Florida Museum of Natural History
Assembling the Tree of Life: Orchidaceae ONCIDIINAE (ORCHIDACEAE) University of Florida Herbarium Home

This study expands our previous work on subtribes Maxillariinae, Oncidiinae, Zygopetalinae, and Stanhopeinae.


INTRODUCTION TO ONCIDIINAE (ORCHIDACEAE)

This web site is based on our paper Neubig, K.M., W. M. Whitten, N. H. Williams, M. A. Blanco, L. Endara, J. G. Burleigh, K. Silvera, J. C. Cushman, & M. W. Chase. 2012. Generic recircumscriptions of Oncidiinae (Orchidaceae: Cymbidieae) based on maximum likelihood analysis of combined DNA datasets. Botanical Journal of the Linnean Society 168: 117-146. For this web site I have modified some of the text, added a few references, and added a large number of photographs. I also provide an interesting method of viewing PHYLOGENETICS, an updated section on REFERENCES, photograhic IMAGES for a number of species, files that can be DOWNLOADED for you to examine on your computer, a list of PEOPLE involved in this project, LINKS to a number of interesting and useful web sites, and ACKNOWLEDGMENTS for all of the people, institutions, and organizations that helped with this project.



In our 2001 paper (Williams, et al., 2001a) we had 77 in-group taxa. We have increased taxon sampling to approximately 600 species (represented by 736 samples). We recovered the same clades as in our 2001 paper, with the following additions and exceptions.

Clade A, the Fernandezia-Ornithocephalus-Telipogon clade was represented by 8 species. We have increased sampling in this group to 96 OTU's and recover the same three major groupings: the Telipogon clade is sister to the Fernandezia clade + the Ornithocephalus clade. Members of the Ornithocephalus and Telipogon clades have four pollinia, whereas traditional members of the Oncidiinae have only two pollinia. Pachyphyllum and Raycadenko are lumped into Fernandezia and all have two pollinia. The presence of four pollinia appears to be a reversal from the two pollinia condition. Hofmeisterella is sister to the remainder of the Telipogon clade. The only chromosome counts that have been reported for members of this clade are for a member of the Ornithocephalus clade (2n=56), the most common number in the Oncidiinae.


Clade B, the Brassia-Miltonia-Aspasia clade, had 13 taxa, which is now increased to 53 OTU's. Within this clade, we recovered the following recognized genera: Miltonia, Systeloglossum, Oliveriana, Cischweinfia, Brassia, and Aspasia. Miltonia (100% BS) is sister to the remainder of this group, with the following relationships all well supported: (Systeloglossum + Oliveriana; 100% BS) is sister to Cischweinfia (100% BS); Ada, Brachtia, and Mesospinidium are embedded in Brassia and that clade is sister to Aspasia (100% BS). Chromosome numbers of members of this clade are usually 2n=56, although a wide range of numbers have been reported for Brassia (2n=18, 50, 52-58, 60), Aspasia (2n=56, 58, 60), and Miltonia (2n=48, 56, 59, 60, 86, 98, 112, although 60 is the most commonly reported number in Miltonia). No chromosome counts have been reported for other members of this clade.


Clade C, the Cyrtochilum clade (86% BS, including Miltoniopsis and Otoglossum) had six taxa in our earlier sampling, which is now increased to 93 OTU's and we recovered the same clades: Otoglossum, Cyrtochilum, Miltonioides, Caucaea, and Cyrtochiloides; although we lacked material of Cyrtochiloides in the first 2001 paper, we included it in the second 2001 paper (Williams, et al., 2001b). The relationships are: Otoglossum (79% BS) sister to the remainder, followed by Cyrtochiloides (100% BS), Miltoniopsis (100% BS), Caucaea (100% BS), and finally Cyrtochilum (99% BS). Our data do not support maintaining Trigonochilum or Dasyglossum as distinct genera.


Clade D in the first paper included Rhynchostele, Erycina, and Tolumnia, but with only 12 taxa and 84% bootstrap support. With increased sampling, Clade D breaks into two groups: the Tolumnia clade (100% BS) which is sister to the "twig-epiphyte" group (100% BS) in Clade E, and the Rhynchostele + Erycina clade (100% BS) which is sister to the "twig-epiphyte + Tolumnia + Gomesa clade (93% BS).


Clade E, originally the "twig-epiphyte" clade plus Zelenkoa plus the Gomesa group (although at that time not recognized as just one genus), has been expanded from a taxon size of 15 to approximately 146 OTU's. The combined older clades D and E are now recognized as being broken into several different groups: (1) the "twig-epiphytes" sister to Tolumnia and several unresolved genera: Solenidium, Capanemia, Zelenkoa, Notyliopsis, and Nohawilliamsia; (2) the "twig-epiphytes" + Tolumnia + the unresolved genera sister to Gomesa; and (3) a Rhynchostele + Erycina clade sister to the above clades.


Clade F, the Oncidium/Odontoglossum clade, with 100% BS support (including Cochlioda, Collare-stuartense, Heteranthocidium, Sigmatostalix, Solenidiopsis, Symphyglossum, Mexicoa, and Miltonioides) has been lumped into one genus (Chase, et al., 2008). As we pointed out earlier, the distinction between Oncidium and Odontoglossum has long been a source of contention. Much of the confusion and causes of concern have been resolved with the earlier transfer of species to other genera, such as Rhynchostele, Otoglossum, Rossioglossum, and Cyrtochilum. Unfortunately, many apparently distinct groups are embedded within Oncidium/Odontoglossum. For example, Cochlioda and Symphyglossum, with their distinctive bird pollinated flowers, are deeply embedded in Oncidium/Odontoglossum. Although it has distinctive oil producing flowers (often quite small), Sigmatostalix is also deeply embedded within Oncidium/Odontoglossum. If we were to try to maintain Oncidium, Odontoglossum, Sigmatostalix, Cochlioda, and Symphyglossum, we would need to erect approximately 20 new genera, none of which we would be able to define on morphological grounds.


Clade G, Lockhartia, remains separate from other groups.


Clade H, the Trichocentrum s.l. group, now includes a number of clades for which we earlier had no material, including the Rossioglossum clade (including Chelyorchis and Ticoglossum) sister to Cuitlauzina (including Osmoglossum and Palumbina), the Grandiphyllum group (the Brazilian "mule-ear" oncidiums; 100% BS), and Saundersia. Although the entire clade has only weak support (57% BS), the individual clades all have strong support. Rossioglossum (100% BS) + Cuitlauzina (100% BS; 97% BS for the two) is sister to the remainder of the group, with a series of grades of Grandiphyllum, Saundersia, and Trichocentrum (100% BS).


Clade I, the Trichopilia clade (100% BS) is the same, although with larger taxon sampling. This clade includes Psychopsis, Psychopsiella, and Trichopilia (including Helcia, Leucohyle, and Neoescobaria).



THE ONCIDIINAE


Oncidiinae (Cymbidieae) are one of the most diverse subtribes of Orchidaceae, with a wide range of floral and vegetative morphologies. They include the greatest diversity of pollination systems and the widest range of chromosome numbers known for Orchidaceae (greater than the rest of the orchid family combined). They also form major components of the Neotropical flora, ranging from sea level to nearly 4,000 meters in the Andes; several species of Brassia, Miltonia, Miltoniopsis, and Oncidium are important ornamental crops. Oncidiinae are members of a Neotropical clade that includes subtribes Coeliopsidinae, Maxillariinae, Stanhopeinae, and Zygopetalinae; these five subtribes are each clearly monophyletic and collectively are sister to Eriopsidinae, but relationships among the five subtribes are still poorly resolved.

Largely following the generic concepts of Chase (2009b), subtribe Oncidiinae includes 62 genera (Table 1 in the PHYLOGENETICS section) and approximately 1,600 species. Prior to molecular phylogenetic studies, subtribal delimitation varied widely, from the relatively broad concept of Dressler (1993) to the narrow concepts of Szlachetko (1995), with the latter splitting out approximately 20 subtribes based largely on column morphology (including their complex pollinaria).

Previous classifications of Oncidiinae were intuitively based mainly on floral morphology and to a lesser extent chromosome number, and all were produced without cladistic methodology (Dressler, 1993; Garay & Stacy, 1974; Senghas, 1997). Recent molecular studies have helped resolve and define Oncidiinae, as well as circumscribe many genera (Chase & Palmer, 1987; Sandoval-Zapotitla, et al., 2010; Williams, et al., 2001a; Williams, Chase, & Whitten, 2001b). Subtribes Ornithocephalinae and Telipogoninae, long held separate on the basis of their four pollinia (vs. two in Oncidiinae), plus the monopodial Pachyphylliinae (two pollinia), were shown to nest within Oncidiinae.

Dressler (1993) emphasized seed characters, velamen type, and number of nodes per pseudobulb in his concepts of Cymbidieae and Maxillarieae. However, molecular data (van den Berg, et al., 2005) indicated that Cymbidieae sensu Chase (Chase, et al., 2003) are likely to be paraphyletic to Maxillarieae, and the two might be regarded as a single tribe. In the current circumscription, Oncidiinae include taxa with both two and four pollinia.

Oncidiinae exhibit an enormous diversity in form and function that makes them attractive subjects for evolutionary studies. Floral size ranges several orders in magnitude, and flowers evolved to utilize a diverse array of pollinators. Floral rewards include nectar, oils, and fragrances, but deceit flowers are the most common pollination strategy.

Chromosome numbers range from the lowest known in orchids of 2n=10 to 2n=168 (Tanaka & Kamemoto, 1984) and genome size spans at least a seven-fold range (Chase, et al., 2005). Vegetatively, plants range from large long-lived perennials with 1 kg pseudobulbs (or more) to highly reduced twig epiphytes the size of a thumbnail with rapid life cycles (several months). Most species are epiphytes and CAM photosynthesis is thought to have arisen repeatedly (Silvera, et al., 2009; Silvera, et al., 2010).

Understanding the evolution of this range of form, function, and biogeographic patterns depends upon a reliable phylogenetic hypothesis of relationships for hundreds of species. Generic boundaries and relationships within Oncidiinae have been highly contentious, and several genera have been viewed as taxa of convenience (non-monophyletic; Garay, 1963). Previous evolutionary studies have been hampered by the choice of non-monophyletic groups and by lack of reliable phylogenetic hypotheses. Our goal is to use combined plastid and nrITS data to produce a densely sampled phylogenetic estimate of relationships within Oncidiinae and to use this to underpin a stable generic classification (Chase, 2009b) that can be used as a framework for more focused studies.



POLLINATION AND FLORAL MIMICRY IN ONCIDIINAE

Historically, many of the difficulties with generic circumscription in Oncidiinae are likely the result of homoplasy and mimicry in flower shape and color. Generic boundaries have long been contentious in both the botanical and horticultural communities (Braem, 2010; Garay, 1963). As in most orchid groups, generic concepts traditionally have emphasized floral characters and neglected vegetative ones. In Oncidiinae, floral traits and pollination systems appear to be especially labile, which has undoubtedly fostered much of the confusion in generic boundaries and resulted in many polyphyletic genera. Floral rewards in Oncidiinae include nectar, oils, and fragrances. Pollen is never offered as a reward, and pseudopollen and resin rewards are unknown in Oncidiinae.

Nectar is a reward for bees, Lepidoptera, and hummingbirds and usually is presented in a nectariferous spur formed by the lip or the adnation of lip and column. However, nectar deceit is common, and the presence of a spur does not always indicate nectar. Relatively few species produce a fragrance reward consisting of monoterpenes, sesquiterpenes, and simple aromatics. These fragrances are collected by male euglossine bees (Apidae: Euglossini) and they are thought to serve a role in sexual selection by female euglossine bees (Bembe, 2004; Eltz, Roubik, & Lunau, 2005; Zimmermann, et al., 2009).

Most Oncidiinae species have flowers that either produce an oil reward or are mimics of oil-producing flowers of Malpighiaceae ("malpighs"; Carmona-Díaz & García-Franco, 2009; Reis, et al., 2000; Reis, et al., 2007; Sigrist & Sazima, 2004; Silvera, 2002). Figure 1. These oil flowers attract a variety of female bees of various sizes of several different genera in the tribes Centridini, Tapinostapidini, and Tetrapediini of the family Apidae (formerly assigned to a separate family, Anthophoridae, and still occasionally referred to as "anthophorid" bees). The female bees collect oil from specialised glands (elaiophores) on the flowers and use the oils as provisions and/or waterproofing for larval cells (Cane, et al., 1983; Melo & Gaglianone, 2005; Roubik, 1989).

Numerous species of Oncidiinae that are putative mimics of malpighs exhibit a suite of characters that include bright yellow (Figure 1) or purple (Figure 2 A-E) flowers, elaiophores consisting of epidermal pads on the lateral lobes of the lip or pads of trichomes on the lip callus and a tabula infrastigmatica (a fleshy ridge at the base of the column that is grasped by the bee's mandibles, freeing their front and middle legs to collect oil). Many Oncidiinae also possess prominent elaiophores (Figure 2 F-J); [e.g., Oncidium cheirophorum Rchb.f., Oncidium sotoanum R. Jiménez & Hágsater, Trichocentrum cavendishianum (Bateman) M.W. Chase & N.H. Williams, various Gomesa spp. (Aliscioni, et al., 2009; Davies & Stpiczynska, 2009; Pansarin, Castro, & Sazima, 2009; Stpiczynska & Davies, 2008; Stpiczynska, Davies, & Gregg, 2007)]. Parra-Tabla et al. (2000) reported that Trichocentrum ascendens (Lindl.) M.W. Chase & N.H. Williams is pollinated primarily by female Trigona bees collecting the oily floral secrections for nest construction. Species with prominent elaiophores represent legitimate oil reward flowers (Figure 2 F-O).We examined floral color and shape convergence in Neotropical plant communities, focusing on certain food-deceptive Oncidiinae orchids (e.g.,Trichocentrum ascendens and Oncidium nebulosum) and rewarding species of Malpighiaceae. We showed that the species from these two distantly related families are often more similar in floral color and shape than expected by chance and propose that a system of multifarious floral mimicry—a form of Batesian mimicry that involves multiple models and is more complex than a simple one model– one mimic system—operates in these orchids (Papadopulos et al., 2013).


Representatives of major clades of core Oncidiinae

Figure 1. Various genera and species of Oncidiinae displaying hypothetical mimicry of yellow Malpighiaceae flowers.

A) Malpighia sp.; B) Psychopsiella limminghei; C) Grandiphyllum auriculatum; D) Trichocentrum nanum; E) Trichocentrum cebolleta; F) Trichocentrum ascendens; G) Rossioglossum ampliatum; H) Lockhartia lepticaula; I) Fernandezia ecuadorensis; J) Vitekorchis excavata; K) Oncidium cultratum; L) Oncidium obryzatum; M) Oncidium sp.; N) Oncidium sphacelatum; O) Oncidium heteranthum; P) Gomesa gardneri; Q) Gomesa insignis; R) Gomesa longipes; S) Otoglossum harlingii; T) Otoglossum scansor; U) Erycina pusilla; V) Nohawilliamsia orthostates; W) Zelenkoa onusta; X) Tolumnia urophylla; Y) Tolumnia quadriloba. Photos by W. Mark Whitten.


Representatives of major clades of core Oncidiinae

Figure 2. Oncidiinae displaying various pollination syndromes.

Row 1 A-E: Hypothetical purple Malpighiaceae mimics. A) Malpighia glabra; B) Oncidium sotoanum; C) Cyrtochilum edwardii; D) Tolumnia hawkesiana; E) Cyrtochilum ioplocon; Rows 2 (all) and 3 (all), Oncidiinae that secrete oil from localized elaiophores: F) Lockhartia longifolia; G-H) Cyrtochilum serratum (arrow denotes elaiophore); I-J) Oncidium cheirophorum (arrow denotes elaiophore); K) Ornithocephalus cochleariformis; L) Ornithocephalus dalstroemii; M) Ornithocephalus dressleri; N) Phymatidium falcifolium; O) Oncidium sp.; (Sigmatostalix clade); Row 4 P-S, putative hummingbird-pollinated species: P) Fernandezia subbiflora; Q) Brassia aurantiaca; R) Brassia andina; S) Oncidium (Cochlioda) beyrodtioides; Row 4 T-U, pseudocopulatory species: T) Tolumnia henekenii; U) Trichoceros antennifer; Row 5 V-Y, species pollinated by nectar-foraging insects: V) Trichocentrum longicalcaratum; W) Comparettia macroplectron; X) Rodriguezia sp.; Y) Trichopilia rostrata; Row 5, Z, floral fragrance reward flower pollinated by male euglossine bees: Z) Macroclinium dalstroemii. Photo E courtesy Guido Deburghgraeve; all others by W. Mark Whitten.


Perhaps a larger percentage of Oncidiinae possess flowers with similar malpigh-mimicking color (bee-UV-green; Powell, 2008), morphology, and tabula infrastigmatica, but lack clearly demonstrable elaiophores. These species represent oil deceit flowers that lure oil-collecting bees but fail to produce a legitimate reward (Figure 1). Oncidiinae floral morphology is probably the result of a complex mixture of Batesian and Müllerian mimicry (Roy & Widmer, 1999). Powell (2008) used spectral reflectance analyses to demonstrate that many Oncidiinae with yellow flowers closely match the color of yellow malpigh flowers [Byrsonima crassifolia (L.) Kunth] and thus satisfy criteria for Batesian mimicry. By mapping these traits onto a phylogenetic tree of Oncidiinae, he estimated at least 14 independent origins of putative malpigh mimicry within Oncidiinae. Carmona-Díaz and García-Franco (2009) demonstrated that the rewardless Trichocentrum cosymbephorum (C. Morren) R. Jiménez & Carnevali is pollinated by the same oil-collecting Centris bees that pollinate Malpighia glabra L., and the orchid has greater reproductive success in the presence of the malpigh than in isolated clumps. Vale, Rojas, and Álvarez (2011) showed that the rewardless and self-incompatible Tolumnia guibertiana (A. Rich.) Braem, endemic to western Cuba, is completely dependent on oil-gathering female bees (Centris poecila Lepeletier) for fruit production. This bee species is also the pollinator of two other yellow-flowered plants in the area: the pollen and oil producing malpigh Stigmaphyllon diversifolium (Kunth) A. Juss. and the polliniferous and buzz-pollinated Ouratea agrophylla (Tiegh.) Urb. (Ochnaceae).

Sazima and Sazima (1988) showed that some eglandular Malpighiaceae (lacking sepalar elaiophores) are possible mimics of glandular forms. There likely are complex mimicry relationships between Malpighiaceae species, oil-producing Oncidiinae, and oil-deceit Oncidiinae. We also suspect that some Oncidiinae mimic oil-producing Calceolaria (Calceolariaceae), since they occur at high elevations where malpighs are absent or rare and Calceolaria species are common. For example, Otoglossum harlingii (Stacy) N.H. Williams & M.W. Chase (Fig. 1 S) bears a striking visual similarity to sympatric species of Calceolaria.

Some oil-secreting species with relatively small, greenish white or yellow flowers (e.g., Ornithocephalus, Phymatidium;Figs. 2 K-N) attract a subset of oil-foraging bees with smaller body size and do not appear to be involved in mimicry.

This extensive homoplasy in oil flower morphology has contributed to grossly polyphyletic classifications of Oncidiinae, especially in clades that contain species with bright yellow "oncidioid" flowers. Floral morphology, including detailed structure of the column (Szlachetko, 1995), is clearly unreliable as the sole basis for generic circumscription. A robust phylogenetic framework based on molecular data can help diagnose polyphyletic groups and inform a new clade-based classification.



MATERIALS AND METHODS

Taxon sampling - Specimens were obtained from wild-collected or cultivated plants (Appendix S1). Most taxon names follow the generic concepts of Chase (2009b). Sampling of Oncidiinae included 738 accessions from a total of 590 ingroup species. We included seven outgroup taxa from other subtribes of Cymbidieae (Cameron, 2004; Cameron, Chase, Whitten, Kores, Jarrell, Albert, Yukawa, Hills, & Goldman, 1999). We were unable to obtain DNA of the following genera: Caluera Dodson & Determann, Centroglossa Barb.Rodr., Cypholoron Dodson & Dressler, Dunstervillea Garay, Platyrhiza Barb.Rodr., Quekettia Lindl., Rauhiella Pabst & Braga, Sanderella Kuntze, Suarezia Dodson, and Thysanoglossa Porto & Brade.

Extractions, amplification, and sequencing - All freshly collected material was preserved in silica gel (Chase & Hills, 1991). Genomic DNA was extracted using a modified cetyl trimethylammonium bromide (CTAB) technique (Doyle & Doyle, 1987), scaled to a 1 mL volume reaction. Approximately 10 mg of dried tissue were ground in 1 mL of CTAB 2X buffer and 2.0 μL of either Β-mercaptoethanol or proteinase-K (25 micrograms/mL; Promega, Inc., Madison, Wisconsin USA). Some total DNAs were then cleaned with QIAquick PCR (Qiagen, Valencia, California, USA) purification columns to remove inhibitory secondary compounds. Amplifications were performed using an Eppendorf Mastercycler EP Gradient S thermocycler (Hauppauge, New York, USA) and Sigma brand reagents (St. Louis, Missouri, USA) in 25 μL volumes with the following reaction components for ITS: 0.5-1.0 μL template DNA (~10-100 ng), 11.0 μL water, 6.5 μL 5M betaine, 2.5 μL 10X buffer, 3.0 μL MgCl2 (25 mM), 0.5 μL of 10 μM dNTPs, 0.5 μL each of 10 μM primers, and 0.5 units Taq DNA polymerase. For the plastid regions, the following reaction components were used: 0.5-1.0 μL template DNA (~10-100 ng), 16-18 μL water, 2.5 μL 10X buffer, 2-3 μL MgCl2 (25 mM), 0.5 μL of 10 μM dNTPs, 0.5 μL each of 10 μM primers, and 0.5 units (0.2 μL) Taq DNA polymerase.

nrITS (ITS 1 + 5.8S rDNA+ ITS 2) - This region was amplified with a touchdown protocol using the parameters 94°C, 2 min; 15X (94°C, 1 min; 76°C, 1 min, reducing 1°C per cycle; 72°C, 1 min); 21X (94°C, 1 min; 59°C, 1 min; 72°C, 1 min); 72°C, 3 min with the primers 17SE and 26SE from Sun et al. (1994). Betaine was added to eliminate secondary structure typical of the ribosomal DNA, so that active ITS copies would predominate in the PCR product. Except for nrITS, all other regions sequenced are plastid regions.

matK-trnK - This region includes the entire matK gene and the flanking 3'trnK spacer and is roughly 1800 base-pairs (bp) in length. This region was amplified with the parameters 94°C, 3 min; 33X (94°C, 45 sec; 60°C, 45 sec; 72°C, 2 min); 72°C, 3 min, with primers -19F (Molvray, Kores, & Chase, 2000) and trnK 2R (Johnson & Soltis, 1994). Internal sequencing primers were matK intF (TGAGCGAACACATTTCTATGG) and matK intR (ATAAGGTTGAAACCAAAAGTG). Some samples were amplified using the primers 56F and 1520R (Whitten, Williams, & Chase, 2000) that yielded a shorter, but nearly complete sequence of the matK exon (missing the 3'trnK spacer).

psaB - This region includes roughly 1700 bp of the protein-coding exon for a subunit of photosystem I. It was amplified with the parameters 94°C, 3 min; 33X (94°C, 30 sec; 55°C, 30 sec; 72°C, 2 min); 72°C, 4 min, with the primers NY159 and NY160 from Cameron (2004).

rbcL - This region (ca. 1350 bp) was amplified with the same parameters as for psaB, but with primers NY35 and NY149 from Cameron (2004).

trnH-psbA - This region includes about 800 bp of the intergenic spacer and a short exon, rps19. This region was amplified with the parameters 94°C, 3 min; 33X (94°C, 1 min; 58°C, 1 min; 72°C, 1 min, 20 sec); 72°C, 6 min, with the primers F and R from Xu et al. (2000).

ycf1- We sequenced two non-contiguous portions of ycf1 (Neubig, Whitten, Carlsward, Blanco, Endara, Williams, & Moore, 2009) including ca. 1200 bp from the 5' end and ca. 1500 bp from the 3' end. Both were amplified using a "touchdown" protocol with the parameters 94°C, 3 min; 8X (94°C, 30 sec; 60-51°C, 1 min; 72°C, 3 min); 30X (94°C, 30 sec; 50°C, 1 min; 72°C, 3 min); 72°C, 3 min. Primers for the 5' portion are 1F (ATGATTTTTAAATCTTTTCTACTAG) and 1200R (TTGTGACATTTCATTGCGTAAAGCCTT). Primers for the 3' portion are 3720F (TACGTATGTAATGAACGAATGG) and 5500R (GCTGTTATTGGCATCAAACCAATAGCG). Additional internal sequencing primers are intF (GATCTGGACCAATGCACATAT T) and intR (TTTGATTGGGATGATCCAAGG).

PCR products were cleaned with Microclean™ (The Gel Company, San Francisco, California, USA) following the manufacturer's protocols, eluted with 50.0 μL of 10 mM Tris-HCl (pH 8.5) and stored at 4°C. Purified PCR products were then cycle-sequenced using the parameters 96°C, 10 sec; 25X (96°C, 10 sec; 50°C, 5 sec; 60°C, 4 min), with a mix of 3.0 μL water, 1.0 μL fluorescent Big Dye dideoxy terminator, 2.0 μL Better Buffer™ (The Gel Company), 1.0 μL template and 0.5 μL primer. Cycle sequencing products were cleaned using ExoSAP™ (USB Corporation, Cleveland, Ohio, USA) following the manufacturer's protocols. Purified cycle sequencing products were directly sequenced on an ABI 377, 3100, or 3130 automated sequencer according to the manufacturer's protocols (Applied Biosystems, Foster City, California, USA). Electropherograms were edited and assembled using Sequencher 4.9™ (GeneCodes, Ann Arbor, Michigan, USA). All sequences were deposited in GenBank (Appendix 1).

Data analyses - We constructed two data matrices. The first included seven DNA regions (nrITS, trnH-psbA, 3'ycf1, 5'ycf1, matK, rbcL, and psaB) for 122 taxa. This smaller restricted data set included two relatively conserved plastid genes (rbcL and psaB) with the goal of providing increased resolution and support for the deeper nodes of the tree. The outgroup for this data set was Rudolfiella Hoehne sp.

The second matrix included five DNA regions (nrITS, trnH-psbA, 5'ycf1, 3'ycf1, and matK) for 737 samples representing approximately 600 species. Outgroup taxa were Eriopsis biloba Lindl., Eulophia graminea Lindl., Cyrtidiorchis stumpflei (Garay) Rauschert, Rudolfiella Hoehne sp., Stanhopea jenishiana F. Kramer ex Rchb.f., and Stanhopea trigrina Bateman ex Lindl. Data matrices are available from the FLMNH public ftp site (ftp://ftp.flmnh.ufl.edu/public/oncids).

Maximum likelihood (ML) phylogenetic analyses were performed on both data sets using RAxML version 7.0.4 (Stamatakis, 2006). We ran analyses that included 1) only ITS, 2) only the plastid loci, and 3) all loci. All ML analyses used the general time-reversible model of evolution (GTR; Tavare, 1986) with among-site rate variation modeled using the "CAT" discrete rate categories option. For the analyses of the plastid loci and all loci, we further partitioned the ML model based on DNA region. Specifically, we estimated substitution model parameters for each region as well as region-specific branch lengths. To find the optimal tree for each data set, we performed 5 runs of the ML tree heuristic search and we performed 200 non-parametric bootstrap replicates to assess clade support in the tree (Felsenstein, 1985).



RESULTS

Seven-locus data set (Fig. 3, Fig. 4). Both the plastid and the nrITS trees recover the same major clades, although there are some differences in the topology along the spines of the trees. There is conflict between the topology of the trees resulting from plastid and nuclear DNA sequences in the relationships of Psychopsis, Psychopsiella, and Trichopilia; Psychopsis and Psychopsiella are strongly supported as sister in the nrITS tree, whereas Psychopsis is strongly supported as sister to Psychopsiella and Trichopilia in the plastid tree. Vitekorchis is isolated in both nuclear and plastid trees. It is weakly supported as sister to Oncidium + all remaining taxa in the plastid tree, but is unresolved at a deeper node in nrITS trees. Tolumnia is strongly supported as sister to Erycina + Rhynchostele in nrITS results, but plastid data place Tolumnia as a well-supported member of a derived clade (Nohawilliamsia to Comparettia). Many of these discrepancies in the deeper nodes may be due to potential alignment error or possible saturation of nrITS. Most of the plastid regions exhibit few or no apparent alignment problems (with the exception of large portions of trnH-psbA that were excluded from the analyses). The combined plastid + nrITS seven-region analysis of 122 taxa (Fig. 4) is largely consistent with the analysis of the larger five-locus data (736 taxa; Figs. 5-12), but the addition of rbcL and psaB data provide slightly more support for the spine of the tree.

Five-locus data set (Figs. 5-12). Many species are represented by two or more samples. In most cases, multiple accessions of a single species form a unique group (e.g., Erycina, Fig. 10). In a few cases, plants from putatively the same species do not fall together (e.g., Cyrtochilum cimiciferum, Fig. 9). Some of these may be due to errors in determinations, but usually these represent taxonomically confusing groups with poorly defined species boundaries. We recognize 62 clades in this tree at the generic level (Table 1, in the PHYLOGENETICS section). All of the clades that we recognize at the generic level are strongly supported, and there is also strong support for nearly all supra-generic nodes in the tree. Exceptions are the monotypic genera Zelenkoa, Notyliopsis, and Nohawilliamsia (Fig. 11). These taxa form a poorly supported grade that is sister to Tolumnia and the twig epiphyte clade (all taxa in Fig. 12). Other genera with weak support for generic topology include Schunkea, Trizeuzis, Seegeriella, and Warmingia.



DISCUSSION

Genera are discussed in order of their appearance in the cladograms (Figs. 5-12). Generic concepts generally follow those of Chase (2009b) with a few exceptions; more detailed information for each genus is presented in that work.


Psychopsis Raf. (5 spp.; Fig. 5) ranges from Costa Rica south through the Andes to Peru. Chase (2005) lumped the monotypic Psychopsiella into Psychopsis on the basis of their sister relationship in the nrITS tree to avoid creation of a monotypic genus, but analysis of the combined data sets place Psychopsiella sister to Trichopilia Lindl. Chromosome numbers also differ for Psychopsis 2n=38 (Dodson, 1957) versus 2n= 56 for Psychopsiella and Trichopilia (Charanasri & Kamemoto, 1975). Both Psychopsiella and Psychopsis have yellow and brown flowers with a tabula infrastigmatica, suggestive of oil-reward flowers, although Dodson (2003) reported pollination of Psychopsis krameriana (Rchb.f.) H.G. Jones by Heliconius butterflies, but his observations have not been replicated.

<i>Psychopsis versteegiana</i>
Psy. versteegiana

<i>Psychopsis</i>
Psy. papilio

<i>Psychopsis</i>
Psy. papilio side view

<i>Psychopsis</i>
Psy. papilio column

<i>Psychopsis papilio</i>
Psychopsis papilio

<i>Psychopsis</i>
Psychopsis comparison of two species


Psychopsiella Lückel & Braem (1 sp.; Fig. 1B; Fig. 5) is monotypic and vegetatively resembles a dwarf Psychopsis, but it lacks the elongate dorsal sepal and petals of the latter. It is restricted to Brazil and has been reported from Venezuela, near Caracas, but this may have been an escape from cultivation. It shares a chromosome number of 2n=56 with its sister, Trichopilia.

Psychopsiella limminghei
Psychopsiella limminghei


Trichopilia Lindl. (approximately 26 spp.; Fig. 2Y ; Fig. 5) is largely characterized by having a lip that enfolds and is fused basally to the column, in some species forming a deep tubular structure suggestive of nectar reward or deceit, although Dodson (1962) reported pollination of one species by fragrance-collecting male euglossine bees. Some species of Cattleya and Sobralia have similar flowers, and they also are visited by male euglossine bees. Vegetatively, plants of Trichopilia are similar to the preceding two genera. Helcia Lindl., Leucohyle Klotzch, and Neoescobaria Garay are embedded within Trichopilia. These differ primarily in the lack of lip/column fusion and have previously been recognized by some authors as members of Trichopilia.

<i>Trichopilia</i>
Trichopilia sp.

<i>Trichopilia xantholeuca</i>
Trichopilia xantholeuca

<i>Trichopilia suavis</i>
Trichopilia suavis

<i>Trichopilia</i>
Trichopilia suavis

<i>Trichopilia</i>
T. (Helcia) sanguinolenta

<i>Trichopilia</i>
T. (Helcia) sanguinolenta

<i>Trichopilia subulata</i>
T. (Leucohyle) subulata

<i>Trichopilia</i>
T. (Leucohyle) subulata

<i>Trichopilia</i>
T. (Leucohyle) subulata


Rossioglossum (Schltr.) Garay & G.C. Kenn. (10 spp.; Fig. 5), as circumscribed here, includes Ticoglossum Lucas Rodr. ex Halb. and Chelyorchis Dressler & N.H. Williams. This genus also includes considerable floral diversity, suggestive of pollination by a variety of bees, but pollination data are mostly lacking. Rossioglossum ampliatum (Lindl.) M.W. Chase & N.H. Williams (Fig. 1G) has numerous bright yellow (bee-UV-green; Powell, 2008) Oncidium-like flowers that are malpigh mimics, whereas other Rossioglossum (e.g., R. insleayi and R. grande) bear relatively few, large flowers barred with yellow and brown. All species share vegetative similarities of rounded, ancipitous pseudobulbs topped by a pair of leathery leaves. Van der Pijl and Dodson (1966) reported pollination of R. grande by Centris bees. Their floral features, particularly the presence of a tabula infrastigmatica, indicates oil-bee pollination, although their floral absorbance has not been investigated. Recognition of Chelyorchis, due to its floral distinctiveness within this clade, would result in a paraphyletic Rossioglossum. The genus ranges mostly from Mexico to Central America, with Chelyorchis pardoi Carnevali & G. A. Romero extending further south to Trinidad and Tobago, Colombia, and Venezuela (Fernandez-Concha, et al., 2009). This species currently lacks a combination in Rossioglossum.

<i>Rossioglossum inslayei</i>
Ros. inslayei

<i>Rossioglossum grande</i>
Ros. grande

<i>Rossioglossum schlieperianum</i>
Ros. schlieperianum

<i>Rossioglossum oerstedii</i>
Ros. (Ticoglossum) oerstedii

<i>Rossioglossum oerstedii</i>
Ros. (Ticoglossum) oerstedii

<i>Rossioglossum oerstedii</i>
Ros. (Ticoglossum) oerstedii

<i>Rossioglossum ampliatum</i>
Ros. (Chelyorchis) ampliatum


Cuitlauzina Lex. (10 spp.; Fig. 5), as circumscribed here, includes Dignathe Lindl., Osmoglossum (Schltr.) Schltr., and Palumbina Rchb.f. and collectively ranges from Mexico to Panama in Central America. Because floral morphology is so divergent within this genus, the close relationships between Cuitlauzina s.s., Palumbina, Dignathe, and Osmoglossum were previously unsuspected, although Ayensu and Williams (1972) showed that Palumbina and Osmoglossum shared some leaf anatomical features, and Williams (1972b) mentioned that there were similarities between the pollinaria of Palumbina and Osmoglossum, as well as vegetative and floral similarities. All four genera were segregated by various authors from Odontoglossum. Cuitlauzina pendula Lex. has a tabula infrastigmatica, but its pollinator is unknown; its color (white or pink) makes it unlikely to be an oil-bee flower. In spite of their gross floral disparity, they share a prominent clinandrial hood and similar pollinarium morphology (Sosa, et al., 2001).

<i>Cuitlauzina</i>
Cuit. pendula white form

<i>Cuitlauzina</i>
Cuit. pendula pink form

<i>Cuitlauzina</i>
Cuit. (Palumbina) candida

<i>Cuitlauzina</i>
Cuit. (Palumbina) candida

<i>Cuitlauzina</i>
Cuit. (Palumbina) candida

<i>Cuitlauzina</i>
Cuit. (Osmoglossum) pulchella


Grandiphyllum Docha Neto (10 spp.; Fig. 1C; Fig. 5; Brazilian "mule-ear" oncidiums) is restricted to Brazil and northern Argentina, and the species were formerly placed as members of two sections of Oncidium. They have large leathery leaves and floral morphology typical of Oncidium with an oil-bearing callus or dense pad of trichomes and a tabula infrastigmatica, but they lack the complex tubularized pollinarium stipe (Chase, 1986b) typical of Oncidium s.s. Except for the placement of Saundersia (Fig. 5), these might have better been included in Trichocentrum; however, this also would have involved transferring Saundersia, which seems to share little with the other two clades.

<i>Grandiphyllum auriculatum</i>
Grandiphyllum auriculatum


Saundersia Rchb.f. (2 spp.; Fig. 5) is restricted to Brazil. These small plants have relatively leathery "mule-ear" leaves and small flowers borne in a dense pendent raceme with a short column that lacks a tabula infrastigmatica. The roots, ovary, and sepals bear dense indumentum, a feature unique within this clade and rare in the entire subtribe (but found in some species of Ornithocephalus, which is not closely related; Fig. 6).

Trichocentrum Poepp. & Endl. (approximately 70 spp.; Figs. 1D,E,F, 2V; Fig. 5), as broadly circumscribed by Chase (2009b), also includes Lophiaris Raf. ("mule-ear" oncidiums), Cohniella Pfitzer ("rat-tail" oncidiums), and Lophiarella Szlach., Mytnik & Romowicz (Trichocentrum microchilum (Bateman ex Lindl.) M.W. Chase & N.H. Williams and T. pumilum (Lindl.) M.W. Chase & N.H. Williams). This clade also includes great floral diversity, but the species are linked by vegetative succulence. The leaves are thick and leathery, and in one clade the leaves are terete ("rat-tail" oncidiums).

Most species have yellow to brown flowers that are either true oil-rewarding or resin-rewarding species: T. stipitatum (Lindl. ex Benth.) M.W. Chase & N.H. Williams, visited by Centris and Paratetrapedia bees (Silvera, 2002); T. ascendens (Lindl.) M.W. Chase & N.H. Williams, pollinated by Trigona and Centris (Parra-Tabla, et al., 2000), and some are oil deceit-flowers. Species of Trichocentrum s.s. typically have a spur (Fig. 2V), although nectar has never been observed. Most Trichocentrum s.s. with spurs might be deceit flowers, attracting nectar-foraging euglossine bees or other long-tongued bees. At least one species, T. tigrinum Linden & Rchb.f., has a strong fragrance and attracts fragrance-collecting male euglossine bees (van der Pijl & Dodson, 1966). Chromosome number varies greatly within this clade, forming a continuum from 2n=24 to 2n=72 that does not correlate well with subclades. Chase and Olmstead (1988) hypothesized that the range of numbers is the result of chromosomal condensation and does not involve polyploidy.

Some workers (Braem, 1993; Christenson, 1999; Fernandez-Concha, et al., 2010) favor a narrow circumscription of Trichocentrum (restricted to those species with a spur) and recognition of Lophiaris and Cohniella. These generic segregates are monophyletic with respect to our molecular data if one species of Lophiarella (T. pumilum (Lindl.) M.W. Chase & N.H. Williams) is included in Lophiaris, but Lophiarella should also include T. flavovirens (L.O.Williams) M.W. Chase & N.H. Williams and T. splendidum (A.Rich. ex Duch.) M.W. Chase & N.H. Williams, if Lophiarella is to be monophyletic. Chase (2009b) argued for lumping all these into a broader Trichocentrum on the basis of pollinarium and vegetative characters (Sandoval-Zapotitla & Terrazas, 2001), which also avoids recognition of a large number of genera.

<i>Trichocentrum splendidum</i>
Tric. (Onc.) splendidum

<i>Trichocentrum splendidum</i>
Tric. (Onc.) splendidum

<i>Trichocentrum carthaginensis</i>
Tric. (Onc.) carthaginensis

<i>Trichocentrum nanum</i>
Tric. (Onc.) nanum

<i>Trichocentrum jonesianum</i>
Tric. (Onc.) jonesianum

<i>Trichocentrum</i>
Tric.(Onc.) sp.

<i>Trichocentrum ascendens</i>
Tric. (Onc.) ascendens

<i>Trichocentrum cebolleta</i>
Tric. (Onc.) cebolleta

<i>Trichocentrum stipitatum</i>
Tric. (Onc.) stipitatum

<i>Trichocentrum lanceanum</i>
Tric. (Onc.) lanceanum

<i>Trichocentrum pumilum</i>
Tric. (Onc.) pumilum

<i>Trichocentrum tigrinum</i>
Tric. tigrinum

<i>Trichocentrum tigrinum</i>
Tric. tigrinum

<i>Trichocentrum brandtiae</i>
Tric. brandtiae

<i>Trichocentrum capistratum</i>
Tric. capistratum

<i>Trichocentrum candidum</i>
Tric. candidum

<i>Trichocentrum panamense</i>
Tric. panamense

<i>Trichocentrum panamense</i>
Tric. panamense

<i>Trichocentrum panamense</i>
Tric. panamense


Lockhartia Hook. (35 spp.; Figs. 1H, 2F; Fig. 5) has confused orchidologists for decades and has been placed in a number of suprageneric taxa. The genus ranges throughout much of the Neotropics. The flowers are mostly bright yellow and bear oil-secreting trichomes, similar to many others in Oncidiinae, but they lack a tabula infrastigmatica. The pollinaria have elongate caudicles that partially replace a stipe (similar to Pachyphyllum = Fernandezia), and all but one species have a "braided" vegetative habit with pseudomonopodial stems lacking pseudobulbs and tightly overlapping, unifacial, non-articulate leaves. The capsules have apical dehiscence instead of lateral dehiscence. These unusual features led some workers to place Lockhartia in a separate subtribe, Lockhartiinae Schltr., but molecular data strongly support its position within Oncidiinae. The unusual vegetative features are best explained as paedomorphic traits common to many seedlings of Oncidiinae (Chase, 1986b). One species (L. genegeorgei D.E. Benn. & Christenson) has prominent pseudobulbs with articulated, bifacial leaves; the lack of paedomorphic traits in this species led Senghas (2001) to describe a new genus, Neobennettia Senghas. We were unable to obtain a DNA sample of this taxon for inclusion in our analyses, but we feel its segregation into a monotypic genus is unwarranted. It may be a natural intergeneric hybrid between Lockhartia (probably L. lepticaula D.E. Benn. & Christenson) and a species of Oncidium or Vitekorchis; the elongate, non-bifid pollinarium stipe of L. genegeorgei is very different from that of other Lockhartias.

<i>Lockhartia oblongicallosa</i>
Lock. oblongicallosa

<i>Lockhartia hercodonta</i>
Lockhartia hercodonta

<i>Lockhartia bennettii</i>
Lockhartia bennettii

<i>Lockhartia acuta</i>
Lockhartia acuta

<i>Lockhartia biserra </i>
Lockhartia biserra

<i>Lockhartia goyazensis</i>
Lockhartia goyazensis

<i>Lockhartia grandibractea</i>
Lock. grandibractea

<i>Lockhartia oerstedii</i>
Lockhartia oerstedii

<i>Lockhartia longifolia</i>
Lockhartia longifolia

<i>Lockhartia chocoensis</i>
Lock. cf. chocoensis

<i>Lockhartia lunifera</i>
Lockhartia lunifera

<i>Lockhartia micrantha</i>
Lockhartia micrantha

<i>Lockhartia obtusata</i>
Lockhartia obtusata

<i>Lockhartia serra</i>
Lockhartia serra

<i>Lockhartia pittieri</i>
Lockhartia pittieri

<i>Lockhartia parthenocomos</i>
Lock. parthenocomos

<i>Lockhartia amoena</i> sp.
Lockhartia amoena

<i>Lockhartia verrucosa</i> sp.
Lockhartia verrucosa

<i>Lockhartia</i> sp.
Lockhartia sp. nov.


The following seven genera include taxa formerly segregated from Oncidiinae as the monopopodial subtribes Pachyphyllinae (pollinia with two long stipes/caudicles) and Ornithocephalinae (four pollinia).

Fernandezia Lindl. (approximately 50 spp.; Figs. 1I, 2P; Fig. 6) has recently been re-circumscribed to include both Pachyphyllum Kunth and Raycadenco Dodson (Chase & Whitten, 2011). The monotypic Raycadenco has yellow and brown flowers with a tabula infrastigmatica typical of many oil-bee pollinated species of Oncidium, but the plants are monopodial (and therefore lack pseudobulbs), a habit shared with others in this clade. Raycadenco is sister to Fernandezia (including Pachyphyllum). These two genera (Fernandezia and Pachyphyllum) were previously distinguished on the basis of flower size and color. Pachyphyllum has tiny white, pink, or yellow flowers for which pollinators are unknown, whereas Fernandezia s.s. has larger flowers that are bright red or orange and are hummingbird pollinated. The two genera are not reciprocally monophyletic in our trees, lending support to our decision to lump Pachyphyllum into Fernandezia. Given the rampant parallelism in floral morphology and in particular the frequent occurrence of oil-bee flowers in Oncidiinae, it makes no sense to keep Raycadenco separate just because it has oil-bee flowers when we disregard different pollination syndromes in other genera (e.g., Cyrtochilum, Gomesa, Oncidium, etc.).

<i>Fernandezia</i>
Frn. subbiflora

<i>Fernandeziasubbiflora</i>
Frn. subbiflora

<i>Fernandezia subbiflora</i>
Frn. subbiflora

<i>Fernandezia ionanthera/i>
Frn. ionanthera

<i>Fernandezia sanguinea/i>
Frn. sanguinea

<i>Pachyphyllum cuencae</i>
Frn. (Pachyphyllum) cuencae

<i>Pachyphyllum hispidulum</i>
Frn. (Pachyphyllum) hispidulum

<i>Pachyphyllum</i> sp.
Frn. (Pachyphyllum) sp.

<i>Fernandezia</i>
Frn. (Pachyphyllum) sp.

<i>Pachyphyllum</i> sp.
Frn. (Pachyphyllum) sp.

<i>Fernandezia</i>
Frn. (Pachyphyllum) sp.

<i>Fernandezia</i>
Frn. (Raycadenco) ecuadorensis


The genera we sampled comprising the former Ornithocephalinae are monophyletic in our trees, although several are represented by only a single sample (Figs. 2K,L M,N; Fig. 6): Phymatidium Lindl. (10 spp.), Zygostates Lindl. (20 spp.), Chytroglossa Rchb.f. (3 spp.), Eloyella P.Ortiz (7 spp.), Hintonella Ames (1 sp.), and Ornithocephalus Hook. (50 spp.). These genera possess tiny green to yellow to white flowers that secrete oil via labellar elaiphores and are pollinated by smaller genera of oil-collecting bees (Buchmann, 1987). Toscano de Brito and Dressler (2000) transferred all species of Sphyrastylis Schltr. into Ornithocephalus, and Dipteranthus Barb. Rodr. is not separable from Zygostates (Chase, 2009b). Genera of the former Ornithocephalinae not sampled in our study include Centroglossa Barb. Rodr. (5 spp.), Caluera Dodson & Determann (3 spp.), Rauhiella Pabst & Braga (3 spp.), Platyrhiza Barb. Rodr. (1 sp.), and Thysanoglossa Porto & Brade (2 spp.). An unpublished analysis of nrITS data (Toscano de Brito, pers. comm.) shows that Centroglossa is embedded within Zygostates, and thus these two should be merged. The new combinations in Zygostates have not been made and Centroglossa was still listed in the Kew World Checklist of Monocotyledons when this manuscript was revised (17 August 2011). His (Toscano de Brito) results also confirm the monophyly and inclusion in this clade of the other four genera.

<i>Ornithocephalus</i> 3264
Ornithocephalus sp. plant

<i>Ornithocephalus</i>
Ornithocephalus sp.

<i>Ornithocephalus</i>
Ornithocephalus sp.

<i>Ornithocephalus</i> sp.
Orn. (Sphyrastyllis) sp.

<i>Ornithocephalus</i>
Orn. (Sphyrastyllis) sp.

<i>Ornithocephalus</i> sp.
Orn. (Sphyrastyllis) sp.

<i>Phymatidium</i>
Phymatidium tillansioides

<i>Ornithocephalus</i>
Chytroglossa marileonniae

<i>Zygostates apiculata</i>
Chytroglossa marileonniae

<i>Zygostates apiculata</i>
Chytroglossa marileonniae

<i>Zygostates apiculata</i>
Chytroglossa marileonniae

<i>Zygostates apiculata</i>
Chytroglossa marileonniae


Hofmeisterella Rchb.f. (1 sp.; Fig. 6), Trichoceros Kunth (9 spp.; Fig. 2U; Fig. 6), and Telipogon Kunth (170 spp.; Fig. 6) include species formerly regarded as subtribe Telipogoninae on the basis of four pollinia (vs. two in Oncidiinae) and pseudocopulatory flowers (that are pollinated by male tachinid flies) with furry columns and lip calli. Within this clade, monotypic Hofmeisterella is sister to Trichoceros (high elevation species with thick, succulent leaves, and pseudobulbs) and Telipogon (intermediate to high elevation species with thin leaves with reduced or absent pseudobulbs). Previous molecular studies of this clade showed that Stellilabium Schltr. is biphyletic and embedded within Telipogon. One Central American clade of Stellilabium is sister to a Central American clade of Telipogon, and these are embedded in a South American grade (Williams, Whitten, & Dressler, 2005).

<i>Hofmeisterella eumicroscopica</i>
Hof. eumicroscopica

<i>Trichoceros</i>
Trichoceros sp.

<i>Trichoceros</i>
Trichoceros antennifer

<i>Telipogon andicola</i>
Telipogon andicola

<i>Telipogon</i> sp.
Telipogon andicola

<i>Telipogon klotzscheanus</i>
Tel. klotzscheanus

<i>Telipogon</i> sp.1
Telipogon sp.1

<i>Telipogon</i> sp.
Telipogon sp. 2

<i>Telipogon</i> sp.
Tel. (Stellilabium) sp.

<i>Telipogon (Stellilabium)</i> sp.
Tel. (Stellilabium) sp.

<i>Telipogon (Stellilabium)</i> sp.
Tel. (Stellilabium) sp. plant


Vitekorchis Romowicz & Szlach. (4 spp; Fig. 1J; Fig. 7) is an Andean genus that is sister to Oncidium in our trees but without strong BS support. The floral similarity to Oncidium and chromosome counts of 2n=56 are evidence supporting their lumping into Oncidium, but without stronger molecular support we prefer to maintain generic status for this clade at present. Their most distinguishing features are relatively large, sharply-ridged pseudobulbs with numerous subtending leaves, massive inflorescences, and small stipes relative to the pollinia. Our circumscription of Vitekorchis differs greatly from that of Szlachetko. His circumscription includes several species that should be retained in Oncidium (O. boothianum Rchb.f., O. iricolor Rchb.f., and O. obryzatum Rchb.f., among others).

<i>Vitekorchis excavata</i>
Vitekorchis excavata


Oncidium Sw. (approximately 520 spp; Figs. 1I-J,O,S; 2B; Fig. 7, Fig. 8), as broadly circumscribed here, includes many previously recognized genera such as Cochlioda Lindl., Mexicoa Garay, Miltonioides Brieger & Lückel, Odontoglossum Kunth, Sigmatostalix Rchb.f., Solenidiopsis Senghas, and Symphyglossum Schltr., as well as a number of recent, minor segregates such as Chamaeleorchis Senghas & Lückel, Collare-stuartense Senghas & Bockemühl, and Heteranthocidium Szlach., Mytnik, & Romowicz. With this broad circumscription, it is the largest genus of the subtribe. Oncidium species range from Mexico and Florida through the Caribbean, Central America south to Bolivia and Peru, with only one species in Brazil (O. baueri Lindl.). There are many chromosome counts of 2n=56 (Tanaka & Kamemoto, 1984).

The circumscription of Oncidium has been highly contentious, especially among horticulturalists. For many years, the angle of attachment of the lip to column was used to distinguish Oncidium from Miltonia and Odontoglossum, but such angles form a continuum and use of this single-character to define genera resulted in highly artificial classifications, as was pointed out by Dressler and Williams (1975). Oncidium is perhaps the best example of our contention that floral morphology must be foregone in Oncidiinae as a basis for generic characters. Floral traits in Oncidiinae are highly plastic and subject to shifts in pollinators. The traditional emphasis on floral features has resulted in many polyphyletic genera. Nearly fifty years ago, Garay (1963) admitted the artificiality of many generic boundaries within Oncidiinae: "To the taxonomist as well as the horticulturalist, it appears to be a serious and unpleasant thought to unite all these genera with Oncidium, but this course seems to be inevitable, since the information gained from experiments in hybridization and from cytological studies strongly points in that direction."

And as Dodson stated in Orchid Flowers: Their Pollination and Evolution, van der Pijl and Dodson, 1966, page 94:

“A point which has been generally overlooked in taxonomy in the orchids is that the characters which result from adaptation to bird-pollination are often striking. These characters are commonly employed by taxonomists in separating genera, with the result that closely related species may be placed in distinct genera. Examples are the Cochlioda-Odontoglossum-Oncidium and the Sophronitis-Laelia-Cattleya complexes where the enormous number of artificial hybrids are mute evidence of the failure of taxonomists to understand the ecological background of speciation in these groups.”

We feel it is better to use vegetative features in combination with a few floral traits to define broader genera. The molecular analyses demonstrate the high levels of homoplasy in pollinator-related traits. Most members of Oncidium s.s. are characterized by flowers that are adapted for pollination by relatively large oil-collecting bees (e.g., Centris), and many species possess prominent elaiophores on the side lobes of the lip together with a tabula infrastigmatica (Figs. 2I,J). Cochlioda and Symphyglossum represent adaptations for hummingbird pollination, with bright red/pink/purple tubular flowers (Fig. 2S).

The lumping of Sigmatostalix into Oncidium seems initially inappropriate, but the vegetative habit of the two taxa differs only in size, and the flowers of Sigmatostalix are diminutive relative to most Oncidium species (Fig. 2O), reflecting adaptations to different groups of smaller oil-collecting bees as pollinators. Although many of the traditionally recognized segregate genera are monophyletic in our trees (e.g., Sigmatostalix, one clade of Odontoglossum), they are embedded within a larger clade of Oncidium species with diverse floral morphologies and pollination systems. Recognition of these segregate genera would require creation of many new genera to maintain monophyly, and these new genera would be difficult or impossible to diagnose using floral or vegetative traits.

<i>Onc. hastatum</i>
Onc. hastatum

<i>Onc. hyphaematicum</i>
Onc. hyphaematicum

<i>Onc. incurvum</i>
Onc. incurvum

<i>Onc. obryzatoides/i>
Onc. obryzatoides

<i>Onc. leuchochilum</i>
Onc. leuchochilum

<i>Onc. ornithorhynchum/i>
Onc. ornithorhynchum

<i>Onc. fuscatum</i>
Onc. fuscatum

<i>Onc. karwinskii</i>
Onc. karwinskii

<i>Onc. maculatum</i>
Onc. maculatum

<i>Onc. ornithorhynchum</i>
Onc. ornithorhynchum

<i>Onc.</i>
Onc. sp.

<i>Onc.</i>
Onc. sp.

<i>Onc.</i>
Onc. (Odont.) sp.

<i>Onc.</i>
Onc. (Odont.) sp.

<i>Onc.</i>
Onc. (Odont.) sp.

<i>Onc. epidendroides</i>
Onc. (Odont.) epidendroides

<i>Onc. cristatellum</i>
Onc. (Odont.) cristatellum

<i>Onc. (Cochlioda)</i>
Onc. (Cochlioda) roseum

<i>Onc. (Cochlioda) noezlianum</i>
Onc. (Cochlioda) noezlianum

<i>Onc. (Cochlioda) vulcanicum</i>
Onc. (Cochlioda) vulcanicum

<i>Onc.</i> sp.
Onc. sp.

<i>Onc. zelenkoanum</i>
Onc. zelenkoanum

<i>Onc. cheirophorum</i>
Onc. cheirophorum

<i>Onc. cheirophorum</i>
Onc. cheirophorum

<i>Onc. boothianum</i>
Onc. boothianum

<i>Onc. portmannii</i>
Onc. (Odont.) portmannii

<i>Onc. schroederianum</i>
Onc. (Odont.) schroederianum

<i>Onc. (Symphyglossum sanguineum) strictum</i>
Onc. (Symphyglossum sanguineum) strictum

<i>Onc. (Sigmatostalix)</i>
Onc. (Sigmatostalix) sp.

<i>Onc. dilitatum</i>
Onc. (Sigmatostalix) dilitatum

<i>Onc. dilitatum</i>
Onc. (Sigmatostalix) dilitatum

<i>Onc. picturatissimum </i>
Onc. (Sigmatostalix) picturatissimum

<i>Onc. adamsii</i>
Onc. (Sigmatostalix) adamsii

<i>Onc. gramineum</i>
Onc. (Sigmatostalix) gramineum

<i>Onc. (Sigmatostalix) hirtzoides</i>
Onc. (Sigmatostalix) hirtzoides

<i>Onc. (Sigmatostalix)</i>
Onc. (Sigmatostalix) sp.

A few species of Oncidium (e.g., O. abortivum Rchb.f., O. echinops Königer, O. heteranthum Poepp. & Endl.; Fig. 7) produce branched inflorescences with terminal normal flowers on the branches, but the proximal flowers are abortive and sterile, consisting of only a cluster of yellow tepals that function as osmophores (W.M. Whitten, pers. obs.). In other species (O. pentadactylon Lindl.), abortive flowers are terminal, with all other proximal flowers being normal. Szlachetko, Mytnik-Ejsmont, & Romowicz (2006) described Heteranthocidium to accommodate these species, but their genus is not monophyletic in our trees. Moreover, several of the 15 species they placed in the genus do not possess dimorphic flowers and are widely scattered in our trees (e.g., O. boothianum, O. exalatum Hágsater, O. fuscans Rchb.f., O. iricolor Rchb.f.). All heteranthous species sampled here form a clade of 16 accessions (O. retusum Lindl. to O. heterodactylum Kraenzl., Fig. 7), but not all the species in this clade bear dimorphic flowers (O. retusum, O. cultratum Lindl., O. lancifolium Lindl. ex Benth.). Species delimitation is difficult within this clade, and there appears to have been multiple loss or gains of the heteranthous trait, coupled with its erratic phenotypic expression.

<i>Onc. abortivum</i>
Onc. abortivum

<i>Onc. lepturum</i>
Onc. lepturum

<i>Onc. ariasii</i>
Onc. ariasii

<i>Onc. ariasii</i>
Onc. ariasii

<i>Onc. cultratum</i>
Onc. cultratum

<i>Onc. echinops</i>
Onc. echinops

<i>Onc. echinops</i>
Onc. echinops

<i>Onc. echinops</i>
Onc. echinops

<i>Onc. heterodactylum</i>
Onc. heterodactylum

<i>Onc. heteranthum</i>
Onc. heteranthum

<i>Onc. lancifolium</i>
Onc. lancifolium


Otoglossum (Schltr.) Garay & Dunst. (13 spp; Figs. 1S,T; Fig. 9) was originally regarded as a subgenus of Odontoglossum by Schlechter, but the floral characters agree most closely with Oncidium. Distribution is primarily Andean, extending north to Costa Rica, with one species on tepuis of the Guyanan shield. It was probably their large, bright reddish-brown flowers and occurrence at higher elevations that caused them to be placed in Odontoglossum. As broadly circumscribed here, Otoglossum includes Oncidium sect. Serpentia (Kraenzl.) Garay, Brevilongium Christenson, and Ecuadorella Dodson & G.A. Romero. Prior to molecular data, a close relationship between Otoglossum s.s. and Oncidium sect. Serpentia was totally unsuspected. Otoglossum s.s. bear many-flowered inflorescences arising laterally from pseudobulbs widely spaced on woody rhizomes (Jenny, 2010), whereas Oncidium sect. Serpentia exhibits a unique vining habit (many meters long) that was interpreted by Christenson (2006) as an indeterminate inflorescence that periodically produces flowering plantlets at the nodes. We regard these elongate, vining structures as stems, not inflorescences, making their habit the same as in Otoglossum s.s. The molecular data strongly support Oncidium sect. Serpentia and Otoglossum s.s as sister taxa, and together they are sister to Otoglossum harlingii (Stacy) N.H. Williams & M.W. Chase, an unusual former Oncidium with an odd upright habit with long internodes and dichotomously forking woody rhizomes. Dodson and Romero created a monotypic genus (Ecuadorella) for this taxon. The inclusion of all these clades in Otoglossum reveals elongate rhizomes as a local synapomorphy for the genus (this trait occurs elsewhere in Oncidiinae, e.g., some species of Cyrtochilum, to which Otoglossum is closely related).

<i>Otoglossum coronarium</i>
Otoglossum coronarium

<i>Otoglossum scansor</i>
Otoglossum scansor

<i>Otoglossum harlingii</i>
Otoglossum harlingii

<i>Otoglossum scansor</i>
Oto. sp. plant

<i>Otoglossum scansor</i>
Oto. sp. plant

<i>Otoglossum</i>sp.
Otoglossum sp.

<i>Otoglossum</i>
Otoglossum sp.

<i>Otoglossum</i>
Otoglossum sp.

<i>Otoglossum</i>sp.
Otoglossum sp.


Cyrtochiloides N.H. Williams & M.W. Chase (4 spp.; Fig. 1J, Fig. 9) flowers have typical Oncidium-like morphology and were considered members of Oncidium until molecular data revealed their distinctiveness (Williams, et al., 2001b). Florally, they are only divergent from Oncidium in their pollinaria with smaller stipes, larger pollinia, and well-developed, stalked caudicles. The generic names alludes to the vegetative similarity of the plants to Cyrtochilum. Cyrtochilum and Cyrtochiloides both have ovoid pseudobulbs rounded in cross-section (not angled) with 2-6 leaf-bearing subtending bracts.

<i>Cyrtochiloides riopalenqueana</i>
Cyrto. riopalenqueana

<i>Cyrtochiloides flexuosum</i>
Cyrto. flexuosum

<i>Cyrtochiloides flexuosum</i>
Cyrto. flexuosum


Miltoniopsis God.-Leb. (5 spp.; Fig. 9) was split from Miltonia and the name reflects their similar floral shapes. The species of Miltoniopsis are distributed from Central America, Venezuela south to Peru, but absent from Brazil, whereas species of Miltonia are predominately Brazilian (and all are non-Andean). The flowers of Miltoniopsis have broad, flat lips, and at least one species is reported to be pollinated by night-flying ptiloglossine bees (Ptiloglossa ducalis Smith; Dodson, 1965), rather than by oil-collecting anthophorid bees.

<i>Miltoniopsis endresii</i>
Miltoniopsis roezlii

<i>Miltoniopsis roezlii</i>
Miltoniopsis roezlii

<i>Miltoniopsis vexillaria</i>
Miltoniopsis vexillaria

<i>Miltoniopsis vexillaria</i>
Miltoniopsis vexillaria

<i>Miltoniopsis</i>
Miltoniopsis bismarkii


Caucaea Schltr. (5-20 spp.; Fig. 9) was previously known as the Oncidium cucullatum Lindl. group, a set of poorly defined, high-elevation Andean species with showy flowers. Their phylogenetic distance from Oncidium and their relationships to the small-flowered, monotypic Caucaea radiata (Lindl.) Mansf. were unsuspected until molecular data revealed their close relationship (Williams, et al., 2001b) and they were lumped into Caucaea. In spite of the floral similarity to Oncidium, they are not closely related. Caucaea is sister to Cyrtochilum, a relationship that was unexpected on the basis of gross floral shape. The two genera do share subtle traits, including pseudobulbs that are rounded (not strongly ancipitous or two-sided) and pollinaria with relatively short stipes and large caudicles. Both genera also occur in cool, high-elevation Andean cloud forests.

<i>Caucaea radiata</i>
Caucaea radiata

<i>Caucaea andigena</i>
Caucaea andigena

<i>Caucaea phalaenopsis</i>
Caucaea phalaenopsis

<i>Caucaea nubigena</i>
Caucaea nubigena

<i>Caucaea rhodosticta</i>
Caucaea rhodosticta

<i>Caucaea olivaceum</i>
Caucaea olivacea

<i>Caucaea sanguinolenta</i>
Caucaea sanguinolenta


Cyrtochilum Kunth (approximately 120 spp.; Figs. 2C,E,G,H; Fig. 9) is restricted to the high Andes of Colombia and Venezuela south to Peru, with a single species, C. meirax (Rchb.f.) Dalström (and a putative related species) occurring in the Caribbean (Dalström, 2001). Many species have long (3-4 m), vining inflorescences and large showy flowers (some with prominent elaiophores; Figs. 2G,H), but a few species have diminutive plants and flowers. Vegetatively, Cyrtochilum are distinguished by dull pseudobulbs that are round or ovoid in cross section with 2-4 apical leaves and 2-6 leaf-bearing bracts and relatively thick roots; in contrast, Oncidium species have glossy, ancipitous (2-edged) pseudobulbs and thin roots (Dalström, 2001). Dalström (2001) and Chase (2009b) discussed the tangled taxonomic history of the genus. Previous workers relied almost exclusively on floral traits, resulting in confusion with concepts of Odontoglossum and Oncidium.

Lindley, in a series of transfers over a period of years (1837-1842) in Sertum Orchidaceum, eventually sunk both Odontoglossum and Cyrtochilum into Oncidium. Kraenzlin resurrected the genus Cyrtochilum in 1922. Dasyglossum Königer & Schildhauer and Trigonochilum Königer & Schildhauer were created to accommodate some of the smaller flowered Cyrtochilum species, but the authors repeatedly transferred taxa between the two genera because they could not decide where they fit on the basis of floral morphology. Senghas (1997) transferred all Dasyglossum into Trigonochilum because he could not reliably distinguish them. Neither genus is monophyletic in our DNA trees. Similarly, Buesiella C. Schweinf., Neodryas Rchb.f., Rusbyella Rolfe ex Rusby, and Siederella Mytnik, Górniak, & Romowicz are simply diminutive and/or brightly colored taxa embedded within Cyrtochilum (Dalström, 2001), probably reflecting a shift in pollinators, although there are few observations of pollination.

<i>Cyrtochilum macranthum</i>
Cyrt. macranthum

<i>Cyrtochilum macranthum<</i>
Cyrt. macranthum

<i>Cyrtochilum macranthum<</i>
Cyrt. macranthum

<i>Cyrtochilum edwardii </i>
Cyrtochilum edwardii

<i>Cyrtochilum ioplocon</i>
Cyrtochilum ioplocon

<i>Cyrtochilum ixioides</i>
Cyrtochilum ixioides

<i>Cyrtochilum leopoldianum</i>
Cyrt. leopoldianum

<i>Cyrtochilum leopoldianum</i>
Cyrt. leopoldianum inflorescence

<i>Cyrtochilum</i> sp.
Cyrt. sp. inflorescence

<i>Cytochilum (Dasyglossum)</i> sp.
Cyrt. (Dasyglossum) sp.

<i>Cytochilum ventilabrum</i>
Cyrtochilum ventilabrum

<i>Cytochilum aureum</i>
Cyrtochilum aureum

<i>Cytochilum meirax</i>
Cyrtochilum meirax

<i>Cytochilum lamelligerum</i>
Cyrt. lamelligerum

<i>Cytochilum halteratum</i>
Cyrt. halteratum

<i>Cytochilum monachicum</i>
Cyrt. monachicum

<i>Cytochilum</i>
Cyrt. sp.

<i>Cytochilum</i>
Cyrt. sp.

<i>Cytochilum</i>
Cyrt. sp.

<i>Cytochilum</i>
Cyrt. sp.

<i>Cytochilum</i>
Cyrt. sp.

<i>Cytochilum gargantua</i>
Cyrt. gargantua

<i>Cytochilum gargantua</i>
Cyrt. gargantua

<i>Cytochilum gargantua</i>
Cyrt. gargantua


Miltonia Lindl. (10 spp.; Fig. 10) occurs in Argentina, Brazil, Paraguay, and Venezuela and is sister to a clade that includes Systeloglossum, Oliveriana, Cischweinfia, Aspasia, and Brassia. Some Miltonia species (e.g., M. regnellii Rchb.f. and M. spectabilis Lindl.) have a short column and a broad, flat lip with a simple, reduced callus, but floral morphology varies a great deal among the species. Miltonia clowesii (Lindl.) Lindl. has typical Oncidium-like oil-bee flowers, whereas M. candida Lindl. and M. russelliana (Lindl.) Lindl. have the lip partly or completely encircling the column, giving them the appearance of a Cischweinfia (suggestive of pollination by nectar-foraging bees). They also have the clinandrial and column arms found in many species of Cischweinfia (see below). Miltonia flavescens (Lindl.) Lindl. on the other hand resembles a species of Brassia in its floral traits, with a similar bilobed lip callus forming a nectar-cavity-like chamber on the lip base and elongate, spidery tepals. The above-mentioned species with the author combination "(Lindl.) Lindl." are due to Lindley considering these to be species of Cyrtochilum or Odontoglossum when he first described them, again an indication of the floral diversity present in a small set of species that forms a clade in our analyses. Like M. clowesii, M. phymatochila (Lindl.) N.H. Williams & M.W. Chase also has typical oncidioid oil-bee flowers with a large complex callus and tabula infrastigmatica. The latter species was transferred from Oncidium to Miltonia by Williams, et al. (2001b) and subsequently transferred to a monotypic genus, Phymatochilum, by Christenson (2005), who cited it as an aberrant member of Miltonia (a virtual "round peg in a square hole"; E. A. Christenson, pers. comm.), but in our view it is no more or less aberrant than the other species with unusual floral traits found in Miltonia.

<i>Miltonia</i>
Miltonia sp.

<i>Miltonia candida</i>
Miltonia candida

<i>Miltonia clowesii</i>
Miltonia clowesii

<i>Miltonia phymatochila</i>
Miltonia phymatochila


Sister to Miltonia is a clade of the following three genera with relatively small flowers that have a prominent clinandrial hood on the column and strongly ancipitous pseudobulbs:

Systeloglossum Schltr. (5 spp.; Fig. 10) has small, purple or yellow-green flowers with a prominent column foot and a simple hinged lip; pollination is presumably by nectar-foraging insects. Szlachetko (2006) created the monotypic Diadeniopsis Szlach. for Systeloglossum bennetii (Garay) Dressler & N.H. Williams. His emphasis on and interpretation of gynostemial structure mistakenly placed it in the twig epiphyte clade as a relative of Comparettia.

<i>Systeloglossum bennetii</i>
Systeloglossum bennetii

<i>Systeloglossum sp.</i>
Systeloglossum sp.


Oliveriana Rchb.f. (6 spp.; Fig. 10) is a high-elevation, Andean genus with relatively flat, open flowers, and Chase (2009b) suggested the flowers are pollinated by hummingbirds on the basis of pollinarium morphology (two widely spaced pollinia with a wedge-shaped viscidium and a bilobed stigma, which are otherwise features found in hummingbird-pollinated Oncidiinae species). Plants are scandent, in contrast to the mostly caespitose habit of other genera in this clade.

<i>Oliveriana</i>
Oliveriana brevilabia

<i>Oliveriana.</i>
Oliveriana egregia

<i>Oliveriana</i>
Oliveriana egregia


Cischweinfia Dressler & N.H. Williams (11 spp.; Fig. 10) grows in middle-elevation forests (to 1,500 m) from Costa Rica to Bolivia. Flowers have a tubular lip enfolding the column and are reportedly pollinated by nectar-seeking euglossine bees (Williams, 1982). Cischweinfia pygmaea (Pupulin, J. Valle, & G. Merino) M.W. Chase has diminutive plants with small flowers and a simple lip. It was originally described as an Ada Lindl., but molecular data from this study clarified its generic placement (Chase & Whitten, 2011).

<i>Cischweinfia popowiana</i>
Cis. popowiana

<i>Cischweinfia pusilla</i>
Cis. pusilla

<i>Cischweinfia pusilla</i>
Cis. pusilla

<i>Cischweinfia rostrata</i>
Cis. rostrata

<i>Cischweinfia parva</i>
Cis. parva

<i>Cischweinfia platychila</i>
Cis. platychila

<i>Cischweinfia dasyandra</i>
Cis. dasyandra

<i>Cischweinfia dasyandra</i>
Cis. dasyandra

<i>Cischweinfia dasyandra</i>
Cis. dasyandra

<i>Cischweinfia dasyandra</i>
Cis. dasyandra

<i>Cischweinfia dasyandra</i>
Cis. dasyandra


Aspasia Lindl. (7 spp.; Fig. 10) ranges from Central America, northern South America and the Andes to coastal Brazil. It is vegetatively similar to Brassia, but the flowers have a flat lip partially adnate to a relatively long column and bent at a right angle, forming a false nectary. Several species are pollinated by euglossine bees, although there may be a mixture of nectar deceit and fragrance reward involved, depending upon the species (Zimmerman & Aide, 1987). Aspasia represents the only known occurrence of male euglossine pollination in this clade (Miltonia to Brassia, Fig. 10).

<i>Aspasia epidendroides</i>
Aspasia epidendroides

<i>Aspasia epidendroides</i>
Asp. epidendroides

<i>Aspasia principissa</i>
Aspasia principissa

<i>Aspasia principissa</i>
Aspasia principissa

<i>Aspasia psittacina</i>
Aspasia psittacina

<i>Aspasia psittacina</i>
Aspasia psittacina

<i>Aspasia lunata</i>
Aspasia lunata

<i>Aspasia lunata</i>
Aspasia lunata

<i>Aspasia lunata</i>
Aspasia lunata


Brassia R.Br. (approximtely 74 spp.; Figs. 2Q,R; Fig. 10) includes Ada Lindl., Brachtia Rchb.f., and Mesospinidium Rchb.f. These genera have been difficult to separate on the basis of floral and vegetative characters. Chase (2009b) treated these separately, but he indicated this to be unsatisfactory. Brachtia (7 spp., Andean) is sister to Brassia s.s. (approximtely 35 spp., Mexico through Central America, Caribbean, to tropical South America). The two genera are vegetatively similar and basic pollinarium and floral structures are similar. They share a simple lip with a pair of small basal keels. They differ mainly in the relative size of the flowers and floral bracts; Brachtia (Fig. 2R) has relatively small flowers with large bracts partially enclosing the flowers. These two genera are sister to Ada (approximtely 35 spp.) and Mesospinidium (approximtely 7 spp.), both ranging from Central America south through the Andes to Bolivia. Ada originally was monotypic and composed of a single hummingbird-pollinated species with bright orange to red flowers (Fig. 2Q), but Williams (1972a) realized that it was morphologically similar to a clade of Brassia (the "glumaceous" Brassias), and he transferred this group into Ada, greatly enlarging the genus. Ada is not monophyletic, with Ada allenii (L.O. Williams ex C. Schweinf.) N.H. Williams sister to Mesospinidium and remaining Ada. Florally, Mesospinidium are small versions of Ada. Given the shared suite of floral morphologies and habits and the aberrant phylogenetic position of Ada allenii, lumping them all into Brassia seems the simplest solution.

<i>Brassia caudata</i>
Brassia caudata

<i>Brassia caudata</i>
Brassia caudata

Brassia gireoudiana
Brassia gireoudiana

<i>Brassia gireoudiana</i>
Brassia gireoudiana

Brassia arcuigera
Brassia arcuigera

Brassia arcuigera
Brassia arcuigera

Brassia maculata
Brassia maculata

Brassia jipijapaensis
Brassia jipijapaensis

Brassia
Brassia (Ada) aurantiaca

Brassia allenii
Brassia (Ada) allenii

Brassia allenii
Brassia (Ada) allenii

Brassia elegantula
Brassia (Ada) elegantula

Brassia chlorops
Brassia (Ada) chlorops

<i>Brassia incantans</i>
Brassia (Mesospinidium) incantans

<i>B. andina</i>
Brassia (Brachtia) andina

<i>B. andina</i>
Brassia (Brachtia) andina


The sister relationship between the following two morphologically divergent genera was unsuspected prior to molecular studies. These genera are remarkably different in size, habit, and floral morphology.

Erycina Lindl. (10 spp.; Fig. 1U; Fig. 10), as broadly defined by Williams et al. (2001a), includes Psygmorchis Dodson & Dressler and monotypic Stacyella Szlach. [=Erycina crista-galli (Rchb. f.) N.H. Williams & M.W. Chase]. All three genera have bright yellow oil reward/deceit flowers and were at one time considered members of Oncidium. Although these three genera could be maintained, we favor lumping them to emphasize their similar floral morphology and modified habit (absence of an apical leaf on pseudobulbs, if pseudobulbs are present). Dodson (2003) reported the pollination of Ery. glossomystax by Epicharis rustica Friese, an oil-collecting bee closely related to the large genus Centris.

<i>Erycina echinata</i>
Erycina echinata

<i>Erycina pusilla</i>
Ery. (Psygmorchis) pusilla

<i>Erycina glossomystax</i>
Ery. (Psygmorchis) glossomystax

<i>Erycina zamorensis</i>
Ery. (Psygmorchis) zamorensis

<i>Erycina</i> plant
Erycina plant

<i>Erycina</i> plant
Erycina plant

<i>Erycina pusilla</i>
Comparison of E. pumilio and E. crista-galli


Rhynchostele Rchb.f. (13 spp.; Fig. 10) as circumscribed here is primarily Mexican and includes Amparoa Schltr. and Mesoglossum Halb.; Cymbiglossum Halb. and Lemboglossum Halb. are later synonyms of Rhynchostele. Lumping of these genera is also supported by anatomical similarities (Rojas Leal, 1993). Most of these species were treated as members of Odontoglossum until split out by Halbinger, first as Cymbiglossum and later as Lemboglossum. Morphological analyses by Soto and coworkers (Soto, Salazar, & Rojas Leal, 1993) revealed a close relationship between these species and the much reduced Rhynchostele pygmaea Rchb.f. They transferred all these taxa into Rhynchostele, a move that is supported by our molecular data.

<i>Rhynchostele cordata</i>
Rhynchostele cordata

<i>Rhynchostele rossii</i>
Rhynchostele rossii

<i>Rhynchostele stellata</i>
Rhynchostele stellata

<i>Rhynchostele maculata</i>
Rhynchostele maculata

<i>Rhynchostele aptera</i>
Rhynchostele aptera

<i>Rhynchostele candidula</i>
Rhynchostele candidula

<i>Rhynchostele cervantesii</i>
Rhynchostele cervantesii

<i>Rhynchostele galleotiana</i>
Rhynchostele galleotiana

<i>Rhynchostele madrensis</i>
Rhynchostele madrensis

<i>Rhynchostele majalis</i>
Rhynchostele majalis


Gomesa R.Br. (approximately 125 spp.; Figs. 1P,Q,R; Fig. 11) as circumscribed here is relatively broad and includes at least 23 other generic concepts (Chase, 2009b; Chase, et al., 2009a) with a great diversity of floral morphology and size. Gomesa has a center of distribution in Brazil, especially the Mata Atlântica, where these species largely replace Oncidium (the genus in which most of them were once included), but it extends to northern Argentina and Amazonian Peru. Nearly all species have fused lateral sepals, a trait that makes them easy to recognize in spite of their floral diversity. In contrast, Oncidium is largely absent from Brazil (O. baueri is the sole representative), and their lateral sepals are usually free. The two genera rarely produce hybrids in horticulture.

Based on the enormous floral diversity within Gomesa, Brazilian and French workers have proposed a number of segregates (Docha Neto, Baptista & Campacci, 2006), but several of these are not monophyletic (e.g., Alatiglossum Baptista, Carenidium Baptista, Coppensia Dumort.). Several recent segregates are monotypic: Campaccia venusta (Drapiez) Baptista, P.A. Harding, & V.P. Castro; Hardingia paranaensis (Kraenzl.) Docha Neto & Baptista (not included in our analyses); and Nitidocidium gracile (Lindl.) F. Barros & V.T. Rodriguez. To make matters worse, Szlachetko and colleagues also segregated a number of genera from this same set of species, often using the same type species but including different sets of species than the Brazilian workers (e.g., Concocidium Romowicz & Szlach. and Carenidium, both based on Oncidium concolor Hook.). Also, Szlachetko (2006) segregated three species of Oncidium as the genus Rhinocerotidium Szlach. (O. longicornu Mutel, O. macronyx Rchb.f., O. rhinoceros Rchb.f.; most workers lump these into a single species). He based the genus mostly upon the large, horn-like lip callus, but the lip callus is perhaps the most variable floral feature within Oncidiinae. These species are closely related to G. varicosa (Lindl.) M.W. Chase & N.H. Williams, a species with a relatively large lip and small callus.

Gomeza
Gomesa sp.

Gomeza
Gomesa sp.

Gomeza longipes
Gomesa longipes

Gomeza longipes
Gomesa longipes

Gomeza Gomeza longipes
Gomesa longipes rear

Gomeza flexuosa
Gomesa flexuosa

Gomeza zappii
Gomesa zappii

Gomeza zappii
Gomesa zappii

Gomeza gardneri
Gomesa gardneri

Gomeza imperatoris
Gomesa imperatoris

Gomesa barbata
Gomesa barbata

Gomesa concolor
Gomesa concolor

Gomesa concolor
Gomesa concolor

Gomesa
Gomesa ranifera

Gomeza radicans
G. (Ornithophora) radicans

Gomeza (Rodrigueziella) gomezoides
Gomesa (Rodrigueziella) gomezoides


Capanemia Barb. Rodr. (7 spp.; Fig. 11) is represented in our analyses by only a single species, Capanemia superflua (Rchb.f.) Garay that is sister to Solenidium. Recent studies have reduced the number of species in the genus, but molecular data are needed to confirm the monophyly of the seven recognized species (Buzatto, et al., 2011; Buzatto, Singer & van den Berg, 2010). The genus is centered in southeastern Brazil, extending to Argentina and Uruguay. Florally, the genus is similar to unrelated Leochilus, but most species do not produce nectar, except C. therezae Barb. Rodr. (Buzatto, et al., in press). Singer and Cocucci (1999) reported visits by halictid bees and vespid wasps. Sanderella also falls here (C. van den Berg, pers. comm.). Morphologically, Sanderella Kuntze is similar to Capanemia (the oldest name) and Sanderella should probably be combined with Capanemia. Its exact status cannot be determined until Sanderella and more species of Capanemia are sampled.

Capanemia micromera
Capanemia micromera

Capanemia micromera
Capanemia micromera

Capanemia micromera
Capanemia micromera


Solenidium Lindl. (3 spp.; Fig. 11) is an Amazonian genus and is similar florally to its sister, Capanemia, bearing small flowers with prominent column wings and an upturned tip of the anther cap; more detailed studies of both may support their combination.

Solenidium portillae
Solenidium portillae


Nohawilliamsia M.W. Chase & Whitten (1 sp.; Fig. 1V, Fig. 11) was created to accommodate this single odd species with no close or clear relatives based on our analyses thus far. It was formerly known as Oncidium pirarense Rchb.f. (syn. O. orthostates Ridl.) from southern Venezuela, Guyana, Suriname, and Brazil (Chase, 2009a; Chase, et al., 2009a; Whitten, 2009). Although the flowers are similar to many yellow flowered species of Oncidium, they lack a tabula infrastigmatica. The leaves have a minutely dentate margin, and plantlets (keikis) are produced on old inflorescences and on top of old pseudobulbs; all of these traits are unusual within Oncidiinae.

Nohawilliamsia pirarensis
Nohawilliamsia pirarensis


Notyliopsis P. Ortiz (1 spp.; Fig. 11) from the wet Colombian Chocó has diminutive flowers that superficially resemble those of Notylia Lindl., but the pseudobulbs are reminiscent of Zelenkoa.

<i>Notyliopsis</i>
N. beatricis plant and multiple inflorescences

<i>Notyliopsis</i>
N. beatricis pseudobulbs

<i>Notyliopsis</i>
N. beatricis pseudobulb

<i>Notyliopsis</i>
N. beatricis inflorescence

<i>Notyliopsis</i>
N. beatricis inflorescence

<i>Notyliopsis</i>
N. beatricis inflorescence

<i>Notyliopsis</i>
N. beatricis flower close up

<i>Notyliopsis</i>
N. beatricis flower close up


Zelenkoa M.W. Chase & N.H. Williams (1 sp.; Fig. 1W; Fig. 11) was long considered an oddity when it was included in Oncidium (often in its own monotypic section), but molecular data revealed its distinctiveness. Like Nohawilliamsia, it also has bright yellow flowers that lack a tabula infrastigmatica. Often epiphytic on cacti in dry coastal forests of Ecuador and Peru, the plants have mottled ovoid pseudobulbs that resemble those of Notyliopsis, which is also a member of this grade relative to Tolumnia and other twig epiphytes.

Zelenkoa onusta
Zelenkoa onusta


Tolumnia Raf. (approximately 40 spp.; Figs. 1X,Y, 2D; Fig. 11) has long been recognized as a distinct group ("equitant" oncidiums) based on their psygmoid fan of succulent leaves and usual absence of pseudobulbs. There is extensive polyploidy within the genus (Braem, 1986) resulting in some conflict between nuclear and plastid phylogenetic trees. Most species have oil-bee flowers that do not secrete oil; pollination by Centris bees is reported for several species (Ackerman, Meléndez-Ackerman, & Salguero Faria, 1997; Nierenberg, 1972). Tolumnia guibertiana (A. Rich.) Braem (endemic to western Cuba) is completely dependent on oil-gathering female bees (Centris poecila Lepeletier) for fruit production (Vale, Rojas, & Álvarez, 2011). Tolumnia henekenii (R.H.Schomb. ex Lindl.) Nir has a furry, insect-like lip and is reportedly pseudocopulatory (Dod, 1976). Braem and Garay have published or resurrected several (often monotypic) segregates based on floral oddities; these include Olgasis Raf., Antillanorchis (Wright ex Griesb.) Garay, Hispaniella Braem, Jamaiciella Braem, Braasiella Braem, Lückel, & Russmann, and Gudrunia Braem. Recognition of all these segregates would require at least a dozen genera to be carved from Tolumnia to maintain monophyly. We feel this is unwarranted. Tolumnia is sister to all others in the remainder of the tree (twig epiphytes), but this relationship is only weakly supported. In contrast to most twig epiphytes, Tolumnia species often occur on the larger axes of trees and live for many years, rather than being restricted to terminal twigs with extremely rapid life cycles, but they also have seeds with pronounced hooks or knob-like extensions (Chase, 1988).

<i>Tolumnia calochila</i>
Tolumnia calochila

<i>Tolumnia hawkesiana</i>
Tolumnia hawkesiana

<i>Tolumnia guianensis</i>
T. guianensis yellow form

<i>Tolumnia guianensis</i>
T. guianensis red form

<i>Tolumnia henekenii</i>
Tolumnia henekenii

<i>Tolumnia quadriloba</i>
Tolumnia quadriloba

<i>Tolumnia gundlachii</i>
Tolumnia gundlachii

<i>Tolumnia caymanensis</i>
Tolumnia caymanensis

<i>Tolumnia triquetra</i>
Tolumnia triquetra

<i>Tolumnia triquetra</i>
Tolumnia triquetra

<i>Tolumnia triquetra</i>
Tolumnia triquetra

<i>Tolumnia gauntlettii</i>
Tolumnia gauntlettii

<i>Tolumnia tetrapetala</i>
Tolumnia guttata

<i>Tolumnia variegata</i>
Tolumnia variegata

<i>Tolumnia gauntlettii</i>
T. variegata Cuban form

<i>Tolumnia jamaicensis</i>
Tolumnia jamaicensis


The twig epiphytes--The clade comprising the remainder of the tree (Plectrophora H.Focke to Notylia Lindl., Fig. 12) has been informally referred to as the "twig epiphyte" clade. Chase (1988) first discussed the morphological and life history features that unite these taxa. Twig epiphytes often grow on the smallest branches (≤2.5 cm) in exposed, high-light zones, have rapid life cycles (often reaching maturity in one season), produce hooks or projections on the seed testa (mostly likely for rapid uptake of water), exhibit psygmoid (paedomorphic) habits, and velamen (root epidermis) cells much longer than wide with evenly spaced secondary thickenings. Not all taxa in this clade are extreme twig epiphytes restricted to terminal twig habitats, but the majority display many of these features. Twig epiphytes occur in other clades of Oncidiinae (e.g., Erycina, Fig. 10), as well as in other subtribes [e.g., Dendrophylax porrectus (Rchb. f.) Carlsward & Whitten, Angraecinae]. None of the genera of the twig epiphyte clade (all genera in Fig. 12) secretes oil or mimic oil flowers. Instead, they attract either nectar-seeking animals or are pollinated by fragrance-collecting male euglossine bees. Suarezia Dodson (1 sp.) was not sampled, but it is presumed to be a member of this clade on the basis of its morphology.

Plectrophora H.Focke (9 spp.; Fig. 12) are diminutive plants with relatively large flowers with a funnel-shaped lip and a sepaline spur without nectar horns. The presence of nectar has not been confirmed, but the flowers likely are pollinated by long-tongued insects seeking nectar.

<i>Plectrophora</i>
Plectrophora sp.

<i>Plectrophora</i>
Plectrophora sp. (cross section of flower)


Leochilus Knowles & Westc. (12 spp.; Fig. 12) are true twig epiphytes, occurring only on small branches and twigs. The small flowers of most species have a simple lip with a shallow nectar cavity at the base. Chase (1986a) reported pollination of two species by nectar-foraging, short-tongued Stelopolybia wasps and Lasioglossum bees. Three other monotypic genera are now included in Leochilus on the basis of their position in phylogenetic studies: Goniochilus Chase, Hybochilus Schltr., and Papperitzia Rchb.f. The floral structure of the first two is similar to that of the other species of Leochilus, but that of Papperitzia is highly divergent. In spite of this, the single species of Papperitzia was originally included in Leochilus.

<i>Leochilus (Papperitzia) leiboldii</i>
Leo. (Papperitzia) leiboldii

<i>Leochilus inconspicuus</i>
Leo. (Hybochilus) inconspicuus

<i>Leochilus</i>
Leochilus sp.

<i>Leochilus</i>
Leochilus sp.


Pterostemma Kraenzl. (2 spp.; Fig. 12) are diminutive Andean twig epiphytes with tiny flowers that are probably bee-pollinated. Their habits are monopodial tufted plants or psymoid fans 1-2 cm in size. The flowers have a dorsal anther with a long stipe and long, forward-sweeping column arms. Both sequence data and morphology confirmed a close relationship of Hirtzia Dodson to Pterostemma, so the two were lumped (Chase, Williams, & Whitten, 2009b).

<i>Pterostemma benzingii</i>
Pter. benzingii

<i>Pterostemma</i>
Pter. benzingii plant


Ionopsis Kunth (3 spp.; Fig. 12) ranges widely throughout the Neotropics. The white to pink flowers have a simple lip with a short sepalar spur without any obvious reward and are probably pollinated by nectar-seeking bees.

<i>Ionopsis satyrioides</i>
Ionopsis satyrioides

<i>Ionopsis utricularioides </i>
Ionopsis utricularioides


Comparettia Poepp. & Endl. (approximately 60 spp.; Fig. 2W; Fig. 12) is broadly circumscribed here to include all species with a sepalar nectar spur(s) furnished by a horn or pair of horns on the column base that secrete nectar. Generic segregates lumped here include Chaenanthe Lindl., Diadenium Poepp. & Endl., Neokoehleria Schltr., Pfitzeria Senghas, Scelochiloides Dodson & M.W. Chase, Scelochilopsis Dodson & M.W. Chase, Scelochilus Klotzsch, and Stigmatorthos M.W. Chase & D.E. Bennett. As more species in this clade were discovered in recent years, generic limits have become more obscure, and amalgamation of all taxa with nectar horns into a single genus seems the best solution. Scelochilus appears not to be monophyletic. There is variation in vegetative habit within this clade from psygmoid fans to caespitose plants with bifacial leaves and pseudobulbs. Some species can begin flowering as juvenile psygmoid plants before transformation into adult plants with pseudobulbs, and damage can cause a reversal to the psygmoid seedling habit. Pollination by hummingbirds (Amazalia sp., Chlorostilbon maugaeus) is documented for C. falcata (Dodson, 1965; Salguero-Faria & Ackerman, 1999). Pollination by butterflies and long-tongued bees seems likely for some taxa.

<i>Comp.(Diadenium) micrantha</i>
Comp.(Diadenium) micrantha

<i>Comp. (Scelochilus) heterophylla</i>
Comp. (Scelochilus) heterophylla

<i>Comp. macroplectron</i>
Comp. macroplectron

<i>Comp. (Scelochilus)tungurahuae</i>
Comp. (Scelochilus) tungurahuae

<i>Comp. (Scelochilus) williamsii</i>
Comp. (Scelochilus) williamsii

<i>Comp. speciosa</i>
Comp. speciosa

<i>Comp. falcata</i> sp.
Comp. falcata


Polyotidium Garay (1 sp.; Fig. 12) is reported only from Ecuador, Venezuela, Brazil, and the Orinoco drainage of Colombia. The 5 mm, fleshy, bright orange flowers have a simple lip and a dorsal anther, suggestive of hummingbird pollination. Its phylogenetic position is unresolved within a strongly supported terminal clade that includes Rodriguezia, Sutrina, Macroclinium, and Notylia.


Sutrina Lindl. (2 spp.; Fig. 12) consists of poorly known species from Amazonian Peru and Bolivia. The yellow-green flowers have simple, linear tepals and lip that do not open widely, forming a tube-like structure. Nothing is known of pollination, but morphology suggests pollination by nectar-foraging insects.


Rodriguezia Ruiz & Pav. (approximately 48 spp.; Fig. 2X; Fig. 12) ranges from Mexico south to Argentina, with one species (R. lanceolata) found on many islands in the Caribbean. The flowers are relatively large, brightly colored, and showy for members of the twig epiphyte clade. The lip is often relatively large and flat, and the lateral sepals are fused along one or both lateral margins to form a curved nectar spur. A projection from the column base secretes nectar into this spur. Reported pollinators include hummingbirds, butterflies, and nectar-foraging bees (Dodson, 1965). There are two strongly supported clades within Rodriguezia, and Chase (2009b) noted the non-monophyletic placement of R. decora Rchb.f. in nrITS trees published in Genera Orchidacearum. This unusual Brazilian species was not included our sampling, but it may warrant generic status. It has long, wiry rhizomes between sympodia and lacks the spur found in other species.

<i>Rodriguezia lanceolata</i>
Rodriguezia lanceolata

<i>Rodriguezia lanceolata</i>
Rodriguezia lanceolata

<i>Rodriguezia lanceolata</i>
Rodriguezia lanceolata

<i>Rodriguezia</i>
Rodriguezia lanceolata

<i>Rodriguezia lanceolata</i>
Rodriguezia lanceolata

<i>Rodriguezia lehmanii</i>
Rodriguezia lehmanii

<i>Rodriguezia venusta</i>
Rodriguezia venusta

<i>Rodriguezia venusta</i>
Rodriguezia venusta

<i>Rodriguezia venusta</i>
Rodriguezia venusta

<i>Rodriguezia batemannii</i>
Rodriguezia batemannii

<i>Rodriguezia batemannii</i>
Rodriguezia sp.

<i>Rodriguezia decora</i>
Rodriguezia decora

<i>Rodriguezia decora</i>
Rodriguezia decora


Schunkea Senghas (1 sp.; Fig. 12) is known only from southeastern Brazil; the small cream colored flowers have an open lip and an unusual pair of downward-pointing arms on the column apex. Nothing is known of pollination. Its placement within this clade is unresolved, and Chase (2009b) hypothesized that it might be related to the monotypic Suarezia from eastern Ecuador. Suarezia was not included in our sampling.


Trizeuxis Lindl. (1 sp.; Fig. 12) is wide ranging from Costa Rica south to Peru and also in eastern Brazil. Its flowers are perhaps the smallest of any Oncidiinae, only 2-3 mm across, yet they are outcrossing and often found growing on twigs of cultivated Citrus L. and Psidium L. Pollinators are unknown, but presumed to be small nectar-foraging insects.

<i>Trizeuxis falcata</i>
Trizeuxis falcata

<i>Trizeuxis falcata</i>
Trizeuxis falcata


Seegeriella Senghas (2 spp.; Fig. 12) is restricted to Argentina and Brazil. Like Trizeuxis, the yellow-green flowers are diminutive with a simple lip and sepals that do not open widely. Pollinators are presumed to be nectar-seeking insects.


The remaining four genera are all pollinated by fragrance-collecting male euglossine bees, and all but Warmingia have a narrow, slit-like stigma, pollinaria with a button-like viscidium and a long, narrow stipe, and pollinia that are dorsiventrally flattened and thin to match the opening of the slit-like stigmatic cavity. The narrow pollinia and stigmatic slit probably act to reduce self-pollination. When first removed by a bee, the pollinia are too wide to fit easily into the narrow stigmatic slit (W.M. Whitten, pers. obs.). The stigma widens after pollinarium removal. Several minutes to hours of drying are required to shrink the pollinia before they will slip into the stigma, during which time the bee is likely to have flown to another plant.

Macradenia R. Br. (10 spp.; Fig. 12) ranges from Mexico south throughout most of lowland South America. The pendent, unbranched inflorescence bears numerous flowers that attract fragrance-collecting male euglossine bees. The anther is terminal and beaked, and the column and lip are twisted, giving the flower a distinct asymmetry unique within Oncidiinae. This asymmetry may be related to pollinarium deposition on the side of the bee's head or eye.

<i>Macradenia</i>
Macradenia brassavolae

<i>Macradenia</i>
Macradenia brassavolae

<i>Macradenia</i>
Macradenia brassavolae

<i>Macradenia</i>
Macradenia sp.


Warmingia Rchb.f. (3 spp.; Fig. 12) has an oddly disjunct distribution, including Costa Rica, southern Ecuador, and Brazil. Pollination has not been reported, but their floral fragrance is similar to some Macroclinium and is produced abundantly during the morning, suggestive of pollination by male euglossine bees.


Macroclinium Barb. Rodr. (approximately 40 spp.; Fig. 2Z; Fig. 12) ranges throughout much of the Neotropics from Mexico south to Peru and Brazil. The plants are diminutive twig epiphytes, and are often found on cultivated Citrus L. and Psidium L. The flowers are similar in morphology and function to its sister genus Notylia, but the two differ in inflorescence and vegetative habit. Macroclinium species often are monopodial, with small psygmoid fans usually lacking pseudobulbs. The inflorescence is pendent, pseudo-umbellate, with clusters of numerous delicate flowers with narrow sepals, petals, and lip. In spite of their small size, the fragrant flowers attract relatively large male euglossine bees. Pollinaria are deposited on the face (frons) of the bee during fragrance collection.

Macroclinium dalstroemii
Macroclinium dalstroemii

Macroclinium
Macroclinium sp.

Macroclinium
Macroclinium sp.


Notylia Lindl. (approximately 60 spp.; Fig. 12) also range throughout much of the Neotropics, similar to its sister, Macroclinium. In contrast to the paedomorphic fans of Macroclinium, plants of Notylia mature to bear spherical or cylindrical and relatively slender pseudobulbs and relatively large conduplicate leaves. The flowers are similar to those of Macroclinium, but are presented evenly spaced on a pendent, usually unbranched raceme. Pollination is also by male euglossine bees, with pollinarium deposition on the frons.

<i>Notylia</i>
Notylia sp.

<i>Notylia latilabia</i>
Notylia latilabia

<i>Notylia</i>
Notylia sp.


Conclusions

This study presents well supported and highly resolved phylogenetic hypotheses of relationships of all major clades within Oncidiinae based on dense taxon sampling. The deeper topology of this tree strongly reflects the emphasis on plastid data. Additional nuclear data sets such as Xdh (Górniak, Paun & Chase, 2010) would be useful to increase support for the topology and improve the resolution of the spine of the tree. Our translation of this tree into a generic classification results in the first classification of Oncidiinae in which the genera are demonstrably monophyletic. Comparison of our trees with previous classifications reveals that most of the taxonomic disputes involve clades that contain large numbers of species with yellow "oncidioid" floral morphology. We hypothesize that widespread mimicry involving Malpighiaceae, Oncidiinae and perhaps Calceolaria (Calceolariaceae) has resulted in extensive homoplasy in gross floral features within Oncidiinae. Previous non-cladistic classifications of Oncidiinae were based largely on floral characters, and the homoplasy in oil flower-related floral traits resulted in non-monophyletic generic concepts. Clades with other pollination syndromes (e.g., nectar reward/deceit or male euglossine fragrance rewards) generally display fewer taxonomic disagreements. The generic scheme presented here paves the way for monographic work and studies of character evolution. Orchid taxonomists may still disagree on which clades to recognize at generic level (e.g., Trichocentrum), but the phylogenetic hypotheses from this study will be useful for framing such debates.


- Top of Page -
HOME | PHYLOGENETICS | REFERENCES | IMAGES | DATABASES | DOWNLOADS | PEOPLE | LINKS | ACKNOWLEDGMENTS

Page developed and maintained by Norris H. Williams.
Please send your comments and suggestions to the webmaster: orchid@[nospam]flmnh.ufl.edu
Page last updated: 5 May 2014