be monophyletic.

Section Calcarifera represents another non-monophyletic group, which is divided into three subclades within Clade D and six subclades within Clade F. Consistent with our results, Clements (2003, 2006) also found that the section was not monophyletic and that its members formed two distinct groups. It is clear that the type species Dendrobium pedicellatum, which was not investigated in this study, belongs to Clade F because a morphologically related species, Dendrobium mutabile, is placed in this clade.

Clade D was shown to also include sections Dolichocentrum, Oxyglossum and Calyptrochilus. Section Dolichocentrum was found to be polyphyletic, with its three species splitting into a lineage represented by Dendrobium furcatum and a subclade (BP100, PP100) made up of Dendrobium auriculatum and D. papilio.

Section Oxyglossum was also found to not be monophyletic. Its six constituent species were divided into two groups, one comprising Dendrobium cyanocentrum, D. masarangense subsp. theionanthum and D. violaceum (BP99, PP100) and the other comprising D. cuthbertsonii, D. laevifolium and D. prasinum, along with D. bracteosum, a member of section Pedilonum (BP73, PP100). Consistent with our results, Clements (2003, 2006) also found that section Oxyglossum was not monophyletic and that its members split into two groups.

Our results suggested that section Calyptrochilus forms a polyphyletic group made up of three subclades: the first comprising Dendrobium aphanochilum and D. glomeratum (BP98, PP100); the second comprising D. lawesii, D. codonosepalum, D. trichostomum, D. mohlianum and D. melinanthum (BP88, PP100); and the third comprising D. erosum.

Section Aporum was found to be polyphyletic, nesting within Clade E along with sections Crumenata and Bolbodium. The members form two subclades: the first comprising nine species (BP83, PP100) and the second comprising Dendrobium calceolum. The monophyly of the first subclade was suggested by Yukawa (2001) and Wongsawad et al. (2001, 2005).

Section Crumenata was also found to nest within Clade E and to not be monophyletic. Its species are divided into two subclades: the first including section Bolbodium and parts of section Aporum (BP92, PP100) and the second representing the earliest diverging clade within Clade E, comprising Dendrobium ephemerum. Analyses by Yukawa (2001) and Clements (2003, 2006) also suggested the first of these subclades to be paraphyletic.

Section Distichophyllae was found to not be monophyletic, with its members splitting into a core clade (Clade H) containing the type species of the section, Dendrobium uniflorum, and the earliest diverging clade within Clade K represented by D. oligophyllum, which is likely a sister group to section Stachyobium. Further, in Clade H, two species of section Conostalix, Dendrobium lobbii and D. pachyglossum, were nested within members of section Distichophyllae. Section Conostalix is probably monophyletic (BP53, PP100).

Members of section Formosae were scattered into four lineages: Clade I, Clade J and two unplaced species, D. jerdonianum and D. trigonopus. Clade I, which contains the type species of the section, D. formosum, comprises only species of section Formosae (BP100, PP100). Clade J similarly comprises only species of section Formosae (BP100, PP100). While the possibility that these four groups represent a monophyletic lineage cannot be ruled out, it is likely that they are polyphyletic. Analyses by Wongsawad et al. (2005) and Clements (2006) also identified two clades within Section Formosae corresponding to Clades I and J.

There is a strong possibility that section Stachyobium forms a monophyletic group within Clade K (BP93, PP100). Consistent with our results, monophyly of section Stachyobium was also suggested by analyses by Yukawa (2001), Wongsawad et al. (2001, 2005), Clements (2006) and Xiang et al. (2013). Furthermore, it is evident that section Stachyobium is a sister group to D. oligophyllum in section Distichophyllae (BP99, PP100).

Section Herbaceae comprises only two species, Dendrobium herbaceum and D. parcum. We found that this section is clearly monophyletic because the two species constitute Clade L (BP96, PP100). Similarly, it is highly probable that section Fytchianthe represents a monophyletic lineage because Dendrobium fytchianum and D. ovatum, members of this section, constitute Clade M (BP100, PP100). While not strongly supported in statistical terms (BP < 50, PP59), Clade M may represent the earliest divergent clade within the Asian clade. Similarly supported by our results, analyses by Wongsawad et al. (2001, 2005) also indicate the possibility that section Fytchianthe is the earliest divergent clade within the Asian clade. Our analyses further indicate that sections Herbaceae and/or Stachyobium may represent the second-earliest divergent clade within the Asian clade after section Fytchianthe.

Evolution of morphological characters

Vegetative stems

The vast majority of species belonging to Clades A, D, F, H, I, J, L and M, and several species in Clades C and K have terete, non-succulent vegetative stems. Results of ancestral state reconstruction analysis showed that this character state represents a plesiomorphy of the Asian clade (Fig. 2A).

Species with a pseudobulbous, heteroblastic, succulent stem in which from one to a few internodes on the upper portions of vegetative stems thicken to become pseudobulbs, account for all species in section Bolbodium and some members of sections Densiflora. Among species analysed in this study, the heteroblastic stem is exhibited by Dendrobium pachyphyllum (section Bolbodium) in Clade E, D. lindleyi, D. jenkinsii and D. cf. jenkinsii (all in section Densiflora) in Clade G. Results of ancestral state reconstruction analysis showed that this character state likely evolved twice in the Asian clade (Fig. 2A).

Dendrobium brymerianum, D. chrysocrepis, D. harveyanum (both in section Dendrobium) and D. chrysotoxum (section Densiflora) in Clade A, D. albosanguineum (section Dendrobium) and D. braianense and D. capillipes (both in section Holochrysa) in Clade B, D. cf. griffithianum, D. farmeri, D. griffithianum, D. guibertii and D. palpebrae (all in section Densiflora) in Clade C, all species of section Oxyglossum in Clade D, D. bellatulum (section Formosae) in Clade I, D. delacourii, D. diodon, D. gregulus, D. porphyrochilum, D. laterale and Dendrobium sp. (all in section Stachyobium) in Clade K and D. trigonopus (section Formosae), an unplaced taxon, have a clavate, non-heteroblastic, succulent stem. Results of ancestral state reconstruction analysis indicated that this character state evolved more than 14 times in the Asian clade (Fig. 2A).

All species of section Crumenata, and Dendrobium hancockii and D. cf. hancockii (both in section Dendrobium) exhibit a basally bulged stem characterized by particular thickening of several internodes positioned on lower portions of vegetative stems. Results of ancestral state reconstruction analysis showed that this character state represents an apomorphy, resulting from two independent evolutionary events within the Asian clade (Fig. 2A).

The stems of the majority of members of section Calcarifera have elliptical cross-sections, while the stems of all members of section Platycaulon are entirely flattened. A few species belonging to sections Breviflores and Densiflora also have flattened stems. Among species analysed in this study, in addition to species of section Platycaulon (Clade F), D. trantuanii (section Breviflores) in Clade A and D. sulcatum (section Densiflora) in Clade C also had flattened stems. Results of ancestral state reconstruction analysis showed that this character state evolved three times in the Asian clade (Fig. 2A).

A further apomorphy of the vegetative stem is thin, wiry stems, observed in all species of sections Aporum and Dolichocentrum and a few species of sections Dendrobium, Calcarifera, Conostalix and Pedilonum. Among species investigated in this study, in addition to species of sections Aporum and Dolichocentrum, thin and wiry stems were shared by Dendrobium leptocladum (section Dendrobium) in Clade A, Dendrobium lancifolium (section Calcarifera) and Dendrobium aff. ramosii (section Pedilonum) in Clade D and D. lobbii (section Conostalix) in Clade H. Results of ancestral state reconstruction analysis suggested that this character state evolved more than seven times in the Asian clade (Fig. 2A). Species belonging to section Aporum have succulent leaves, while other species with thin and wiry stems have fleshy roots. This suggests that, in species with thin and wiry stems, the components responsible for water storage have shifted from the stems to the leaves or roots.

Shoot architecture

The majority of species in the genus Dendrobium exhibit an architecture ubiquitously observed in perennial herbs in which vegetative shoots emerge in a repeated sympodial branching pattern from nodes on basal portions of the vegetative stem and produce roots also at the stem base. However, all species of section Herbaceae and a few species of sections Dendrobium, Calcarifera, Dolichocentrum and Crumenata produce vegetative shoots in a sympodial branching pattern from nodes on the upper parts of the stem, forming ramified stems, whereby roots do not grow from basal portions of new shoots. Among the species examined in this study, in addition to species of section Herbaceae, this architecture was observed in Dendrobiuim falconeri and D. hancockii (both in section Dendrobium) in Clade A; D. chameleon (section Calcarifera), D. furcatum (section Dolichocentrum) and D. victoria-reginae (section Calcarifera) in Clade D; and D. junceum (section Crumenata) in Clade E. This character state is not present in Clade M, which likely represents the earliest divergent clade within the Asian clade. In addition, given that this character state is only observed in a small number of species in a few lineages of both Asian and Australasian clades (Wood 2006), it is likely that ramified stems represent an apomorphy for the Asian clade, resulting from at least four distinct evolutionary events.

Although reproductive shoots are produced by axillary branching from vegetative shoots in most Dendrobium species, all species of sections Stachyobium, Herbaceae and Fytchianthe have terminal buds of vegetative shoots that develop into reproductive shoots. In this study, such a terminal inflorescence was observed in all species belonging to Clades M, L and K, which likely represent the earliest divergent clades within the Asian clade. Dressler (1981) proposed that terminal inflorescence is historically primary, being found in the most primitive orchid. Given that members of the genus Epigeneium, which represents the earliest divergence within subtribe Dendrobiinae, also exhibit a terminal inflorescence, terminal inflorescences may represent a plesiomorphy for the Asian clade.

Roots

Given that almost all Asian clade species have roots with a smooth, shiny white surface and that the majority of the sister Australasian clade species also have smooth roots, it appears that smooth roots represent a plesiomorphy for the Asian clade. In contrast, a number of species belonging to Clade F (Dendrobium annae, D. cumulatum, D. hymenophyllum, D. intricatum, D. mutabile, D. panduriferum, D. profusum and D. sanguinolentum in section Calcarifera and D. platygastrium, D. speckmaieri and D. ypsilon in section Platycaulon) have verrucose roots. All members of section Brevisaccata in the Australasian clade and some species of the genus Epigeneium also exhibit this feature. Such roots certainly evolved three times in the subtribe Dendrobiinae. However, this rough surface feature has no known function, except for increasing the absorptive surface of the root (Wood 2006).

Leaves

While almost all species belonging to the Asian clade have conduplicate leaves with both adaxial and abaxial surfaces, all species in section Aporum and some species in section Crumenata have laterally flattened or terete leaves in which the adaxial surface has been reduced, leaving a unifacial leaf consisting only of the abaxial surface. This feature is accompanied by thickening of the mesophyll, and these character states represent a xerophytic adaptation. In fact, the species with unifacial leaves generally grow as epiphytes in dry habitats of tropical Asia (Yukawa and Uehara 1996). Among the species analysed in this study, all species in section Aporum and section Crumenata species Dendrobium acinaciforme, D. macfarlanei, D. goldfinchii, D. philippinense, D. junceum and D. aff. modestum in Clade E exhibit a unifacial leaf. Since evolution of a unifacial leaf occurred only in Clade E in the Asian clade, it represents an apomorphy resulting from one or a few evolutionary events.

Members of the genus Dendrobium exhibit two shedding patterns: a deciduous pattern, in which leaf life-span is shorter than 1 year and stems lose all leaves temporarily, and an evergreen pattern, in which leaf life-span is longer than 1 year and stems retain their leaves at all times. While the vast majority of species are evergreen, all species in sections Herbaceae, Fytchianthe and Stachyobium and some members in sections Dendrobium and Holochrysa are deciduous. Among species investigated in this study, in addition to species of sections Herbaceae, Fytchianthe and Stachyobium, section Dendrobium species D. aphyllum, D. crystallinum, D. pendulum, D. wardianum, D. amoenum, D. anosmum, D. bensoniae, D. crepidatum, D. devonianum, D. gratiosissimum, D. lituiflorum, D. parishii, D. primulinum, D. transparens, D. unicum and D. dixanthum (section Holochrysa) in Clade A, along with D. albosanguineum (section Dendrobium), D. capillipes (section Holochrysa) in Clade B and D. senile (section Dendrobium), with unresolved phylogenetic position, are all deciduous. Furthermore, while members of the Australasian clade as well as Epigeneium and Bulbophyllum, allied genera to Dendrobium, have evergreen leaves, given that species of sections Fytchianthe, Herbaceae and Stachyobium, which likely are the earliest divergent clades within the Asian clade, are deciduous, deciduousness may represent a plesiomorphy for the Asian clade.

Results of ancestral state reconstruction analysis showed that a stem with more than five leaves represents a plesiomorphy of the Asian clade and that reduction of leaf number likely evolved more than 12 times (Fig. 2B). As shown in Fig. 2A and B, the species with less than six leaves on each stem usually exhibit pseudobulbous, heteroblastic, succulent or clavate, non-heteroblastic, succulent stems except for D. aff. densiflorum, D. amabile, D. densiflorum, D. sulcatum and D. thyrsiflorum (all in section Densiflora) in Clade C. Furthermore, leaves of these species are either succulent or deciduous. Yukawa and Uehara (1996) and Wood (2006) suggested that these combinations of character states show adaptation to a xeric environment.

Schlechter (1912), in his infrageneric classification system of Dendrobium, emphasized the presence or absence of leaf sheaths as a cardinal diagnostic character at the subgenus level. In species with the above-mentioned succulent stems producing a small number of leaves in the upper part, leaf sheaths do not develop. However, in species that produce non-succulent stems, leaf sheaths do not develop in the uppermost leaves, but do develop in lower leaves. Consequently, reduction of the leaf sheath in the uppermost leaves is a common character state both in the succulent and non-succulent stem species, and it is not appropriate to use the presence or absence of leaf sheaths as a character for classification at least in the Asian clade.

Anatomical characters of vegetative organs

Yukawa and Uehara (1996) demonstrated that modifications in size and number of parenchymatous cells of vegetative stems substantially contributed to vegetative diversification in Dendrobium. This observation implies that a simple structural adjustment can result in a major modification of growth habit in this group. They also found that several anatomical characters associated with xeromorphy, such as a thick outer wall of epidermal cells on the stem surface and a thick sclerenchymatous cap on vascular bundles in the stem evolved in members of sections Aporum and Crumenata. In other words, these character states are likely to have evolved in the common ancestor of Clade E. Acquisition of xeromorphic anatomical characters in this clade may have facilitated evolution of unique vegetative characters suited to dry environments such as a succulent, laterally flattened leaf and a wiry or basally bulged stem.

Namba and Lin (1981a, b), Singh (1986) and Morris et al. (1996) investigated the anatomical characters of roots of Dendrobium. Morris et al. demonstrated that root characteristics including the number of layers in the velamina are not useful characters to use in determining sectional relationships.

Reproductive organs

All species of section Densiflora and most members of section Dendrobium have velvety lips with many papillae on the surface. Among species analysed in this study, all Clade A species with the exception of D. unicum and D. lamyaiae in section Dendrobium, all Clade B, C and G species, and D. gibsonii exhibit this character state. When Dendrobium flowers are stained with neutral red, which is taken up by osmophores, or odour-producing cells (Stern et al. 1986), the papillae of the above-listed species are strongly stained (Yukawa 1993). Consequently, it is likely that the papillae on the lip surface accumulate substances responsible for fragrance (Müller 1935). Observation of the near-ultraviolet (near-UV) reflectance of the flowers using near-UV reflectance photography reveals that the parts with papillae absorb near-UV, while other areas of the perianth lobes reflect near-UV, resulting in distinct high-contrast patterns (Yukawa 1993; Indsto and Weston 2000). Given that this combination of characteristics is typical of bee-pollinated flowers, we demonstrate that the dense packing of papillae on lip surfaces represents a character state that evolved in adaptation to bee pollination. The only Dendrobium species producing lips with dense papillae for which pollinators have been observed is D. anosmum, which two species of Apis are known to pollinate (Burkill 1919). Since lip surfaces with dense papillae are not observed in species of the Australasian clade or in Clades K, L and M, which likely represent the earliest divergent clades within the Asian clade, this character state was probably acquired after the divergence of the Asian clade. However, it is unclear how many evolutionary events have resulted in the character state, because phylogenetic relationships of Clades A, B, C and G have not been resolved.

Meanwhile, specialization has occurred in terms of flowering behaviour. Reproductive shoots of Dendrobium generally branch sympodially from vegetative shoots and disappear immediately after flowering. However, in all species in Clade E, namely, members of sections Aporum, Bolbodium and Crumenata, the reproductive shoots remain alive even after flowering and continue to flower repeatedly. There is a high probability that the repeated flowering behaviour was acquired by a common ancestor of Clade E. Further, all species of section Bolbodium and some species of sections Crumenata and Aporum in this clade exhibit gregarious flowering in which, after the flower has differentiated, its growth is suspended at the bud stage and all flowers begin to grow and then bloom at once when the temperature decreases to a certain level or when the difference between daytime and night-time temperatures narrows to a certain point (Seifriz 1923; Coster 1925; Gerlach 1992). This behaviour is observed, for example, in D. pachyphyllum (section Bolbodium), D. crumenatum (section Crumenata), D. ephemerum (section Crumenata), D. acerosum (section Aporum) and D. spatella (section Aporum). It is possible that acquisition of repeated flowering in the common ancestor of Clade E provided further character evolution that increased plant attractiveness to pollinators, because it enabled an individual plant to have numerous buds at once. Acquisition of repeated flowering thus may trigger the evolution of gregarious flowering.

Reappraisal of infrageneric classification

There is no single criterion, but rather, several options for deciding what level of monophyletic unit warrants designation as a genus. In order to fulfil the phylogenetic consistency and conservation of a widely accepted concept of the genus Dendrobium, Yukawa et al. (1993, 1996, 1999, 2000), Yukawa (2001), Burke et al. (2008), Adams (2011) and Schuiteman (2011) have recommended expanding the genus Dendrobium to include the genera Cadetia, Diplocaulobium and Flickingeria. In contrast, Clements (2003, 2006) subdivided Dendrobium and established many new genera to recover the monophyly of the genus. The advantages and disadvantages of these two approaches were discussed by Yukawa et al. (1999) and Schuiteman and Adams (2011), who concluded that the former approach has greater merit. Here, we follow the former approach and examine the infrageneric classification system of Dendrobium proposed by Wood (2006).

We demonstrated that a number of sections used by Wood (2006) are not monophyletic. In order to restore the monophyly of these groups, a revised system at the section level is necessary. Clades A through M defined in this study were all strongly supported, indicating that they are stable units whose monophyly will continue to be supported even if the sampling intensity and gene regions analysed were increased. Furthermore, given that considerable numbers of clades roughly correspond to traditionally used sections, it would seem reasonable to assign, for the most part, Clades A through M to the section level.

We found that in Clade A, section Dendrobium is a paraphyletic group comprising all species in sections Breviflores and Stuposa, most species in section Holochrysa and D. chrysotoxum in section Densiflora. If we recognize these heterogeneous elements at the section level, section Dendrobium also must be subdivided. As sections Breviflores, Stuposa, Holochrysa and D. chrysotoxum in Densiflora do not deviate from a broad concept of section Dendrobium, it would be appropriate to redefine section Dendrobium as the range of Clade A and to include sections Breviflores and Stuposa, and most species in section Holochrysa, and D. chrysotoxum of section Densiflora in section Dendrobium. Schuiteman (2011) also suggested that section Breviflores is polyphyletic and should be included in section Dendrobium. Further, Xiang et al. (2013) proposed to subsume these three sections and D. chrysotoxum into section Dendrobium. Our analyses indicated that D. gibsonii (section Holochrysa) and D. senile (section Dendrobium) may form a monophyletic group with Clade A (BP < 50, PP100). At present, it is appropriate to include these two species in section Dendrobium.

Clade B can be characterized by the following combination of characters: clavate, non-heteroblastic, succulent stems; long flowering stems produced from the upper part of vegetative stems; lip with numerous papillae on the adaxial surface; tapered anther cap; and leaf sheath with brownish margins. Among these characters, leaf sheath with brownish margins represents synapomorphy of this clade. While Clade B itself is strongly supported (BP100, PP100), its relationship to other clades remains to be resolved. Therefore, it would seem appropriate to treat Clade B as an independent section. Clade B is comprised of members of sections Dendrobium and Holochrysa, both of whose type species belong to Clade A. Given that no name corresponding to the three species constituting Clade B currently exists, a new section name will have to be assigned.

Clade C comprises sections Densiflora and Amblyanthus, both of which are not monophyletic in the clade. Dendrobium microglaphys of section Amblyanthus occupies a nested position among members of section Densiflora and lacks scales outside the perianth lobes that characterize section Amblyanthus; instead, it has reduced leaf sheaths and pendulous flowering stems, both of which are diagnostic characters of section Densiflora. Consequently, it is reasonable to transfer D. microglaphys to section Densiflora. Meanwhile, the members of the basal subclade of Clade C, which includes Dendrobium melanostictum, the type species of section Amblyanthus, share features such as scales outside the perianth lobes, short racemes, and a peculiar scaly covering of the flowers and a backward-pointing lip appendage. Therefore, this subclade corresponds to section Amblyanthus and the rest of Clade C can be defined as section Densiflora.

Clade D comprises sections Pedilonum, Calcarifera, Oxyglossum, Calyptrochilus and Dolichocentrum. Since sections Pedilonum, Calcarifera, Oxyglossum and Dolichocentrum are not monophyletic, redefinition of these sections is necessary. If we conserve these sections, establishment of a large number of new sections is required. To avoid this complexity, we suggest designating Clade D as a single section. For this purpose, the section names Pedilonum and Calcarifera are not appropriate since their type species belong to Clade F. Therefore, the oldest available section name for Clade D is Calyptrochilus. An elongated mentum characterizes all species in this clade, while this character also defines Clade F and several species in other clades.

Clade E comprises sections Aporum, Crumenata and Bolbodium, the first two of which were found to not be monophyletic. As in the case of Clade D, conserving these two section names requires the establishment of multiple new sections. Thus, it is better to treat Clade E as a single section. The oldest available name corresponding to this clade is section Aporum. Clade E species have the following synapomorphies: a thick outer wall of epidermal cells on the stem surface and a thick sclerenchymatous cap on vascular bundles in the stem (Yukawa and Uehara 1996), and repeated flowering behaviour from the same reproductive shoot. Schuiteman (2011) also suggested that sections Aporum, Crumenata and Bolbodium should be treated as a single section.

Clade F comprises sections Pedilonum, Calcarifera and Platycaulon, the first two of which were found to not be monophyletic. Again, to avoid complexity by naming many new sections, it is reasonable to designate Clade F as a single section. The oldest available name corresponding to this clade is section Pedilonum. A shared character for Clade F is an elongated mentum, but this character also appeared in other clades, as mentioned previously.

The three species constituting Clade G have previously been treated as section Densiflora. Since we redefine section Densiflora as a subclade in Clade C where the type species is included, this section name cannot be used in Clade G. While the monophyly of this clade is strongly supported (BP100, PP100), its relationship to other clades remains to be resolved. Consequently, it is appropriate to deal with Clade G as an independent section. Given that no name corresponding to the three species constituting Clade G currently exists, a new section name will have to be assigned. Xiang et al. (2013) also suggested a separate status for a clade comprising D. jenkinsii and D. lindleyi. Members of this clade share the following combination of characters: pseudobulbous, heteroblastic, succulent stem, single leaf on the stem, and a few long flowering stems from the apical part of the vegetative stem. Among these characters, a single leaf represents the synapomorphy of this clade.

Clade H consists of members of sections Distichophyllae and Conostalix. As the latter section is nested within the first, the first section becomes paraphyletic. Except for thinner vegetative stems in section Conostalix, there is no obvious character to distinguish the two sections. Therefore, we deem it appropriate to redefine section Distichophyllae as the range of Clade H.

Clades I and J as well as unplaced species D. jerdonianum and D. trigonopus compose section Formosae. In addition to the definite monophyletic status of Clades I and J, it cannot be ruled out that some or all clades of this section along with section Distichophyllae form a monophyletic group. Since further analyses may resolve the ambiguous relationships of these clades, we suspend taxonomic treatment of section Formosae for the moment.

Clade K comprises section Stachyobium and D. oligophyllum of section Distichophyllae, which is the earliest divergent lineage in this clade. Re-examination of D. oligophyllum showed that the morphological characters of this species are consistent with the definition of section Stachyobium except for the long life-span of the leaf in this species. We thus transfer D. oligophyllum to section Stachyobium, and by this treatment, Clade K coincides with section Stachyobium. While the species of section Stachyobium share terminal inflorescences, we did not identify any synapomorphies of this section.

Clades L and M correspond to sections Herbaceae and Fytchianthe, respectively, endorsing the validity of the current definition of these sections. The species of section Herbaceae are characterized by ramified stems, terminal inflorescences and deciduous leaves. The species of section Fytchianthe are characterized by terminal inflorescences and deciduous leaves.

The scope of this study did not include the Australasian clade, the other major clade in genus Dendrobium. While a number of molecular phylogenetic studies have been conducted for this clade as well (Y. T. unpubl. res.; Yukawa et al. 1993, 1996, 2000; Yukawa 2001; Clements 2003, 2006; Burke et al. 2008), no attempt has been made to comprehensively analyse this clade as a whole. A consistent infrageneric classification of this genus should be proposed after combining the results of a phylogenetic analysis of both Asian and Australasian clades.

Sources of Funding

This work was partially supported by a Grant-in-Aid for Scientific Research (24370040 to T.Y.) from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

Contributions by the Authors

T.Y., T.T., T.H., S.K. and H.I. conceived and designed the study. T.Y., T.T., P.W., A.S., T.N., S.S., S.K. and N.S. performed sequence and phylogenetic analyses. T.Y. and T.T. primarily wrote the manuscript with contributions from all the authors.

Conflicts of Interest Statement

None declared.

Supporting Information

The following Supporting Information is available in the online version of this article –

Figure S1. Consensus phylogram obtained from 100 754 Bayesian trees based on ITS sequences. Values below and above branches indicate percentage bootstrap values from maximum parsimony analysis and Bayesian posterior probabilities, respectively.

Figure S2. Consensus phylogram obtained from 119 935 Bayesian trees based on matK sequences. Values below and above branches indicate percentage bootstrap values from maximum parsimony analysis and Bayesian posterior probabilities, respectively.

Appendix 1.

List of Dendrobium taxa and GenBank IDs.

Species	Section	Voucher	ITS	matK
D. acerosum	Aporum	Yukawa04-32	AB847639	AB847678
D. acinaciforme	Crumenata	TBG120483	AB847640	AB847679
D. aduncum	Breviflores	TBG134842	AB593484	AB847680
D. aff. anosmum	Dendrobium	Ting H19	AB593490	AB847686
D. aff. densiflorum	Densiflora	TBG159449	AB593485	AB847681
D. aff. infundibulum	Formosae	TBG128905	AB593486	AB847682
D. aff. modestum	Crumenata	TBG119268	AB847642	AB847683
D. aff. moniliforme	Dendrobium	TBG124453	AB593488	AB847684
D. aff. ramosii	Pedilonum	HBG4335	AB593492	AB847687
D. aff. signatum	Dendrobium	TBG118279	AB593489	AB847685
D. alaticaulinum	Pedilonum	TBG135586	AB593493	AB847688
D. albosanguineum	Dendrobium	TBG122507	AB593494	AB847689
D. amabile	Densiflora	TBG111421	AB593495	AB847690
D. amethystoglossum	Calcarifera	TBG120501	AB593496	AB847691
D. amoenum	Dendrobium	TBG159411	AB593497	AB847692
D. annae	Calcarifera	TBG111422	AB593498	AB847693
D. anosmum	Dendrobium	TBG118511	AB593499	AB847694
D. aphanochilum	Calyptrochilus	TBG142252	AB593500	AB847695
D. aphrodite	Dendrobium	TBG122797	AB593501	AB847696
D. aphyllum	Dendrobium	TBG122508	AB593539	AB847736
D. atavus	Holochrysa	TBG159413	AB593502	AB847697
D. auriculatum	Dolichocentrum	TBG120492	AB593503	AB847698
D. austrocaledonicum	Distichophyllae	TBG130055	AB593504	AB847699
D. auyongii	Aporum	TBG123341	AB847643	AB847700
D. bellatulum	Formosae	TBG133256	AB847644	AB847701
D. bensoniae	Dendrobium	TBG128899	AB593505	AB847702
D. bifarium	Distichophyllae	TBG159414	AB593506	AB847703
D. bifurcatum	Breviflores	TBG157293	AB593507	AB847704
D. bracteosum	Pedilonum	TBG102743	AB593509	AB847705
D. braianense	Holochrysa	TBG159424	AB593510	AB847706
D. brymerianum	Dendrobium	TBG118826	AB593511	AB847707
D. calcaratum	Pedilonum	TBG125164	AB593512	AB847708
D. calceolum	Aporum	TBG122494	AB593513	AB847709
D. caliculi-mentum	Pedilonum	TBG118610	AB593514	AB847710
D. capillipes	Holochrysa	TBG128878	AB593515	AB847711
D. cariniferum	Formosae	TBG128900	AB847645	AB847712
D. catenatum	Dendrobium	TBG159450	AB593517	AB847713
D. catillare	Pedilonum	TBG124769	AB847646	AB847714
D. cf. bensoniae	Dendrobium	036193-B	AB847647	AB847715
D. cf. candidum	Dendrobium	29746	AB593519	AB847716
D. cf. gratiosissimum	Dendrobium	TBG126713	AB593522	AB847719
D. cf. griffithianum	Densiflora	TBG128876	AB593523	AB847720
D. cf. hancockii	Dendrobium	TBG142251	AB593524	AB847721
D. cf. jenkinsii	Densiflora	TBG129904	AB593525	AB847722
D. cf. microglaphys	Amblyanthus	TBG116027	AB593526	AB847723
D. cf. primulinum	Dendrobium	TBG118293	AB593521	AB847718
D. cf. roseipes	Pedilonum	TBG159460	AB847648	AB847717
D. chameleon	Calcarifera	TBG135506	AB593527	AB847724
D. chrysanthum	Dendrobium	TBG124319	AB593529	AB847725
D. chryseum	Holochrysa	TBG127382	AB593530	AB847726
D. chrysocrepis	Dendrobium	TBG128879	AB593531	AB847727
D. chrysotoxum	Densiflora	TBG118321	AB593533	AB847728
D. codonosepalum	Calyptrochilus	TBG137050	AB847649	AB847729
D. compactum	Stachyobium	TBG133060	AB847650	AB847730
D. crepidatum	Dendrobium	TBG128871	AB593534	AB847731
D. cruentum	Formosae	TBG134572	AB593536	AB847733
D. crumenatum	Crumenata	TBG115833	AB593537	AB847734
D. crystallinum	Dendrobium	TNS8502972	AB593538	AB847735
D. cumulatum	Calcarifera	TBG159418	AB593541	AB847737
D. cuthbertsonii	Oxyglossum	TBG116108	AB593542	AB847738
D. cyanocentrum	Oxyglossum	TBG159420	AB593543	AB847739
D. dearei	Formosae	TBG144587	AB847651	AB847740
D. delacourii	Stachyobium	TBG100237	AB593545	AB847741
D. densiflorum	Densiflora	TBG159421	AB593546	AB847742
D. denudans	Stachyobium	TBG132760	AB593547	AB847743
D. devonianum	Dendrobium	TBG124383	AB593548	AB847744
D. diodon	Stachyobium	TBG116089	AB593550	AB847745
D. distichum	Aporum	TBG137051	AB593551	AB847746
D. dixanthum	Holochrysa	TBG128877	AB593552	AB847747
D. draconis	Formosae	TBG118276	AB593553	AB847748
D. ellipsophyllum	Distichophyllae	TBG126635	AB593554	AB847749
D. ephemerum	Crumenata	TBG159426	AB593555	AB847750
D. eriiflorum	Stachyobium	TBG140557	AB593556	AB847751
D. erosum	Calyptrochilus	TBG159429	AB593557	AB847752
D. eserre	Stachyobium	TBG133168	AB593558	AB847753
D. faciferum	Crumenata	YukawaDNA2134	AB847652	AB847754
D. fairchildiae	Calcarifera	TBG120506	AB593559	AB847755
D. falconeri	Dendrobium	TBG128903	AB593560	AB847756
D. farmeri	Densiflora	TBG159430	AB593561	AB847757
D. fimbriatum	Holochrysa	84386	AB593562	AB847758
D. findlayanum	Dendrobium	TBG128904	AB593563	AB847759
D. formosum	Formosae	TBG128902	AB593564	AB847760
D. friedericksianum	Dendrobium	TBG124322	AB593565	AB847761
D. furcatum	Dolichocentrum	TBG116070	AB593566	AB847762
D. fytchianum	Fytchianthe	TBG128868	AB593567	AB847763
D. gibsonii	Holochrysa	TBG129876	AB593568	AB847764
D. glomeratum	Calyptrochilus	TBG138022	AB593535	AB847732
D. goldfinchii	Crumenata	Yukawa 97-2008	AB593569	AB847765
D. goldschmidtianum	Pedilonum	TBG159434	AB593570	AB847766
D. gratiosissimum	Dendrobium	TBG126651	AB593571	AB847767
D. gregulus	Stachyobium	TBG132862	AB593572	AB847768
D. griffithianum	Densiflora	TBG138067	AB593573	AB847769
D. guibertii	Densiflora	TBG132542	AB593574	AB847770
D. hancockii	Dendrobium	TBG122506	AB593575	AB847771
D. harveyanum	Dendrobium	TBG133184	AB593576	AB847772
D. hasseltii	Pedilonum	TBG136022	AB593577	AB847773
D. hemimelanoglossum	Stachyobium	TBG133257	AB593578	AB847774
D. henryi	Holochrysa	TBG127415	AB593579	AB847775
D. herbaceum	Herbaceae	TBG156719	AB847654	AB847776
D. hercoglossum_1	Breviflores	TBG118850	AB593580	AB847777
D. hercoglossum_2	Breviflores	TBG124432	AB593581	AB847778
D. heterocarpum	Dendrobium	TBG116506	AB593582	AB847779
D. hookerianum	Dendrobium	TBG133007	AB593584	AB847780
D. hymenophyllum	Calcarifera	TBG140599	AB847655	AB847781
D. intricatum	Calcarifera	TBG137294	AB593586	AB847782
D. jacobsonii	Pedilonum	TBG133080	AB593587	AB847783
D. jenkinsii	Densiflora	TBG129880	AB593589	AB847784
D. jerdonianum	Formosae	TBG156717	AB847656	AB847785
D. junceum	Crumenata	TBG119079	AB593590	AB847786
D. kentrophyllum	Aporum	TBG102615	AB593591	AB847787
D. kontumense	Formosae	TBG126642	AB593592	AB847788
D. kraemeri	Pedilonum	TBG123234	AB847657	AB847789
D. kratense	Stachyobium	TBG126712	AB847658	AB847790
D. laevifolium	Oxyglossum	TBG142428	AB593593	AB847791
D. lamyaiae	Dendrobium	TBG116046	AB593595	AB847792
D. lancifolium	Calcarifera	TBG142427	AB847659	AB847793
D. laterale	Stachyobium	TNS8501719	AB847660	AB847794
D. lawesii	Calyptrochilus	TBG140275	AB593596	AB847795
D. leonis	Aporum	84206	AB593597	AB847796
D. leptocladum	Dendrobium	TNS8502939	AB593598	AB847797
D. linawianum	Dendrobium	TBG142429	AB593599	AB847798
D. lindleyi	Densiflora	TBG115834	AB593600	AB847799
D. linguella	Breviflores	TBG115831	AB593601	AB847800
D. lituiflorum	Dendrobium	TBG128908	AB593602	AB847801
D. lobbii	Conostalix	NSW426014	AB593603	AB847802
D. loddigesii	Dendrobium	74531	AB593604	AB847803
D. lohohense	Holochrysa	TBG132863	AB593605	AB847804
D. longicornu	Formosae	TBG122802	AB847661	AB847805
D. lowii	Formosae	TBG134848	AB593606	AB847806
D. luteolum	Dendrobium	TBG122798	AB593607	AB847807
D. maccarthiae	Dendrobium	TBG128803	AB593608	AB847808
D. macfarlanei	Crumenata	TBG122496	AB847662	AB847809
D. macrostachyum	Dendrobium	TBG123340	AB847663	AB847810
D. masarangense subsp. theionanthum	Oxyglossum	Tajima 27	AB593609	AB847811
D. melanostictum	Amblyanthus	TBG123873	AB593610	AB847812
D. melinanthum	Calyptrochilus	TBG159436	AB593611	AB847813
D. microglaphys	Amblyanthus	TBG135481	AB593612	AB847814
D. mohlianum	Calyptrochilus	TBG124783	AB593613	AB847815
D. moniliforme	Dendrobium	TBG115845	AB593614	AB847816
D. morrisonii	Pedilonum	Yukawa 97-2045	AB847664	AB847817
D. moschatum	Holochrysa	56113	AB593616	AB847818
D. mutabile	Calcarifera	TBG159437	AB593617	AB847819
D. nemorale	Distichophyllae	TBG159440	AB593618	AB847820
D. nobile	Dendrobium	TBG128809	AB593619	AB847821
D. ochreatum	Dendrobium	TBG128865	AB593621	AB847822
D. okinawense	Dendrobium	TBG115938	AB593622	AB847823
D. oligophyllum	Distichophyllae	TBG128728	AB847665	AB847824
D. pachyglossum	Conostalix	TBG124380	AB593623	AB847825
D. pachyphyllum	Bolbodium	TBG126623	AB593624	AB847826
D. palpebrae_1	Densiflora	TBG159441	AB593625	AB847827
D. palpebrae_2	Densiflora	TBG159442	AB593626	AB847828
D. palpebrae_3	Densiflora	TBG118313	AB593627	AB847829
D. panduriferum	Calcarifera	TBG144531	AB847666	AB847830
D. papilio	Dolichocentrum	TBG120493	AB847667	AB847831
D. parciflorum	Aporum	TBG118261	AB593628	AB847832
D. parcum	Herbaceae	TBG129884	AB593629	AB847833
D. parishii	Dendrobium	TBG159443	AB593630	AB847834
D. parthenium	Formosae	TBG136015	AB847668	AB847835
D. patentilobum	Aporum	TBG137060	AB847669	AB847836
D. pendulum	Dendrobium	TBG128867	AB593633	AB847837
D. philippinense	Crumenata	TBG142250	AB593634	AB847838
D. platygastrium	Platycaulon	TBG142432	AB593635	AB847839
D. polyanthum	Dendrobium	TBG159417	AB593636	AB847840
D. porphyrochilum	Stachyobium	TBG141468	AB593637	AB847841
D. praecinctum	Stuposa	TBG116030	AB593638	AB847842
D. prasinum	Oxyglossum	TBG119108	AB593639	AB847843
D. prianganense	Calcarifera	TBG134835	AB593640	AB847844
D. primulinum	Dendrobium	TBG159445	AB593641	AB847845
D. profusum	Calcarifera	TBG133266	AB593642	AB847846
D. pulchellum	Holochrysa	TBG118088	AB593643	AB847847
D. rarum	Pedilonum	TBG124851	AB593644	AB847848
D. regium	Dendrobium	TNS8501314	AB593645	AB847849
D. rhombeum_1	Dendrobium	TBG135976	AB593646	AB847850
D. rhombeum_2	Dendrobium	TBG142436	AB593647	AB847851
D. roseiodorum	Formosae	TBG118270	AB593648	AB847852
D. roseipes	Pedilonum	TBG142437	AB593649	AB847853
D. rosellum	Aporum	TBG133320	AB593650	AB847854
D. ruckeri	Dendrobium	TBG159446	AB593651	AB847855
D. sanderae	Formosae	TBG120503	AB593654	AB847856
D. sanguinolentum	Calcarifera	TBG159447	AB593655	AB847857
D. scabrilingue	Formosae	TBG119087	AB593656	AB847858
D. schuetzei	Formosae	TBG119088	AB593658	AB847859
D. scoriarum	Dendrobium	YukawaDNA0381	AB593659	AB847860
D. secundum	Pedilonum	TBG100260	AB593660	AB847862
D. senile	Dendrobium	TBG119090	AB593661	AB847863
D. signatum	Dendrobium	TBG133598	AB593662	AB847864
D. singkawangense	Formosae	TBG141081	AB593663	AB847865
D. smillieae	Pedilonum	TBG1402699	AB593664	AB847866
D. sp.	Stachyobium	YukawaDNA0749	AB847670	AB847861
D. spatella	Aporum	TBG84203	AB847671	AB847867
D. speckmaieri	Platycaulon	TBG136027	AB593665	AB847868
D. spectatissimum	Formosae	TBG133877	AB593666	AB847869
D. stockeri	Amblyanthus	TBG130082	AB593667	AB847870
D. stuposum	Stuposa	TBG134843	AB593668	AB847871
D. sulcatum	Densiflora	TBG159451	AB593670	AB847872
D. sutepense	Formosae	TNS8503543	AB593671	AB847873
D. suzukii	Formosae	TNS9518319	AB593672	AB847874
D. tetrachromum	Dendrobium	TBG137068	AB847672	AB847875
D. thyrsiflorum	Densiflora	TBG159453	AB593674	AB847876
D. tobaense	Formosae	TBG126681	AB593677	AB847877
D. tortile	Dendrobium	TBG119099	AB593678	AB847878
D. trankimianum	Formosae	TBG127512	AB847673	AB847879
D. transparens	Dendrobium	TBG126731	AB593679	AB847880
D. trantuanii	Breviflores	TBG157296	AB847674	AB847881
D. trichostomum	Calyptrochilus	TBG142265	AB593680	AB847882
D. trigonopus	Formosae	TBG119100	AB847675	AB847883
D. unicum	Dendrobium	TBG119101	AB593682	AB847884
D. uniflorum	Distichophyllae	TBG120499	AB593683	AB847885
D. venustum	Stachyobium	TBG129885	AB847676	AB847886
D. victoriae-reginae	Calcarifera	TBG141156	AB593684	AB847887
D. violaceum	Oxyglossum	TBG100203	AB593685	AB847888
D. wardianum	Dendrobium	TBG126641	AB593686	AB847889
D. ypsilon	Platycaulon	TBG118514	AB593688	AB847890