Abstract

INTRODUCTION

Oncidium Sw. (Oncidiinae; Cymbidieae; Orchidaceae) is one of the most conspicuous genera of neotropical orchids. Their mostly bright-yellow flowers often marked with brown are commonly called lluvia de oro (golden rain) in Spanish or ‘dancing ladies’ in English. They are mostly epiphytic and frequently seen in cultivation. In its broadest circumscription, Oncidium has over 400 species. Delimitation of Oncidium and related genera has been the subject of much controversy and inconsistency (Dressler and Williams, 1975; Garay and Stacy, 1974). In simplest terms, Oncidium has always been a genus of convenience, but if a set of circumscribing characters were to be proposed these would include: a spurless (non-rewarding) flower with the lip (labellum) attached at a 90° or greater angle to the column and a complex tuberculate callus and tabula infrastigmatica (a fleshy structure on the base of the column; Fig. 1). Given the clear arbitrariness of the primary character, that of the angle of lip/column attachment and heterogeneity of vegetative traits and chromosome numbers in Oncidium when circumscribed using this character (Chase, 1986, 1987), no one was at all surprised that the early restriction endonuclease studies of plastid DNA showed that Oncidium was polyphyletic (Chase and Palmer, 1992). Subsequent studies of plastid and nuclear DNA sequences (Williams et al., 2001a, b) have further detailed the extent of problems with delimitation of Oncidium, and the process of generic recircumscription to achieve monophyly has been started, resulting in the description of Cyrtochiloides N.H.Williams & M.W.Chase and Zelenkoa M.W.Chase & N.H.Williams, resurrection of Cyrtochilum Kunth, and expansion of Caucaea Schltr., Miltonia Lindl., Otoglossum Garay & Dunsterville and Trichocentrum Poeppig & Endl. We have also amalgamated into Oncidium the other genera of Oncidiinae that are embedded in that clade; these include Cochlioda, Odontoglossum and Symphyglossum (Chase et al., 2008). This is the fourth installment of a series of papers recircumscribing genera in Oncidiinae. Here the focus is on the Brazilian species that have previously been included in Oncidium.

Open in a separate window

Fig. 1.

Floral morphology of Gomesa (Oncidium/Klaberiella) longipes (Oncidiinae; Orchidaceae). This species has been treated as a member of O. section Barbata (Garay and Stacy, 1974) and has a synsepal that is fused to about one-third of its length. Scale bars: 0·5 cm (upper), 0·1 cm (lower). (Drawing: Rodrigo B. Singer).

In their synopsis of Oncidium, Garay and Stacy (1974) recognized several sections of O. subgenus Oncidium that were either endemic to Brazil (often most diverse in the Mata Atlantica vegetation) or had their greatest species diversity there. These include nearly all sections of the genus with a synsepal (fused lateral sepals): O. sections Barbata Lindl., Concoloria Kraenzl. pro parte, Crispa Rchb.f., Rhinocerotes Garay & Stacy, Synsepala Pfitzer and Waluewa (Regel) Schltr. In addition, some species of O. section Paucituberculata Lindl., which some authors have recognized as O. section Ranifera Kraenzl., also have a synsepal. Garay and Stacy (1974) also included in O. section Paucituberculata species with thickened leaves and reduced pseudobulbs, such as O. hians Lindl., that are morphologically similar to the species of O. section Pulvinata, the Brazilian mule-ear species of Oncidium (named for their long, flat, thickened leaves and reduced pseudobulbs; these species are now placed in Grandiphyllum Docha Neto) that have entirely free lateral sepals. Species of O. section Ranifera (e.g. O. raniferum Lindl., O. paranaense Kraenzl. and O. hookeri Rolfe) have larger pseudobulbs (relative to their size) and thin leaves compared with O. hians.

Some species of O. section Verrucituberculata Lindl., such as O. batemanianum Parm ex Knowl. & Westc., have a synsepal and are principally distributed in Brazil, whereas others lack the synsepal and are found outside of Brazil (e.g. O. auriferum Rchb.f. from Colombia). The key character of O. section Verrucituberculata (sensu Garay and Stacy) is the presence of a few tubercles isolated on the lip away from the main callus, but these species are otherwise similar to species in other sections of the genus. Oncidium section Concoloria also contains some non-Brazilian species, such as O. hyphaematicum Rchb.f. and O. brachyandrum Lindl., that lack a synsepal but do have the section-defining lip without lateral lobes and an even number of parallel tubercles. Another ‘odd-ball’ Brazilian species, O. venustum Drapiez (synonym O. trulliferum Lindl.), is vegetatively similar to the members of O. section Waluewa, but it lacks the floral traits typical of this group (a pilose column and synsepal). It was treated as a member of O. section Rostrata by Garay and Stacy (1974) because it has a beaked column, but it is out of place in this group vegetatively and geographically (it is the sole species of this section found in Brazil). Chiron (2008) recently transferred this species to Baptistonia.

In addition, there are a number of other principally Brazilian genera that have a synsepal; these are Baptistonia Barb.Rodr. (monospecific or recently with additional species transferred from Oncidium; Chiron and Castro Neto, 2004), Binotia Rolfe (monospecific), the newly described Carriella V.P.Castro & K.G.Lacerda (monospecific; Castro and Lacerda, 2005, 2006), Gomesa Lindl., Ornithophora Barb.Rodr. (monospecific), Rodrigueziella Kuntze (synonym Theodorea Barb.Rodr. non Cass.; two species) and Rodrigueziopsis Schltr (two species). Both Coppensia Dumort. and Waluewa Regel were based on species with a synsepal, and these species were included by Garay and Stacy (1974) in O. sections Synsepala and Waluewa, respectively. A synsepal is relatively rare in Oncidiinae outside Brazil, but this trait does appear in Caucaea Rchb.f. (Andes), Symphyglossum Schltr. (Andes), Systeloglossum Schltr. (Andes to Central America) and Tolumnia Raf. (Caribbean), which were found in previously published studies not to be related to the Brazilian groups with this trait (Williams et al., 2001a, b). In addition to a synsepal, the Brazilian groups fold the perianth parts differently in the bud (the lip is often pushed through the tepals before they unfold), hold their buds on a flexed rather than straight pedicle and have a different inflorescence structure in which the bracts (one at each node) are not closely held to the developing raceme, such that their tips project outward and away from the apex of the raceme, whereas in groups of Oncidium related to the type species, tips of these bracts are tightly appressed to the emerging raceme. These traits also vary among species, and ones such as the protruding lip are much more pronounced in species with large lips, such as O. varicosum Lindl. and O. insigne (Rolfe) Christenson. Most Brazilian groups also have a glaucous cast to their leaves so that even in the vegetative state most species can be separated from those of Oncidium proper.

All these differences from the type group of Oncidium, which is diverse in Andean South America, Central America, the Caribbean and Mexico rather than Brazil, had been previously noted, but the floral morphology of all groups included in Oncidium is more or less stereotypical, mostly bright yellow with brown spotting/bars on sepals and petals with the typical lip/column arrangement. The Brazilian groups also have the most common chromosome number in Oncidium and Oncidiinae, 2n = 56, but they were known not to produce hybrids when crossed to other sections of the genus (Sanford, 1964, 1967). However, interfertility is not useful in determining generic limits in Orchidaceae, as Garay and Stacy (1974) pointed out: ‘The fact that one Oncidium species will or will not cross with another Oncidium species is quite inconsequential as far as this whole alliance is concerned, for genetic compatibility among species is the rule rather than the exception in the entire orchid family. The mere placing of a pollinium from a given plant onto the stigma of another does not convey any information to the vector whether it be a human being or a member of the insect world.’

In earlier studies (Williams et al., 2001a, b), two approaches were used to investigate generic relationships in Oncidiinae: (1) a combined data set of all DNA regions from the plastid and nuclear genomes; and (2) separate analyses of each region to determine if they produce congruent results. The first matrix was assembled to provide well-supported (by the bootstrap; Felsenstein, 1985) estimates of relationships because none of the individual matrices on their own provided enough variable sites to obtain clear patterns. This same procedure is used in this paper but to matK are added three other plastid regions, two portions of the highly variable ycf1 gene (5′ and 3′, 1098 and 1569 bp, respectively) and the psbA-trnH spacer, to further increase the number of variable sites included in analyses.

MATERIALS AND METHODS

Total DNA was extracted from fresh or silica-gel dried samples of leaves or flowers (Chase and Hills, 1991) using a modified 2X CTAB protocol (Doyle and Doyle, 1987), except that some of the samples were purified by caesium chloride/ethidium bromide density-dependent centrifugation (1·5 g mL⁻¹). All samples were vouchered by herbarium specimens or pickled flowers (Appendix).

Production of DNA sequences for matK and ITS used protocols and primers described in Williams et al. (2001a) and Whitten et al. (2000). Primers and protocols for ycf1 and psbA-trnH are described in Neubig et al. (2009) and Xu et al. (2000), respectively. Raw electropherograms were assembled and checked for agreement of both strands using Sequencher 4·8 (Genecodes, Inc.). Sequences were aligned manually in PAUP* version 4·0b10 (Swofford, 2001) following the guidelines for insertion of gaps described by Kelchner (2000). No computer algorithm is capable of estimating insertion/deletion events for the different categories of change in DNA sequences; only the human eye can accomplish this task, so no computerized algorithm was used to align the sequences. Data matrices are available from M.W.C. and N.H.W. upon request (gro.wek@esahc.m; ude.lfu.hnmlf@dihcro); all sequences have been deposited in GenBank (Appendix).

Parsimony analyses

The following heuristic tree search strategy in PAUP* was used for all analyses. In a first phase, 1000 replicates of random-taxon entry order were used with tree-bisection-reconnection (TBR) swapping and a tree limit of ten trees per step to eliminate having to swap on large numbers of trees. The shortest trees found in these 1000 replicates were then used as starting trees in another round of swapping with no tree limit, and all shortest trees found were swapped on to completion. To produce estimates of internal support, 500 replicates of the bootstrap were used with the following settings: simple addition, TBR swapping and a tree limit of ten trees per step. The bootstrap consensus tree and one of the individual trees with branch lengths proportional to the amount of change are illustrated.

Searches were performed on three matrices: (1) a combined plastid matrix for 60 taxa for which all data were present for all taxa (Fig. 2A; but Cariella colorata Königer & J.G.Weinm.bis is missing 5′ ycf1); (2) ITS for 59 taxa (Gomesa handroi (Hoehne) Pabst was missing for ITS; Fig. 2B); and (3) a combined matrix of for all 60 taxa (G. handroi and G. colorata for which ITS and 5′ ycf1 were coded as missing data, respectively). For plastid regions, there are no reasons to expect incongruence for different parts of that non-recombining, uniparentally inherited genome so these were directly combined without performing separate analyses. For the various parts of the ITS/5·8S nrDNA region, a similar argument can be made, although it is possible for other portions of the large ribosomal subunit to exhibit differential patterns of inheritance, e.g. in Caulerpa (Durand et al., 2002) and Drosophila (Polanco et al., 1998). Direct combination of the plastid and nuclear ribosomal regions and their degree of congruence in Oncidiinae have been covered in Williams et al. (2001a, b). No cases of strongly supported incongruence that could be attributed to genealogical discordance were detected here (Fig. 2), so direct combination is appropriate. The positions of Oncidium (Carenidium) gracile Lindl. are discordant in the plastid and ITS results, but there was no strong support for its position in either analysis.

Open in a separate window

Fig. 2.

Comparison of bootstrap consensus trees from separate plastid (A) and nr ITS (B) analyses; bootstrap percentages greater than 50 are indicated above branches. Single trees with proportional branch lengths (DELTRAN optimization) are illustrated in the uppermost corners to demonstrate low levels of divergence within some clades. Alternative recent circumscriptions are indicated in brackets. The type species of Oncidium s.s. is O. altissimum (seventh species from the bottom).

RESULTS

Due to similarity of the separate ITS and plastid trees (Fig. 2), with the noted exception only relationships observed in the combined analysis (Fig. 3) will be described. The ITS matrix had 869 characters, of which 212 (24 %) were potentially parsimony-informative; the 17 ITS trees had 750 steps with a consistency index (CI) of 0·59 (including autapomorphies) and a retention index (RI) of 0·77. The combined plastid matrix had 5184 characters, of which 445 (9 %) were potentially parsimony-informative; the seven plastid trees had 1677 steps with a CI of 0·68 and an RI of 0·79. Analysis of the combined matrix produced three trees of 2464 steps with a CI of 0·65 and an RI of 0·77. The species representing Oncidium sections Barbata, Concoloria, Crispa, Ranifera, Rhinocerotes, Synsepala and Waluewa (sections are indicated to the right of the cladogram; Fig. 3) and most of the Brazilian species of Verrucituberculata form a clade [100 % bootstrap percentage (BP)] in which Baptistonia, Carriella, Gomesa, Ornithophora, Rodrigueziella and Rodrigueziopsis are embedded (Fig. 3). The larger clade is well supported (100 BP) as distinct from the type section of Oncidium (the type is the Caribbean species, O. altissimum Sw.; Figs 2 and ​and3),3), and species representing other major clades of Oncidiinae are well supported in intercalating positions.

Open in a separate window

Fig. 3.

Bootstrap consensus tree from the combined analysis of plastid and nrITS DNA; bootstrap percentages greater than 50 are indicated above branches. A single tree with proportional branch lengths (DELTRAN optimization) is illustrated in the upper left-hand corner to demonstrate low levels of divergence within some clades. Previously recognized genera and the sectional classification of sensu Garay and Stacy (1974) are indicated by the bar on the right. The type species of Oncidium s.s. is O. altissimum (seventh species from the bottom).

In Fig. 2, the generic assignments of these taxa in the system of Docha Neto et al. (2006) and Chiron and Castro Neto (2004) are indicated in brackets. Baptistonia may be monophyletic if Carriella is sister to the rest of the species (but it is unresolved in our analyses; Figs 2 and ​and3),3), but none of the other recently described genera with two or more species is monophyletic. The sections of Oncidium sensu Garay and Stacy (1974) are indicated on Fig. 3; none is monophyletic; if Baptistonia echinata Barb.Rodr., O. venustum Drapiez and O. sarcodes Lindl. (section Crispa) are included, O. section Waluewa is monophyletic. There are few differences in the ITS and plastid sequences between species in this clade, e.g. the sequence for B. echinata is identical to that of Oncidium lietzei Regel and Oncidium widgrenii Lindl. Similarly low levels of variation are observed within several other species clusters: these are centred around Coppensia barbaceniae (Lindl.) Campacci, Coppensia varicosa (Lindl.) Campacci, Rhinocidium macronyx (Rchb.f.) Baptista and Brasilidium crispum (Lodd.) Carpacci [=Anettea imperatoris-maximiliani (Rchb.f.) Szlach. & Mytnik.].

DISCUSSION

Nearly all species of Oncidiinae are pollinated by insects that receive no reward for visiting these plants (pollinator relationships are reviewed by Chase in Pridgeon et al., 2009). Exceptions are the members of the twig epiphyte clade, many of which have a spur or cavity into which nectar is secreted by a horn at the base of the lip or simply the lip base itself (e.g. Comparettia Poeppig & Endl., Rodriguezia Ruiz & Pavón and Leochilus Knowles & Westc.). Floral fragrances may be collected from some genera by male euglossine bees (e.g. Notylia Lindl.). Some taxa in the Brazilian group (e.g. species of O. sections Ranifera, Rhinocerotes and Waluewa) appear to produce varying amounts of oil on the lip callus, which may be collected by Centris bees that mix it with pollen collected from other species and feed it to their larvae (no orchid pollen has ever been observed to be used in this manner; van der Pijl and Dodson, 1966). There is sufficient oil present on lips of some species to collect it with a capillary tube (Faria, 2004), and thus it is possible that this is offered as a reward for pollination. In many Oncidium species, we question whether there is sufficient oil present to function as anything other than an attractant, in which case this would be another instance of deceit pollination, but this situation requires further study. The structure of the oil present in these species has been worked out to be acylglycerols in Oncidium pubes Lindl. and free fatty acids, non-volatile isoprenoids (triterpene squalene) and steroids (stigmasterol and clionasterol) in O. flexuosum Sims (Reis et al., 2000; Faria, 2004). Studies in the field are needed to characterize this pollination syndrome more accurately.

Other research (Powell, 2008) has shown that the yellow-flowered species with typical Oncidium morphology (regardless of whether they are members of type group of Oncidium or the Gomesa clade) have floral UV-absorption patterns that mimic those of similarly coloured species of Malpighiaceae, which have sepalar glands that produce oil collected by Centris bees. No other yellow flowers have been shown to have this same pattern of light absorbance (Powell, 2008). Thus, we may have a continuum of taxa that mimick malpigh flowers in colour but offer no reward to some that appear to be producing enough oil to reward a pollinator. Some of the latter, such as O. raniferum Lindl., are not the typical ‘malpigh yellow’, but UV absorption patterns in these species have yet to be studied. Observations of oil bees visiting Oncidium species (reviewed in van der Cingel, 2001) have recorded that these bees use their mandibles to grab the tabula infrastigmatica and then attempt to scrap oil from the lip callus with their legs, paralleling the situation in malpigh species in which bees use their mandibles to hold onto the banner petal (the one at 90° to the rest of the petals) and scrape oil from sepalar glands with their legs. This observation explains the presence of the tabula infrastigmatica, which otherwise appears to be an enigmatic structure of Oncidium.

In any case, the typical malpigh-mimicking species of Oncidiinae have floral traits associated with Oncidium, and thus it is clear that Oncidium as a genus has historically been defined by a floral syndrome related to its pollination biology. It should therefore come as no great surprise that such floral traits are subject to convergence, and the taxa they characterize are unlikely to be monophyletic. Given that these species are all members of a single subtribe and closely related, then such convergence would be difficult to separate from cases in which similarity of floral morphology was due to common ancestry. Floral traits in Oncidiinae should largely be ignored when considering generic limits, and instead chromosome number and vegetative characters should be emphasized. For example, when identifying the species to be circumscribed into a resurrected concept of Cyrtochilum (Williams et al., 2001a), floral traits had to be disregarded; these had led previous taxonomists [e.g. Garay and Stacy (1974) among many] to place these species in several genera (e.g. Buesiella C.Schweinf., Oliveriana Rchb.f., Oncidium and Odontoglossum Kunth). Instead, we emphasized vegetative traits, in this case pseudobulbs that were round in cross-section (rather than laterally compressed as in most Oncidiinae) with many leaves subtending pseudobulbs rather than one or two as in the great majority of Oncidiinae (Williams et al., 2001a). Previous authors had used other floral traits (e.g. a trilobed lip in which the apical lobe was smaller than the side lobes and often reflexed) in circumscribing Cyrtochilum (Kraenzlin, 1922), but this also did not produce a natural group. For example, it included two species of the mule-ear group, Trichocentrum: T. flavovirens (L.O.Williams) M.W.Chase & N.H.Williams and T. microchilum (Bateman ex Lindl.) M.W.Chase & N.H.Williams. When recircumscribing Trichocentrum, lower chromosome numbers (2n = 42 or less rather than 2n = 56) and reduced pseudobulbs with succulent leaves were emphasized (Williams et al., 2001b). A parallel to this situation is to be found in Moraea (Iridaceae), in which Goldblatt (1998) had to downplay floral traits in circumscribing genera. In many groups of plants, pollination syndromes confound efforts to create a taxonomic scheme based on monophyletic groups.

We propose here to merge the principally Brazilian groups of Oncidium with Gomesa, which we are certain will come as surprise to many people familiar with Gomesa. Although vegetatively compatible, flowers of Gomesa are completely different in morphology (the lip parallels the column, as in many species of Odontoglossum) and colour (mostly pale green); Dressler (1981) reported that Gomesa is pollinated by bees, but did not specify which sort. Gomesa has a synsepal, the leaf character (a glaucous surface), outwardly pointing inflorescence bracts and flexed pedicels, and in spite of general appearances it is fairly easy to diagnose Gomesa as a member of this expanded genus based on these characters.

Gomesa had previously been considered to be closely related to Oncidium (Pabst and Dungs, 1977), but of course saying this does not help much since the latter is polymorphic and polyphyletic. Dressler (1981) considered that Gomesa might be congeneric with Ada Lindl., Brachtia Rchb.f. and Mesospinidium Rchb.f., presumably because of the similar lip/column relationships of these taxa. A relationship to Binotia Rolfe, Rodrigueziella and Rodrigueziopsis is easy to imagine because of their floral similarity to Gomesa, although vegetatively all three are distinctive due to the relatively elongate rhizome between growths and their reduced size.

Oncidium brunleesianum Rchb.f. (syn Baptistonia echinata) is a particularly unusual member of Oncidium section Waluewa; it has a lip that projects forward with the broad side lobes enveloping the column, which are unusual traits for a member of Oncidium. This trait led Szlachetko (1995) to consider it to be a member of his subtribe Trichopiliinae and not at all a close relative of either Gomesa or Oncidium. Carriella colorata (Königer & J.G.Weinm.bis) V.P.Castro & K.G.Lacerda, although described in Oncidium (Königer and Weinmann, 2003), was segregated from Oncidium because it is similar to Baptistonia, but it differs in floral traits as well, the justification given for not including it in an expanded concept of Baptistonia (Castro and Lacerda, 2005). In terms of habit, B. echinata and Carriella colorata are typical members of O. section Waluewa, and recently Chiron (2008) transferred the latter to Baptistonia.

Chiron and Castro Neto (2004) transferred O. sarcodes (section Crispa) to Baptistonia; its floral morphology is ‘intermediate’ between O. sections Waluewa and Crispa, but the present analysis shows that Chiron and Castro Neto were correct; it is a member of O. section Waluewa. Garay and Stacy (1974) considered O. venustum (as O. trulliferum Lindl., a well-known later synonym) a member of O. section Rostrata because it has the typical beaked column of this section, but this structure is a part of some oil-bee syndromes and also occurs in otherwise unrelated groups of orchids with other pollination syndromes (e.g. Phalaenopsis Blume, tribe Vandeae). The present results show that O. venustum is also embedded in the Gomesa clade, perhaps sister to Baptistonia s.l. with high bootstrap support (Fig. 2). Chiron (2008) also has recently transferred this species to Baptistonia.

Ornithophora radicans (Rchb.f.) Garay & Pabst was long been considered problematic and perhaps to have a relationship to Sigmatostalix Rchb.f. (Reichenbach originally placed it in that genus). Production of oil and pollination by Centris bees is accompanied by reduction in size of both flowers (which makes sense given the small sizes of the bees) and plants, as well as a long, beak-like column apex, which curves out over the lip callus. Why this syndrome should involve reduced vegetative stature is unclear, but G. radicans is a good example of this (the bees are small, so floral reduction makes sense). Most species of Sigmatostalix (now included in Oncidium s.s.; Chase et al., 2008) are similarly small. A relationship to the Brazilian sections of Oncidium was never envisioned for this species; it was always viewed as either a member of Sigmatostalix or the sole member of Ornithophora. There are a number of similarities in the beak-like column of this species to those of Sigmatostalix and Ornithocephalus, which share this type of the oil-bee syndrome. In 2003, Castro Neto and Chiron described a new species in Oncidium section Rostrata, O. itapetingense V.P.Castro & Chiron, which Baptista (in Docha Neto et al., 2006) transferred to Carenidium. The floral morphology of this species, with a beaked rostellum, is highly reminiscent of G. radicans, and if material of this species becomes available we predict that it will fall as sister to G. radicans.

We do not propose a sectional classification of Gomesa here because we believe that this serves little purpose in groups in which parallelism is so rife; in practical terms this means that neat boxes upon which such classifications have been typically founded do not exist. Some of the sections of Garay and Stacy (1974) do hold up reasonably well. For example, O. section Waluewa has all of the species included there by Garay and Stacy, but Baptistonia echinata, Carriella colorata, O. sarcodes and O. venustum also belong here. However, other sections like O. section Concoloria are grossly polyphyletic. Characters for O. section Concoloria are a simple lip with a callus composed of a series or parallel keels and no side lobes, and some of the species included here by Garay and Stacy (1974) are members of Miltonia or Oncidium s.s. Of the ten species Garay and Stacy included here, only O. dasytyle Rchb.f., O. concolor Hook. and O. gracile Lindl. fall in the Gomesa clade; the first two form a sister pair embedded among members of O. section Crispa, whereas the last is related to O. section Barbata (Fig. 3). To a large extent, knowing something about phylogenetic relationships makes natural subgeneric categories either extremely difficult to create (as well as for which to write keys) or it renders them artificial due to abandoning monophyly as the primary guiding principle. Looking at the patterns of relationships in Figs. 2 and ​and3,3, one can see that there are well-supported subclades within Gomesa s.l., but it is not possible to identify clear sets of characters that could be used to serve as the basis for constructing a workable classification within the clade. Thus, we abandon the sections of Garay and Stacy (1974) as unworkable and prefer not to set up such a system ourselves. Szlachetko and colleagues transferred sections of Oncidium sensu Garay and Stacy (1974) to segregate genera: in the case of Brazilian groups these were O. section Crispa to Anettea Szlach. & Mytnik, O. section Concoloria to Concocidium Romowicz & Szlach. and O. section Rhinocerotes to Rhinocerotidium Szlach. (Romowicz and Szlachetko, 2006; Szlachetko, 2006; Szlachetko and Mytnik-Ejsmont, 2006). Also in 2006, Campacci (in Docha Neto et al., 2006) established the genus Brasilidium for the species of section Crispa, but he included a different set of species in his genus than Szlachetko and Mytnik-Ejsmont (2006) included in Anettea (both cited the type species as Oncidium crispum, so they are taxonomic synonyms in spite of differing species composition). For example, Szlachetko and Mytnik-Ejsmont (2006) included O. sarcodes in Anettea, but Chiron and Castro Neto (2004) had transferred this species to Baptistonia; Campacci did not include O. sarcodes in Brasilidium in parallel with the treatment by Chiron and Castro Neto (2004). Similarly, Baptista described Carenidium based on the same type species as Romowicz and Szlachetko (2006) had used for Concocidium, the type of Oncidium section Concoloria (O. concolor), but they did not circumscribe the species of these genera in the same way; Romowicz and Szlachetko (2006) included there as well species such as O. hyphaematicum Rchb.f., which is a member of Oncidium s.s. Baptista at least excluded the non-Brazilian species from Carenidium, but he included in his concept two species, O. hookeri Rolfe and O. raniferum Lindl. (O. section Raniferum sensu Garay and Stacy), that are unrelated to the rest of Carenidium (Figs 2 and ​and3).3). This problem was rectified by Chiron and Castro Neto (2006), who segregated this group into Menezesiella. The species of Menezesiella are indeed not related to those in Carenidium, but instead they are embedded in Coppensia (Fig. 3).

Docha Neto et al. (2006) described several other new genera to accommodate these mostly Brazilian groups, many of which belong to the Gomesa clade, but these new genera are not monophyletic, even though they stated that phylogeny should be the basis of taxonomic practice. For example, in Brasilidium, Campacci did not include O. concolor or O. dasytyle, even though both are members of this clade in the present analyses; instead Baptista placed them in Carenidium. Campacci transferred all names he made in Ampliglossum to Coppensia when he (probably) mistakenly thought that O. varicosum was related to O. bicolor. In the present results, O. section Synsepala is not monophyletic, and given the relationships shown in Figs 2 and ​and3,3, Ampliglossum could still apply to a group of species related to O. varicosum because they are not related to the group in which we presume O. bicolor would fall (the clade containing O. warmingii, in which the rest of the species of O. section Synsepala fall). Oncidium varicosum, O. welteri Pabst and O. insigne (Rolfe) Christenson would then be the members of Ampliglossum, which are well supported as related to O. macronyx Rchb.f., for which Baptista (in Docha Neto et al., 2006) described Rhinocidium (Figs 2 and ​and3)3) and Szlachetko (2006) later described Rhinocerotidium. The potential for confusion caused by differing generic concepts of these various authors boggles even the minds of people who are reasonably familiar with these taxa.

However, the story of confusion does not end here. Soon after Baptista (in Docha Neto et al., 2006) described a new genus, Alatiglossum Baptista, for the former members of O. section Barbata, Castro Neto and Martins Catharino (2006) desribed two other genera, Kleberiella V.P.Castro & Cath. and Neoruschia Cath. & V.P.Castro, based on O. uniflorum Booth ex Lindl. and O. cogniauxianum Schltr., respectively, which had been included in Alatiglossum by Baptista. Also in 2006, Chiron and Castro Neto described Menezesiella Chiron & V.P.Castro for three of the species, O. hookeri, O. loefgrenii Cogn. and O. raniferum, that Baptista had included in Carenidium. Chiron and Castro Neto were correct that these species are not related to those of O. section Concoloria, and if we are to use the segregate genera, then use of this name would be appropriate. Finally (we hope) in 2008, Castro Neto and colleagues (including Chiron) transferred another three species to Menezesiella, even though they must certainly have been aware of these at the time they described the genus (Castro Neto et al., 2008). One of these newly transferred species was M. calimaniorum (Guiard) V.P.Castro & G.F.Carr, which was the type species for Castroa Guiard, published in 2006 with a photograph of the species in the same journal in which Chiron (who is also the chief editor) and Castro described Menezesiella (Guiard, 2006); from the photograph it is clear that Castroa calimaniora is a member of Menezesiella. Some workers might prefer to continue the process started by Chiron and Castro Neto (2004, 2005a, b) and Docha Neto et al. (2006) in which successively new sets of genera would be established for the Brazilian members of Oncidium, but once we publish our results it seems certain that someone will decide to make further changes to those that have already been published and another new wave of synonyms will appear. From a simple inspection of the results presented here (Figs 2 and ​and3),3), to produce a reasonable (monophyletic) system of classification at least 15 genera must be recognized (some not yet described; e.g. a new genus would be needed for Oncidium gracile). If such fine splitting at the generic level were to be extended to the whole of Oncidiinae, then there would be double the number of genera to be recognized, from 70 in Pridgeon et al. (2009) to well over 120. We prefer instead to recognize larger and inevitably more heterogeneous generic limits rather than narrower but more homogeneous genera because this ultimately makes the classification easier for most people to use; only experts are able to tell apart the narrowly circumscribed taxa in the Gomesa clade. It is telling that the Brazilians who claimed to know these taxa well did not manage to circumscribe monophyletic genera and have published successive transfers of the same species into new genera. Furthermore, Szlachetko and colleagues fared no better with their reliance on column morphology in establishing monophyletic taxa. In effect, all these workers have relied upon characters that are highly integrated into pollination syndromes; this reliance prevented them from understanding relationships and erecting natural taxa.

In previous classifications, there were two other species of Oncidium that are known to occur in Brazil. Oncidium phymatochilum Lindl. is a vegetatively anomalous species, but it has been shown by Williams et al. (2001b) to be sister to the species of Miltonia, into which these authors transferred it. Morphologically, this species is no more similar to Miltonia than it is to Oncidium, but Williams et al. (2001b) stated that monotypic genera are a redundancy in classification and therefore decided to lump it with its sister taxon Miltonia. Later, Christenson (2005) transferred it into its own genus, Phymatochilum Christenson; in any case, it is no longer a member of Oncidium. The other species, Oncidium orthostates Ridl., occurs from the tepuis of the Guyana Highlands in Venezuela (Amazonas and Bolivar) into Brazil (Amazonas, Goiás, Rondônia, Roraima and Tocantins). This is a peculiar distribution for a member of Oncidium or Gomesa; there are no other members of these two genera in that region. It is also a peculiar species in Oncidium or Gomesa due to absence of a tabula infrastigmatica, although it otherwise looks enough like section Synsepala for Campacci (in Docha Neto et al., 2006) to transfer it into Ampliglossum and later into Coppensia. In the present analyses, this species falls neither with Oncidium nor Gomesa, but rather in an isolated position between these two major clades (Fig. 3). Florally, it is similar to Zelenkoa onusta (Lindl.) M.W.Chase & N.H.Willaims, but it differs in habit and does not fall with that species in any of the present analyses. A new genus for it is described below. With the removal of this species from Oncidium, O. baueri becomes the only member of Oncidium in Brazil.

TAXONOMIC TREATMENT

Gomesa R.Br., Bot. Mag. t. 1748. 1815. Type species: Gomesa recurva R.Br.

Synonyms:

Coppensia Dumort., Nouv. Mém. Acad. Roy. Sci. Bruxelles 9: 10. 1835. Type species: Coppensia bifolia (Sims) Dumort. (basionym: Oncidium bifolium Sims).

Baptistonia Barb.Rodr., Gen. Sp. Orchid. 1: 95. 1877. Type species: Baptistonia echinata Barb.Rodr.

Theodorea Barb.Rodr. (non Cass.), Gen. Sp. Orchid. 1: 144. 1877. Type species: Theodorea gomezoides Barb.Rodr.; this is a later homonym of Theodorea Cass., a genus of Asteraceae.

Ornithophora Barb.Rodr., Gen. Sp. Orchid. 2: 225. 1882. Type species: Ornithophora quadricolor Barb.Rodr. (= Ornithophora radicans (Rchb.f.) Garay & Pabst).

Waluewa Regel, Trudy Imp. S.-Peterburgsk. Bot. Sada 11: 309. 1890. Type species: Waluewa pulchella Regel.

Rodrigueziella Kuntze, Rev. Gen. 649. 1891. Type species: Rodrigueziella gomezoides (Barb.Rodr.) Kuntze (basionym: Theodorea gomezoides Barb.Rodr.).

Binotia Rolfe, Orchid Rev. 13: 296. 1905. Type species: Binotia brasiliensis Rolfe.

Rodrigueziopsis Schltr., Repert. Spec. Nov. Regni Veg. 16: 427. 1920. Type species: Rodrigueziopsis eleutherosepala (Barb.Rodr.) Schltr.

Hellerorchis A.D.Hawkes, Orchid J. 3: 275. 1959. Type species: Hellerorchis gomezoides (Barb.Rodr.) A.D.Hawkes (basionym: Rodrigueziella gomezoides Barb.Rodr.).

Carria V.P.Castro & K.G.Lacerda (non Gardn.), Orchids 74: 694. 2005. Type species: Carria colorata (Königer & J.G.Weinm.bis) V.P.Castro & K.G.Lacerda (basionym: Oncidium coloratum Königer & J.G.Weinm.bis). Carria Castro & Lacerda is a later homonym of Carria Gardn., a genus in Ternstroemiaceae.

Carriella V.P.Castro & K.G.Lacerda, Icon. Orchid. Brasil. 2, t. 123. 2006. Type species: Carriella colorata (Königer & J.G.Weinm.bis) V.P.Castro & K.G.Lacerda (basionym: Oncidium coloratum Königer & J.G.Weinm.bis).

Alatiglossum D.H.Baptista, Colet. Orq. Bras. 3: 87. 2006. Type species: Alatiglossum barbatum (Lindl.) D.H.Baptista (basionym: Oncidium barbatum Lindl.).

Ampliglossum Campacci, Colet. Orq. Bras. 3: 83. 2006. Type species: Ampliglossum varicosum Campacci (basionym: Oncidium varicosum Lindl. & Paxton).

Anettea Szlach. & Mytnik, Pol. J. Bot. 51: 49. 2006. Type species: Anettea crispa (Lodd.) Szlach. & Mytnik (basionym: Oncidium crispum Lodd.).

Brasilidium Campacci, Colet. Orq. Bras. 3: 78. 2006. Type species: Brasilidium crispum Campacci (basionym: Oncidium crispum Lindl.).

Carenidium D.H.Baptista, Colet. Orq. Bras. 3: 90. 2006. Type species: Carenidium concolor (Hook.) D.H.Baptista (basionym: Oncidium concolor Hook.).

Castroa Guiard, Richardiana 6: 162. 2006. Type species: Castroa calimaniana Guiard.

Concocidium Romowicz & Szlach., Pol. Bot. J. 51: 44. 2006. Type species: Concocidium concolor (Hook.) Romowicz & Szlach. (basionym: Oncidium concolor Hook.).

Kleberiella V.P.Castro & Catharino, Richardiana 6: 158. 2006. Type species: Kleberiella uniflora (Booth ex Lindl.) V.P.Castro & Catharino (basionym Oncidium uniflorum Booth ex Lindl.).

Menezesiella Chiron & V.P.Castro, Richardiana 6: 103. 2006. Type species: Menezesiella ranifera (Lindl.) Chiron & V.P.Castro (basionym: Oncidium raniferum Lindl.).

Neoruschia Catharino & V.P.Castro, Richardiana 6: 58. 2006. Type species: Neoruschia cogniauxiana (Schltr.) Catharino & V.P.Castro (basionym: Oncidium cogniauxianum Schltr.).

Rhinocidium D.H.Baptista, Colet. Orq. Bras. 3: 93. 2006. Type species: Rhinocidium longicornu (Mutel) D.H.Baptista (basionym: Oncidium longicornu Mutel).

Rhinocerotidium Szlach., Pol. Bot. J. 51: 40. 2006. Type species: Rhinocerotidium longicornu (Mutel) Szlach. (basionym: Oncidium longicornu Mutel).

Gomesa ×amicta (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium × amictum Lindl., Bot. Reg. t. 66. 1847. (G. sarcodes × G. lietzei).

Gomesa ×colnagoi (Pabst) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium × colnagoi Pabst, Bradea 2: 4. 1976. (G. forbesii × G. zappii).

Gomesa ×lita (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium × litum Rchb.f., Gard. Chron., n.s. 20: 3328. 1883.

Gomesa ×regentii (V.P.Castro and Chiron) M.W.Chase & N.H.Williams, comb. nov.; basionym: Baptisonia × regentii V.P.Castro & Chiron, Richardiana 5: 187. 2005.

Gomesa ×scullyi (Pabst & Mello) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium × scullyi Pabst & Mello, Bradea 2: 187(-188), fig. D. 1977. (G. curta×G. gravesiana).

Gomesa ×wheatleyana (W.H.Gower) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium × wheatleyanum W.H.Gower, The Garden, 154: 227. 1893. (G. imperatoris-maximimiliana × G. dasytyle).

Gomesa adamantina (Marçal & Cath.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium adamantinum Marçal & Cath., Richardiana 6: 119. 2006.

Gomesa albinoi (Schltr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium albinoi Schltr., Repert. Spec. Nov. Regni Veg. 23: 63. 1926.

Gomesa barbaceniae (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium barbaceniae Lindl., Fol. Orch. Oncid. 6: 32. 1855.

Gomesa barbata (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium barbatum Lindl., Coll. Bot. t. 27. 1821.

Gomesa batemanniana (Parm. ex Knowles & Westc.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium batemannianum Parm. ex Knowles & Westc., Fl. Cab. 3: 183. 1840.

Gomesa bicolor (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium bicolor Lindl., Bot. Reg. 66. 1843.

Gomesa bifolia (Sims) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium bifolium Sims, Bot. Mag. 36: t. 1491. 1812.

Gomesa blanchetii (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium blancheti Rchb.f., Linnaea 22: 845. 1850.

Gomesa bohnkiana (V.P.Castro & G.F.Carr) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium bohnkianum V.P.Castro & G.F.Carr, Orchid Digest 70: 88. 2006.

Gomesa brasiliensis (Rolfe) M.W.Chase & N.H.Williams, comb. nov.; basionym: Binotia brasiliensis Rolfe, Orchid Rev. 13: 296. 1905.

Gomesa brieniana (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium brienianum Rchb.f. Gard. Chron. nov. ser. 15: 40. 1881.

Gomesa brunnipetala (Barb.Rodr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium brunnipetalum Barb.Rodr., Gen. Sp. Orchid. 2: 190. 1881.

Gomesa caldensis (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium caldense Rchb.f., Linnaea, 22: 846. 1849.

Gomesa calimaniana (Guiard) M.W.Chase & N.H.Williams, comb. nov.; basionym: Castroa calimaniana Guiard, Richardiana 6: 162. 2006.

Gomesa chapadensis (V.P.Castro & Campacci) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium chapadense V.P.Castro & Campacci, Orquidário 6: 150. 1992.

Gomesa chrysoptera (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium macropetalum var. chrysopterum Lindl., Fol. Orch. Oncid. 16. 1855.

Gomesa chrysopterantha (Lückel) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium chrysopteranthum Lückel, Orchidee (Hamburg) 49: 43. 1998.

Gomesa chrysorhapis (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium chrysorhapis Rchb.f. Gard. Chron. ser. 3, 3: 72. 1888. Probably a synonym of G. cornigera.

Gomesa chrysothyrsus (Rchb.f. ex Warn.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium chrysothyrsus Rchb.f. ex Warn., Sel. Orch. Ser. II. t. 5. 1864.

Gomesa ciliata (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium ciliatum Lindl., Gen. Sp. Orch. 200. 1833.

Gomesa cogniauxiana (Schltr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium cogniauxianum Schltr., Repert. Spec. Nov. Regni Veg. Beih. 9: 175. 1921.

Gomesa colorata (Königer & J.G.Weinm.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium coloratum Königer & J.G.Weinm.bis.

Gomesa concolor (Hook.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium concolor Hook., Bot. Mag. t. 3752. 1839.

Gomesa cornigera (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium cornigerum Lindl., Edwards's Bot. Reg. 18: t. 1542. 1832.

Gomesa croesus (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium croesus Rchb.f., Hamb. Garten-Blumenzeitung 13: 314. 1857.

Gomesa cruciata (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium cruciatum Rchb.f., Gard. Chron. 1: 138. 1878.

Gomesa cuneata (Scheidw.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium cuneatum Scheidw., Allg. Gartenzeitung 10: 309. 1842.

Gomesa curta (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium curtum Lindl., Edwards's Bot. Reg. 33: t. 68. 1847.

Gomesa damacenoi (Chiron & V.P.Castro) M.W.Chase & N.H.Williams, comb. nov.; basionym: Baptisonia damacenoi Chiron & V.P.Castro, Richardiana 4: 124. 2004.

Gomesa dasytyle (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium dasytyle Rchb.f., Gard. Chron. 253. 1873.

Gomesa discifera (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium disciferum Lindl., Fol. Orch. Oncid. 24. 1855.

Gomesa doeringii (Hoehne) M.W.Chase & N.H.Williams, comb. nov.; basionym: Theodorea doeringii Hoehne, Arq. Bot. Estado São Paulo, n.s., f.m., 1: 86. 1942.

Gomesa doniana (Batem. ex W.Baxt.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium donianum Batem. ex W.Baxt., J.C. Loudon Suppl. Hort. Brit. 3: 598. 1850.

Gomesa echinata (Barb.Rodr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Baptistonia echinata Barb.Rodr., Gen. Sp. Orchid. 1: 95. 1877.

Gomesa edmundoi (Pabst) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium edmundoi Pabst, Bradea 1: 172, tab. 1, fig. B. 1972.

Gomesa eleutherosepala (Barb.Rodr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Rodriguezia eleutherosepala Barb.Rodr., Gen. Spec. Orchid. 2: 240. 1881.

Gomesa emilii (Schltr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium emilii Schltr., Repert. Spec. Nov. Regni Veg. 16: 449. 1920.

Gomesa enderiana (Hort. ex Gard. Chron.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium enderianum Hort. ex Gard. Chron. 2: 75. 1892.

Gomesa eurycline (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium eurycline Rchb.f., Gard. Chron. 2: 812. 1883.

Gomesa flexuosa (Sims) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium flexuosum Sims, Bot. Mag. t. 2203. 1821.

Gomesa forbesii (Hook.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium forbesii Hook., Bot. Mag. 65: t. 3705. 1839.

Gomesa fuscans (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium fuscans Rchb.f., Otia Bot. Hamb. 86. 1881.

Gomesa fuscopetala (Hoehne) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium macropetalum Lindl. var. fuscopetalum Hoehne, Relat. Commiss. Linhas. Telegr. Estrateg. Matto Grosso Amazonas 5: 61, pl. 56, f. 2. 1910.

Gomesa gardneri (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium gardneri Lindl., in Hook., Lond. J. Bot. 2: 662. 1843.

Gomesa gilva (Vell.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Epidendrum gilvum Vell., Fl. Flum. lx. t. 33. 1831.

Gomesa gracilis (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium gracile Lindl., Edwards's Bot. Reg. 23: t. 1920. 1837.

Gomesa gracillima (Schltr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium gracillimum Cogn., in Mart. Fl. Bras. 3(6): 401. 1906.

Gomesa gravesiana (Rolfe) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium gravesianum Rolfe, Gard. Chron. ser. 3, 11: 650. 1892.

Gomesa gutfreundiana (Chiron & V.P.Castro) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium gutfreundianum Chiron & V.P.Castro, Richardiana 3: 122. 2003.

Gomesa hookeri (Rolfe) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium hookeri Rolfe, Gard. Chron. ser. 3, 2: 520. 1941.

Gomesa hydrophila (Barb.Rodr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium hydrophilum Barb.Rodr., Gen. Sp. Orchid. 1: 92. 1877.

Gomesa insignis (Rolfe) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium varicosum Lindl. var. insigne (Rolfe), Orchid Rev. 6: 27. 1898.

Gomesa imperatoris-maximiliani (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium imperatoris-maxiliani Rchb.f., in H.R.vanFernseeWawra, Bot. Ergebn.: 154. 1866. Synonym: Oncidium crispum Lodd. Gen. Sp. Orch. Pl. 197. 1833. Note: the epithet crispa is already occupied in Gomesa, so we have used a later name for this species in Gomesa.

Gomesa isoptera (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium isopterum Lindl., Edwards's Bot. Reg. 23: t. 1920. 1837.

Gomesa itapetingensis (V.P.Castro & Chiron) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium itapetingense V.P.Castro & Chiron, Richardiana 3: 25. 2003.

Gomesa jucunda (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Mesospinidium jucundum Rchb.f., Gard. Chron. (n. s.) 6: 580. 1876.

Gomesa kautskyi (Pabst) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium kautskyi Pabst, Bradea 2(14): 90, t. 2, E. 1976.

Gomesa leinigii (Pabst) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium leinigii Pabst, Bradea 1(20): 182, t. 3, f. B. 1972.

Gomesa lietzei (Regel) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium lietzei Regel, Act. Hort. Petrop. 7: 387. 1880.

Gomesa loefgrenii (Cogn.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium loefgrenii Cogn., in Mart. Fl. Bras. 3, pt. 6: 381, pl. 90, fig. 1. 1905.

Gomesa longicornu (Mutel) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium longicornu Mutel, Mem. Scarpe. 13. 1838.

Gomesa longipes (Lindl. & Paxt.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium longipes Lindl. & Paxt., Fl. Gard. 1: 46. 1850.

Gomesa macronyx (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium macronyx Rchb.f., Otia Bot. Hamb. 95. 1881.

Gomesa macropetala (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium macropetalum Lindl., Sert. Orch. sub t. 48. 1841.

Gomesa majevskyi (Toscano & V.P.Castro) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium majevskyi Toscano & V.P.Castro, Bradea 3: 349. 1983.

Gomesa mandonii (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium mandonii Rchb.f., Xenia Orch. 3: 21. 1878.

Gomesa marshalliana (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium marshallianum Rchb.f., Gard. Chron. 1866: 682. 1866.

Gomesa martiana (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium martianum Lindl., Edwards's Bot. Reg. 23: t. 1920. 1837.

Gomesa megaloptera (Kraenzl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium megalopterum Kraenzl., in Engler, Pflanzenr. 4, Fam. 50: 156. 1922.

Gomesa microphyta (Barb.Rodr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Rodriguezia microphyta Barb.Rodr., Rev. Engenh. 3: 144. 1881.

Gomesa micropogon (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium micropogon Rchb.f., in Bonplandia (Hannover) 2: 90. 1854.

Gomesa montana (Barb.Rodr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium montanum Barb.Rodr., Gen. Sp. Orch. Nov. 1: 93. 1877.

Gomesa neoparanaensis (Chiron & V.P.Castro) M.W.Chase & N.H.Williams, comb. nov.; basionym: Menezesiella neoparanaensis Chiron & V.P.Castro, Richardiana 8: 51. 2008.

Gomesa nitida (Barb.Rodr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium nitidum Barb.Rodr., Gen. Sp. Orch. Nov. 2: 193. 1882.

Gomesa ouricanensis (V.P.Castro & Campacci) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium ouricanense V.P.Castro & Campacci, Bol. CAOB 4: 17. 1992.

Gomesa pabstii (Campacci & C.Espejo) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium pabstii Campacci & C.Espejo, Orquideophilo 6: 36. 1998.

Gomesa paranaensis (Kraenzl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium paranaense Kraenzl., in Svensk. Vet. Akad. Handl. XLVI, 10: 84. 1911.

Gomesa paranapiacabensis (Hoehne) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium paranapiacabense Hoehne, Arq. Bot. Estado São Paulo, new ser. 1: 21, tab. 14. 1938.

Gomesa pardoglossa (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium pardoglossum Rchb.f., Gard. Chron. 1: 516. 1886.

Gomesa pectoralis (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium pectorale Lindl., Sert. Orch. t. 39. 1840.

Gomesa petropolitana (Pabst) M. W. Chase and N. H. Williams, comb. nov.; basionym: Rodrigueziella petropolitana Pabst, Bradea 22: 87. 1976.

Gomesa pirarensis (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium pirarense Rchb.f., Linnaea 22: 846. 1850.

Gomesa praetexta (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium praetextum Rchb.f., Gard. Chron. 1873: 1206. 1873.

Gomesa psyche (Schltr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium psyche Schltr., Repert. Spec. Nov. Regni Veg. 17: 16. 1921.

Gomesa pubes (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium pubes Lindl., Bot. Reg. t. 1007. 1826.

Gomesa pulchella (Regel) M.W.Chase & N.H.Williams, comb. nov.; basionym: Waluewa pulchella Regel, Trudy Imp. S.-Peterburgsk. Bot. Sada 11: 309. 1890. Synonym: Oncidium waluewa Schltr., Repert. Spec. Nov. Regni Veg. 15: 215. 1918. Note: in Oncidium, the epithet pulchella is occupied, so Schlechter published a new name for this species, but in Gomesa it must have its original epithet.

Gomesa radicans (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Sigmatostalix radicans Rchb.f., Walp. Ann. 6: 859. 1849.

Gomesa ramosa (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium ramosum Lindl., Edwards's Bot. Reg. 23: t. 1920. 1837.

Gomesa ranifera (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium raniferum Lindl., Bot. Reg. sub t. 1920, t. 48. 1838.

Gomesa reducta (Kraenzl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium reductum Kraenzl., Repert. Spec. Nov. Regni Veg. 26: 344. 1929.

Gomesa reichertii (L.C.Menezes & V.P.Castro) M.W.Chase & N.H.Williams, comb. nov.; basionym: Coppensia reichertii L.C.Menezes & V.P.Castro, Richardiana 8: 21. 2008.

Gomesa remotiflora (Garay) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium remotiflorum Garay, Taxon 19: 454. 1970.

Gomesa rhinoceros (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium rhinoceros Rchb.f., Bot. Zeit. 14: 514. 1856.

Gomesa riograndensis (Cogn.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium riograndense Cogn., Fl. Bras. (Martius) 3, pt. 6: 446, pl. 89, fig. 2. 1906.

Gomesa riviereana (St.-Lég.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium riviereanum St.-Lég., Rev. Hort. Belge Étrangère, 30: 144; 1904.

Gomesa salesopolitana (V.P.Castro & Chiron) M.W.Chase & N.H.Williams, comb. nov.; basionym: Menezesiella salesopolitana V.P.Castro & Chiron, Richardiana 6: 200. 2006.

Gomesa sarcodes (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium sarcodes Lindl., Journ. Hort. Soc. 4: 266. 1849.

Gomesa sellowii (Cogn.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium sellowii Cogn., in Mart. Fl. Bras. 3, pt. 6: 404. 1906.

Gomesa silvana (V.P.Castro & Campacci) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium silvanum V.P.Castro & Campacci, Orquidário 5: 22. 1991.

Gomesa sincorana (Campacci & Cath.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Coppensia sincorana Campacci & Cath., Bol. CAOP 64: 124. 2006.

Gomesa spiloptera (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium spilopterum Lindl., Edwards's Bot. Reg. 30 Misc.: 75. 1844.

Gomesa trichodes (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium trichodes Lindl., Fol. Orch. Oncid. 6: 15. 1855.

Gomesa truncata (Pabst) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium truncatum Pabst, Orquidea (Rio de Janeiro) 17: 45. 1955.

Gomesa uhlii (Chiron & V.P.Castro) M.W.Chase & N.H.Williams, comb. nov.; basionym: Baptistonia uhlii Chiron & V.P.Castro, Richardiana 6: 21. 2006.

Gomesa unicolor (Rolfe) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium unicolor Rolfe, Orchid Rev. 1: 266. 1893.

Gomesa uniflora (Booth ex Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium uniflorum Booth ex Lindl., Edwards's Bot. Reg. 29: t. 43. 1843.

Gomesa varicosa (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium varicosum Lindl., Edwards's Bot. Reg. 23: t. 1920. 1837.

Gomesa velteniana (V.P.Castro & Chiron) M.W.Chase & N.H.Williams, comb. nov.; basionym: Baptistonia velteniana V.P.Castro & Chiron, Richardiana 5: 80. 2005.

Gomesa venusta (Drapiez) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium venustum Drapiez, Hort. Belge 3: 28. 1836.

Gomesa viperina (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium viperinum Lindl., Gen. Sp. Orch. 197. 1833.

Gomesa warmingii (Rchb.f.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium warmingii Rchb.f., Otia Bot. Hamb. 86. 1881.

Gomesa welteri (Pabst) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium welteri Pabst, Arch. Jard. Bot. Rio de Janeiro 14: 25, tab. 7B. 1956.

Gomesa widgreni (Lindl.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium widgreni Lindl., Fol. Orch. Oncid. 6: 17. 1855.

Gomesa williamsii (Schltr.) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium williamsii Schltr., Repert. Spec. Nov. Regni Veg. 10: 459. 1912.

Gomesa zappii (Pabst) M.W.Chase & N.H.Williams, comb. nov.; basionym: Oncidium zappii Pabst, Bradea 2(22): 3. 1976.

Not transferred: Oncidium kraenzlinianum Cogn. because it does not appear to be a member of this group, even though Garay put it in O. section Synsepala. From its description, it could be Caucaea radiata.

We describe below a new genus for the singular species, Oncidium orthostates Ridl., which differs in floral traits from all other genera of Oncidiinae. This species has a peculiar distribution in the Guyana Highlands of Venezuela, Guyana and Brazil. The species is sister to clades of other genera and is not resolved as sister to any other single taxon. It is most similar florally to Zelenkoa onusta (Lindl.) M.W.Chase & N.H.Williams., the single species of that genus, but it never comes out exclusively with it. We name the new genus in honour of Norris H. Williams, who has spent his botanical career studying members of Oncidiinae and pollination biology of neotropical orchids.

Nohawilliamsia M.W. Chase & Whitten, gen. nov.

Plantae terrestres vel lithophyticae, parvae, pseudobulbi lati oblongi, 1 vel 2 bracteis foliiferis subtenti, folio apicali 1 vel 2. Folia crassa. Inflorescentia lateralis, racemosa, plerumque non ramosa, floribus numerosis. Structura floris Oncidio similis, sed tabula infrastigmatica carenti atque columna brevi. Pollinia dua.

Small, perennial, caespitose herbs, terrestrial to lithophytic, with pseudobulbs clustered on a short rhizome. Pseudobulbs oblong to elliptical-oblong in outline, weakly ancipitous, weakly ridged longitudinally, usually with 1–2 terminal leaves, composed of a single internode, lower portions concealed by 3–4 sheathing bracts, the uppermost 1–2 with a lamina. Leaves elliptic-lanceolate, coriaceous, glabrous, conduplicate, margins entire, eventually deciduous. Inflorescences produced laterally from the base of the pseudobulb, subtended by a sheathing bract, much longer than leaves, racemose (rarely sparsely branched), many-flowered, in some cases producing plantlets at nodes after flowering. Flowers showy, resupinate, pedicellate ovary twisted, glabrous. Sepals free, more or less equal in size to the petals, shortly elliptic, dorsally carinate and shortly acuminate. Petals free, same shape as sepals; lip broadly attached to column, trilobed with the apical lobe much larger than the lateral lobes and apically cleft, with a basal trilobed callus. Column shorter than dorsal sepal, swollen apically, with a pair of lateral wings on the sides of the stigma; stigma round; anther terminal operculate, incumbent, one-celled; pollinarium with two, hard, waxy pollinia with an obvious, abaxial suture, attached to the head of an elongate stipe by irregularly shaped viscin (caudicles); viscidium oval in outline. Capsules and seeds not seen. Seedlings not seen.

There is at this time only a single species in this genus, and we here transfer it formally to Nohawilliamsia.

Nohawilliamsia orthostates (Ridl.) M.W.Chase & Whitten, comb. nov.; basionym: Oncidium orthostates Ridl., Timehri 5: 204. 1886.

APPENDIX

LITERATURE CITED